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 1.1.1a and 1.1.1b [xx XXX xxxx]
12 *) Remove the 'dist' target and add a tarball building script. The
13 'dist' target has fallen out of use, and it shouldn't be
14 necessary to configure just to create a source distribution.
17 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
19 *) Timing vulnerability in DSA signature generation
21 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
22 timing side channel attack. An attacker could use variations in the signing
23 algorithm to recover the private key.
25 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
29 *) Timing vulnerability in ECDSA signature generation
31 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
32 timing side channel attack. An attacker could use variations in the signing
33 algorithm to recover the private key.
35 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
39 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
40 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
41 are retained for backwards compatibility.
44 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
45 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
46 of two gigabytes and the error handling improved.
48 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
49 categorized as a normal bug, not a security issue, because the DRBG reseeds
50 automatically and is fully functional even without additional randomness
51 provided by the application.
53 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
55 *) Add a new ClientHello callback. Provides a callback interface that gives
56 the application the ability to adjust the nascent SSL object at the
57 earliest stage of ClientHello processing, immediately after extensions have
58 been collected but before they have been processed. In particular, this
59 callback can adjust the supported TLS versions in response to the contents
63 *) Add SM2 base algorithm support.
66 *) s390x assembly pack: add (improved) hardware-support for the following
67 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
68 aes-cfb/cfb8, aes-ecb.
71 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
72 parameter is no longer accepted, as it leads to a corrupt table. NULL
73 pem_str is reserved for alias entries only.
76 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
77 step for prime curves. The new implementation is based on formulae from
78 differential addition-and-doubling in homogeneous projective coordinates
79 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
80 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
81 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
82 to work in projective coordinates.
83 [Billy Bob Brumley, Nicola Tuveri]
85 *) Change generating and checking of primes so that the error rate of not
86 being prime depends on the intended use based on the size of the input.
87 For larger primes this will result in more rounds of Miller-Rabin.
88 The maximal error rate for primes with more than 1080 bits is lowered
90 [Kurt Roeckx, Annie Yousar]
92 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
95 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
96 moving between systems, and to avoid confusion when a Windows build is
97 done with mingw vs with MSVC. For POSIX installs, there's still a
98 symlink or copy named 'tsget' to avoid that confusion as well.
101 *) Revert blinding in ECDSA sign and instead make problematic addition
102 length-invariant. Switch even to fixed-length Montgomery multiplication.
105 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
106 step for binary curves. The new implementation is based on formulae from
107 differential addition-and-doubling in mixed Lopez-Dahab projective
108 coordinates, modified to independently blind the operands.
109 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
111 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
112 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
113 EC_METHODs to implement their own specialized "ladder step", to take
114 advantage of more favorable coordinate systems or more efficient
115 differential addition-and-doubling algorithms.
116 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
118 *) Modified the random device based seed sources to keep the relevant
119 file descriptors open rather than reopening them on each access.
120 This allows such sources to operate in a chroot() jail without
121 the associated device nodes being available. This behaviour can be
122 controlled using RAND_keep_random_devices_open().
125 *) Numerous side-channel attack mitigations have been applied. This may have
126 performance impacts for some algorithms for the benefit of improved
127 security. Specific changes are noted in this change log by their respective
131 *) AIX shared library support overhaul. Switch to AIX "natural" way of
132 handling shared libraries, which means collecting shared objects of
133 different versions and bitnesses in one common archive. This allows to
134 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
135 doesn't affect the way 3rd party applications are linked, only how
136 multi-version installation is managed.
139 *) Make ec_group_do_inverse_ord() more robust and available to other
140 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
141 mitigations are applied to the fallback BN_mod_inverse().
142 When using this function rather than BN_mod_inverse() directly, new
143 EC cryptosystem implementations are then safer-by-default.
146 *) Add coordinate blinding for EC_POINT and implement projective
147 coordinate blinding for generic prime curves as a countermeasure to
148 chosen point SCA attacks.
149 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
151 *) Add blinding to ECDSA and DSA signatures to protect against side channel
152 attacks discovered by Keegan Ryan (NCC Group).
155 *) Enforce checking in the pkeyutl command line app to ensure that the input
156 length does not exceed the maximum supported digest length when performing
157 a sign, verify or verifyrecover operation.
160 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
161 I/O in combination with something like select() or poll() will hang. This
162 can be turned off again using SSL_CTX_clear_mode().
163 Many applications do not properly handle non-application data records, and
164 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
165 around the problems in those applications, but can also break some.
166 It's recommended to read the manpages about SSL_read(), SSL_write(),
167 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
168 SSL_CTX_set_read_ahead() again.
171 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
172 now allow empty (zero character) pass phrases.
175 *) Apply blinding to binary field modular inversion and remove patent
176 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
179 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
180 binary and prime elliptic curves.
183 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
184 constant time fixed point multiplication.
187 *) Revise elliptic curve scalar multiplication with timing attack
188 defenses: ec_wNAF_mul redirects to a constant time implementation
189 when computing fixed point and variable point multiplication (which
190 in OpenSSL are mostly used with secret scalars in keygen, sign,
191 ECDH derive operations).
192 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
195 *) Updated CONTRIBUTING
198 *) Updated DRBG / RAND to request nonce and additional low entropy
199 randomness from the system.
200 [Matthias St. Pierre]
202 *) Updated 'openssl rehash' to use OpenSSL consistent default.
205 *) Moved the load of the ssl_conf module to libcrypto, which helps
206 loading engines that libssl uses before libssl is initialised.
209 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
212 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
213 [Ingo Schwarze, Rich Salz]
215 *) Added output of accepting IP address and port for 'openssl s_server'
218 *) Added a new API for TLSv1.3 ciphersuites:
219 SSL_CTX_set_ciphersuites()
220 SSL_set_ciphersuites()
223 *) Memory allocation failures consistenly add an error to the error
227 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
228 in libcrypto when run as setuid/setgid.
231 *) Load any config file by default when libssl is used.
234 *) Added new public header file <openssl/rand_drbg.h> and documentation
235 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
236 [Matthias St. Pierre]
238 *) QNX support removed (cannot find contributors to get their approval
239 for the license change).
242 *) TLSv1.3 replay protection for early data has been implemented. See the
243 SSL_read_early_data() man page for further details.
246 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
247 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
248 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
249 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
250 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
251 configuration has been separated out. See the ciphers man page or the
252 SSL_CTX_set_ciphersuites() man page for more information.
255 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
256 in responder mode now supports the new "-multi" option, which
257 spawns the specified number of child processes to handle OCSP
258 requests. The "-timeout" option now also limits the OCSP
259 responder's patience to wait to receive the full client request
260 on a newly accepted connection. Child processes are respawned
261 as needed, and the CA index file is automatically reloaded
262 when changed. This makes it possible to run the "ocsp" responder
263 as a long-running service, making the OpenSSL CA somewhat more
264 feature-complete. In this mode, most diagnostic messages logged
265 after entering the event loop are logged via syslog(3) rather than
269 *) Added support for X448 and Ed448. Heavily based on original work by
273 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
274 objects loaded. This adds the functions OSSL_STORE_expect() and
275 OSSL_STORE_find() as well as needed tools to construct searches and
276 get the search data out of them.
279 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
280 version of OpenSSL should review their configuration settings to ensure
281 that they are still appropriate for TLSv1.3. For further information see:
282 https://wiki.openssl.org/index.php/TLS1.3
285 *) Grand redesign of the OpenSSL random generator
287 The default RAND method now utilizes an AES-CTR DRBG according to
288 NIST standard SP 800-90Ar1. The new random generator is essentially
289 a port of the default random generator from the OpenSSL FIPS 2.0
290 object module. It is a hybrid deterministic random bit generator
291 using an AES-CTR bit stream and which seeds and reseeds itself
292 automatically using trusted system entropy sources.
294 Some of its new features are:
295 o Support for multiple DRBG instances with seed chaining.
296 o The default RAND method makes use of a DRBG.
297 o There is a public and private DRBG instance.
298 o The DRBG instances are fork-safe.
299 o Keep all global DRBG instances on the secure heap if it is enabled.
300 o The public and private DRBG instance are per thread for lock free
302 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
304 *) Changed Configure so it only says what it does and doesn't dump
305 so much data. Instead, ./configdata.pm should be used as a script
306 to display all sorts of configuration data.
309 *) Added processing of "make variables" to Configure.
312 *) Added SHA512/224 and SHA512/256 algorithm support.
315 *) The last traces of Netware support, first removed in 1.1.0, have
319 *) Get rid of Makefile.shared, and in the process, make the processing
320 of certain files (rc.obj, or the .def/.map/.opt files produced from
321 the ordinal files) more visible and hopefully easier to trace and
322 debug (or make silent).
325 *) Make it possible to have environment variable assignments as
326 arguments to config / Configure.
329 *) Add multi-prime RSA (RFC 8017) support.
332 *) Add SM3 implemented according to GB/T 32905-2016
333 [ Jack Lloyd <jack.lloyd@ribose.com>,
334 Ronald Tse <ronald.tse@ribose.com>,
335 Erick Borsboom <erick.borsboom@ribose.com> ]
337 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
338 as documented in RFC6066.
339 Based on a patch from Tomasz Moń
340 [Filipe Raimundo da Silva]
342 *) Add SM4 implemented according to GB/T 32907-2016.
343 [ Jack Lloyd <jack.lloyd@ribose.com>,
344 Ronald Tse <ronald.tse@ribose.com>,
345 Erick Borsboom <erick.borsboom@ribose.com> ]
347 *) Reimplement -newreq-nodes and ERR_error_string_n; the
348 original author does not agree with the license change.
351 *) Add ARIA AEAD TLS support.
354 *) Some macro definitions to support VS6 have been removed. Visual
355 Studio 6 has not worked since 1.1.0
358 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
359 without clearing the errors.
362 *) Add "atfork" functions. If building on a system that without
363 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
364 requirements. The RAND facility now uses/requires this.
370 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
371 not possible to disable entirely. However, it's still possible to
372 disable the console reading UI method, UI_OpenSSL() (use UI_null()
375 To disable, configure with 'no-ui-console'. 'no-ui' is still
376 possible to use as an alias. Check at compile time with the
377 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
378 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
381 *) Add a STORE module, which implements a uniform and URI based reader of
382 stores that can contain keys, certificates, CRLs and numerous other
383 objects. The main API is loosely based on a few stdio functions,
384 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
385 OSSL_STORE_error and OSSL_STORE_close.
386 The implementation uses backends called "loaders" to implement arbitrary
387 URI schemes. There is one built in "loader" for the 'file' scheme.
390 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
391 then adjusted to work on FreeBSD 8.4 as well.
392 Enable by configuring with 'enable-devcryptoeng'. This is done by default
393 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
396 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
397 util/mkerr.pl, which is adapted to allow those prefixes, leading to
398 error code calls like this:
400 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
402 With this change, we claim the namespaces OSSL and OPENSSL in a manner
403 that can be encoded in C. For the foreseeable future, this will only
405 [Richard Levitte and Tim Hudson]
407 *) Removed BSD cryptodev engine.
410 *) Add a build target 'build_all_generated', to build all generated files
411 and only that. This can be used to prepare everything that requires
412 things like perl for a system that lacks perl and then move everything
413 to that system and do the rest of the build there.
416 *) In the UI interface, make it possible to duplicate the user data. This
417 can be used by engines that need to retain the data for a longer time
418 than just the call where this user data is passed.
421 *) Ignore the '-named_curve auto' value for compatibility of applications
423 [Tomas Mraz <tmraz@fedoraproject.org>]
425 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
426 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
427 alerts across multiple records (some of which could be empty). In practice
428 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
429 prohibts this altogether and other libraries (BoringSSL, NSS) do not
430 support this at all. Supporting it adds significant complexity to the
431 record layer, and its removal is unlikely to cause inter-operability
435 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
436 with Z. These are meant to replace LONG and ZLONG and to be size safe.
437 The use of LONG and ZLONG is discouraged and scheduled for deprecation
441 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
442 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
443 [Richard Levitte, Andy Polyakov]
445 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
449 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
450 platform rather than 'mingw'.
453 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
454 success if they are asked to add an object which already exists
455 in the store. This change cascades to other functions which load
456 certificates and CRLs.
459 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
460 facilitate stack unwinding even from assembly subroutines.
463 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
464 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
467 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
468 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
469 which is the minimum version we support.
472 *) Certificate time validation (X509_cmp_time) enforces stricter
473 compliance with RFC 5280. Fractional seconds and timezone offsets
474 are no longer allowed.
477 *) Add support for ARIA
480 *) s_client will now send the Server Name Indication (SNI) extension by
481 default unless the new "-noservername" option is used. The server name is
482 based on the host provided to the "-connect" option unless overridden by
486 *) Add support for SipHash
489 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
490 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
491 prevent issues where no progress is being made and the peer continually
492 sends unrecognised record types, using up resources processing them.
495 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
496 using the algorithm defined in
497 https://www.akkadia.org/drepper/SHA-crypt.txt
500 *) Heartbeat support has been removed; the ABI is changed for now.
501 [Richard Levitte, Rich Salz]
503 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
506 *) The RSA "null" method, which was partially supported to avoid patent
507 issues, has been replaced to always returns NULL.
511 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
513 *) Client DoS due to large DH parameter
515 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
516 malicious server can send a very large prime value to the client. This will
517 cause the client to spend an unreasonably long period of time generating a
518 key for this prime resulting in a hang until the client has finished. This
519 could be exploited in a Denial Of Service attack.
521 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
525 *) Cache timing vulnerability in RSA Key Generation
527 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
528 a cache timing side channel attack. An attacker with sufficient access to
529 mount cache timing attacks during the RSA key generation process could
530 recover the private key.
532 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
533 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
537 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
538 parameter is no longer accepted, as it leads to a corrupt table. NULL
539 pem_str is reserved for alias entries only.
542 *) Revert blinding in ECDSA sign and instead make problematic addition
543 length-invariant. Switch even to fixed-length Montgomery multiplication.
546 *) Change generating and checking of primes so that the error rate of not
547 being prime depends on the intended use based on the size of the input.
548 For larger primes this will result in more rounds of Miller-Rabin.
549 The maximal error rate for primes with more than 1080 bits is lowered
551 [Kurt Roeckx, Annie Yousar]
553 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
556 *) Add blinding to ECDSA and DSA signatures to protect against side channel
557 attacks discovered by Keegan Ryan (NCC Group).
560 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
561 now allow empty (zero character) pass phrases.
564 *) Certificate time validation (X509_cmp_time) enforces stricter
565 compliance with RFC 5280. Fractional seconds and timezone offsets
566 are no longer allowed.
569 *) Fixed a text canonicalisation bug in CMS
571 Where a CMS detached signature is used with text content the text goes
572 through a canonicalisation process first prior to signing or verifying a
573 signature. This process strips trailing space at the end of lines, converts
574 line terminators to CRLF and removes additional trailing line terminators
575 at the end of a file. A bug in the canonicalisation process meant that
576 some characters, such as form-feed, were incorrectly treated as whitespace
577 and removed. This is contrary to the specification (RFC5485). This fix
578 could mean that detached text data signed with an earlier version of
579 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
580 signed with a fixed OpenSSL may fail to verify with an earlier version of
581 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
582 and use the "-binary" flag (for the "cms" command line application) or set
583 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
586 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
588 *) Constructed ASN.1 types with a recursive definition could exceed the stack
590 Constructed ASN.1 types with a recursive definition (such as can be found
591 in PKCS7) could eventually exceed the stack given malicious input with
592 excessive recursion. This could result in a Denial Of Service attack. There
593 are no such structures used within SSL/TLS that come from untrusted sources
594 so this is considered safe.
596 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
601 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
603 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
604 effectively reduced to only comparing the least significant bit of each
605 byte. This allows an attacker to forge messages that would be considered as
606 authenticated in an amount of tries lower than that guaranteed by the
607 security claims of the scheme. The module can only be compiled by the
608 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
610 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
615 *) Add a build target 'build_all_generated', to build all generated files
616 and only that. This can be used to prepare everything that requires
617 things like perl for a system that lacks perl and then move everything
618 to that system and do the rest of the build there.
621 *) Backport SSL_OP_NO_RENGOTIATION
623 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
624 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
625 changes this is no longer possible in 1.1.0. Therefore the new
626 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
627 1.1.0 to provide equivalent functionality.
629 Note that if an application built against 1.1.0h headers (or above) is run
630 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
631 accepted but nothing will happen, i.e. renegotiation will not be prevented.
634 *) Removed the OS390-Unix config target. It relied on a script that doesn't
638 *) rsaz_1024_mul_avx2 overflow bug on x86_64
640 There is an overflow bug in the AVX2 Montgomery multiplication procedure
641 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
642 Analysis suggests that attacks against RSA and DSA as a result of this
643 defect would be very difficult to perform and are not believed likely.
644 Attacks against DH1024 are considered just feasible, because most of the
645 work necessary to deduce information about a private key may be performed
646 offline. The amount of resources required for such an attack would be
647 significant. However, for an attack on TLS to be meaningful, the server
648 would have to share the DH1024 private key among multiple clients, which is
649 no longer an option since CVE-2016-0701.
651 This only affects processors that support the AVX2 but not ADX extensions
652 like Intel Haswell (4th generation).
654 This issue was reported to OpenSSL by David Benjamin (Google). The issue
655 was originally found via the OSS-Fuzz project.
659 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
661 *) bn_sqrx8x_internal carry bug on x86_64
663 There is a carry propagating bug in the x86_64 Montgomery squaring
664 procedure. No EC algorithms are affected. Analysis suggests that attacks
665 against RSA and DSA as a result of this defect would be very difficult to
666 perform and are not believed likely. Attacks against DH are considered just
667 feasible (although very difficult) because most of the work necessary to
668 deduce information about a private key may be performed offline. The amount
669 of resources required for such an attack would be very significant and
670 likely only accessible to a limited number of attackers. An attacker would
671 additionally need online access to an unpatched system using the target
672 private key in a scenario with persistent DH parameters and a private
673 key that is shared between multiple clients.
675 This only affects processors that support the BMI1, BMI2 and ADX extensions
676 like Intel Broadwell (5th generation) and later or AMD Ryzen.
678 This issue was reported to OpenSSL by the OSS-Fuzz project.
682 *) Malformed X.509 IPAddressFamily could cause OOB read
684 If an X.509 certificate has a malformed IPAddressFamily extension,
685 OpenSSL could do a one-byte buffer overread. The most likely result
686 would be an erroneous display of the certificate in text format.
688 This issue was reported to OpenSSL by the OSS-Fuzz project.
692 Changes between 1.1.0e and 1.1.0f [25 May 2017]
694 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
695 platform rather than 'mingw'.
698 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
699 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
700 which is the minimum version we support.
703 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
705 *) Encrypt-Then-Mac renegotiation crash
707 During a renegotiation handshake if the Encrypt-Then-Mac extension is
708 negotiated where it was not in the original handshake (or vice-versa) then
709 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
710 and servers are affected.
712 This issue was reported to OpenSSL by Joe Orton (Red Hat).
716 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
718 *) Truncated packet could crash via OOB read
720 If one side of an SSL/TLS path is running on a 32-bit host and a specific
721 cipher is being used, then a truncated packet can cause that host to
722 perform an out-of-bounds read, usually resulting in a crash.
724 This issue was reported to OpenSSL by Robert Święcki of Google.
728 *) Bad (EC)DHE parameters cause a client crash
730 If a malicious server supplies bad parameters for a DHE or ECDHE key
731 exchange then this can result in the client attempting to dereference a
732 NULL pointer leading to a client crash. This could be exploited in a Denial
735 This issue was reported to OpenSSL by Guido Vranken.
739 *) BN_mod_exp may produce incorrect results on x86_64
741 There is a carry propagating bug in the x86_64 Montgomery squaring
742 procedure. No EC algorithms are affected. Analysis suggests that attacks
743 against RSA and DSA as a result of this defect would be very difficult to
744 perform and are not believed likely. Attacks against DH are considered just
745 feasible (although very difficult) because most of the work necessary to
746 deduce information about a private key may be performed offline. The amount
747 of resources required for such an attack would be very significant and
748 likely only accessible to a limited number of attackers. An attacker would
749 additionally need online access to an unpatched system using the target
750 private key in a scenario with persistent DH parameters and a private
751 key that is shared between multiple clients. For example this can occur by
752 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
753 similar to CVE-2015-3193 but must be treated as a separate problem.
755 This issue was reported to OpenSSL by the OSS-Fuzz project.
759 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
761 *) ChaCha20/Poly1305 heap-buffer-overflow
763 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
764 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
765 crash. This issue is not considered to be exploitable beyond a DoS.
767 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
771 *) CMS Null dereference
773 Applications parsing invalid CMS structures can crash with a NULL pointer
774 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
775 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
776 structure callback if an attempt is made to free certain invalid encodings.
777 Only CHOICE structures using a callback which do not handle NULL value are
780 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
784 *) Montgomery multiplication may produce incorrect results
786 There is a carry propagating bug in the Broadwell-specific Montgomery
787 multiplication procedure that handles input lengths divisible by, but
788 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
789 and DH private keys are impossible. This is because the subroutine in
790 question is not used in operations with the private key itself and an input
791 of the attacker's direct choice. Otherwise the bug can manifest itself as
792 transient authentication and key negotiation failures or reproducible
793 erroneous outcome of public-key operations with specially crafted input.
794 Among EC algorithms only Brainpool P-512 curves are affected and one
795 presumably can attack ECDH key negotiation. Impact was not analyzed in
796 detail, because pre-requisites for attack are considered unlikely. Namely
797 multiple clients have to choose the curve in question and the server has to
798 share the private key among them, neither of which is default behaviour.
799 Even then only clients that chose the curve will be affected.
801 This issue was publicly reported as transient failures and was not
802 initially recognized as a security issue. Thanks to Richard Morgan for
803 providing reproducible case.
807 *) Removed automatic addition of RPATH in shared libraries and executables,
808 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
811 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
813 *) Fix Use After Free for large message sizes
815 The patch applied to address CVE-2016-6307 resulted in an issue where if a
816 message larger than approx 16k is received then the underlying buffer to
817 store the incoming message is reallocated and moved. Unfortunately a
818 dangling pointer to the old location is left which results in an attempt to
819 write to the previously freed location. This is likely to result in a
820 crash, however it could potentially lead to execution of arbitrary code.
822 This issue only affects OpenSSL 1.1.0a.
824 This issue was reported to OpenSSL by Robert Święcki.
828 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
830 *) OCSP Status Request extension unbounded memory growth
832 A malicious client can send an excessively large OCSP Status Request
833 extension. If that client continually requests renegotiation, sending a
834 large OCSP Status Request extension each time, then there will be unbounded
835 memory growth on the server. This will eventually lead to a Denial Of
836 Service attack through memory exhaustion. Servers with a default
837 configuration are vulnerable even if they do not support OCSP. Builds using
838 the "no-ocsp" build time option are not affected.
840 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
844 *) SSL_peek() hang on empty record
846 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
847 sends an empty record. This could be exploited by a malicious peer in a
848 Denial Of Service attack.
850 This issue was reported to OpenSSL by Alex Gaynor.
854 *) Excessive allocation of memory in tls_get_message_header() and
855 dtls1_preprocess_fragment()
857 A (D)TLS message includes 3 bytes for its length in the header for the
858 message. This would allow for messages up to 16Mb in length. Messages of
859 this length are excessive and OpenSSL includes a check to ensure that a
860 peer is sending reasonably sized messages in order to avoid too much memory
861 being consumed to service a connection. A flaw in the logic of version
862 1.1.0 means that memory for the message is allocated too early, prior to
863 the excessive message length check. Due to way memory is allocated in
864 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
865 to service a connection. This could lead to a Denial of Service through
866 memory exhaustion. However, the excessive message length check still takes
867 place, and this would cause the connection to immediately fail. Assuming
868 that the application calls SSL_free() on the failed connection in a timely
869 manner then the 21Mb of allocated memory will then be immediately freed
870 again. Therefore the excessive memory allocation will be transitory in
871 nature. This then means that there is only a security impact if:
873 1) The application does not call SSL_free() in a timely manner in the event
874 that the connection fails
876 2) The application is working in a constrained environment where there is
877 very little free memory
879 3) The attacker initiates multiple connection attempts such that there are
880 multiple connections in a state where memory has been allocated for the
881 connection; SSL_free() has not yet been called; and there is insufficient
882 memory to service the multiple requests.
884 Except in the instance of (1) above any Denial Of Service is likely to be
885 transitory because as soon as the connection fails the memory is
886 subsequently freed again in the SSL_free() call. However there is an
887 increased risk during this period of application crashes due to the lack of
888 memory - which would then mean a more serious Denial of Service.
890 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
891 (CVE-2016-6307 and CVE-2016-6308)
894 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
895 had to be removed. Primary reason is that vendor assembler can't
896 assemble our modules with -KPIC flag. As result it, assembly
897 support, was not even available as option. But its lack means
898 lack of side-channel resistant code, which is incompatible with
899 security by todays standards. Fortunately gcc is readily available
900 prepackaged option, which we firmly point at...
903 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
905 *) Windows command-line tool supports UTF-8 opt-in option for arguments
906 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
907 (to any value) allows Windows user to access PKCS#12 file generated
908 with Windows CryptoAPI and protected with non-ASCII password, as well
909 as files generated under UTF-8 locale on Linux also protected with
913 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
914 have been disabled by default and removed from DEFAULT, just like RC4.
915 See the RC4 item below to re-enable both.
918 *) The method for finding the storage location for the Windows RAND seed file
919 has changed. First we check %RANDFILE%. If that is not set then we check
920 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
921 all else fails we fall back to C:\.
924 *) The EVP_EncryptUpdate() function has had its return type changed from void
925 to int. A return of 0 indicates and error while a return of 1 indicates
929 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
930 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
931 off the constant time implementation for RSA, DSA and DH have been made
932 no-ops and deprecated.
935 *) Windows RAND implementation was simplified to only get entropy by
936 calling CryptGenRandom(). Various other RAND-related tickets
938 [Joseph Wylie Yandle, Rich Salz]
940 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
941 and OPENSSL_LH_, respectively. The old names are available
942 with API compatibility. They new names are now completely documented.
945 *) Unify TYPE_up_ref(obj) methods signature.
946 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
947 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
948 int (instead of void) like all others TYPE_up_ref() methods.
949 So now these methods also check the return value of CRYPTO_atomic_add(),
950 and the validity of object reference counter.
951 [fdasilvayy@gmail.com]
953 *) With Windows Visual Studio builds, the .pdb files are installed
954 alongside the installed libraries and executables. For a static
955 library installation, ossl_static.pdb is the associate compiler
956 generated .pdb file to be used when linking programs.
959 *) Remove openssl.spec. Packaging files belong with the packagers.
962 *) Automatic Darwin/OSX configuration has had a refresh, it will now
963 recognise x86_64 architectures automatically. You can still decide
964 to build for a different bitness with the environment variable
965 KERNEL_BITS (can be 32 or 64), for example:
967 KERNEL_BITS=32 ./config
971 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
972 256 bit AES and HMAC with SHA256.
975 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
978 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
981 *) To enable users to have their own config files and build file templates,
982 Configure looks in the directory indicated by the environment variable
983 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
984 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
985 name and is used as is.
988 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
989 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
990 X509_CERT_FILE_CTX was removed.
993 *) "shared" builds are now the default. To create only static libraries use
994 the "no-shared" Configure option.
997 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
998 All of these option have not worked for some while and are fundamental
1002 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1003 global cleanup functions are no longer required because they are handled
1004 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1005 Explicitly de-initing can cause problems (e.g. where a library that uses
1006 OpenSSL de-inits, but an application is still using it). The affected
1007 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1008 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1009 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1010 COMP_zlib_cleanup().
1013 *) --strict-warnings no longer enables runtime debugging options
1014 such as REF_DEBUG. Instead, debug options are automatically
1015 enabled with '--debug' builds.
1016 [Andy Polyakov, Emilia Käsper]
1018 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1019 have been moved out of the public header files. New functions for managing
1020 these have been added.
1023 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1024 objects have been moved out of the public header files. New
1025 functions for managing these have been added.
1028 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1029 have been moved out of the public header files. New functions for managing
1030 these have been added.
1033 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1034 moved out of the public header files. New functions for managing these
1038 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1041 *) Removed the mk1mf build scripts.
1044 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1045 it is always safe to #include a header now.
1048 *) Removed the aged BC-32 config and all its supporting scripts
1051 *) Removed support for Ultrix, Netware, and OS/2.
1054 *) Add support for HKDF.
1055 [Alessandro Ghedini]
1057 *) Add support for blake2b and blake2s
1060 *) Added support for "pipelining". Ciphers that have the
1061 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1062 encryptions/decryptions simultaneously. There are currently no built-in
1063 ciphers with this property but the expectation is that engines will be able
1064 to offer it to significantly improve throughput. Support has been extended
1065 into libssl so that multiple records for a single connection can be
1066 processed in one go (for >=TLS 1.1).
1069 *) Added the AFALG engine. This is an async capable engine which is able to
1070 offload work to the Linux kernel. In this initial version it only supports
1071 AES128-CBC. The kernel must be version 4.1.0 or greater.
1074 *) OpenSSL now uses a new threading API. It is no longer necessary to
1075 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1076 are two supported threading models: pthreads and windows threads. It is
1077 also possible to configure OpenSSL at compile time for "no-threads". The
1078 old threading API should no longer be used. The functions have been
1079 replaced with "no-op" compatibility macros.
1080 [Alessandro Ghedini, Matt Caswell]
1082 *) Modify behavior of ALPN to invoke callback after SNI/servername
1083 callback, such that updates to the SSL_CTX affect ALPN.
1086 *) Add SSL_CIPHER queries for authentication and key-exchange.
1089 *) Changes to the DEFAULT cipherlist:
1090 - Prefer (EC)DHE handshakes over plain RSA.
1091 - Prefer AEAD ciphers over legacy ciphers.
1092 - Prefer ECDSA over RSA when both certificates are available.
1093 - Prefer TLSv1.2 ciphers/PRF.
1094 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1098 *) Change the ECC default curve list to be this, in order: x25519,
1099 secp256r1, secp521r1, secp384r1.
1102 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1103 disabled by default. They can be re-enabled using the
1104 enable-weak-ssl-ciphers option to Configure.
1107 *) If the server has ALPN configured, but supports no protocols that the
1108 client advertises, send a fatal "no_application_protocol" alert.
1109 This behaviour is SHALL in RFC 7301, though it isn't universally
1110 implemented by other servers.
1113 *) Add X25519 support.
1114 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1115 for public and private key encoding using the format documented in
1116 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1117 key generation and key derivation.
1119 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1123 *) Deprecate SRP_VBASE_get_by_user.
1124 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1125 In order to fix an unavoidable memory leak (CVE-2016-0798),
1126 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1127 seed, even if the seed is configured.
1129 Users should use SRP_VBASE_get1_by_user instead. Note that in
1130 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1131 also that even though configuring the SRP seed attempts to hide
1132 invalid usernames by continuing the handshake with fake
1133 credentials, this behaviour is not constant time and no strong
1134 guarantees are made that the handshake is indistinguishable from
1135 that of a valid user.
1138 *) Configuration change; it's now possible to build dynamic engines
1139 without having to build shared libraries and vice versa. This
1140 only applies to the engines in engines/, those in crypto/engine/
1141 will always be built into libcrypto (i.e. "static").
1143 Building dynamic engines is enabled by default; to disable, use
1144 the configuration option "disable-dynamic-engine".
1146 The only requirements for building dynamic engines are the
1147 presence of the DSO module and building with position independent
1148 code, so they will also automatically be disabled if configuring
1149 with "disable-dso" or "disable-pic".
1151 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1152 are also taken away from openssl/opensslconf.h, as they are
1156 *) Configuration change; if there is a known flag to compile
1157 position independent code, it will always be applied on the
1158 libcrypto and libssl object files, and never on the application
1159 object files. This means other libraries that use routines from
1160 libcrypto / libssl can be made into shared libraries regardless
1161 of how OpenSSL was configured.
1163 If this isn't desirable, the configuration options "disable-pic"
1164 or "no-pic" can be used to disable the use of PIC. This will
1165 also disable building shared libraries and dynamic engines.
1168 *) Removed JPAKE code. It was experimental and has no wide use.
1171 *) The INSTALL_PREFIX Makefile variable has been renamed to
1172 DESTDIR. That makes for less confusion on what this variable
1173 is for. Also, the configuration option --install_prefix is
1177 *) Heartbeat for TLS has been removed and is disabled by default
1178 for DTLS; configure with enable-heartbeats. Code that uses the
1179 old #define's might need to be updated.
1180 [Emilia Käsper, Rich Salz]
1182 *) Rename REF_CHECK to REF_DEBUG.
1185 *) New "unified" build system
1187 The "unified" build system is aimed to be a common system for all
1188 platforms we support. With it comes new support for VMS.
1190 This system builds supports building in a different directory tree
1191 than the source tree. It produces one Makefile (for unix family
1192 or lookalikes), or one descrip.mms (for VMS).
1194 The source of information to make the Makefile / descrip.mms is
1195 small files called 'build.info', holding the necessary
1196 information for each directory with source to compile, and a
1197 template in Configurations, like unix-Makefile.tmpl or
1200 With this change, the library names were also renamed on Windows
1201 and on VMS. They now have names that are closer to the standard
1202 on Unix, and include the major version number, and in certain
1203 cases, the architecture they are built for. See "Notes on shared
1204 libraries" in INSTALL.
1206 We rely heavily on the perl module Text::Template.
1209 *) Added support for auto-initialisation and de-initialisation of the library.
1210 OpenSSL no longer requires explicit init or deinit routines to be called,
1211 except in certain circumstances. See the OPENSSL_init_crypto() and
1212 OPENSSL_init_ssl() man pages for further information.
1215 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1216 "peer" argument is now expected to be a BIO_ADDR object.
1218 *) Rewrite of BIO networking library. The BIO library lacked consistent
1219 support of IPv6, and adding it required some more extensive
1220 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1221 which hold all types of addresses and chains of address information.
1222 It also introduces a new API, with functions like BIO_socket,
1223 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1224 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1225 have been adapted accordingly.
1228 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1232 *) CRIME protection: disable compression by default, even if OpenSSL is
1233 compiled with zlib enabled. Applications can still enable compression
1234 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1235 using the SSL_CONF library to configure compression.
1238 *) The signature of the session callback configured with
1239 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1240 was explicitly marked as 'const unsigned char*' instead of
1244 *) Always DPURIFY. Remove the use of uninitialized memory in the
1245 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1248 *) Removed many obsolete configuration items, including
1249 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1250 MD2_CHAR, MD2_INT, MD2_LONG
1252 IDEA_SHORT, IDEA_LONG
1253 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1254 [Rich Salz, with advice from Andy Polyakov]
1256 *) Many BN internals have been moved to an internal header file.
1257 [Rich Salz with help from Andy Polyakov]
1259 *) Configuration and writing out the results from it has changed.
1260 Files such as Makefile include/openssl/opensslconf.h and are now
1261 produced through general templates, such as Makefile.in and
1262 crypto/opensslconf.h.in and some help from the perl module
1265 Also, the center of configuration information is no longer
1266 Makefile. Instead, Configure produces a perl module in
1267 configdata.pm which holds most of the config data (in the hash
1268 table %config), the target data that comes from the target
1269 configuration in one of the Configurations/*.conf files (in
1273 *) To clarify their intended purposes, the Configure options
1274 --prefix and --openssldir change their semantics, and become more
1275 straightforward and less interdependent.
1277 --prefix shall be used exclusively to give the location INSTALLTOP
1278 where programs, scripts, libraries, include files and manuals are
1279 going to be installed. The default is now /usr/local.
1281 --openssldir shall be used exclusively to give the default
1282 location OPENSSLDIR where certificates, private keys, CRLs are
1283 managed. This is also where the default openssl.cnf gets
1285 If the directory given with this option is a relative path, the
1286 values of both the --prefix value and the --openssldir value will
1287 be combined to become OPENSSLDIR.
1288 The default for --openssldir is INSTALLTOP/ssl.
1290 Anyone who uses --openssldir to specify where OpenSSL is to be
1291 installed MUST change to use --prefix instead.
1294 *) The GOST engine was out of date and therefore it has been removed. An up
1295 to date GOST engine is now being maintained in an external repository.
1296 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1297 support for GOST ciphersuites (these are only activated if a GOST engine
1301 *) EGD is no longer supported by default; use enable-egd when
1303 [Ben Kaduk and Rich Salz]
1305 *) The distribution now has Makefile.in files, which are used to
1306 create Makefile's when Configure is run. *Configure must be run
1307 before trying to build now.*
1310 *) The return value for SSL_CIPHER_description() for error conditions
1314 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1316 Obtaining and performing DNSSEC validation of TLSA records is
1317 the application's responsibility. The application provides
1318 the TLSA records of its choice to OpenSSL, and these are then
1319 used to authenticate the peer.
1321 The TLSA records need not even come from DNS. They can, for
1322 example, be used to implement local end-entity certificate or
1323 trust-anchor "pinning", where the "pin" data takes the form
1324 of TLSA records, which can augment or replace verification
1325 based on the usual WebPKI public certification authorities.
1328 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1329 continues to support deprecated interfaces in default builds.
1330 However, applications are strongly advised to compile their
1331 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1332 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1333 or the 1.1.0 releases.
1335 In environments in which all applications have been ported to
1336 not use any deprecated interfaces OpenSSL's Configure script
1337 should be used with the --api=1.1.0 option to entirely remove
1338 support for the deprecated features from the library and
1339 unconditionally disable them in the installed headers.
1340 Essentially the same effect can be achieved with the "no-deprecated"
1341 argument to Configure, except that this will always restrict
1342 the build to just the latest API, rather than a fixed API
1345 As applications are ported to future revisions of the API,
1346 they should update their compile-time OPENSSL_API_COMPAT define
1347 accordingly, but in most cases should be able to continue to
1348 compile with later releases.
1350 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1351 0x10000000L and 0x00908000L, respectively. However those
1352 versions did not support the OPENSSL_API_COMPAT feature, and
1353 so applications are not typically tested for explicit support
1354 of just the undeprecated features of either release.
1357 *) Add support for setting the minimum and maximum supported protocol.
1358 It can bet set via the SSL_set_min_proto_version() and
1359 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1360 MaxProtocol. It's recommended to use the new APIs to disable
1361 protocols instead of disabling individual protocols using
1362 SSL_set_options() or SSL_CONF's Protocol. This change also
1363 removes support for disabling TLS 1.2 in the OpenSSL TLS
1364 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1367 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1370 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1371 and integrates ECDSA and ECDH functionality into EC. Implementations can
1372 now redirect key generation and no longer need to convert to or from
1375 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1376 include the ec.h header file instead.
1379 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1380 ciphers who are no longer supported and drops support the ephemeral RSA key
1381 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1384 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1385 opaque. For HMAC_CTX, the following constructors and destructors
1388 HMAC_CTX *HMAC_CTX_new(void);
1389 void HMAC_CTX_free(HMAC_CTX *ctx);
1391 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1392 destroy such methods has been added. See EVP_MD_meth_new(3) and
1393 EVP_CIPHER_meth_new(3) for documentation.
1396 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1397 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1398 EVP_MD_CTX_reset() should be called instead to reinitialise
1399 an already created structure.
1400 2) For consistency with the majority of our object creators and
1401 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1402 EVP_MD_CTX_(new|free). The old names are retained as macros
1403 for deprecated builds.
1406 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1407 cryptographic operations to be performed asynchronously as long as an
1408 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1409 further details. Libssl has also had this capability integrated with the
1410 introduction of the new mode SSL_MODE_ASYNC and associated error
1411 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1412 pages. This work was developed in partnership with Intel Corp.
1415 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1416 always enabled now. If you want to disable the support you should
1417 exclude it using the list of supported ciphers. This also means that the
1418 "-no_ecdhe" option has been removed from s_server.
1421 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1422 SSL_{CTX_}set1_curves() which can set a list.
1425 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1426 curve you want to support using SSL_{CTX_}set1_curves().
1429 *) State machine rewrite. The state machine code has been significantly
1430 refactored in order to remove much duplication of code and solve issues
1431 with the old code (see ssl/statem/README for further details). This change
1432 does have some associated API changes. Notably the SSL_state() function
1433 has been removed and replaced by SSL_get_state which now returns an
1434 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1435 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1439 *) All instances of the string "ssleay" in the public API were replaced
1440 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1441 Some error codes related to internal RSA_eay API's were renamed.
1444 *) The demo files in crypto/threads were moved to demo/threads.
1447 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1449 [Matt Caswell, Rich Salz]
1451 *) New ASN.1 embed macro.
1453 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1454 structure is not allocated: it is part of the parent. That is instead of
1462 This reduces memory fragmentation and make it impossible to accidentally
1463 set a mandatory field to NULL.
1465 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1466 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1467 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1471 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1474 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1475 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1476 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1477 DES and RC4 ciphersuites.
1480 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1481 This changes the decoding behaviour for some invalid messages,
1482 though the change is mostly in the more lenient direction, and
1483 legacy behaviour is preserved as much as possible.
1486 *) Fix no-stdio build.
1487 [ David Woodhouse <David.Woodhouse@intel.com> and also
1488 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1490 *) New testing framework
1491 The testing framework has been largely rewritten and is now using
1492 perl and the perl modules Test::Harness and an extended variant of
1493 Test::More called OpenSSL::Test to do its work. All test scripts in
1494 test/ have been rewritten into test recipes, and all direct calls to
1495 executables in test/Makefile have become individual recipes using the
1496 simplified testing OpenSSL::Test::Simple.
1498 For documentation on our testing modules, do:
1500 perldoc test/testlib/OpenSSL/Test/Simple.pm
1501 perldoc test/testlib/OpenSSL/Test.pm
1505 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1506 are used; the latter aborts on memory leaks (usually checked on exit).
1507 Some undocumented "set malloc, etc., hooks" functions were removed
1508 and others were changed. All are now documented.
1511 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1513 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1515 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1516 from RFC4279, RFC4785, RFC5487, RFC5489.
1518 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1519 original RSA_PSK patch.
1522 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1523 era flag was never set throughout the codebase (only read). Also removed
1524 SSL3_FLAGS_POP_BUFFER which was only used if
1525 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1528 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1529 to be "oneline" instead of "compat".
1532 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1533 not aware of clients that still exhibit this bug, and the workaround
1534 hasn't been working properly for a while.
1537 *) The return type of BIO_number_read() and BIO_number_written() as well as
1538 the corresponding num_read and num_write members in the BIO structure has
1539 changed from unsigned long to uint64_t. On platforms where an unsigned
1540 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1544 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1545 OpenSSL without support for them. It also means that maintaining
1546 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1547 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1550 *) Removed support for the two export grade static DH ciphersuites
1551 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1552 were newly added (along with a number of other static DH ciphersuites) to
1553 1.0.2. However the two export ones have *never* worked since they were
1554 introduced. It seems strange in any case to be adding new export
1555 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1558 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1559 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1560 and turned into macros which simply call the new preferred function names
1561 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1562 should use the new names instead. Also as part of this change the ssl23.h
1563 header file has been removed.
1566 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1567 code and the associated standard is no longer considered fit-for-purpose.
1570 *) RT2547 was closed. When generating a private key, try to make the
1571 output file readable only by the owner. This behavior change might
1572 be noticeable when interacting with other software.
1574 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1578 *) Added HTTP GET support to the ocsp command.
1581 *) Changed default digest for the dgst and enc commands from MD5 to
1585 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1588 *) Added support for TLS extended master secret from
1589 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1590 initial patch which was a great help during development.
1593 *) All libssl internal structures have been removed from the public header
1594 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1595 now redundant). Users should not attempt to access internal structures
1596 directly. Instead they should use the provided API functions.
1599 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1600 Access to deprecated functions can be re-enabled by running config with
1601 "enable-deprecated". In addition applications wishing to use deprecated
1602 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1603 will, by default, disable some transitive includes that previously existed
1604 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1607 *) Added support for OCB mode. OpenSSL has been granted a patent license
1608 compatible with the OpenSSL license for use of OCB. Details are available
1609 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1610 for OCB can be removed by calling config with no-ocb.
1613 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1614 compatible client hello.
1617 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1618 done while fixing the error code for the key-too-small case.
1619 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1621 *) CA.sh has been removed; use CA.pl instead.
1624 *) Removed old DES API.
1627 *) Remove various unsupported platforms:
1633 Sinix/ReliantUNIX RM400
1638 16-bit platforms such as WIN16
1641 *) Clean up OPENSSL_NO_xxx #define's
1642 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1643 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1644 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1645 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1646 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1647 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1648 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1649 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1650 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1651 Remove MS_STATIC; it's a relic from platforms <32 bits.
1654 *) Cleaned up dead code
1655 Remove all but one '#ifdef undef' which is to be looked at.
1658 *) Clean up calling of xxx_free routines.
1659 Just like free(), fix most of the xxx_free routines to accept
1660 NULL. Remove the non-null checks from callers. Save much code.
1663 *) Add secure heap for storage of private keys (when possible).
1664 Add BIO_s_secmem(), CBIGNUM, etc.
1665 Contributed by Akamai Technologies under our Corporate CLA.
1668 *) Experimental support for a new, fast, unbiased prime candidate generator,
1669 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1670 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1672 *) New output format NSS in the sess_id command line tool. This allows
1673 exporting the session id and the master key in NSS keylog format.
1674 [Martin Kaiser <martin@kaiser.cx>]
1676 *) Harmonize version and its documentation. -f flag is used to display
1678 [mancha <mancha1@zoho.com>]
1680 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1681 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1682 [mancha <mancha1@zoho.com>]
1684 *) Fix some double frees. These are not thought to be exploitable.
1685 [mancha <mancha1@zoho.com>]
1687 *) A missing bounds check in the handling of the TLS heartbeat extension
1688 can be used to reveal up to 64k of memory to a connected client or
1691 Thanks for Neel Mehta of Google Security for discovering this bug and to
1692 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1693 preparing the fix (CVE-2014-0160)
1694 [Adam Langley, Bodo Moeller]
1696 *) Fix for the attack described in the paper "Recovering OpenSSL
1697 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1698 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1699 http://eprint.iacr.org/2014/140
1701 Thanks to Yuval Yarom and Naomi Benger for discovering this
1702 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1703 [Yuval Yarom and Naomi Benger]
1705 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1706 this fixes a limitation in previous versions of OpenSSL.
1709 *) Experimental encrypt-then-mac support.
1711 Experimental support for encrypt then mac from
1712 draft-gutmann-tls-encrypt-then-mac-02.txt
1714 To enable it set the appropriate extension number (0x42 for the test
1715 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1717 For non-compliant peers (i.e. just about everything) this should have no
1720 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1724 *) Add EVP support for key wrapping algorithms, to avoid problems with
1725 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1726 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1727 algorithms and include tests cases.
1730 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1734 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1735 MGF1 digest and OAEP label.
1738 *) Make openssl verify return errors.
1739 [Chris Palmer <palmer@google.com> and Ben Laurie]
1741 *) New function ASN1_TIME_diff to calculate the difference between two
1742 ASN1_TIME structures or one structure and the current time.
1745 *) Update fips_test_suite to support multiple command line options. New
1746 test to induce all self test errors in sequence and check expected
1750 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1751 sign or verify all in one operation.
1754 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1755 test programs and fips_test_suite. Includes functionality to parse
1756 the minimal script output of fipsalgest.pl directly.
1759 *) Add authorisation parameter to FIPS_module_mode_set().
1762 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1765 *) Use separate DRBG fields for internal and external flags. New function
1766 FIPS_drbg_health_check() to perform on demand health checking. Add
1767 generation tests to fips_test_suite with reduced health check interval to
1768 demonstrate periodic health checking. Add "nodh" option to
1769 fips_test_suite to skip very slow DH test.
1772 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1776 *) More extensive health check for DRBG checking many more failure modes.
1777 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1778 combination: call this in fips_test_suite.
1781 *) Add support for canonical generation of DSA parameter 'g'. See
1784 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1785 POST to handle HMAC cases.
1788 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1789 to return numerical and string versions of the FIPS module number.
1792 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1793 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1794 outside the validated module in the FIPS capable OpenSSL.
1797 *) Minor change to DRBG entropy callback semantics. In some cases
1798 there is no multiple of the block length between min_len and
1799 max_len. Allow the callback to return more than max_len bytes
1800 of entropy but discard any extra: it is the callback's responsibility
1801 to ensure that the extra data discarded does not impact the
1802 requested amount of entropy.
1805 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1806 information in FIPS186-3, SP800-57 and SP800-131A.
1809 *) CCM support via EVP. Interface is very similar to GCM case except we
1810 must supply all data in one chunk (i.e. no update, final) and the
1811 message length must be supplied if AAD is used. Add algorithm test
1815 *) Initial version of POST overhaul. Add POST callback to allow the status
1816 of POST to be monitored and/or failures induced. Modify fips_test_suite
1817 to use callback. Always run all selftests even if one fails.
1820 *) XTS support including algorithm test driver in the fips_gcmtest program.
1821 Note: this does increase the maximum key length from 32 to 64 bytes but
1822 there should be no binary compatibility issues as existing applications
1823 will never use XTS mode.
1826 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1827 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1828 performs algorithm blocking for unapproved PRNG types. Also do not
1829 set PRNG type in FIPS_mode_set(): leave this to the application.
1830 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1831 the standard OpenSSL PRNG: set additional data to a date time vector.
1834 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1835 This shouldn't present any incompatibility problems because applications
1836 shouldn't be using these directly and any that are will need to rethink
1837 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1840 *) Extensive self tests and health checking required by SP800-90 DRBG.
1841 Remove strength parameter from FIPS_drbg_instantiate and always
1842 instantiate at maximum supported strength.
1845 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1848 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1851 *) New function DH_compute_key_padded() to compute a DH key and pad with
1852 leading zeroes if needed: this complies with SP800-56A et al.
1855 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1856 anything, incomplete, subject to change and largely untested at present.
1859 *) Modify fipscanisteronly build option to only build the necessary object
1860 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1863 *) Add experimental option FIPSSYMS to give all symbols in
1864 fipscanister.o and FIPS or fips prefix. This will avoid
1865 conflicts with future versions of OpenSSL. Add perl script
1866 util/fipsas.pl to preprocess assembly language source files
1867 and rename any affected symbols.
1870 *) Add selftest checks and algorithm block of non-fips algorithms in
1871 FIPS mode. Remove DES2 from selftests.
1874 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1875 return internal method without any ENGINE dependencies. Add new
1876 tiny fips sign and verify functions.
1879 *) New build option no-ec2m to disable characteristic 2 code.
1882 *) New build option "fipscanisteronly". This only builds fipscanister.o
1883 and (currently) associated fips utilities. Uses the file Makefile.fips
1884 instead of Makefile.org as the prototype.
1887 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1888 Update fips_gcmtest to use IV generator.
1891 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1892 setting output buffer to NULL. The *Final function must be
1893 called although it will not retrieve any additional data. The tag
1894 can be set or retrieved with a ctrl. The IV length is by default 12
1895 bytes (96 bits) but can be set to an alternative value. If the IV
1896 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1900 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1901 underlying do_cipher function handles all cipher semantics itself
1902 including padding and finalisation. This is useful if (for example)
1903 an ENGINE cipher handles block padding itself. The behaviour of
1904 do_cipher is subtly changed if this flag is set: the return value
1905 is the number of characters written to the output buffer (zero is
1906 no longer an error code) or a negative error code. Also if the
1907 input buffer is NULL and length 0 finalisation should be performed.
1910 *) If a candidate issuer certificate is already part of the constructed
1911 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1914 *) Improve forward-security support: add functions
1916 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1917 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1919 for use by SSL/TLS servers; the callback function will be called whenever a
1920 new session is created, and gets to decide whether the session may be
1921 cached to make it resumable (return 0) or not (return 1). (As by the
1922 SSL/TLS protocol specifications, the session_id sent by the server will be
1923 empty to indicate that the session is not resumable; also, the server will
1924 not generate RFC 4507 (RFC 5077) session tickets.)
1926 A simple reasonable callback implementation is to return is_forward_secure.
1927 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1928 by the SSL/TLS server library, indicating whether it can provide forward
1930 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1932 *) New -verify_name option in command line utilities to set verification
1936 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1937 Add CMAC pkey methods.
1940 *) Experimental renegotiation in s_server -www mode. If the client
1941 browses /reneg connection is renegotiated. If /renegcert it is
1942 renegotiated requesting a certificate.
1945 *) Add an "external" session cache for debugging purposes to s_server. This
1946 should help trace issues which normally are only apparent in deployed
1947 multi-process servers.
1950 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1951 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1952 BIO_set_cipher() and some obscure PEM functions were changed so they
1953 can now return an error. The RAND changes required a change to the
1954 RAND_METHOD structure.
1957 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1958 a gcc attribute to warn if the result of a function is ignored. This
1959 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1960 whose return value is often ignored.
1963 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1964 These allow SCTs (signed certificate timestamps) to be requested and
1965 validated when establishing a connection.
1966 [Rob Percival <robpercival@google.com>]
1968 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1970 *) Prevent padding oracle in AES-NI CBC MAC check
1972 A MITM attacker can use a padding oracle attack to decrypt traffic
1973 when the connection uses an AES CBC cipher and the server support
1976 This issue was introduced as part of the fix for Lucky 13 padding
1977 attack (CVE-2013-0169). The padding check was rewritten to be in
1978 constant time by making sure that always the same bytes are read and
1979 compared against either the MAC or padding bytes. But it no longer
1980 checked that there was enough data to have both the MAC and padding
1983 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1987 *) Fix EVP_EncodeUpdate overflow
1989 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1990 Base64 encoding of binary data. If an attacker is able to supply very large
1991 amounts of input data then a length check can overflow resulting in a heap
1994 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1995 the PEM_write_bio* family of functions. These are mainly used within the
1996 OpenSSL command line applications, so any application which processes data
1997 from an untrusted source and outputs it as a PEM file should be considered
1998 vulnerable to this issue. User applications that call these APIs directly
1999 with large amounts of untrusted data may also be vulnerable.
2001 This issue was reported by Guido Vranken.
2005 *) Fix EVP_EncryptUpdate overflow
2007 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2008 is able to supply very large amounts of input data after a previous call to
2009 EVP_EncryptUpdate() with a partial block then a length check can overflow
2010 resulting in a heap corruption. Following an analysis of all OpenSSL
2011 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2012 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2013 the first called function after an EVP_EncryptInit(), and therefore that
2014 specific call must be safe. The second form is where the length passed to
2015 EVP_EncryptUpdate() can be seen from the code to be some small value and
2016 therefore there is no possibility of an overflow. Since all instances are
2017 one of these two forms, it is believed that there can be no overflows in
2018 internal code due to this problem. It should be noted that
2019 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2020 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2021 of these calls have also been analysed too and it is believed there are no
2022 instances in internal usage where an overflow could occur.
2024 This issue was reported by Guido Vranken.
2028 *) Prevent ASN.1 BIO excessive memory allocation
2030 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2031 a short invalid encoding can cause allocation of large amounts of memory
2032 potentially consuming excessive resources or exhausting memory.
2034 Any application parsing untrusted data through d2i BIO functions is
2035 affected. The memory based functions such as d2i_X509() are *not* affected.
2036 Since the memory based functions are used by the TLS library, TLS
2037 applications are not affected.
2039 This issue was reported by Brian Carpenter.
2045 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2046 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2047 in arbitrary stack data being returned in the buffer.
2049 This issue was reported by Guido Vranken.
2053 *) Modify behavior of ALPN to invoke callback after SNI/servername
2054 callback, such that updates to the SSL_CTX affect ALPN.
2057 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2061 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2062 methods are enabled and ssl2 is disabled the methods return NULL.
2065 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2067 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2068 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2069 provide any "EXPORT" or "LOW" strength ciphers.
2072 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2073 is by default disabled at build-time. Builds that are not configured with
2074 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2075 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2076 will need to explicitly call either of:
2078 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2080 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2082 as appropriate. Even if either of those is used, or the application
2083 explicitly uses the version-specific SSLv2_method() or its client and
2084 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2085 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2086 ciphers, and SSLv2 56-bit DES are no longer available.
2090 *) Fix a double-free in DSA code
2092 A double free bug was discovered when OpenSSL parses malformed DSA private
2093 keys and could lead to a DoS attack or memory corruption for applications
2094 that receive DSA private keys from untrusted sources. This scenario is
2097 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2102 *) Disable SRP fake user seed to address a server memory leak.
2104 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2106 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2107 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2108 was changed to ignore the "fake user" SRP seed, even if the seed
2111 Users should use SRP_VBASE_get1_by_user instead. Note that in
2112 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2113 also that even though configuring the SRP seed attempts to hide
2114 invalid usernames by continuing the handshake with fake
2115 credentials, this behaviour is not constant time and no strong
2116 guarantees are made that the handshake is indistinguishable from
2117 that of a valid user.
2121 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2123 In the BN_hex2bn function the number of hex digits is calculated using an
2124 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2125 large values of |i| this can result in |bn_expand| not allocating any
2126 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2127 field as NULL leading to a subsequent NULL ptr deref. For very large values
2128 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2129 In this case memory is allocated to the internal BIGNUM data field, but it
2130 is insufficiently sized leading to heap corruption. A similar issue exists
2131 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2132 is ever called by user applications with very large untrusted hex/dec data.
2133 This is anticipated to be a rare occurrence.
2135 All OpenSSL internal usage of these functions use data that is not expected
2136 to be untrusted, e.g. config file data or application command line
2137 arguments. If user developed applications generate config file data based
2138 on untrusted data then it is possible that this could also lead to security
2139 consequences. This is also anticipated to be rare.
2141 This issue was reported to OpenSSL by Guido Vranken.
2145 *) Fix memory issues in BIO_*printf functions
2147 The internal |fmtstr| function used in processing a "%s" format string in
2148 the BIO_*printf functions could overflow while calculating the length of a
2149 string and cause an OOB read when printing very long strings.
2151 Additionally the internal |doapr_outch| function can attempt to write to an
2152 OOB memory location (at an offset from the NULL pointer) in the event of a
2153 memory allocation failure. In 1.0.2 and below this could be caused where
2154 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2155 could be in processing a very long "%s" format string. Memory leaks can
2158 The first issue may mask the second issue dependent on compiler behaviour.
2159 These problems could enable attacks where large amounts of untrusted data
2160 is passed to the BIO_*printf functions. If applications use these functions
2161 in this way then they could be vulnerable. OpenSSL itself uses these
2162 functions when printing out human-readable dumps of ASN.1 data. Therefore
2163 applications that print this data could be vulnerable if the data is from
2164 untrusted sources. OpenSSL command line applications could also be
2165 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2166 as command line arguments.
2168 Libssl is not considered directly vulnerable. Additionally certificates etc
2169 received via remote connections via libssl are also unlikely to be able to
2170 trigger these issues because of message size limits enforced within libssl.
2172 This issue was reported to OpenSSL Guido Vranken.
2176 *) Side channel attack on modular exponentiation
2178 A side-channel attack was found which makes use of cache-bank conflicts on
2179 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2180 of RSA keys. The ability to exploit this issue is limited as it relies on
2181 an attacker who has control of code in a thread running on the same
2182 hyper-threaded core as the victim thread which is performing decryptions.
2184 This issue was reported to OpenSSL by Yuval Yarom, The University of
2185 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2186 Nadia Heninger, University of Pennsylvania with more information at
2187 http://cachebleed.info.
2191 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2192 if no keysize is specified with default_bits. This fixes an
2193 omission in an earlier change that changed all RSA/DSA key generation
2194 apps to use 2048 bits by default.
2197 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2198 *) DH small subgroups
2200 Historically OpenSSL only ever generated DH parameters based on "safe"
2201 primes. More recently (in version 1.0.2) support was provided for
2202 generating X9.42 style parameter files such as those required for RFC 5114
2203 support. The primes used in such files may not be "safe". Where an
2204 application is using DH configured with parameters based on primes that are
2205 not "safe" then an attacker could use this fact to find a peer's private
2206 DH exponent. This attack requires that the attacker complete multiple
2207 handshakes in which the peer uses the same private DH exponent. For example
2208 this could be used to discover a TLS server's private DH exponent if it's
2209 reusing the private DH exponent or it's using a static DH ciphersuite.
2211 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2212 TLS. It is not on by default. If the option is not set then the server
2213 reuses the same private DH exponent for the life of the server process and
2214 would be vulnerable to this attack. It is believed that many popular
2215 applications do set this option and would therefore not be at risk.
2217 The fix for this issue adds an additional check where a "q" parameter is
2218 available (as is the case in X9.42 based parameters). This detects the
2219 only known attack, and is the only possible defense for static DH
2220 ciphersuites. This could have some performance impact.
2222 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2223 default and cannot be disabled. This could have some performance impact.
2225 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2229 *) SSLv2 doesn't block disabled ciphers
2231 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2232 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2233 been disabled, provided that the SSLv2 protocol was not also disabled via
2236 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2237 and Sebastian Schinzel.
2241 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2243 *) BN_mod_exp may produce incorrect results on x86_64
2245 There is a carry propagating bug in the x86_64 Montgomery squaring
2246 procedure. No EC algorithms are affected. Analysis suggests that attacks
2247 against RSA and DSA as a result of this defect would be very difficult to
2248 perform and are not believed likely. Attacks against DH are considered just
2249 feasible (although very difficult) because most of the work necessary to
2250 deduce information about a private key may be performed offline. The amount
2251 of resources required for such an attack would be very significant and
2252 likely only accessible to a limited number of attackers. An attacker would
2253 additionally need online access to an unpatched system using the target
2254 private key in a scenario with persistent DH parameters and a private
2255 key that is shared between multiple clients. For example this can occur by
2256 default in OpenSSL DHE based SSL/TLS ciphersuites.
2258 This issue was reported to OpenSSL by Hanno Böck.
2262 *) Certificate verify crash with missing PSS parameter
2264 The signature verification routines will crash with a NULL pointer
2265 dereference if presented with an ASN.1 signature using the RSA PSS
2266 algorithm and absent mask generation function parameter. Since these
2267 routines are used to verify certificate signature algorithms this can be
2268 used to crash any certificate verification operation and exploited in a
2269 DoS attack. Any application which performs certificate verification is
2270 vulnerable including OpenSSL clients and servers which enable client
2273 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2277 *) X509_ATTRIBUTE memory leak
2279 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2280 memory. This structure is used by the PKCS#7 and CMS routines so any
2281 application which reads PKCS#7 or CMS data from untrusted sources is
2282 affected. SSL/TLS is not affected.
2284 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2289 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2290 This changes the decoding behaviour for some invalid messages,
2291 though the change is mostly in the more lenient direction, and
2292 legacy behaviour is preserved as much as possible.
2295 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2297 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2299 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2301 *) Alternate chains certificate forgery
2303 During certificate verification, OpenSSL will attempt to find an
2304 alternative certificate chain if the first attempt to build such a chain
2305 fails. An error in the implementation of this logic can mean that an
2306 attacker could cause certain checks on untrusted certificates to be
2307 bypassed, such as the CA flag, enabling them to use a valid leaf
2308 certificate to act as a CA and "issue" an invalid certificate.
2310 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2314 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2316 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2317 incompatibility in the handling of HMAC. The previous ABI has now been
2321 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2323 *) Malformed ECParameters causes infinite loop
2325 When processing an ECParameters structure OpenSSL enters an infinite loop
2326 if the curve specified is over a specially malformed binary polynomial
2329 This can be used to perform denial of service against any
2330 system which processes public keys, certificate requests or
2331 certificates. This includes TLS clients and TLS servers with
2332 client authentication enabled.
2334 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2338 *) Exploitable out-of-bounds read in X509_cmp_time
2340 X509_cmp_time does not properly check the length of the ASN1_TIME
2341 string and can read a few bytes out of bounds. In addition,
2342 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2345 An attacker can use this to craft malformed certificates and CRLs of
2346 various sizes and potentially cause a segmentation fault, resulting in
2347 a DoS on applications that verify certificates or CRLs. TLS clients
2348 that verify CRLs are affected. TLS clients and servers with client
2349 authentication enabled may be affected if they use custom verification
2352 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2353 independently by Hanno Böck.
2357 *) PKCS7 crash with missing EnvelopedContent
2359 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2360 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2361 with missing content and trigger a NULL pointer dereference on parsing.
2363 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2364 structures from untrusted sources are affected. OpenSSL clients and
2365 servers are not affected.
2367 This issue was reported to OpenSSL by Michal Zalewski (Google).
2371 *) CMS verify infinite loop with unknown hash function
2373 When verifying a signedData message the CMS code can enter an infinite loop
2374 if presented with an unknown hash function OID. This can be used to perform
2375 denial of service against any system which verifies signedData messages using
2377 This issue was reported to OpenSSL by Johannes Bauer.
2381 *) Race condition handling NewSessionTicket
2383 If a NewSessionTicket is received by a multi-threaded client when attempting to
2384 reuse a previous ticket then a race condition can occur potentially leading to
2385 a double free of the ticket data.
2389 *) Only support 256-bit or stronger elliptic curves with the
2390 'ecdh_auto' setting (server) or by default (client). Of supported
2391 curves, prefer P-256 (both).
2394 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2396 *) ClientHello sigalgs DoS fix
2398 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2399 invalid signature algorithms extension a NULL pointer dereference will
2400 occur. This can be exploited in a DoS attack against the server.
2402 This issue was was reported to OpenSSL by David Ramos of Stanford
2405 [Stephen Henson and Matt Caswell]
2407 *) Multiblock corrupted pointer fix
2409 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2410 feature only applies on 64 bit x86 architecture platforms that support AES
2411 NI instructions. A defect in the implementation of "multiblock" can cause
2412 OpenSSL's internal write buffer to become incorrectly set to NULL when
2413 using non-blocking IO. Typically, when the user application is using a
2414 socket BIO for writing, this will only result in a failed connection.
2415 However if some other BIO is used then it is likely that a segmentation
2416 fault will be triggered, thus enabling a potential DoS attack.
2418 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2422 *) Segmentation fault in DTLSv1_listen fix
2424 The DTLSv1_listen function is intended to be stateless and processes the
2425 initial ClientHello from many peers. It is common for user code to loop
2426 over the call to DTLSv1_listen until a valid ClientHello is received with
2427 an associated cookie. A defect in the implementation of DTLSv1_listen means
2428 that state is preserved in the SSL object from one invocation to the next
2429 that can lead to a segmentation fault. Errors processing the initial
2430 ClientHello can trigger this scenario. An example of such an error could be
2431 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2434 This issue was reported to OpenSSL by Per Allansson.
2438 *) Segmentation fault in ASN1_TYPE_cmp fix
2440 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2441 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2442 certificate signature algorithm consistency this can be used to crash any
2443 certificate verification operation and exploited in a DoS attack. Any
2444 application which performs certificate verification is vulnerable including
2445 OpenSSL clients and servers which enable client authentication.
2449 *) Segmentation fault for invalid PSS parameters fix
2451 The signature verification routines will crash with a NULL pointer
2452 dereference if presented with an ASN.1 signature using the RSA PSS
2453 algorithm and invalid parameters. Since these routines are used to verify
2454 certificate signature algorithms this can be used to crash any
2455 certificate verification operation and exploited in a DoS attack. Any
2456 application which performs certificate verification is vulnerable including
2457 OpenSSL clients and servers which enable client authentication.
2459 This issue was was reported to OpenSSL by Brian Carpenter.
2463 *) ASN.1 structure reuse memory corruption fix
2465 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2466 memory corruption via an invalid write. Such reuse is and has been
2467 strongly discouraged and is believed to be rare.
2469 Applications that parse structures containing CHOICE or ANY DEFINED BY
2470 components may be affected. Certificate parsing (d2i_X509 and related
2471 functions) are however not affected. OpenSSL clients and servers are
2476 *) PKCS7 NULL pointer dereferences fix
2478 The PKCS#7 parsing code does not handle missing outer ContentInfo
2479 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2480 missing content and trigger a NULL pointer dereference on parsing.
2482 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2483 otherwise parse PKCS#7 structures from untrusted sources are
2484 affected. OpenSSL clients and servers are not affected.
2486 This issue was reported to OpenSSL by Michal Zalewski (Google).
2490 *) DoS via reachable assert in SSLv2 servers fix
2492 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2493 servers that both support SSLv2 and enable export cipher suites by sending
2494 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2496 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2497 (OpenSSL development team).
2501 *) Empty CKE with client auth and DHE fix
2503 If client auth is used then a server can seg fault in the event of a DHE
2504 ciphersuite being selected and a zero length ClientKeyExchange message
2505 being sent by the client. This could be exploited in a DoS attack.
2509 *) Handshake with unseeded PRNG fix
2511 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2512 with an unseeded PRNG. The conditions are:
2513 - The client is on a platform where the PRNG has not been seeded
2514 automatically, and the user has not seeded manually
2515 - A protocol specific client method version has been used (i.e. not
2516 SSL_client_methodv23)
2517 - A ciphersuite is used that does not require additional random data from
2518 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2520 If the handshake succeeds then the client random that has been used will
2521 have been generated from a PRNG with insufficient entropy and therefore the
2522 output may be predictable.
2524 For example using the following command with an unseeded openssl will
2525 succeed on an unpatched platform:
2527 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2531 *) Use After Free following d2i_ECPrivatekey error fix
2533 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2534 could cause a use after free condition. This, in turn, could cause a double
2535 free in several private key parsing functions (such as d2i_PrivateKey
2536 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2537 for applications that receive EC private keys from untrusted
2538 sources. This scenario is considered rare.
2540 This issue was discovered by the BoringSSL project and fixed in their
2545 *) X509_to_X509_REQ NULL pointer deref fix
2547 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2548 the certificate key is invalid. This function is rarely used in practice.
2550 This issue was discovered by Brian Carpenter.
2554 *) Removed the export ciphers from the DEFAULT ciphers
2557 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2559 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2560 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2561 So far those who have to target multiple platforms would compromise
2562 and argue that binary targeting say ARMv5 would still execute on
2563 ARMv8. "Universal" build resolves this compromise by providing
2564 near-optimal performance even on newer platforms.
2567 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2568 (other platforms pending).
2569 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2571 *) Add support for the SignedCertificateTimestampList certificate and
2572 OCSP response extensions from RFC6962.
2575 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2576 for corner cases. (Certain input points at infinity could lead to
2577 bogus results, with non-infinity inputs mapped to infinity too.)
2580 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2581 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2582 common cases are optimized and there still is room for further
2583 improvements. Vector Permutation AES for Altivec is also added.
2586 *) Add support for little-endian ppc64 Linux target.
2587 [Marcelo Cerri (IBM)]
2589 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2590 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2591 are optimized and there still is room for further improvements.
2592 Both 32- and 64-bit modes are supported.
2593 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2595 *) Improved ARMv7 NEON support.
2598 *) Support for SPARC Architecture 2011 crypto extensions, first
2599 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2600 SHA256/512, MD5, GHASH and modular exponentiation.
2601 [Andy Polyakov, David Miller]
2603 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2605 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2607 *) Support for new and upcoming Intel processors, including AVX2,
2608 BMI and SHA ISA extensions. This includes additional "stitched"
2609 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2612 This work was sponsored by Intel Corp.
2615 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2616 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2617 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2620 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2621 this fixes a limitation in previous versions of OpenSSL.
2624 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2625 MGF1 digest and OAEP label.
2628 *) Add EVP support for key wrapping algorithms, to avoid problems with
2629 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2630 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2631 algorithms and include tests cases.
2634 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2636 [Douglas E. Engert, Steve Henson]
2638 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2639 difference in days and seconds between two tm or ASN1_TIME structures.
2642 *) Add -rev test option to s_server to just reverse order of characters
2643 received by client and send back to server. Also prints an abbreviated
2644 summary of the connection parameters.
2647 *) New option -brief for s_client and s_server to print out a brief summary
2648 of connection parameters.
2651 *) Add callbacks for arbitrary TLS extensions.
2652 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2654 *) New option -crl_download in several openssl utilities to download CRLs
2655 from CRLDP extension in certificates.
2658 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2661 *) New function X509_CRL_diff to generate a delta CRL from the difference
2662 of two full CRLs. Add support to "crl" utility.
2665 *) New functions to set lookup_crls function and to retrieve
2666 X509_STORE from X509_STORE_CTX.
2669 *) Print out deprecated issuer and subject unique ID fields in
2673 *) Extend OCSP I/O functions so they can be used for simple general purpose
2674 HTTP as well as OCSP. New wrapper function which can be used to download
2675 CRLs using the OCSP API.
2678 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2681 *) SSL_CONF* functions. These provide a common framework for application
2682 configuration using configuration files or command lines.
2685 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2686 message callback and prints the results. Needs compile time option
2687 "enable-ssl-trace". New options to s_client and s_server to enable
2691 *) New ctrl and macro to retrieve supported points extensions.
2692 Print out extension in s_server and s_client.
2695 *) New functions to retrieve certificate signature and signature
2699 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2703 *) New Suite B modes for TLS code. These use and enforce the requirements
2704 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2705 only use Suite B curves. The Suite B modes can be set by using the
2706 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2709 *) New chain verification flags for Suite B levels of security. Check
2710 algorithms are acceptable when flags are set in X509_verify_cert.
2713 *) Make tls1_check_chain return a set of flags indicating checks passed
2714 by a certificate chain. Add additional tests to handle client
2715 certificates: checks for matching certificate type and issuer name
2719 *) If an attempt is made to use a signature algorithm not in the peer
2720 preference list abort the handshake. If client has no suitable
2721 signature algorithms in response to a certificate request do not
2722 use the certificate.
2725 *) If server EC tmp key is not in client preference list abort handshake.
2728 *) Add support for certificate stores in CERT structure. This makes it
2729 possible to have different stores per SSL structure or one store in
2730 the parent SSL_CTX. Include distinct stores for certificate chain
2731 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2732 to build and store a certificate chain in CERT structure: returning
2733 an error if the chain cannot be built: this will allow applications
2734 to test if a chain is correctly configured.
2736 Note: if the CERT based stores are not set then the parent SSL_CTX
2737 store is used to retain compatibility with existing behaviour.
2741 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2742 mask based on the current session, check mask when sending client
2743 hello and checking the requested ciphersuite.
2746 *) New ctrls to retrieve and set certificate types in a certificate
2747 request message. Print out received values in s_client. If certificate
2748 types is not set with custom values set sensible values based on
2749 supported signature algorithms.
2752 *) Support for distinct client and server supported signature algorithms.
2755 *) Add certificate callback. If set this is called whenever a certificate
2756 is required by client or server. An application can decide which
2757 certificate chain to present based on arbitrary criteria: for example
2758 supported signature algorithms. Add very simple example to s_server.
2759 This fixes many of the problems and restrictions of the existing client
2760 certificate callback: for example you can now clear an existing
2761 certificate and specify the whole chain.
2764 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2765 the certificate can be used for (if anything). Set valid_flags field
2766 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2767 to have similar checks in it.
2769 Add new "cert_flags" field to CERT structure and include a "strict mode".
2770 This enforces some TLS certificate requirements (such as only permitting
2771 certificate signature algorithms contained in the supported algorithms
2772 extension) which some implementations ignore: this option should be used
2773 with caution as it could cause interoperability issues.
2776 *) Update and tidy signature algorithm extension processing. Work out
2777 shared signature algorithms based on preferences and peer algorithms
2778 and print them out in s_client and s_server. Abort handshake if no
2779 shared signature algorithms.
2782 *) Add new functions to allow customised supported signature algorithms
2783 for SSL and SSL_CTX structures. Add options to s_client and s_server
2787 *) New function SSL_certs_clear() to delete all references to certificates
2788 from an SSL structure. Before this once a certificate had been added
2789 it couldn't be removed.
2792 *) Integrate hostname, email address and IP address checking with certificate
2793 verification. New verify options supporting checking in openssl utility.
2796 *) Fixes and wildcard matching support to hostname and email checking
2797 functions. Add manual page.
2798 [Florian Weimer (Red Hat Product Security Team)]
2800 *) New functions to check a hostname email or IP address against a
2801 certificate. Add options x509 utility to print results of checks against
2805 *) Fix OCSP checking.
2806 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2808 *) Initial experimental support for explicitly trusted non-root CAs.
2809 OpenSSL still tries to build a complete chain to a root but if an
2810 intermediate CA has a trust setting included that is used. The first
2811 setting is used: whether to trust (e.g., -addtrust option to the x509
2815 *) Add -trusted_first option which attempts to find certificates in the
2816 trusted store even if an untrusted chain is also supplied.
2819 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2820 platform support for Linux and Android.
2823 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2826 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2827 When in FIPS mode the approved implementations are used as normal,
2828 when not in FIPS mode the internal unapproved versions are used instead.
2829 This means that the FIPS capable OpenSSL isn't forced to use the
2830 (often lower performance) FIPS implementations outside FIPS mode.
2833 *) Transparently support X9.42 DH parameters when calling
2834 PEM_read_bio_DHparameters. This means existing applications can handle
2835 the new parameter format automatically.
2838 *) Initial experimental support for X9.42 DH parameter format: mainly
2839 to support use of 'q' parameter for RFC5114 parameters.
2842 *) Add DH parameters from RFC5114 including test data to dhtest.
2845 *) Support for automatic EC temporary key parameter selection. If enabled
2846 the most preferred EC parameters are automatically used instead of
2847 hardcoded fixed parameters. Now a server just has to call:
2848 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2849 support ECDH and use the most appropriate parameters.
2852 *) Enhance and tidy EC curve and point format TLS extension code. Use
2853 static structures instead of allocation if default values are used.
2854 New ctrls to set curves we wish to support and to retrieve shared curves.
2855 Print out shared curves in s_server. New options to s_server and s_client
2856 to set list of supported curves.
2859 *) New ctrls to retrieve supported signature algorithms and
2860 supported curve values as an array of NIDs. Extend openssl utility
2861 to print out received values.
2864 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2865 between NIDs and the more common NIST names such as "P-256". Enhance
2866 ecparam utility and ECC method to recognise the NIST names for curves.
2869 *) Enhance SSL/TLS certificate chain handling to support different
2870 chains for each certificate instead of one chain in the parent SSL_CTX.
2873 *) Support for fixed DH ciphersuite client authentication: where both
2874 server and client use DH certificates with common parameters.
2877 *) Support for fixed DH ciphersuites: those requiring DH server
2881 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2883 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2884 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2885 X509_CINF_get_signature were reverted post internal team review.
2887 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2889 *) Build fixes for the Windows and OpenVMS platforms
2890 [Matt Caswell and Richard Levitte]
2892 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2894 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2895 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2896 dereference. This could lead to a Denial Of Service attack. Thanks to
2897 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2901 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2902 dtls1_buffer_record function under certain conditions. In particular this
2903 could occur if an attacker sent repeated DTLS records with the same
2904 sequence number but for the next epoch. The memory leak could be exploited
2905 by an attacker in a Denial of Service attack through memory exhaustion.
2906 Thanks to Chris Mueller for reporting this issue.
2910 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2911 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2912 method would be set to NULL which could later result in a NULL pointer
2913 dereference. Thanks to Frank Schmirler for reporting this issue.
2917 *) Abort handshake if server key exchange message is omitted for ephemeral
2920 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2921 reporting this issue.
2925 *) Remove non-export ephemeral RSA code on client and server. This code
2926 violated the TLS standard by allowing the use of temporary RSA keys in
2927 non-export ciphersuites and could be used by a server to effectively
2928 downgrade the RSA key length used to a value smaller than the server
2929 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2930 INRIA or reporting this issue.
2934 *) Fixed issue where DH client certificates are accepted without verification.
2935 An OpenSSL server will accept a DH certificate for client authentication
2936 without the certificate verify message. This effectively allows a client to
2937 authenticate without the use of a private key. This only affects servers
2938 which trust a client certificate authority which issues certificates
2939 containing DH keys: these are extremely rare and hardly ever encountered.
2940 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2945 *) Ensure that the session ID context of an SSL is updated when its
2946 SSL_CTX is updated via SSL_set_SSL_CTX.
2948 The session ID context is typically set from the parent SSL_CTX,
2949 and can vary with the CTX.
2952 *) Fix various certificate fingerprint issues.
2954 By using non-DER or invalid encodings outside the signed portion of a
2955 certificate the fingerprint can be changed without breaking the signature.
2956 Although no details of the signed portion of the certificate can be changed
2957 this can cause problems with some applications: e.g. those using the
2958 certificate fingerprint for blacklists.
2960 1. Reject signatures with non zero unused bits.
2962 If the BIT STRING containing the signature has non zero unused bits reject
2963 the signature. All current signature algorithms require zero unused bits.
2965 2. Check certificate algorithm consistency.
2967 Check the AlgorithmIdentifier inside TBS matches the one in the
2968 certificate signature. NB: this will result in signature failure
2969 errors for some broken certificates.
2971 Thanks to Konrad Kraszewski from Google for reporting this issue.
2973 3. Check DSA/ECDSA signatures use DER.
2975 Re-encode DSA/ECDSA signatures and compare with the original received
2976 signature. Return an error if there is a mismatch.
2978 This will reject various cases including garbage after signature
2979 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2980 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2981 (negative or with leading zeroes).
2983 Further analysis was conducted and fixes were developed by Stephen Henson
2984 of the OpenSSL core team.
2989 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2990 results on some platforms, including x86_64. This bug occurs at random
2991 with a very low probability, and is not known to be exploitable in any
2992 way, though its exact impact is difficult to determine. Thanks to Pieter
2993 Wuille (Blockstream) who reported this issue and also suggested an initial
2994 fix. Further analysis was conducted by the OpenSSL development team and
2995 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2996 the OpenSSL core team.
3000 *) Do not resume sessions on the server if the negotiated protocol
3001 version does not match the session's version. Resuming with a different
3002 version, while not strictly forbidden by the RFC, is of questionable
3003 sanity and breaks all known clients.
3004 [David Benjamin, Emilia Käsper]
3006 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3007 early CCS messages during renegotiation. (Note that because
3008 renegotiation is encrypted, this early CCS was not exploitable.)
3011 *) Tighten client-side session ticket handling during renegotiation:
3012 ensure that the client only accepts a session ticket if the server sends
3013 the extension anew in the ServerHello. Previously, a TLS client would
3014 reuse the old extension state and thus accept a session ticket if one was
3015 announced in the initial ServerHello.
3017 Similarly, ensure that the client requires a session ticket if one
3018 was advertised in the ServerHello. Previously, a TLS client would
3019 ignore a missing NewSessionTicket message.
3022 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3024 *) SRTP Memory Leak.
3026 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3027 sends a carefully crafted handshake message, to cause OpenSSL to fail
3028 to free up to 64k of memory causing a memory leak. This could be
3029 exploited in a Denial Of Service attack. This issue affects OpenSSL
3030 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3031 whether SRTP is used or configured. Implementations of OpenSSL that
3032 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3034 The fix was developed by the OpenSSL team.
3038 *) Session Ticket Memory Leak.
3040 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3041 integrity of that ticket is first verified. In the event of a session
3042 ticket integrity check failing, OpenSSL will fail to free memory
3043 causing a memory leak. By sending a large number of invalid session
3044 tickets an attacker could exploit this issue in a Denial Of Service
3049 *) Build option no-ssl3 is incomplete.
3051 When OpenSSL is configured with "no-ssl3" as a build option, servers
3052 could accept and complete a SSL 3.0 handshake, and clients could be
3053 configured to send them.
3055 [Akamai and the OpenSSL team]
3057 *) Add support for TLS_FALLBACK_SCSV.
3058 Client applications doing fallback retries should call
3059 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3061 [Adam Langley, Bodo Moeller]
3063 *) Add additional DigestInfo checks.
3065 Re-encode DigestInto in DER and check against the original when
3066 verifying RSA signature: this will reject any improperly encoded
3067 DigestInfo structures.
3069 Note: this is a precautionary measure and no attacks are currently known.
3073 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3075 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3076 SRP code can be overrun an internal buffer. Add sanity check that
3077 g, A, B < N to SRP code.
3079 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3080 Group for discovering this issue.
3084 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3085 TLS 1.0 instead of higher protocol versions when the ClientHello message
3086 is badly fragmented. This allows a man-in-the-middle attacker to force a
3087 downgrade to TLS 1.0 even if both the server and the client support a
3088 higher protocol version, by modifying the client's TLS records.
3090 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3091 researching this issue.
3095 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3096 to a denial of service attack. A malicious server can crash the client
3097 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3098 ciphersuite and sending carefully crafted handshake messages.
3100 Thanks to Felix Gröbert (Google) for discovering and researching this
3105 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3106 to leak memory. This can be exploited through a Denial of Service attack.
3107 Thanks to Adam Langley for discovering and researching this issue.
3111 *) An attacker can force openssl to consume large amounts of memory whilst
3112 processing DTLS handshake messages. This can be exploited through a
3113 Denial of Service attack.
3114 Thanks to Adam Langley for discovering and researching this issue.
3118 *) An attacker can force an error condition which causes openssl to crash
3119 whilst processing DTLS packets due to memory being freed twice. This
3120 can be exploited through a Denial of Service attack.
3121 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3126 *) If a multithreaded client connects to a malicious server using a resumed
3127 session and the server sends an ec point format extension it could write
3128 up to 255 bytes to freed memory.
3130 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3135 *) A malicious server can crash an OpenSSL client with a null pointer
3136 dereference (read) by specifying an SRP ciphersuite even though it was not
3137 properly negotiated with the client. This can be exploited through a
3138 Denial of Service attack.
3140 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3141 discovering and researching this issue.
3145 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3146 X509_name_oneline, X509_name_print_ex et al. to leak some information
3147 from the stack. Applications may be affected if they echo pretty printing
3148 output to the attacker.
3150 Thanks to Ivan Fratric (Google) for discovering this issue.
3152 [Emilia Käsper, and Steve Henson]
3154 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3155 for corner cases. (Certain input points at infinity could lead to
3156 bogus results, with non-infinity inputs mapped to infinity too.)
3159 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3161 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3162 handshake can force the use of weak keying material in OpenSSL
3163 SSL/TLS clients and servers.
3165 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3166 researching this issue. (CVE-2014-0224)
3167 [KIKUCHI Masashi, Steve Henson]
3169 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3170 OpenSSL DTLS client the code can be made to recurse eventually crashing
3173 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3175 [Imre Rad, Steve Henson]
3177 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3178 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3179 client or server. This is potentially exploitable to run arbitrary
3180 code on a vulnerable client or server.
3182 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3183 [Jüri Aedla, Steve Henson]
3185 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3186 are subject to a denial of service attack.
3188 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3189 this issue. (CVE-2014-3470)
3190 [Felix Gröbert, Ivan Fratric, Steve Henson]
3192 *) Harmonize version and its documentation. -f flag is used to display
3194 [mancha <mancha1@zoho.com>]
3196 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3197 in i2d_ECPrivateKey.
3198 [mancha <mancha1@zoho.com>]
3200 *) Fix some double frees. These are not thought to be exploitable.
3201 [mancha <mancha1@zoho.com>]
3203 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3205 *) A missing bounds check in the handling of the TLS heartbeat extension
3206 can be used to reveal up to 64k of memory to a connected client or
3209 Thanks for Neel Mehta of Google Security for discovering this bug and to
3210 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3211 preparing the fix (CVE-2014-0160)
3212 [Adam Langley, Bodo Moeller]
3214 *) Fix for the attack described in the paper "Recovering OpenSSL
3215 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3216 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3217 http://eprint.iacr.org/2014/140
3219 Thanks to Yuval Yarom and Naomi Benger for discovering this
3220 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3221 [Yuval Yarom and Naomi Benger]
3223 *) TLS pad extension: draft-agl-tls-padding-03
3225 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3226 TLS client Hello record length value would otherwise be > 255 and
3227 less that 512 pad with a dummy extension containing zeroes so it
3228 is at least 512 bytes long.
3230 [Adam Langley, Steve Henson]
3232 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3234 *) Fix for TLS record tampering bug. A carefully crafted invalid
3235 handshake could crash OpenSSL with a NULL pointer exception.
3236 Thanks to Anton Johansson for reporting this issues.
3239 *) Keep original DTLS digest and encryption contexts in retransmission
3240 structures so we can use the previous session parameters if they need
3241 to be resent. (CVE-2013-6450)
3244 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3245 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3246 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3247 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3248 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3249 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3250 [Rob Stradling, Adam Langley]
3252 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3254 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3255 supporting platforms or when small records were transferred.
3256 [Andy Polyakov, Steve Henson]
3258 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3260 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3262 This addresses the flaw in CBC record processing discovered by
3263 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3264 at: http://www.isg.rhul.ac.uk/tls/
3266 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3267 Security Group at Royal Holloway, University of London
3268 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3269 Emilia Käsper for the initial patch.
3271 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3273 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3274 ciphersuites which can be exploited in a denial of service attack.
3275 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3276 and detecting this bug and to Wolfgang Ettlinger
3277 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3281 *) Return an error when checking OCSP signatures when key is NULL.
3282 This fixes a DoS attack. (CVE-2013-0166)
3285 *) Make openssl verify return errors.
3286 [Chris Palmer <palmer@google.com> and Ben Laurie]
3288 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3289 the right response is stapled. Also change SSL_get_certificate()
3290 so it returns the certificate actually sent.
3291 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3292 [Rob Stradling <rob.stradling@comodo.com>]
3294 *) Fix possible deadlock when decoding public keys.
3297 *) Don't use TLS 1.0 record version number in initial client hello
3301 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3303 *) Sanity check record length before skipping explicit IV in TLS
3304 1.2, 1.1 and DTLS to fix DoS attack.
3306 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3307 fuzzing as a service testing platform.
3311 *) Initialise tkeylen properly when encrypting CMS messages.
3312 Thanks to Solar Designer of Openwall for reporting this issue.
3315 *) In FIPS mode don't try to use composite ciphers as they are not
3319 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3321 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3322 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3323 mean any application compiled against OpenSSL 1.0.0 headers setting
3324 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3325 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3326 0x10000000L Any application which was previously compiled against
3327 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3328 will need to be recompiled as a result. Letting be results in
3329 inability to disable specifically TLS 1.1 and in client context,
3330 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3333 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3334 disable just protocol X, but all protocols above X *if* there are
3335 protocols *below* X still enabled. In more practical terms it means
3336 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3337 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3338 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3342 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3344 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3345 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3346 in CRYPTO_realloc_clean.
3348 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3349 issue and to Adam Langley <agl@chromium.org> for fixing it.
3351 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3353 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3356 *) Workarounds for some broken servers that "hang" if a client hello
3357 record length exceeds 255 bytes.
3359 1. Do not use record version number > TLS 1.0 in initial client
3360 hello: some (but not all) hanging servers will now work.
3361 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3362 the number of ciphers sent in the client hello. This should be
3363 set to an even number, such as 50, for example by passing:
3364 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3365 Most broken servers should now work.
3366 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3367 TLS 1.2 client support entirely.
3370 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3373 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3375 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3376 STRING form instead of a DigestInfo.
3379 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3380 and the RSA_sign/RSA_verify functions. This was made more apparent when
3381 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3382 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3383 the correct format in RSA_verify so both forms transparently work.
3386 *) Some servers which support TLS 1.0 can choke if we initially indicate
3387 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3388 encrypted premaster secret. As a workaround use the maximum permitted
3389 client version in client hello, this should keep such servers happy
3390 and still work with previous versions of OpenSSL.
3393 *) Add support for TLS/DTLS heartbeats.
3394 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3396 *) Add support for SCTP.
3397 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3399 *) Improved PRNG seeding for VOS.
3400 [Paul Green <Paul.Green@stratus.com>]
3402 *) Extensive assembler packs updates, most notably:
3404 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3405 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3406 - x86_64: bit-sliced AES implementation;
3407 - ARM: NEON support, contemporary platforms optimizations;
3408 - s390x: z196 support;
3409 - *: GHASH and GF(2^m) multiplication implementations;
3413 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3414 (removal of unnecessary code)
3415 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3417 *) Add TLS key material exporter from RFC 5705.
3420 *) Add DTLS-SRTP negotiation from RFC 5764.
3423 *) Add Next Protocol Negotiation,
3424 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3425 disabled with a no-npn flag to config or Configure. Code donated
3427 [Adam Langley <agl@google.com> and Ben Laurie]
3429 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3430 NIST-P256, NIST-P521, with constant-time single point multiplication on
3431 typical inputs. Compiler support for the nonstandard type __uint128_t is
3432 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3433 Code made available under Apache License version 2.0.
3435 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3436 line to include this in your build of OpenSSL, and run "make depend" (or
3437 "make update"). This enables the following EC_METHODs:
3439 EC_GFp_nistp224_method()
3440 EC_GFp_nistp256_method()
3441 EC_GFp_nistp521_method()
3443 EC_GROUP_new_by_curve_name() will automatically use these (while
3444 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3446 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3448 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3449 all platforms. Move ssize_t definition from e_os.h to the public
3450 header file e_os2.h as it now appears in public header file cms.h
3453 *) New -sigopt option to the ca, req and x509 utilities. Additional
3454 signature parameters can be passed using this option and in
3458 *) Add RSA PSS signing function. This will generate and set the
3459 appropriate AlgorithmIdentifiers for PSS based on those in the
3460 corresponding EVP_MD_CTX structure. No application support yet.
3463 *) Support for companion algorithm specific ASN1 signing routines.
3464 New function ASN1_item_sign_ctx() signs a pre-initialised
3465 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3466 the appropriate parameters.
3469 *) Add new algorithm specific ASN1 verification initialisation function
3470 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3471 handling will be the same no matter what EVP_PKEY_METHOD is used.
3472 Add a PSS handler to support verification of PSS signatures: checked
3473 against a number of sample certificates.
3476 *) Add signature printing for PSS. Add PSS OIDs.
3477 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3479 *) Add algorithm specific signature printing. An individual ASN1 method
3480 can now print out signatures instead of the standard hex dump.
3482 More complex signatures (e.g. PSS) can print out more meaningful
3483 information. Include DSA version that prints out the signature
3487 *) Password based recipient info support for CMS library: implementing
3491 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3492 neatly separates the code into cipher and PBE sections and is required
3493 for some algorithms that split PBES2 into separate pieces (such as
3494 password based CMS).
3497 *) Session-handling fixes:
3498 - Fix handling of connections that are resuming with a session ID,
3499 but also support Session Tickets.
3500 - Fix a bug that suppressed issuing of a new ticket if the client
3501 presented a ticket with an expired session.
3502 - Try to set the ticket lifetime hint to something reasonable.
3503 - Make tickets shorter by excluding irrelevant information.
3504 - On the client side, don't ignore renewed tickets.
3505 [Adam Langley, Bodo Moeller (Google)]
3507 *) Fix PSK session representation.
3510 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3512 This work was sponsored by Intel.
3515 *) Add GCM support to TLS library. Some custom code is needed to split
3516 the IV between the fixed (from PRF) and explicit (from TLS record)
3517 portions. This adds all GCM ciphersuites supported by RFC5288 and
3518 RFC5289. Generalise some AES* cipherstrings to include GCM and
3519 add a special AESGCM string for GCM only.
3522 *) Expand range of ctrls for AES GCM. Permit setting invocation
3523 field on decrypt and retrieval of invocation field only on encrypt.
3526 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3527 As required by RFC5289 these ciphersuites cannot be used if for
3528 versions of TLS earlier than 1.2.
3531 *) For FIPS capable OpenSSL interpret a NULL default public key method
3532 as unset and return the appropriate default but do *not* set the default.
3533 This means we can return the appropriate method in applications that
3534 switch between FIPS and non-FIPS modes.
3537 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3538 ENGINE is used then we cannot handle that in the FIPS module so we
3539 keep original code iff non-FIPS operations are allowed.
3542 *) Add -attime option to openssl utilities.
3543 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3545 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3548 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3549 FIPS EC methods unconditionally for now.
3552 *) New build option no-ec2m to disable characteristic 2 code.
3555 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3556 all cases can be covered as some introduce binary incompatibilities.
3559 *) Redirect RSA operations to FIPS module including keygen,
3560 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3563 *) Add similar low level API blocking to ciphers.
3566 *) Low level digest APIs are not approved in FIPS mode: any attempt
3567 to use these will cause a fatal error. Applications that *really* want
3568 to use them can use the private_* version instead.
3571 *) Redirect cipher operations to FIPS module for FIPS builds.
3574 *) Redirect digest operations to FIPS module for FIPS builds.
3577 *) Update build system to add "fips" flag which will link in fipscanister.o
3578 for static and shared library builds embedding a signature if needed.
3581 *) Output TLS supported curves in preference order instead of numerical
3582 order. This is currently hardcoded for the highest order curves first.
3583 This should be configurable so applications can judge speed vs strength.
3586 *) Add TLS v1.2 server support for client authentication.
3589 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3593 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3594 FIPS modules versions.
3597 *) Add TLS v1.2 client side support for client authentication. Keep cache
3598 of handshake records longer as we don't know the hash algorithm to use
3599 until after the certificate request message is received.
3602 *) Initial TLS v1.2 client support. Add a default signature algorithms
3603 extension including all the algorithms we support. Parse new signature
3604 format in client key exchange. Relax some ECC signing restrictions for
3605 TLS v1.2 as indicated in RFC5246.
3608 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3609 to new signature format when needed using client digest preference.
3610 All server ciphersuites should now work correctly in TLS v1.2. No client
3611 support yet and no support for client certificates.
3614 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3615 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3616 ciphersuites. At present only RSA key exchange ciphersuites work with
3617 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3618 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3619 and version checking.
3622 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3623 with this defined it will not be affected by any changes to ssl internal
3624 structures. Add several utility functions to allow openssl application
3625 to work with OPENSSL_NO_SSL_INTERN defined.
3628 *) A long standing patch to add support for SRP from EdelWeb (Peter
3629 Sylvester and Christophe Renou) was integrated.
3630 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3631 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3634 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3637 *) Permit abbreviated handshakes when renegotiating using the function
3638 SSL_renegotiate_abbreviated().
3639 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3641 *) Add call to ENGINE_register_all_complete() to
3642 ENGINE_load_builtin_engines(), so some implementations get used
3643 automatically instead of needing explicit application support.
3646 *) Add support for TLS key exporter as described in RFC5705.
3647 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3649 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3650 a few changes are required:
3652 Add SSL_OP_NO_TLSv1_1 flag.
3653 Add TLSv1_1 methods.
3654 Update version checking logic to handle version 1.1.
3655 Add explicit IV handling (ported from DTLS code).
3656 Add command line options to s_client/s_server.
3659 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3661 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3662 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3663 content decryption and always return the same error. Note: this attack
3664 needs on average 2^20 messages so it only affects automated senders. The
3665 old behaviour can be re-enabled in the CMS code by setting the
3666 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3667 an MMA defence is not necessary.
3668 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3669 this issue. (CVE-2012-0884)
3672 *) Fix CVE-2011-4619: make sure we really are receiving a
3673 client hello before rejecting multiple SGC restarts. Thanks to
3674 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3677 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3679 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3680 Thanks to Antonio Martin, Enterprise Secure Access Research and
3681 Development, Cisco Systems, Inc. for discovering this bug and
3682 preparing a fix. (CVE-2012-0050)
3685 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3687 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3688 of the Vaudenay padding oracle attack on CBC mode encryption
3689 which enables an efficient plaintext recovery attack against
3690 the OpenSSL implementation of DTLS. Their attack exploits timing
3691 differences arising during decryption processing. A research
3692 paper describing this attack can be found at:
3693 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3694 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3695 Security Group at Royal Holloway, University of London
3696 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3697 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3698 for preparing the fix. (CVE-2011-4108)
3699 [Robin Seggelmann, Michael Tuexen]
3701 *) Clear bytes used for block padding of SSL 3.0 records.
3703 [Adam Langley (Google)]
3705 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3706 Kadianakis <desnacked@gmail.com> for discovering this issue and
3707 Adam Langley for preparing the fix. (CVE-2011-4619)
3708 [Adam Langley (Google)]
3710 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3711 [Andrey Kulikov <amdeich@gmail.com>]
3713 *) Prevent malformed RFC3779 data triggering an assertion failure.
3714 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3715 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3716 [Rob Austein <sra@hactrn.net>]
3718 *) Improved PRNG seeding for VOS.
3719 [Paul Green <Paul.Green@stratus.com>]
3721 *) Fix ssl_ciph.c set-up race.
3722 [Adam Langley (Google)]
3724 *) Fix spurious failures in ecdsatest.c.
3725 [Emilia Käsper (Google)]
3727 *) Fix the BIO_f_buffer() implementation (which was mixing different
3728 interpretations of the '..._len' fields).
3729 [Adam Langley (Google)]
3731 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3732 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3733 threads won't reuse the same blinding coefficients.
3735 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3736 lock to call BN_BLINDING_invert_ex, and avoids one use of
3737 BN_BLINDING_update for each BN_BLINDING structure (previously,
3738 the last update always remained unused).
3739 [Emilia Käsper (Google)]
3741 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3742 [Bob Buckholz (Google)]
3744 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3746 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3747 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3748 [Kaspar Brand <ossl@velox.ch>]
3750 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3751 for multi-threaded use of ECDH. (CVE-2011-3210)
3752 [Adam Langley (Google)]
3754 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3757 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3758 signature public key algorithm by using OID xref utilities instead.
3759 Before this you could only use some ECC ciphersuites with SHA1 only.
3762 *) Add protection against ECDSA timing attacks as mentioned in the paper
3763 by Billy Bob Brumley and Nicola Tuveri, see:
3765 http://eprint.iacr.org/2011/232.pdf
3767 [Billy Bob Brumley and Nicola Tuveri]
3769 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3771 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3772 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3774 *) Fix bug in string printing code: if *any* escaping is enabled we must
3775 escape the escape character (backslash) or the resulting string is
3779 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3781 *) Disable code workaround for ancient and obsolete Netscape browsers
3782 and servers: an attacker can use it in a ciphersuite downgrade attack.
3783 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3786 *) Fixed J-PAKE implementation error, originally discovered by
3787 Sebastien Martini, further info and confirmation from Stefan
3788 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3791 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3793 *) Fix extension code to avoid race conditions which can result in a buffer
3794 overrun vulnerability: resumed sessions must not be modified as they can
3795 be shared by multiple threads. CVE-2010-3864
3798 *) Fix WIN32 build system to correctly link an ENGINE directory into
3802 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3804 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3806 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3808 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3810 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3811 context. The operation can be customised via the ctrl mechanism in
3812 case ENGINEs want to include additional functionality.
3815 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3818 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3819 output hashes compatible with older versions of OpenSSL.
3820 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3822 *) Fix compression algorithm handling: if resuming a session use the
3823 compression algorithm of the resumed session instead of determining
3824 it from client hello again. Don't allow server to change algorithm.
3827 *) Add load_crls() function to apps tidying load_certs() too. Add option
3828 to verify utility to allow additional CRLs to be included.
3831 *) Update OCSP request code to permit adding custom headers to the request:
3832 some responders need this.
3835 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3837 [Julia Lawall <julia@diku.dk>]
3839 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3840 needlessly dereferenced structures, used obsolete functions and
3841 didn't handle all updated verify codes correctly.
3844 *) Disable MD2 in the default configuration.
3847 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3848 indicate the initial BIO being pushed or popped. This makes it possible
3849 to determine whether the BIO is the one explicitly called or as a result
3850 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3851 it handles reference counts correctly and doesn't zero out the I/O bio
3852 when it is not being explicitly popped. WARNING: applications which
3853 included workarounds for the old buggy behaviour will need to be modified
3854 or they could free up already freed BIOs.
3857 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3858 renaming to all platforms (within the 0.9.8 branch, this was
3859 done conditionally on Netware platforms to avoid a name clash).
3860 [Guenter <lists@gknw.net>]
3862 *) Add ECDHE and PSK support to DTLS.
3863 [Michael Tuexen <tuexen@fh-muenster.de>]
3865 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3869 *) Add "missing" function EVP_MD_flags() (without this the only way to
3870 retrieve a digest flags is by accessing the structure directly. Update
3871 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3872 or cipher is registered as in the "from" argument. Print out all
3873 registered digests in the dgst usage message instead of manually
3874 attempting to work them out.
3877 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3878 this allows the use of compression and extensions. Change default cipher
3879 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3880 by default unless an application cipher string requests it.
3883 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3884 key ids to find matching certificates and keys but some PKCS#12 files
3885 don't follow the (somewhat unwritten) rules and this strategy fails.
3886 Now just gather all certificates together and the first private key
3887 then look for the first certificate that matches the key.
3890 *) Support use of registered digest and cipher names for dgst and cipher
3891 commands instead of having to add each one as a special case. So now
3898 openssl dgst -sha256 foo
3900 and this works for ENGINE based algorithms too.
3904 *) Update Gost ENGINE to support parameter files.
3905 [Victor B. Wagner <vitus@cryptocom.ru>]
3907 *) Support GeneralizedTime in ca utility.
3908 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3910 *) Enhance the hash format used for certificate directory links. The new
3911 form uses the canonical encoding (meaning equivalent names will work
3912 even if they aren't identical) and uses SHA1 instead of MD5. This form
3913 is incompatible with the older format and as a result c_rehash should
3914 be used to rebuild symbolic links.
3917 *) Make PKCS#8 the default write format for private keys, replacing the
3918 traditional format. This form is standardised, more secure and doesn't
3919 include an implicit MD5 dependency.
3922 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3923 committed to OpenSSL should pass this lot as a minimum.
3926 *) Add session ticket override functionality for use by EAP-FAST.
3927 [Jouni Malinen <j@w1.fi>]
3929 *) Modify HMAC functions to return a value. Since these can be implemented
3930 in an ENGINE errors can occur.
3933 *) Type-checked OBJ_bsearch_ex.
3936 *) Type-checked OBJ_bsearch. Also some constification necessitated
3937 by type-checking. Still to come: TXT_DB, bsearch(?),
3938 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3942 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3943 seconds to a tm structure directly, instead of going through OS
3944 specific date routines. This avoids any issues with OS routines such
3945 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3946 and X509_time_adj_ex() to cover the extended range. The existing
3947 X509_time_adj() is still usable and will no longer have any date issues.
3950 *) Delta CRL support. New use deltas option which will attempt to locate
3951 and search any appropriate delta CRLs available.
3953 This work was sponsored by Google.
3956 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3957 code and add additional score elements. Validate alternate CRL paths
3958 as part of the CRL checking and indicate a new error "CRL path validation
3959 error" in this case. Applications wanting additional details can use
3960 the verify callback and check the new "parent" field. If this is not
3961 NULL CRL path validation is taking place. Existing applications won't
3962 see this because it requires extended CRL support which is off by
3965 This work was sponsored by Google.
3968 *) Support for freshest CRL extension.
3970 This work was sponsored by Google.
3973 *) Initial indirect CRL support. Currently only supported in the CRLs
3974 passed directly and not via lookup. Process certificate issuer
3975 CRL entry extension and lookup CRL entries by bother issuer name
3976 and serial number. Check and process CRL issuer entry in IDP extension.
3978 This work was sponsored by Google.
3981 *) Add support for distinct certificate and CRL paths. The CRL issuer
3982 certificate is validated separately in this case. Only enabled if
3983 an extended CRL support flag is set: this flag will enable additional
3984 CRL functionality in future.
3986 This work was sponsored by Google.
3989 *) Add support for policy mappings extension.
3991 This work was sponsored by Google.
3994 *) Fixes to pathlength constraint, self issued certificate handling,
3995 policy processing to align with RFC3280 and PKITS tests.
3997 This work was sponsored by Google.
4000 *) Support for name constraints certificate extension. DN, email, DNS
4001 and URI types are currently supported.
4003 This work was sponsored by Google.
4006 *) To cater for systems that provide a pointer-based thread ID rather
4007 than numeric, deprecate the current numeric thread ID mechanism and
4008 replace it with a structure and associated callback type. This
4009 mechanism allows a numeric "hash" to be extracted from a thread ID in
4010 either case, and on platforms where pointers are larger than 'long',
4011 mixing is done to help ensure the numeric 'hash' is usable even if it
4012 can't be guaranteed unique. The default mechanism is to use "&errno"
4013 as a pointer-based thread ID to distinguish between threads.
4015 Applications that want to provide their own thread IDs should now use
4016 CRYPTO_THREADID_set_callback() to register a callback that will call
4017 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4019 Note that ERR_remove_state() is now deprecated, because it is tied
4020 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4021 to free the current thread's error state should be replaced by
4022 ERR_remove_thread_state(NULL).
4024 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4025 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4026 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4027 application was previously providing a numeric thread callback that
4028 was inappropriate for distinguishing threads, then uniqueness might
4029 have been obtained with &errno that happened immediately in the
4030 intermediate development versions of OpenSSL; this is no longer the
4031 case, the numeric thread callback will now override the automatic use
4033 [Geoff Thorpe, with help from Bodo Moeller]
4035 *) Initial support for different CRL issuing certificates. This covers a
4036 simple case where the self issued certificates in the chain exist and
4037 the real CRL issuer is higher in the existing chain.
4039 This work was sponsored by Google.
4042 *) Removed effectively defunct crypto/store from the build.
4045 *) Revamp of STACK to provide stronger type-checking. Still to come:
4046 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4047 ASN1_STRING, CONF_VALUE.
4050 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4051 RAM on SSL connections. This option can save about 34k per idle SSL.
4054 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4055 STACK, TXT_DB, bsearch, qsort.
4058 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4059 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4060 support for data, signedData, compressedData, digestedData and
4061 encryptedData, envelopedData types included. Scripts to check against
4062 RFC4134 examples draft and interop and consistency checks of many
4063 content types and variants.
4066 *) Add options to enc utility to support use of zlib compression BIO.
4069 *) Extend mk1mf to support importing of options and assembly language
4070 files from Configure script, currently only included in VC-WIN32.
4071 The assembly language rules can now optionally generate the source
4072 files from the associated perl scripts.
4075 *) Implement remaining functionality needed to support GOST ciphersuites.
4076 Interop testing has been performed using CryptoPro implementations.
4077 [Victor B. Wagner <vitus@cryptocom.ru>]
4079 *) s390x assembler pack.
4082 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4086 *) Implement Opaque PRF Input TLS extension as specified in
4087 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4088 official specification yet and no extension type assignment by
4089 IANA exists, this extension (for now) will have to be explicitly
4090 enabled when building OpenSSL by providing the extension number
4091 to use. For example, specify an option
4093 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4095 to the "config" or "Configure" script to enable the extension,
4096 assuming extension number 0x9527 (which is a completely arbitrary
4097 and unofficial assignment based on the MD5 hash of the Internet
4098 Draft). Note that by doing so, you potentially lose
4099 interoperability with other TLS implementations since these might
4100 be using the same extension number for other purposes.
4102 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4103 opaque PRF input value to use in the handshake. This will create
4104 an internal copy of the length-'len' string at 'src', and will
4105 return non-zero for success.
4107 To get more control and flexibility, provide a callback function
4110 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4111 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4115 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4118 Callback function 'cb' will be called in handshakes, and is
4119 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4120 Argument 'arg' is for application purposes (the value as given to
4121 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4122 be provided to the callback function). The callback function
4123 has to return non-zero to report success: usually 1 to use opaque
4124 PRF input just if possible, or 2 to enforce use of the opaque PRF
4125 input. In the latter case, the library will abort the handshake
4126 if opaque PRF input is not successfully negotiated.
4128 Arguments 'peerinput' and 'len' given to the callback function
4129 will always be NULL and 0 in the case of a client. A server will
4130 see the client's opaque PRF input through these variables if
4131 available (NULL and 0 otherwise). Note that if the server
4132 provides an opaque PRF input, the length must be the same as the
4133 length of the client's opaque PRF input.
4135 Note that the callback function will only be called when creating
4136 a new session (session resumption can resume whatever was
4137 previously negotiated), and will not be called in SSL 2.0
4138 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4139 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4140 for applications that need to enforce opaque PRF input.
4144 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4147 [Victor B. Wagner <vitus@cryptocom.ru>]
4149 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4150 RFC4507bis. The encrypted ticket format is an encrypted encoded
4151 SSL_SESSION structure, that way new session features are automatically
4154 If a client application caches session in an SSL_SESSION structure
4155 support is transparent because tickets are now stored in the encoded
4158 The SSL_CTX structure automatically generates keys for ticket
4159 protection in servers so again support should be possible
4160 with no application modification.
4162 If a client or server wishes to disable RFC4507 support then the option
4163 SSL_OP_NO_TICKET can be set.
4165 Add a TLS extension debugging callback to allow the contents of any client
4166 or server extensions to be examined.
4168 This work was sponsored by Google.
4171 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4172 OpenSSL should now compile cleanly on gcc 4.2
4173 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4175 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4176 support including streaming MAC support: this is required for GOST
4177 ciphersuite support.
4178 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4180 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4181 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4182 to output in BER and PEM format.
4185 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4186 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4187 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4188 ENGINE support for HMAC keys which are unextractable. New -mac and
4189 -macopt options to dgst utility.
4192 *) New option -sigopt to dgst utility. Update dgst to use
4193 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4194 alternative signing parameters such as X9.31 or PSS in the dgst
4198 *) Change ssl_cipher_apply_rule(), the internal function that does
4199 the work each time a ciphersuite string requests enabling
4200 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4201 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4202 the order of disabled ciphersuites such that those ciphersuites
4203 that most recently went from enabled to disabled not only stay
4204 in order with respect to each other, but also have higher priority
4205 than other disabled ciphersuites the next time ciphersuites are
4208 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4209 the same ciphersuites as with "HIGH" alone, but in a specific
4210 order where the PSK ciphersuites come first (since they are the
4211 most recently disabled ciphersuites when "HIGH" is parsed).
4213 Also, change ssl_create_cipher_list() (using this new
4214 functionality) such that between otherwise identical
4215 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4219 *) Change ssl_create_cipher_list() so that it automatically
4220 arranges the ciphersuites in reasonable order before starting
4221 to process the rule string. Thus, the definition for "DEFAULT"
4222 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4223 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4224 This makes it much easier to arrive at a reasonable default order
4225 in applications for which anonymous ciphers are OK (meaning
4226 that you can't actually use DEFAULT).
4227 [Bodo Moeller; suggested by Victor Duchovni]
4229 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4230 processing) into multiple integers instead of setting
4231 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4232 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4233 (These masks as well as the individual bit definitions are hidden
4234 away into the non-exported interface ssl/ssl_locl.h, so this
4235 change to the definition of the SSL_CIPHER structure shouldn't
4236 affect applications.) This give us more bits for each of these
4237 categories, so there is no longer a need to coagulate AES128 and
4238 AES256 into a single algorithm bit, and to coagulate Camellia128
4239 and Camellia256 into a single algorithm bit, which has led to all
4242 Thus, among other things, the kludge introduced in 0.9.7m and
4243 0.9.8e for masking out AES256 independently of AES128 or masking
4244 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4246 With the change, we also introduce new ciphersuite aliases that
4247 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4251 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4252 Use the leftmost N bytes of the signature input if the input is
4253 larger than the prime q (with N being the size in bytes of q).
4256 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4257 it yet and it is largely untested.
4260 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4263 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4264 some compilers (gcc 4.2 and later) reject their use. Safestack is
4265 reimplemented. Update ASN1 to avoid use of legacy functions.
4268 *) Win32/64 targets are linked with Winsock2.
4271 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4272 to external functions. This can be used to increase CRL handling
4273 efficiency especially when CRLs are very large by (for example) storing
4274 the CRL revoked certificates in a database.
4277 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4278 new CRLs added to a directory can be used. New command line option
4279 -verify_return_error to s_client and s_server. This causes real errors
4280 to be returned by the verify callback instead of carrying on no matter
4281 what. This reflects the way a "real world" verify callback would behave.
4284 *) GOST engine, supporting several GOST algorithms and public key formats.
4285 Kindly donated by Cryptocom.
4288 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4289 partitioned by DP are handled but no indirect CRL or reason partitioning
4290 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4291 selected via a scoring technique which handles IDP and AKID in CRLs.
4294 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4295 will ultimately be used for all verify operations: this will remove the
4296 X509_STORE dependency on certificate verification and allow alternative
4297 lookup methods. X509_STORE based implementations of these two callbacks.
4300 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4301 Modify get_crl() to find a valid (unexpired) CRL if possible.
4304 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4305 this would be called X509_CRL_cmp() but that name is already used by
4306 a function that just compares CRL issuer names. Cache several CRL
4307 extensions in X509_CRL structure and cache CRLDP in X509.
4310 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4311 this maps equivalent X509_NAME structures into a consistent structure.
4312 Name comparison can then be performed rapidly using memcmp().
4315 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4319 *) Allow digests to supply their own micalg string for S/MIME type using
4320 the ctrl EVP_MD_CTRL_MICALG.
4323 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4324 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4325 ctrl. It can then customise the structure before and/or after signing
4329 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4330 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4331 to free up any added signature OIDs.
4334 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4335 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4336 digest and cipher tables. New options added to openssl utility:
4337 list-message-digest-algorithms and list-cipher-algorithms.
4340 *) Change the array representation of binary polynomials: the list
4341 of degrees of non-zero coefficients is now terminated with -1.
4342 Previously it was terminated with 0, which was also part of the
4343 value; thus, the array representation was not applicable to
4344 polynomials where t^0 has coefficient zero. This change makes
4345 the array representation useful in a more general context.
4348 *) Various modifications and fixes to SSL/TLS cipher string
4349 handling. For ECC, the code now distinguishes between fixed ECDH
4350 with RSA certificates on the one hand and with ECDSA certificates
4351 on the other hand, since these are separate ciphersuites. The
4352 unused code for Fortezza ciphersuites has been removed.
4354 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4355 (not "ECDHE"). For consistency with the code for DH
4356 certificates, use of ECDH certificates is now considered ECDH
4357 authentication, not RSA or ECDSA authentication (the latter is
4358 merely the CA's signing algorithm and not actively used in the
4361 The temporary ciphersuite alias "ECCdraft" is no longer
4362 available, and ECC ciphersuites are no longer excluded from "ALL"
4363 and "DEFAULT". The following aliases now exist for RFC 4492
4364 ciphersuites, most of these by analogy with the DH case:
4366 kECDHr - ECDH cert, signed with RSA
4367 kECDHe - ECDH cert, signed with ECDSA
4368 kECDH - ECDH cert (signed with either RSA or ECDSA)
4369 kEECDH - ephemeral ECDH
4370 ECDH - ECDH cert or ephemeral ECDH
4376 AECDH - anonymous ECDH
4377 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4381 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4382 Use correct micalg parameters depending on digest(s) in signed message.
4385 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4386 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4389 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4390 an engine to register a method. Add ENGINE lookups for methods and
4391 functional reference processing.
4394 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4395 EVP_{Sign,Verify}* which allow an application to customise the signature
4399 *) New -resign option to smime utility. This adds one or more signers
4400 to an existing PKCS#7 signedData structure. Also -md option to use an
4401 alternative message digest algorithm for signing.
4404 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4405 create PKCS7 structures containing multiple signers. Update smime
4406 application to support multiple signers.
4409 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4413 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4414 Reorganize PBE internals to lookup from a static table using NIDs,
4415 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4416 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4417 PRF which will be automatically used with PBES2.
4420 *) Replace the algorithm specific calls to generate keys in "req" with the
4424 *) Update PKCS#7 enveloped data routines to use new API. This is now
4425 supported by any public key method supporting the encrypt operation. A
4426 ctrl is added to allow the public key algorithm to examine or modify
4427 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4431 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4432 a default digest type to use. In most cases this will be SHA1 but some
4433 algorithms (such as GOST) need to specify an alternative digest. The
4434 return value indicates how strong the preference is 1 means optional and
4435 2 is mandatory (that is it is the only supported type). Modify
4436 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4437 use the default md. Update openssl utilities to use the default digest
4438 type for signing if it is not explicitly indicated.
4441 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4442 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4443 signing method from the key type. This effectively removes the link
4444 between digests and public key types.
4447 *) Add an OID cross reference table and utility functions. Its purpose is to
4448 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4449 rsaEncryption. This will allow some of the algorithm specific hackery
4450 needed to use the correct OID to be removed.
4453 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4454 structures for PKCS7_sign(). They are now set up by the relevant public
4458 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4461 *) Add support for key derivation (agreement) in the API, DH method and
4465 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4466 public and private key formats. As a side effect these add additional
4467 command line functionality not previously available: DSA signatures can be
4468 generated and verified using pkeyutl and DH key support and generation in
4473 [Oliver Tappe <zooey@hirschkaefer.de>]
4475 *) New make target "install_html_docs" installs HTML renditions of the
4477 [Oliver Tappe <zooey@hirschkaefer.de>]
4479 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4480 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4481 support key and parameter generation and add initial key generation
4482 functionality for RSA.
4485 *) Add functions for main EVP_PKEY_method operations. The undocumented
4486 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4487 EVP_PKEY_{encrypt,decrypt}_old.
4490 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4491 key API, doesn't do much yet.
4494 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4495 public key algorithms. New option to openssl utility:
4496 "list-public-key-algorithms" to print out info.
4499 *) Implement the Supported Elliptic Curves Extension for
4500 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4503 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4504 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4507 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4508 utilities such as rsa, dsa, dsaparam etc except they process any key
4512 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4513 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4514 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4518 *) Initial support for pluggable public key ASN1.
4519 De-spaghettify the public key ASN1 handling. Move public and private
4520 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4521 algorithm specific handling to a single module within the relevant
4522 algorithm directory. Add functions to allow (near) opaque processing
4523 of public and private key structures.
4526 *) Implement the Supported Point Formats Extension for
4527 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4530 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4531 for the psk identity [hint] and the psk callback functions to the
4532 SSL_SESSION, SSL and SSL_CTX structure.
4535 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4539 SSL_CTX_use_psk_identity_hint
4540 SSL_get_psk_identity_hint
4541 SSL_get_psk_identity
4542 SSL_use_psk_identity_hint
4544 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4546 *) Add RFC 3161 compliant time stamp request creation, response generation
4547 and response verification functionality.
4548 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4550 *) Add initial support for TLS extensions, specifically for the server_name
4551 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4552 have new members for a host name. The SSL data structure has an
4553 additional member SSL_CTX *initial_ctx so that new sessions can be
4554 stored in that context to allow for session resumption, even after the
4555 SSL has been switched to a new SSL_CTX in reaction to a client's
4556 server_name extension.
4558 New functions (subject to change):
4560 SSL_get_servername()
4561 SSL_get_servername_type()
4564 New CTRL codes and macros (subject to change):
4566 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4567 - SSL_CTX_set_tlsext_servername_callback()
4568 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4569 - SSL_CTX_set_tlsext_servername_arg()
4570 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4572 openssl s_client has a new '-servername ...' option.
4574 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4575 '-key2 ...', '-servername_fatal' (subject to change). This allows
4576 testing the HostName extension for a specific single host name ('-cert'
4577 and '-key' remain fallbacks for handshakes without HostName
4578 negotiation). If the unrecognized_name alert has to be sent, this by
4579 default is a warning; it becomes fatal with the '-servername_fatal'
4582 [Peter Sylvester, Remy Allais, Christophe Renou]
4584 *) Whirlpool hash implementation is added.
4587 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4588 bn(64,32). Because of instruction set limitations it doesn't have
4589 any negative impact on performance. This was done mostly in order
4590 to make it possible to share assembler modules, such as bn_mul_mont
4591 implementations, between 32- and 64-bit builds without hassle.
4594 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4595 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4599 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4600 dedicated Montgomery multiplication procedure, is introduced.
4601 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4602 "64-bit" performance on certain 32-bit targets.
4605 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4606 in SSL structures. New SSL ctrl to set maximum send fragment size.
4607 Save memory by setting the I/O buffer sizes dynamically instead of
4608 using the maximum available value.
4611 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4612 in addition to the text details.
4615 *) Very, very preliminary EXPERIMENTAL support for printing of general
4616 ASN1 structures. This currently produces rather ugly output and doesn't
4617 handle several customised structures at all.
4620 *) Integrated support for PVK file format and some related formats such
4621 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4622 these in the 'rsa' and 'dsa' utilities.
4625 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4628 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4629 place for the (very old) "NETSCAPE" format certificates which are now
4630 handled using new ASN1 code equivalents.
4633 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4634 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4635 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4638 *) Modify CRL distribution points extension code to print out previously
4639 unsupported fields. Enhance extension setting code to allow setting of
4643 *) Add print and set support for Issuing Distribution Point CRL extension.
4646 *) Change 'Configure' script to enable Camellia by default.
4649 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4651 *) When rejecting SSL/TLS records due to an incorrect version number, never
4652 update s->server with a new major version number. As of
4653 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4654 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4655 the previous behavior could result in a read attempt at NULL when
4656 receiving specific incorrect SSL/TLS records once record payload
4657 protection is active. (CVE-2010-0740)
4658 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4660 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4661 could be crashed if the relevant tables were not present (e.g. chrooted).
4662 [Tomas Hoger <thoger@redhat.com>]
4664 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4666 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4667 [Martin Olsson, Neel Mehta]
4669 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4670 accommodate for stack sorting, always a write lock!).
4673 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4674 excessive delays in the RAND_poll(): over a minute. As a workaround
4675 include a time check in the inner Heap32Next loop too.
4678 *) The code that handled flushing of data in SSL/TLS originally used the
4679 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4680 the problem outlined in PR#1949. The fix suggested there however can
4681 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4682 of Apache). So instead simplify the code to flush unconditionally.
4683 This should be fine since flushing with no data to flush is a no op.
4686 *) Handle TLS versions 2.0 and later properly and correctly use the
4687 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4688 off ancient servers have a habit of sticking around for a while...
4691 *) Modify compression code so it frees up structures without using the
4692 ex_data callbacks. This works around a problem where some applications
4693 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4694 restarting) then use compression (e.g. SSL with compression) later.
4695 This results in significant per-connection memory leaks and
4696 has caused some security issues including CVE-2008-1678 and
4700 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4701 change when encrypting or decrypting.
4704 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4705 connect and renegotiate with servers which do not support RI.
4706 Until RI is more widely deployed this option is enabled by default.
4709 *) Add "missing" ssl ctrls to clear options and mode.
4712 *) If client attempts to renegotiate and doesn't support RI respond with
4713 a no_renegotiation alert as required by RFC5746. Some renegotiating
4714 TLS clients will continue a connection gracefully when they receive
4715 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4716 waiting for a server hello which it will never receive. Now we treat a
4717 received no_renegotiation alert as a fatal error. This is because
4718 applications requesting a renegotiation might well expect it to succeed
4719 and would have no code in place to handle the server denying it so the
4720 only safe thing to do is to terminate the connection.
4723 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4724 peer supports secure renegotiation and 0 otherwise. Print out peer
4725 renegotiation support in s_client/s_server.
4728 *) Replace the highly broken and deprecated SPKAC certification method with
4729 the updated NID creation version. This should correctly handle UTF8.
4732 *) Implement RFC5746. Re-enable renegotiation but require the extension
4733 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4734 turns out to be a bad idea. It has been replaced by
4735 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4736 SSL_CTX_set_options(). This is really not recommended unless you
4737 know what you are doing.
4738 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4740 *) Fixes to stateless session resumption handling. Use initial_ctx when
4741 issuing and attempting to decrypt tickets in case it has changed during
4742 servername handling. Use a non-zero length session ID when attempting
4743 stateless session resumption: this makes it possible to determine if
4744 a resumption has occurred immediately after receiving server hello
4745 (several places in OpenSSL subtly assume this) instead of later in
4749 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4750 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4751 fixes for a few places where the return code is not checked
4753 [Julia Lawall <julia@diku.dk>]
4755 *) Add --strict-warnings option to Configure script to include devteam
4756 warnings in other configurations.
4759 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4760 makes it possible to install openssl libraries in locations which
4761 have names other than "lib", for example "/usr/lib64" which some
4763 [Steve Henson, based on patch from Jeremy Utley]
4765 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4766 X690 8.9.12 and can produce some misleading textual output of OIDs.
4767 [Steve Henson, reported by Dan Kaminsky]
4769 *) Delete MD2 from algorithm tables. This follows the recommendation in
4770 several standards that it is not used in new applications due to
4771 several cryptographic weaknesses. For binary compatibility reasons
4772 the MD2 API is still compiled in by default.
4775 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4779 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4780 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4782 [Guenter <lists@gknw.net>]
4784 *) Fix the server certificate chain building code to use X509_verify_cert(),
4785 it used to have an ad-hoc builder which was unable to cope with anything
4786 other than a simple chain.
4787 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4789 *) Don't check self signed certificate signatures in X509_verify_cert()
4790 by default (a flag can override this): it just wastes time without
4791 adding any security. As a useful side effect self signed root CAs
4792 with non-FIPS digests are now usable in FIPS mode.
4795 *) In dtls1_process_out_of_seq_message() the check if the current message
4796 is already buffered was missing. For every new message was memory
4797 allocated, allowing an attacker to perform an denial of service attack
4798 with sending out of seq handshake messages until there is no memory
4799 left. Additionally every future message was buffered, even if the
4800 sequence number made no sense and would be part of another handshake.
4801 So only messages with sequence numbers less than 10 in advance will be
4802 buffered. (CVE-2009-1378)
4803 [Robin Seggelmann, discovered by Daniel Mentz]
4805 *) Records are buffered if they arrive with a future epoch to be
4806 processed after finishing the corresponding handshake. There is
4807 currently no limitation to this buffer allowing an attacker to perform
4808 a DOS attack with sending records with future epochs until there is no
4809 memory left. This patch adds the pqueue_size() function to determine
4810 the size of a buffer and limits the record buffer to 100 entries.
4812 [Robin Seggelmann, discovered by Daniel Mentz]
4814 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4815 parent structure is freed. (CVE-2009-1379)
4818 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4819 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4822 [Ilya O. <vrghost@gmail.com>]
4824 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4826 *) Disable renegotiation completely - this fixes a severe security
4827 problem (CVE-2009-3555) at the cost of breaking all
4828 renegotiation. Renegotiation can be re-enabled by setting
4829 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4830 run-time. This is really not recommended unless you know what
4834 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4836 *) Don't set val to NULL when freeing up structures, it is freed up by
4837 underlying code. If sizeof(void *) > sizeof(long) this can result in
4838 zeroing past the valid field. (CVE-2009-0789)
4839 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4841 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4842 checked correctly. This would allow some invalid signed attributes to
4843 appear to verify correctly. (CVE-2009-0591)
4844 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4846 *) Reject UniversalString and BMPString types with invalid lengths. This
4847 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4848 a legal length. (CVE-2009-0590)
4851 *) Set S/MIME signing as the default purpose rather than setting it
4852 unconditionally. This allows applications to override it at the store
4856 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4857 to handle some structures.
4860 *) Improve efficiency of mem_gets: don't search whole buffer each time
4862 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4864 *) New -hex option for openssl rand.
4867 *) Print out UTF8String and NumericString when parsing ASN1.
4870 *) Support NumericString type for name components.
4873 *) Allow CC in the environment to override the automatically chosen
4874 compiler. Note that nothing is done to ensure flags work with the
4878 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4880 *) Properly check EVP_VerifyFinal() and similar return values
4882 [Ben Laurie, Bodo Moeller, Google Security Team]
4884 *) Enable TLS extensions by default.
4887 *) Allow the CHIL engine to be loaded, whether the application is
4888 multithreaded or not. (This does not release the developer from the
4889 obligation to set up the dynamic locking callbacks.)
4890 [Sander Temme <sander@temme.net>]
4892 *) Use correct exit code if there is an error in dgst command.
4893 [Steve Henson; problem pointed out by Roland Dirlewanger]
4895 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4896 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4899 *) Add experimental JPAKE support, including demo authentication in
4900 s_client and s_server.
4903 *) Set the comparison function in v3_addr_canonize().
4904 [Rob Austein <sra@hactrn.net>]
4906 *) Add support for XMPP STARTTLS in s_client.
4907 [Philip Paeps <philip@freebsd.org>]
4909 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4910 to ensure that even with this option, only ciphersuites in the
4911 server's preference list will be accepted. (Note that the option
4912 applies only when resuming a session, so the earlier behavior was
4913 just about the algorithm choice for symmetric cryptography.)
4916 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4918 *) Fix NULL pointer dereference if a DTLS server received
4919 ChangeCipherSpec as first record (CVE-2009-1386).
4922 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4923 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4926 *) The fix in 0.9.8c that supposedly got rid of unsafe
4927 double-checked locking was incomplete for RSA blinding,
4928 addressing just one layer of what turns out to have been
4929 doubly unsafe triple-checked locking.
4931 So now fix this for real by retiring the MONT_HELPER macro
4932 in crypto/rsa/rsa_eay.c.
4934 [Bodo Moeller; problem pointed out by Marius Schilder]
4936 *) Various precautionary measures:
4938 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4940 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4941 (NB: This would require knowledge of the secret session ticket key
4942 to exploit, in which case you'd be SOL either way.)
4944 - Change bn_nist.c so that it will properly handle input BIGNUMs
4945 outside the expected range.
4947 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4950 [Neel Mehta, Bodo Moeller]
4952 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4953 the load fails. Useful for distros.
4954 [Ben Laurie and the FreeBSD team]
4956 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4959 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4962 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4964 This work was sponsored by Logica.
4967 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4968 keystores. Support for SSL/TLS client authentication too.
4969 Not compiled unless enable-capieng specified to Configure.
4971 This work was sponsored by Logica.
4974 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4975 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4976 attribute creation routines such as certificate requests and PKCS#12
4980 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4982 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4983 handshake which could lead to a client crash as found using the
4984 Codenomicon TLS test suite (CVE-2008-1672)
4985 [Steve Henson, Mark Cox]
4987 *) Fix double free in TLS server name extensions which could lead to
4988 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4991 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4993 Clear the error queue to ensure that error entries left from
4994 older function calls do not interfere with the correct operation.
4995 [Lutz Jaenicke, Erik de Castro Lopo]
4997 *) Remove root CA certificates of commercial CAs:
4999 The OpenSSL project does not recommend any specific CA and does not
5000 have any policy with respect to including or excluding any CA.
5001 Therefore it does not make any sense to ship an arbitrary selection
5002 of root CA certificates with the OpenSSL software.
5005 *) RSA OAEP patches to fix two separate invalid memory reads.
5006 The first one involves inputs when 'lzero' is greater than
5007 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5008 before the beginning of from). The second one involves inputs where
5009 the 'db' section contains nothing but zeroes (there is a one-byte
5010 invalid read after the end of 'db').
5011 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5013 *) Partial backport from 0.9.9-dev:
5015 Introduce bn_mul_mont (dedicated Montgomery multiplication
5016 procedure) as a candidate for BIGNUM assembler implementation.
5017 While 0.9.9-dev uses assembler for various architectures, only
5018 x86_64 is available by default here in the 0.9.8 branch, and
5019 32-bit x86 is available through a compile-time setting.
5021 To try the 32-bit x86 assembler implementation, use Configure
5022 option "enable-montasm" (which exists only for this backport).
5024 As "enable-montasm" for 32-bit x86 disclaims code stability
5025 anyway, in this constellation we activate additional code
5026 backported from 0.9.9-dev for further performance improvements,
5027 namely BN_from_montgomery_word. (To enable this otherwise,
5028 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5030 [Andy Polyakov (backport partially by Bodo Moeller)]
5032 *) Add TLS session ticket callback. This allows an application to set
5033 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5034 values. This is useful for key rollover for example where several key
5035 sets may exist with different names.
5038 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5039 This was broken until now in 0.9.8 releases, such that the only way
5040 a registered ENGINE could be used (assuming it initialises
5041 successfully on the host) was to explicitly set it as the default
5042 for the relevant algorithms. This is in contradiction with 0.9.7
5043 behaviour and the documentation. With this fix, when an ENGINE is
5044 registered into a given algorithm's table of implementations, the
5045 'uptodate' flag is reset so that auto-discovery will be used next
5046 time a new context for that algorithm attempts to select an
5048 [Ian Lister (tweaked by Geoff Thorpe)]
5050 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5051 implementation in the following ways:
5053 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5056 Lack of BER streaming support means one pass streaming processing is
5057 only supported if data is detached: setting the streaming flag is
5058 ignored for embedded content.
5060 CMS support is disabled by default and must be explicitly enabled
5061 with the enable-cms configuration option.
5064 *) Update the GMP engine glue to do direct copies between BIGNUM and
5065 mpz_t when openssl and GMP use the same limb size. Otherwise the
5066 existing "conversion via a text string export" trick is still used.
5067 [Paul Sheer <paulsheer@gmail.com>]
5069 *) Zlib compression BIO. This is a filter BIO which compressed and
5070 uncompresses any data passed through it.
5073 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5074 RFC3394 compatible AES key wrapping.
5077 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5078 sets string data without copying. X509_ALGOR_set0() and
5079 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5080 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5081 from an X509_ATTRIBUTE structure optionally checking it occurs only
5082 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5086 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5087 to get the expected BN_FLG_CONSTTIME behavior.
5088 [Bodo Moeller (Google)]
5092 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5093 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5094 - added some more tests to do_tests.pl
5095 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5096 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5097 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5098 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5099 - various changes to netware.pl to enable gcc-cross builds on Win32
5101 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5102 - various changes to fix missing prototype warnings
5103 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5104 - added AES, WHIRLPOOL and CPUID assembler code to build files
5105 - added missing AES assembler make rules to mk1mf.pl
5106 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5107 [Guenter Knauf <eflash@gmx.net>]
5109 *) Implement certificate status request TLS extension defined in RFC3546.
5110 A client can set the appropriate parameters and receive the encoded
5111 OCSP response via a callback. A server can query the supplied parameters
5112 and set the encoded OCSP response in the callback. Add simplified examples
5113 to s_client and s_server.
5116 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5118 *) Fix various bugs:
5119 + Binary incompatibility of ssl_ctx_st structure
5120 + DTLS interoperation with non-compliant servers
5121 + Don't call get_session_cb() without proposed session
5122 + Fix ia64 assembler code
5123 [Andy Polyakov, Steve Henson]
5125 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5127 *) DTLS Handshake overhaul. There were longstanding issues with
5128 OpenSSL DTLS implementation, which were making it impossible for
5129 RFC 4347 compliant client to communicate with OpenSSL server.
5130 Unfortunately just fixing these incompatibilities would "cut off"
5131 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5132 server keeps tolerating non RFC compliant syntax. The opposite is
5133 not true, 0.9.8f client can not communicate with earlier server.
5134 This update even addresses CVE-2007-4995.
5137 *) Changes to avoid need for function casts in OpenSSL: some compilers
5138 (gcc 4.2 and later) reject their use.
5139 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5142 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5143 RFC4507bis. The encrypted ticket format is an encrypted encoded
5144 SSL_SESSION structure, that way new session features are automatically
5147 If a client application caches session in an SSL_SESSION structure
5148 support is transparent because tickets are now stored in the encoded
5151 The SSL_CTX structure automatically generates keys for ticket
5152 protection in servers so again support should be possible
5153 with no application modification.
5155 If a client or server wishes to disable RFC4507 support then the option
5156 SSL_OP_NO_TICKET can be set.
5158 Add a TLS extension debugging callback to allow the contents of any client
5159 or server extensions to be examined.
5161 This work was sponsored by Google.
5164 *) Add initial support for TLS extensions, specifically for the server_name
5165 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5166 have new members for a host name. The SSL data structure has an
5167 additional member SSL_CTX *initial_ctx so that new sessions can be
5168 stored in that context to allow for session resumption, even after the
5169 SSL has been switched to a new SSL_CTX in reaction to a client's
5170 server_name extension.
5172 New functions (subject to change):
5174 SSL_get_servername()
5175 SSL_get_servername_type()
5178 New CTRL codes and macros (subject to change):
5180 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5181 - SSL_CTX_set_tlsext_servername_callback()
5182 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5183 - SSL_CTX_set_tlsext_servername_arg()
5184 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5186 openssl s_client has a new '-servername ...' option.
5188 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5189 '-key2 ...', '-servername_fatal' (subject to change). This allows
5190 testing the HostName extension for a specific single host name ('-cert'
5191 and '-key' remain fallbacks for handshakes without HostName
5192 negotiation). If the unrecognized_name alert has to be sent, this by
5193 default is a warning; it becomes fatal with the '-servername_fatal'
5196 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5198 *) Add AES and SSE2 assembly language support to VC++ build.
5201 *) Mitigate attack on final subtraction in Montgomery reduction.
5204 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5205 (which previously caused an internal error).
5208 *) Squeeze another 10% out of IGE mode when in != out.
5211 *) AES IGE mode speedup.
5212 [Dean Gaudet (Google)]
5214 *) Add the Korean symmetric 128-bit cipher SEED (see
5215 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5216 add SEED ciphersuites from RFC 4162:
5218 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5219 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5220 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5221 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5223 To minimize changes between patchlevels in the OpenSSL 0.9.8
5224 series, SEED remains excluded from compilation unless OpenSSL
5225 is configured with 'enable-seed'.
5226 [KISA, Bodo Moeller]
5228 *) Mitigate branch prediction attacks, which can be practical if a
5229 single processor is shared, allowing a spy process to extract
5230 information. For detailed background information, see
5231 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5232 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5233 and Necessary Software Countermeasures"). The core of the change
5234 are new versions BN_div_no_branch() and
5235 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5236 respectively, which are slower, but avoid the security-relevant
5237 conditional branches. These are automatically called by BN_div()
5238 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5239 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5240 remove a conditional branch.
5242 BN_FLG_CONSTTIME is the new name for the previous
5243 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5244 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5245 in the exponent causes BN_mod_exp_mont() to use the alternative
5246 implementation in BN_mod_exp_mont_consttime().) The old name
5247 remains as a deprecated alias.
5249 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5250 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5251 constant-time implementations for more than just exponentiation.
5252 Here too the old name is kept as a deprecated alias.
5254 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5255 the BN_BLINDING structure gets an independent copy of the
5256 modulus. This means that the previous "BIGNUM *m" argument to
5257 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5258 essentially becomes "const BIGNUM *m", although we can't actually
5259 change this in the header file before 0.9.9. It allows
5260 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5261 enable BN_FLG_CONSTTIME.
5263 [Matthew D Wood (Intel Corp)]
5265 *) In the SSL/TLS server implementation, be strict about session ID
5266 context matching (which matters if an application uses a single
5267 external cache for different purposes). Previously,
5268 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5269 set. This did ensure strict client verification, but meant that,
5270 with applications using a single external cache for quite
5271 different requirements, clients could circumvent ciphersuite
5272 restrictions for a given session ID context by starting a session
5273 in a different context.
5276 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5277 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5278 authentication-only ciphersuites.
5281 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5282 not complete and could lead to a possible single byte overflow
5283 (CVE-2007-5135) [Ben Laurie]
5285 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5287 *) Since AES128 and AES256 (and similarly Camellia128 and
5288 Camellia256) share a single mask bit in the logic of
5289 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5290 kludge to work properly if AES128 is available and AES256 isn't
5291 (or if Camellia128 is available and Camellia256 isn't).
5294 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5295 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5296 When a point or a seed is encoded in a BIT STRING, we need to
5297 prevent the removal of trailing zero bits to get the proper DER
5298 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5299 of a NamedBitList, for which trailing 0 bits need to be removed.)
5302 *) Have SSL/TLS server implementation tolerate "mismatched" record
5303 protocol version while receiving ClientHello even if the
5304 ClientHello is fragmented. (The server can't insist on the
5305 particular protocol version it has chosen before the ServerHello
5306 message has informed the client about his choice.)
5309 *) Add RFC 3779 support.
5310 [Rob Austein for ARIN, Ben Laurie]
5312 *) Load error codes if they are not already present instead of using a
5313 static variable. This allows them to be cleanly unloaded and reloaded.
5314 Improve header file function name parsing.
5317 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5318 or CAPABILITY handshake as required by RFCs.
5321 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5323 *) Introduce limits to prevent malicious keys being able to
5324 cause a denial of service. (CVE-2006-2940)
5325 [Steve Henson, Bodo Moeller]
5327 *) Fix ASN.1 parsing of certain invalid structures that can result
5328 in a denial of service. (CVE-2006-2937) [Steve Henson]
5330 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5331 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5333 *) Fix SSL client code which could crash if connecting to a
5334 malicious SSLv2 server. (CVE-2006-4343)
5335 [Tavis Ormandy and Will Drewry, Google Security Team]
5337 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5338 match only those. Before that, "AES256-SHA" would be interpreted
5339 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5340 the same strength classification in 0.9.7h) as we currently only
5341 have a single AES bit in the ciphersuite description bitmap.
5342 That change, however, also applied to ciphersuite strings such as
5343 "RC4-MD5" that intentionally matched multiple ciphersuites --
5344 namely, SSL 2.0 ciphersuites in addition to the more common ones
5345 from SSL 3.0/TLS 1.0.
5347 So we change the selection algorithm again: Naming an explicit
5348 ciphersuite selects this one ciphersuite, and any other similar
5349 ciphersuite (same bitmap) from *other* protocol versions.
5350 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5351 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5353 Since SSL 2.0 does not have any ciphersuites for which the
5354 128/256 bit distinction would be relevant, this works for now.
5355 The proper fix will be to use different bits for AES128 and
5356 AES256, which would have avoided the problems from the beginning;
5357 however, bits are scarce, so we can only do this in a new release
5358 (not just a patchlevel) when we can change the SSL_CIPHER
5359 definition to split the single 'unsigned long mask' bitmap into
5360 multiple values to extend the available space.
5364 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5366 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5367 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5369 *) Add AES IGE and biIGE modes.
5372 *) Change the Unix randomness entropy gathering to use poll() when
5373 possible instead of select(), since the latter has some
5374 undesirable limitations.
5375 [Darryl Miles via Richard Levitte and Bodo Moeller]
5377 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5378 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5379 cannot be implicitly activated as part of, e.g., the "AES" alias.
5380 However, please upgrade to OpenSSL 0.9.9[-dev] for
5381 non-experimental use of the ECC ciphersuites to get TLS extension
5382 support, which is required for curve and point format negotiation
5383 to avoid potential handshake problems.
5386 *) Disable rogue ciphersuites:
5388 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5389 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5390 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5392 The latter two were purportedly from
5393 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5396 Also deactivate the remaining ciphersuites from
5397 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5398 unofficial, and the ID has long expired.
5401 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5402 dual-core machines) and other potential thread-safety issues.
5405 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5406 versions), which is now available for royalty-free use
5407 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5408 Also, add Camellia TLS ciphersuites from RFC 4132.
5410 To minimize changes between patchlevels in the OpenSSL 0.9.8
5411 series, Camellia remains excluded from compilation unless OpenSSL
5412 is configured with 'enable-camellia'.
5415 *) Disable the padding bug check when compression is in use. The padding
5416 bug check assumes the first packet is of even length, this is not
5417 necessarily true if compression is enabled and can result in false
5418 positives causing handshake failure. The actual bug test is ancient
5419 code so it is hoped that implementations will either have fixed it by
5420 now or any which still have the bug do not support compression.
5423 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5425 *) When applying a cipher rule check to see if string match is an explicit
5426 cipher suite and only match that one cipher suite if it is.
5429 *) Link in manifests for VC++ if needed.
5430 [Austin Ziegler <halostatue@gmail.com>]
5432 *) Update support for ECC-based TLS ciphersuites according to
5433 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5434 TLS extensions, which are supported starting with the 0.9.9
5435 branch, not in the OpenSSL 0.9.8 branch).
5438 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5439 opaque EVP_CIPHER_CTX handling.
5442 *) Fixes and enhancements to zlib compression code. We now only use
5443 "zlib1.dll" and use the default __cdecl calling convention on Win32
5444 to conform with the standards mentioned here:
5445 http://www.zlib.net/DLL_FAQ.txt
5446 Static zlib linking now works on Windows and the new --with-zlib-include
5447 --with-zlib-lib options to Configure can be used to supply the location
5448 of the headers and library. Gracefully handle case where zlib library
5452 *) Several fixes and enhancements to the OID generation code. The old code
5453 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5454 handle numbers larger than ULONG_MAX, truncated printing and had a
5455 non standard OBJ_obj2txt() behaviour.
5458 *) Add support for building of engines under engine/ as shared libraries
5459 under VC++ build system.
5462 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5463 Hopefully, we will not see any false combination of paths any more.
5466 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5468 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5469 (part of SSL_OP_ALL). This option used to disable the
5470 countermeasure against man-in-the-middle protocol-version
5471 rollback in the SSL 2.0 server implementation, which is a bad
5472 idea. (CVE-2005-2969)
5474 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5475 for Information Security, National Institute of Advanced Industrial
5476 Science and Technology [AIST], Japan)]
5478 *) Add two function to clear and return the verify parameter flags.
5481 *) Keep cipherlists sorted in the source instead of sorting them at
5482 runtime, thus removing the need for a lock.
5485 *) Avoid some small subgroup attacks in Diffie-Hellman.
5486 [Nick Mathewson and Ben Laurie]
5488 *) Add functions for well-known primes.
5491 *) Extended Windows CE support.
5492 [Satoshi Nakamura and Andy Polyakov]
5494 *) Initialize SSL_METHOD structures at compile time instead of during
5495 runtime, thus removing the need for a lock.
5498 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5499 attempting to decrypt each encrypted key in turn. Add support to
5503 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5505 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5508 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5511 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5512 key into the same file any more.
5515 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5518 *) Add -utf8 command line and config file option to 'ca'.
5519 [Stefan <stf@udoma.org]
5521 *) Removed the macro des_crypt(), as it seems to conflict with some
5522 libraries. Use DES_crypt().
5525 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5526 involves renaming the source and generated shared-libs for
5527 both. The engines will accept the corrected or legacy ids
5528 ('ncipher' and '4758_cca' respectively) when binding. NB,
5529 this only applies when building 'shared'.
5530 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5532 *) Add attribute functions to EVP_PKEY structure. Modify
5533 PKCS12_create() to recognize a CSP name attribute and
5534 use it. Make -CSP option work again in pkcs12 utility.
5537 *) Add new functionality to the bn blinding code:
5538 - automatic re-creation of the BN_BLINDING parameters after
5539 a fixed number of uses (currently 32)
5540 - add new function for parameter creation
5541 - introduce flags to control the update behaviour of the
5542 BN_BLINDING parameters
5543 - hide BN_BLINDING structure
5544 Add a second BN_BLINDING slot to the RSA structure to improve
5545 performance when a single RSA object is shared among several
5549 *) Add support for DTLS.
5550 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5552 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5553 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5556 *) Remove buggy and incomplete DH cert support from
5557 ssl/ssl_rsa.c and ssl/s3_both.c
5560 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5561 the apps/openssl applications.
5564 *) Compile clean with "-Wall -Wmissing-prototypes
5565 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5566 DEBUG_SAFESTACK must also be set.
5569 *) Change ./Configure so that certain algorithms can be disabled by default.
5570 The new counterpiece to "no-xxx" is "enable-xxx".
5572 The patented RC5 and MDC2 algorithms will now be disabled unless
5573 "enable-rc5" and "enable-mdc2", respectively, are specified.
5575 (IDEA remains enabled despite being patented. This is because IDEA
5576 is frequently required for interoperability, and there is no license
5577 fee for non-commercial use. As before, "no-idea" can be used to
5578 avoid this algorithm.)
5582 *) Add processing of proxy certificates (see RFC 3820). This work was
5583 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5584 EGEE (Enabling Grids for E-science in Europe).
5587 *) RC4 performance overhaul on modern architectures/implementations, such
5588 as Intel P4, IA-64 and AMD64.
5591 *) New utility extract-section.pl. This can be used specify an alternative
5592 section number in a pod file instead of having to treat each file as
5593 a separate case in Makefile. This can be done by adding two lines to the
5596 =for comment openssl_section:XXX
5598 The blank line is mandatory.
5602 *) New arguments -certform, -keyform and -pass for s_client and s_server
5603 to allow alternative format key and certificate files and passphrase
5607 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5608 update associated structures and add various utility functions.
5610 Add new policy related verify parameters, include policy checking in
5611 standard verify code. Enhance 'smime' application with extra parameters
5612 to support policy checking and print out.
5615 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5616 Nehemiah processors. These extensions support AES encryption in hardware
5617 as well as RNG (though RNG support is currently disabled).
5618 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5620 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5623 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5624 [Andy Polyakov and a number of other people]
5626 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5627 implementation contributed by IBM.
5628 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5630 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5631 exponent rather than 'unsigned long'. There is a corresponding change to
5632 the new 'rsa_keygen' element of the RSA_METHOD structure.
5633 [Jelte Jansen, Geoff Thorpe]
5635 *) Functionality for creating the initial serial number file is now
5636 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5638 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5639 number file to 1, which is bound to cause problems. To avoid
5640 the problems while respecting compatibility between different 0.9.7
5641 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5642 CA.pl for serial number initialization. With the new release 0.9.8,
5643 we can fix the problem directly in the 'ca' utility.)
5646 *) Reduced header interdependencies by declaring more opaque objects in
5647 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5648 give fewer recursive includes, which could break lazy source code - so
5649 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5650 developers should define this symbol when building and using openssl to
5651 ensure they track the recommended behaviour, interfaces, [etc], but
5652 backwards-compatible behaviour prevails when this isn't defined.
5655 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5658 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5659 This will generate a random key of the appropriate length based on the
5660 cipher context. The EVP_CIPHER can provide its own random key generation
5661 routine to support keys of a specific form. This is used in the des and
5662 3des routines to generate a key of the correct parity. Update S/MIME
5663 code to use new functions and hence generate correct parity DES keys.
5664 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5665 valid (weak or incorrect parity).
5668 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5669 as looking them up. This is useful when the verified structure may contain
5670 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5671 present unless the new PKCS7_NO_CRL flag is asserted.
5674 *) Extend ASN1 oid configuration module. It now additionally accepts the
5677 shortName = some long name, 1.2.3.4
5680 *) Reimplemented the BN_CTX implementation. There is now no more static
5681 limitation on the number of variables it can handle nor the depth of the
5682 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5683 information can now expand as required, and rather than having a single
5684 static array of bignums, BN_CTX now uses a linked-list of such arrays
5685 allowing it to expand on demand whilst maintaining the usefulness of
5686 BN_CTX's "bundling".
5689 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5690 to allow all RSA operations to function using a single BN_CTX.
5693 *) Preliminary support for certificate policy evaluation and checking. This
5694 is initially intended to pass the tests outlined in "Conformance Testing
5695 of Relying Party Client Certificate Path Processing Logic" v1.07.
5698 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5699 remained unused and not that useful. A variety of other little bignum
5700 tweaks and fixes have also been made continuing on from the audit (see
5704 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5705 associated ASN1, EVP and SSL functions and old ASN1 macros.
5708 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5709 and this should never fail. So the return value from the use of
5710 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5711 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5714 *) BN_CTX_get() should return zero-valued bignums, providing the same
5715 initialised value as BN_new().
5716 [Geoff Thorpe, suggested by Ulf Möller]
5718 *) Support for inhibitAnyPolicy certificate extension.
5721 *) An audit of the BIGNUM code is underway, for which debugging code is
5722 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5723 is considered valid when processing BIGNUMs, and causes execution to
5724 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5725 further steps are taken to deliberately pollute unused data in BIGNUM
5726 structures to try and expose faulty code further on. For now, openssl will
5727 (in its default mode of operation) continue to tolerate the inconsistent
5728 forms that it has tolerated in the past, but authors and packagers should
5729 consider trying openssl and their own applications when compiled with
5730 these debugging symbols defined. It will help highlight potential bugs in
5731 their own code, and will improve the test coverage for OpenSSL itself. At
5732 some point, these tighter rules will become openssl's default to improve
5733 maintainability, though the assert()s and other overheads will remain only
5734 in debugging configurations. See bn.h for more details.
5735 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5737 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5738 that can only be obtained through BN_CTX_new() (which implicitly
5739 initialises it). The presence of this function only made it possible
5740 to overwrite an existing structure (and cause memory leaks).
5743 *) Because of the callback-based approach for implementing LHASH as a
5744 template type, lh_insert() adds opaque objects to hash-tables and
5745 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5746 to clean up those corresponding objects before destroying the hash table
5747 (and losing the object pointers). So some over-zealous constifications in
5748 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5749 objects as "const" and the lh_doall[_arg] callback wrappers are not
5750 prototyped to have "const" restrictions on the object pointers they are
5751 given (and so aren't required to cast them away any more).
5754 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5755 (speed) prefers to use its own implementation. The two implementations
5756 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5757 its object type properly exposed (MS_TM) instead of casting to/from "char
5758 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5759 aren't necessarily the greatest nomenclatures - but this is what was used
5760 internally to the implementation so I've used that for now.
5763 *) Ensure that deprecated functions do not get compiled when
5764 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5765 the self-tests were still using deprecated key-generation functions so
5766 these have been updated also.
5769 *) Reorganise PKCS#7 code to separate the digest location functionality
5770 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5771 New function PKCS7_set_digest() to set the digest type for PKCS#7
5772 digestedData type. Add additional code to correctly generate the
5773 digestedData type and add support for this type in PKCS7 initialization
5777 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5778 structure of type "other".
5781 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5782 sure the loop does correctly stop and breaking ("division by zero")
5783 modulus operations are not performed. The (pre-generated) prime
5784 table crypto/bn/bn_prime.h was already correct, but it could not be
5785 re-generated on some platforms because of the "division by zero"
5786 situation in the script.
5787 [Ralf S. Engelschall]
5789 *) Update support for ECC-based TLS ciphersuites according to
5790 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5791 SHA-1 now is only used for "small" curves (where the
5792 representation of a field element takes up to 24 bytes); for
5793 larger curves, the field element resulting from ECDH is directly
5794 used as premaster secret.
5795 [Douglas Stebila (Sun Microsystems Laboratories)]
5797 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5798 curve secp160r1 to the tests.
5799 [Douglas Stebila (Sun Microsystems Laboratories)]
5801 *) Add the possibility to load symbols globally with DSO.
5802 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5804 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5805 control of the error stack.
5808 *) Add support for STORE in ENGINE.
5811 *) Add the STORE type. The intention is to provide a common interface
5812 to certificate and key stores, be they simple file-based stores, or
5813 HSM-type store, or LDAP stores, or...
5814 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5817 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5818 pass a list of arguments to any function as well as provide a way
5819 for a function to pass data back to the caller.
5822 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5823 works like BUF_strdup() but can be used to duplicate a portion of
5824 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5828 *) Add the function sk_find_ex() which works like sk_find(), but will
5829 return an index to an element even if an exact match couldn't be
5830 found. The index is guaranteed to point at the element where the
5831 searched-for key would be inserted to preserve sorting order.
5834 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5835 takes an extra flags argument for optional functionality. Currently,
5836 the following flags are defined:
5838 OBJ_BSEARCH_VALUE_ON_NOMATCH
5839 This one gets OBJ_bsearch_ex() to return a pointer to the first
5840 element where the comparing function returns a negative or zero
5843 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5844 This one gets OBJ_bsearch_ex() to return a pointer to the first
5845 element where the comparing function returns zero. This is useful
5846 if there are more than one element where the comparing function
5850 *) Make it possible to create self-signed certificates with 'openssl ca'
5851 in such a way that the self-signed certificate becomes part of the
5852 CA database and uses the same mechanisms for serial number generation
5853 as all other certificate signing. The new flag '-selfsign' enables
5854 this functionality. Adapt CA.sh and CA.pl.in.
5857 *) Add functionality to check the public key of a certificate request
5858 against a given private. This is useful to check that a certificate
5859 request can be signed by that key (self-signing).
5862 *) Make it possible to have multiple active certificates with the same
5863 subject in the CA index file. This is done only if the keyword
5864 'unique_subject' is set to 'no' in the main CA section (default
5865 if 'CA_default') of the configuration file. The value is saved
5866 with the database itself in a separate index attribute file,
5867 named like the index file with '.attr' appended to the name.
5870 *) Generate multi-valued AVAs using '+' notation in config files for
5874 *) Support for nameConstraints certificate extension.
5877 *) Support for policyConstraints certificate extension.
5880 *) Support for policyMappings certificate extension.
5883 *) Make sure the default DSA_METHOD implementation only uses its
5884 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5885 and change its own handlers to be NULL so as to remove unnecessary
5886 indirection. This lets alternative implementations fallback to the
5887 default implementation more easily.
5890 *) Support for directoryName in GeneralName related extensions
5894 *) Make it possible to link applications using Makefile.shared.
5895 Make that possible even when linking against static libraries!
5898 *) Support for single pass processing for S/MIME signing. This now
5899 means that S/MIME signing can be done from a pipe, in addition
5900 cleartext signing (multipart/signed type) is effectively streaming
5901 and the signed data does not need to be all held in memory.
5903 This is done with a new flag PKCS7_STREAM. When this flag is set
5904 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5905 is done after the data is output (and digests calculated) in
5906 SMIME_write_PKCS7().
5909 *) Add full support for -rpath/-R, both in shared libraries and
5910 applications, at least on the platforms where it's known how
5914 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5915 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5916 will now compute a table of multiples of the generator that
5917 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5918 faster (notably in the case of a single point multiplication,
5919 scalar * generator).
5920 [Nils Larsch, Bodo Moeller]
5922 *) IPv6 support for certificate extensions. The various extensions
5923 which use the IP:a.b.c.d can now take IPv6 addresses using the
5924 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5928 *) Added an ENGINE that implements RSA by performing private key
5929 exponentiations with the GMP library. The conversions to and from
5930 GMP's mpz_t format aren't optimised nor are any montgomery forms
5931 cached, and on x86 it appears OpenSSL's own performance has caught up.
5932 However there are likely to be other architectures where GMP could
5933 provide a boost. This ENGINE is not built in by default, but it can be
5934 specified at Configure time and should be accompanied by the necessary
5935 linker additions, eg;
5936 ./config -DOPENSSL_USE_GMP -lgmp
5939 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5940 testing availability of engines with "-t" - the old behaviour is
5941 produced by increasing the feature's verbosity with "-tt".
5944 *) ECDSA routines: under certain error conditions uninitialized BN objects
5945 could be freed. Solution: make sure initialization is performed early
5946 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5950 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5951 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5952 software implementations. For DSA and DH, parameter generation can
5953 also be overridden by providing the appropriate method callbacks.
5956 *) Change the "progress" mechanism used in key-generation and
5957 primality testing to functions that take a new BN_GENCB pointer in
5958 place of callback/argument pairs. The new API functions have "_ex"
5959 postfixes and the older functions are reimplemented as wrappers for
5960 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5961 declarations of the old functions to help (graceful) attempts to
5962 migrate to the new functions. Also, the new key-generation API
5963 functions operate on a caller-supplied key-structure and return
5964 success/failure rather than returning a key or NULL - this is to
5965 help make "keygen" another member function of RSA_METHOD etc.
5967 Example for using the new callback interface:
5969 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5973 BN_GENCB_set(&my_cb, my_callback, my_arg);
5975 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5976 /* For the meaning of a, b in calls to my_callback(), see the
5977 * documentation of the function that calls the callback.
5978 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5979 * my_callback should return 1 if it wants BN_is_prime_ex()
5980 * to continue, or 0 to stop.
5985 *) Change the ZLIB compression method to be stateful, and make it
5986 available to TLS with the number defined in
5987 draft-ietf-tls-compression-04.txt.
5990 *) Add the ASN.1 structures and functions for CertificatePair, which
5991 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5993 CertificatePair ::= SEQUENCE {
5994 forward [0] Certificate OPTIONAL,
5995 reverse [1] Certificate OPTIONAL,
5996 -- at least one of the pair shall be present -- }
5998 Also implement the PEM functions to read and write certificate
5999 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6001 This needed to be defined, mostly for the sake of the LDAP
6002 attribute crossCertificatePair, but may prove useful elsewhere as
6006 *) Make it possible to inhibit symlinking of shared libraries in
6007 Makefile.shared, for Cygwin's sake.
6010 *) Extend the BIGNUM API by creating a function
6011 void BN_set_negative(BIGNUM *a, int neg);
6012 and a macro that behave like
6013 int BN_is_negative(const BIGNUM *a);
6015 to avoid the need to access 'a->neg' directly in applications.
6018 *) Implement fast modular reduction for pseudo-Mersenne primes
6019 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6020 EC_GROUP_new_curve_GFp() will now automatically use this
6022 [Nils Larsch <nla@trustcenter.de>]
6024 *) Add new lock type (CRYPTO_LOCK_BN).
6027 *) Change the ENGINE framework to automatically load engines
6028 dynamically from specific directories unless they could be
6029 found to already be built in or loaded. Move all the
6030 current engines except for the cryptodev one to a new
6032 The engines in engines/ are built as shared libraries if
6033 the "shared" options was given to ./Configure or ./config.
6034 Otherwise, they are inserted in libcrypto.a.
6035 /usr/local/ssl/engines is the default directory for dynamic
6036 engines, but that can be overridden at configure time through
6037 the usual use of --prefix and/or --openssldir, and at run
6038 time with the environment variable OPENSSL_ENGINES.
6039 [Geoff Thorpe and Richard Levitte]
6041 *) Add Makefile.shared, a helper makefile to build shared
6042 libraries. Adapt Makefile.org.
6045 *) Add version info to Win32 DLLs.
6046 [Peter 'Luna' Runestig" <peter@runestig.com>]
6048 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6049 can be added using this API to created arbitrary PKCS#12
6050 files while avoiding the low level API.
6052 New options to PKCS12_create(), key or cert can be NULL and
6053 will then be omitted from the output file. The encryption
6054 algorithm NIDs can be set to -1 for no encryption, the mac
6055 iteration count can be set to 0 to omit the mac.
6057 Enhance pkcs12 utility by making the -nokeys and -nocerts
6058 options work when creating a PKCS#12 file. New option -nomac
6059 to omit the mac, NONE can be set for an encryption algorithm.
6060 New code is modified to use the enhanced PKCS12_create()
6061 instead of the low level API.
6064 *) Extend ASN1 encoder to support indefinite length constructed
6065 encoding. This can output sequences tags and octet strings in
6066 this form. Modify pk7_asn1.c to support indefinite length
6067 encoding. This is experimental and needs additional code to
6068 be useful, such as an ASN1 bio and some enhanced streaming
6071 Extend template encode functionality so that tagging is passed
6072 down to the template encoder.
6075 *) Let 'openssl req' fail if an argument to '-newkey' is not
6076 recognized instead of using RSA as a default.
6079 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6080 As these are not official, they are not included in "ALL";
6081 the "ECCdraft" ciphersuite group alias can be used to select them.
6082 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6084 *) Add ECDH engine support.
6085 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6087 *) Add ECDH in new directory crypto/ecdh/.
6088 [Douglas Stebila (Sun Microsystems Laboratories)]
6090 *) Let BN_rand_range() abort with an error after 100 iterations
6091 without success (which indicates a broken PRNG).
6094 *) Change BN_mod_sqrt() so that it verifies that the input value
6095 is really the square of the return value. (Previously,
6096 BN_mod_sqrt would show GIGO behaviour.)
6099 *) Add named elliptic curves over binary fields from X9.62, SECG,
6100 and WAP/WTLS; add OIDs that were still missing.
6102 [Sheueling Chang Shantz and Douglas Stebila
6103 (Sun Microsystems Laboratories)]
6105 *) Extend the EC library for elliptic curves over binary fields
6106 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6109 EC_GF2m_simple_method
6113 EC_GROUP_new_curve_GF2m
6114 EC_GROUP_set_curve_GF2m
6115 EC_GROUP_get_curve_GF2m
6116 EC_POINT_set_affine_coordinates_GF2m
6117 EC_POINT_get_affine_coordinates_GF2m
6118 EC_POINT_set_compressed_coordinates_GF2m
6120 Point compression for binary fields is disabled by default for
6121 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6124 As binary polynomials are represented as BIGNUMs, various members
6125 of the EC_GROUP and EC_POINT data structures can be shared
6126 between the implementations for prime fields and binary fields;
6127 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6128 are essentially identical to their ..._GFp counterparts.
6129 (For simplicity, the '..._GFp' prefix has been dropped from
6130 various internal method names.)
6132 An internal 'field_div' method (similar to 'field_mul' and
6133 'field_sqr') has been added; this is used only for binary fields.
6135 [Sheueling Chang Shantz and Douglas Stebila
6136 (Sun Microsystems Laboratories)]
6138 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6139 through methods ('mul', 'precompute_mult').
6141 The generic implementations (now internally called 'ec_wNAF_mul'
6142 and 'ec_wNAF_precomputed_mult') remain the default if these
6143 methods are undefined.
6145 [Sheueling Chang Shantz and Douglas Stebila
6146 (Sun Microsystems Laboratories)]
6148 *) New function EC_GROUP_get_degree, which is defined through
6149 EC_METHOD. For curves over prime fields, this returns the bit
6150 length of the modulus.
6152 [Sheueling Chang Shantz and Douglas Stebila
6153 (Sun Microsystems Laboratories)]
6155 *) New functions EC_GROUP_dup, EC_POINT_dup.
6156 (These simply call ..._new and ..._copy).
6158 [Sheueling Chang Shantz and Douglas Stebila
6159 (Sun Microsystems Laboratories)]
6161 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6162 Polynomials are represented as BIGNUMs (where the sign bit is not
6163 used) in the following functions [macros]:
6166 BN_GF2m_sub [= BN_GF2m_add]
6167 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6168 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6169 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6171 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6172 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6173 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6174 BN_GF2m_cmp [= BN_ucmp]
6176 (Note that only the 'mod' functions are actually for fields GF(2^m).
6177 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6179 For some functions, an the irreducible polynomial defining a
6180 field can be given as an 'unsigned int[]' with strictly
6181 decreasing elements giving the indices of those bits that are set;
6182 i.e., p[] represents the polynomial
6183 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6185 p[0] > p[1] > ... > p[k] = 0.
6186 This applies to the following functions:
6191 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6192 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6194 BN_GF2m_mod_sqrt_arr
6195 BN_GF2m_mod_solve_quad_arr
6199 Conversion can be performed by the following functions:
6204 bntest.c has additional tests for binary polynomial arithmetic.
6206 Two implementations for BN_GF2m_mod_div() are available.
6207 The default algorithm simply uses BN_GF2m_mod_inv() and
6208 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6209 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6210 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6212 [Sheueling Chang Shantz and Douglas Stebila
6213 (Sun Microsystems Laboratories)]
6215 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6216 functionality is disabled at compile-time.
6217 [Douglas Stebila <douglas.stebila@sun.com>]
6219 *) Change default behaviour of 'openssl asn1parse' so that more
6220 information is visible when viewing, e.g., a certificate:
6222 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6223 mode the content of non-printable OCTET STRINGs is output in a
6224 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6225 avoid the appearance of a printable string.
6226 [Nils Larsch <nla@trustcenter.de>]
6228 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6230 EC_GROUP_set_asn1_flag()
6231 EC_GROUP_get_asn1_flag()
6232 EC_GROUP_set_point_conversion_form()
6233 EC_GROUP_get_point_conversion_form()
6234 These control ASN1 encoding details:
6235 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6236 has been set to OPENSSL_EC_NAMED_CURVE.
6237 - Points are encoded in uncompressed form by default; options for
6238 asn1_for are as for point2oct, namely
6239 POINT_CONVERSION_COMPRESSED
6240 POINT_CONVERSION_UNCOMPRESSED
6241 POINT_CONVERSION_HYBRID
6243 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6246 EC_GROUP_get0_seed()
6247 EC_GROUP_get_seed_len()
6248 This is used only for ASN1 purposes (so far).
6249 [Nils Larsch <nla@trustcenter.de>]
6251 *) Add 'field_type' member to EC_METHOD, which holds the NID
6252 of the appropriate field type OID. The new function
6253 EC_METHOD_get_field_type() returns this value.
6254 [Nils Larsch <nla@trustcenter.de>]
6259 EC_POINT_point2hex()
6260 EC_POINT_hex2point()
6261 providing useful interfaces to EC_POINT_point2oct() and
6262 EC_POINT_oct2point().
6263 [Nils Larsch <nla@trustcenter.de>]
6265 *) Change internals of the EC library so that the functions
6266 EC_GROUP_set_generator()
6267 EC_GROUP_get_generator()
6268 EC_GROUP_get_order()
6269 EC_GROUP_get_cofactor()
6270 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6271 to methods, which would lead to unnecessary code duplication when
6272 adding different types of curves.
6273 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6275 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6276 arithmetic, and such that modified wNAFs are generated
6277 (which avoid length expansion in many cases).
6280 *) Add a function EC_GROUP_check_discriminant() (defined via
6281 EC_METHOD) that verifies that the curve discriminant is non-zero.
6283 Add a function EC_GROUP_check() that makes some sanity tests
6284 on a EC_GROUP, its generator and order. This includes
6285 EC_GROUP_check_discriminant().
6286 [Nils Larsch <nla@trustcenter.de>]
6288 *) Add ECDSA in new directory crypto/ecdsa/.
6290 Add applications 'openssl ecparam' and 'openssl ecdsa'
6291 (these are based on 'openssl dsaparam' and 'openssl dsa').
6293 ECDSA support is also included in various other files across the
6294 library. Most notably,
6295 - 'openssl req' now has a '-newkey ecdsa:file' option;
6296 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6297 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6298 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6299 them suitable for ECDSA where domain parameters must be
6300 extracted before the specific public key;
6301 - ECDSA engine support has been added.
6302 [Nils Larsch <nla@trustcenter.de>]
6304 *) Include some named elliptic curves, and add OIDs from X9.62,
6305 SECG, and WAP/WTLS. Each curve can be obtained from the new
6307 EC_GROUP_new_by_curve_name(),
6308 and the list of available named curves can be obtained with
6309 EC_get_builtin_curves().
6310 Also add a 'curve_name' member to EC_GROUP objects, which can be
6312 EC_GROUP_set_curve_name()
6313 EC_GROUP_get_curve_name()
6314 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6316 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6317 was actually never needed) and in BN_mul(). The removal in BN_mul()
6318 required a small change in bn_mul_part_recursive() and the addition
6319 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6320 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6321 bn_sub_words() and bn_add_words() except they take arrays with
6325 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6327 *) Cleanse PEM buffers before freeing them since they may contain
6329 [Benjamin Bennett <ben@psc.edu>]
6331 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6332 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6333 authentication-only ciphersuites.
6336 *) Since AES128 and AES256 share a single mask bit in the logic of
6337 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6338 kludge to work properly if AES128 is available and AES256 isn't.
6341 *) Expand security boundary to match 1.1.1 module.
6344 *) Remove redundant features: hash file source, editing of test vectors
6345 modify fipsld to use external fips_premain.c signature.
6348 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6349 run algorithm test programs.
6352 *) Make algorithm test programs more tolerant of whitespace.
6355 *) Have SSL/TLS server implementation tolerate "mismatched" record
6356 protocol version while receiving ClientHello even if the
6357 ClientHello is fragmented. (The server can't insist on the
6358 particular protocol version it has chosen before the ServerHello
6359 message has informed the client about his choice.)
6362 *) Load error codes if they are not already present instead of using a
6363 static variable. This allows them to be cleanly unloaded and reloaded.
6366 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6368 *) Introduce limits to prevent malicious keys being able to
6369 cause a denial of service. (CVE-2006-2940)
6370 [Steve Henson, Bodo Moeller]
6372 *) Fix ASN.1 parsing of certain invalid structures that can result
6373 in a denial of service. (CVE-2006-2937) [Steve Henson]
6375 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6376 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6378 *) Fix SSL client code which could crash if connecting to a
6379 malicious SSLv2 server. (CVE-2006-4343)
6380 [Tavis Ormandy and Will Drewry, Google Security Team]
6382 *) Change ciphersuite string processing so that an explicit
6383 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6384 will no longer include "AES128-SHA"), and any other similar
6385 ciphersuite (same bitmap) from *other* protocol versions (so that
6386 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6387 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6388 changes from 0.9.8b and 0.9.8d.
6391 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6393 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6394 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6396 *) Change the Unix randomness entropy gathering to use poll() when
6397 possible instead of select(), since the latter has some
6398 undesirable limitations.
6399 [Darryl Miles via Richard Levitte and Bodo Moeller]
6401 *) Disable rogue ciphersuites:
6403 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6404 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6405 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6407 The latter two were purportedly from
6408 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6411 Also deactivate the remaining ciphersuites from
6412 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6413 unofficial, and the ID has long expired.
6416 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6417 dual-core machines) and other potential thread-safety issues.
6420 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6422 *) Adapt fipsld and the build system to link against the validated FIPS
6423 module in FIPS mode.
6426 *) Fixes for VC++ 2005 build under Windows.
6429 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6430 from a Windows bash shell such as MSYS. It is autodetected from the
6431 "config" script when run from a VC++ environment. Modify standard VC++
6432 build to use fipscanister.o from the GNU make build.
6435 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6437 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6438 The value now differs depending on if you build for FIPS or not.
6439 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6440 safely run with a non-FIPSed libcrypto, as it may crash because of
6441 the difference induced by this change.
6444 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6446 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6447 (part of SSL_OP_ALL). This option used to disable the
6448 countermeasure against man-in-the-middle protocol-version
6449 rollback in the SSL 2.0 server implementation, which is a bad
6450 idea. (CVE-2005-2969)
6452 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6453 for Information Security, National Institute of Advanced Industrial
6454 Science and Technology [AIST], Japan)]
6456 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6457 mainly for FIPS compliance and not fully integrated at this stage.
6460 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6461 the exponentiation using a fixed-length exponent. (Otherwise,
6462 the information leaked through timing could expose the secret key
6463 after many signatures; cf. Bleichenbacher's attack on DSA with
6467 *) Make a new fixed-window mod_exp implementation the default for
6468 RSA, DSA, and DH private-key operations so that the sequence of
6469 squares and multiplies and the memory access pattern are
6470 independent of the particular secret key. This will mitigate
6471 cache-timing and potential related attacks.
6473 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6474 and this is automatically used by BN_mod_exp_mont() if the new flag
6475 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6476 will use this BN flag for private exponents unless the flag
6477 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6478 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6480 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6482 *) Change the client implementation for SSLv23_method() and
6483 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6484 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6485 (Previously, the SSL 2.0 backwards compatible Client Hello
6486 message format would be used even with SSL_OP_NO_SSLv2.)
6489 *) Add support for smime-type MIME parameter in S/MIME messages which some
6493 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6494 a threadsafe manner. Modify rsa code to use new function and add calls
6495 to dsa and dh code (which had race conditions before).
6498 *) Include the fixed error library code in the C error file definitions
6499 instead of fixing them up at runtime. This keeps the error code
6500 structures constant.
6503 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6505 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6508 *) Fixes for newer kerberos headers. NB: the casts are needed because
6509 the 'length' field is signed on one version and unsigned on another
6510 with no (?) obvious way to tell the difference, without these VC++
6511 complains. Also the "definition" of FAR (blank) is no longer included
6512 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6513 some needed definitions.
6516 *) Undo Cygwin change.
6519 *) Added support for proxy certificates according to RFC 3820.
6520 Because they may be a security thread to unaware applications,
6521 they must be explicitly allowed in run-time. See
6522 docs/HOWTO/proxy_certificates.txt for further information.
6525 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6527 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6528 server and client random values. Previously
6529 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6530 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6532 This change has negligible security impact because:
6534 1. Server and client random values still have 24 bytes of pseudo random
6537 2. Server and client random values are sent in the clear in the initial
6540 3. The master secret is derived using the premaster secret (48 bytes in
6541 size for static RSA ciphersuites) as well as client server and random
6544 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6547 [Stephen Henson, reported by UK NISCC]
6549 *) Use Windows randomness collection on Cygwin.
6552 *) Fix hang in EGD/PRNGD query when communication socket is closed
6553 prematurely by EGD/PRNGD.
6554 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6556 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6559 *) Back-port of selected performance improvements from development
6560 branch, as well as improved support for PowerPC platforms.
6563 *) Add lots of checks for memory allocation failure, error codes to indicate
6564 failure and freeing up memory if a failure occurs.
6565 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6567 *) Add new -passin argument to dgst.
6570 *) Perform some character comparisons of different types in X509_NAME_cmp:
6571 this is needed for some certificates that re-encode DNs into UTF8Strings
6572 (in violation of RFC3280) and can't or won't issue name rollover
6576 *) Make an explicit check during certificate validation to see that
6577 the CA setting in each certificate on the chain is correct. As a
6578 side effect always do the following basic checks on extensions,
6579 not just when there's an associated purpose to the check:
6581 - if there is an unhandled critical extension (unless the user
6582 has chosen to ignore this fault)
6583 - if the path length has been exceeded (if one is set at all)
6584 - that certain extensions fit the associated purpose (if one has
6588 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6590 *) Avoid a race condition when CRLs are checked in a multi threaded
6591 environment. This would happen due to the reordering of the revoked
6592 entries during signature checking and serial number lookup. Now the
6593 encoding is cached and the serial number sort performed under a lock.
6594 Add new STACK function sk_is_sorted().
6597 *) Add Delta CRL to the extension code.
6600 *) Various fixes to s3_pkt.c so alerts are sent properly.
6601 [David Holmes <d.holmes@f5.com>]
6603 *) Reduce the chances of duplicate issuer name and serial numbers (in
6604 violation of RFC3280) using the OpenSSL certificate creation utilities.
6605 This is done by creating a random 64 bit value for the initial serial
6606 number when a serial number file is created or when a self signed
6607 certificate is created using 'openssl req -x509'. The initial serial
6608 number file is created using 'openssl x509 -next_serial' in CA.pl
6609 rather than being initialized to 1.
6612 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6614 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6615 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6616 [Joe Orton, Steve Henson]
6618 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6620 [Joe Orton, Steve Henson]
6622 *) Make it possible to have multiple active certificates with the same
6623 subject in the CA index file. This is done only if the keyword
6624 'unique_subject' is set to 'no' in the main CA section (default
6625 if 'CA_default') of the configuration file. The value is saved
6626 with the database itself in a separate index attribute file,
6627 named like the index file with '.attr' appended to the name.
6630 *) X509 verify fixes. Disable broken certificate workarounds when
6631 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6632 keyUsage extension present. Don't accept CRLs with unhandled critical
6633 extensions: since verify currently doesn't process CRL extensions this
6634 rejects a CRL with *any* critical extensions. Add new verify error codes
6638 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6639 A clarification of RFC2560 will require the use of OCTET STRINGs and
6640 some implementations cannot handle the current raw format. Since OpenSSL
6641 copies and compares OCSP nonces as opaque blobs without any attempt at
6642 parsing them this should not create any compatibility issues.
6645 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6646 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6647 this HMAC (and other) operations are several times slower than OpenSSL
6651 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6652 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6654 *) Use the correct content when signing type "other".
6657 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6659 *) Fix various bugs revealed by running the NISCC test suite:
6661 Stop out of bounds reads in the ASN1 code when presented with
6662 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6664 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6666 If verify callback ignores invalid public key errors don't try to check
6667 certificate signature with the NULL public key.
6671 *) New -ignore_err option in ocsp application to stop the server
6672 exiting on the first error in a request.
6675 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6676 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6680 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6681 extra data after the compression methods not only for TLS 1.0
6682 but also for SSL 3.0 (as required by the specification).
6683 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6685 *) Change X509_certificate_type() to mark the key as exported/exportable
6686 when it's 512 *bits* long, not 512 bytes.
6689 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6690 blocks during encryption.
6693 *) Various fixes to base64 BIO and non blocking I/O. On write
6694 flushes were not handled properly if the BIO retried. On read
6695 data was not being buffered properly and had various logic bugs.
6696 This also affects blocking I/O when the data being decoded is a
6700 *) Various S/MIME bugfixes and compatibility changes:
6701 output correct application/pkcs7 MIME type if
6702 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6703 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6704 of files as .eml work). Correctly handle very long lines in MIME
6708 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6710 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6711 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6712 a protocol version number mismatch like a decryption error
6713 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6716 *) Turn on RSA blinding by default in the default implementation
6717 to avoid a timing attack. Applications that don't want it can call
6718 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6719 They would be ill-advised to do so in most cases.
6720 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6722 *) Change RSA blinding code so that it works when the PRNG is not
6723 seeded (in this case, the secret RSA exponent is abused as
6724 an unpredictable seed -- if it is not unpredictable, there
6725 is no point in blinding anyway). Make RSA blinding thread-safe
6726 by remembering the creator's thread ID in rsa->blinding and
6727 having all other threads use local one-time blinding factors
6728 (this requires more computation than sharing rsa->blinding, but
6729 avoids excessive locking; and if an RSA object is not shared
6730 between threads, blinding will still be very fast).
6733 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6734 ENGINE as defaults for all supported algorithms irrespective of
6735 the 'flags' parameter. 'flags' is now honoured, so applications
6736 should make sure they are passing it correctly.
6739 *) Target "mingw" now allows native Windows code to be generated in
6740 the Cygwin environment as well as with the MinGW compiler.
6743 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6745 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6746 via timing by performing a MAC computation even if incorrect
6747 block cipher padding has been found. This is a countermeasure
6748 against active attacks where the attacker has to distinguish
6749 between bad padding and a MAC verification error. (CVE-2003-0078)
6751 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6752 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6753 Martin Vuagnoux (EPFL, Ilion)]
6755 *) Make the no-err option work as intended. The intention with no-err
6756 is not to have the whole error stack handling routines removed from
6757 libcrypto, it's only intended to remove all the function name and
6758 reason texts, thereby removing some of the footprint that may not
6759 be interesting if those errors aren't displayed anyway.
6761 NOTE: it's still possible for any application or module to have it's
6762 own set of error texts inserted. The routines are there, just not
6763 used by default when no-err is given.
6766 *) Add support for FreeBSD on IA64.
6767 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6769 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6770 Kerberos function mit_des_cbc_cksum(). Before this change,
6771 the value returned by DES_cbc_cksum() was like the one from
6772 mit_des_cbc_cksum(), except the bytes were swapped.
6773 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6775 *) Allow an application to disable the automatic SSL chain building.
6776 Before this a rather primitive chain build was always performed in
6777 ssl3_output_cert_chain(): an application had no way to send the
6778 correct chain if the automatic operation produced an incorrect result.
6780 Now the chain builder is disabled if either:
6782 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6784 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6786 The reasoning behind this is that an application would not want the
6787 auto chain building to take place if extra chain certificates are
6788 present and it might also want a means of sending no additional
6789 certificates (for example the chain has two certificates and the
6793 *) Add the possibility to build without the ENGINE framework.
6794 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6796 *) Under Win32 gmtime() can return NULL: check return value in
6797 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6800 *) DSA routines: under certain error conditions uninitialized BN objects
6801 could be freed. Solution: make sure initialization is performed early
6802 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6803 Nils Larsch <nla@trustcenter.de> via PR#459)
6806 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6807 checked on reconnect on the client side, therefore session resumption
6808 could still fail with a "ssl session id is different" error. This
6809 behaviour is masked when SSL_OP_ALL is used due to
6810 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6811 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6812 followup to PR #377.
6815 *) IA-32 assembler support enhancements: unified ELF targets, support
6816 for SCO/Caldera platforms, fix for Cygwin shared build.
6819 *) Add support for FreeBSD on sparc64. As a consequence, support for
6820 FreeBSD on non-x86 processors is separate from x86 processors on
6821 the config script, much like the NetBSD support.
6822 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6824 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6826 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6829 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6830 code (06) was taken as the first octet of the session ID and the last
6831 octet was ignored consequently. As a result SSLv2 client side session
6832 caching could not have worked due to the session ID mismatch between
6834 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6838 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6839 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6843 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6844 seems that in spite of existing for more than a year, many application
6845 author have done nothing to provide the necessary callbacks, which
6846 means that this particular engine will not work properly anywhere.
6847 This is a very unfortunate situation which forces us, in the name
6848 of usability, to give the hw_ncipher.c a static lock, which is part
6850 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6851 appear in 0.9.8 or later. We EXPECT application authors to have
6852 dealt properly with this when 0.9.8 is released (unless we actually
6853 make such changes in the libcrypto locking code that changes will
6854 have to be made anyway).
6857 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6858 octets have been read, EOF or an error occurs. Without this change
6859 some truncated ASN1 structures will not produce an error.
6862 *) Disable Heimdal support, since it hasn't been fully implemented.
6863 Still give the possibility to force the use of Heimdal, but with
6864 warnings and a request that patches get sent to openssl-dev.
6867 *) Add the VC-CE target, introduce the WINCE sysname, and add
6868 INSTALL.WCE and appropriate conditionals to make it build.
6869 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6871 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6872 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6873 edit numbers of the version.
6874 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6876 *) Introduce safe string copy and catenation functions
6877 (BUF_strlcpy() and BUF_strlcat()).
6878 [Ben Laurie (CHATS) and Richard Levitte]
6880 *) Avoid using fixed-size buffers for one-line DNs.
6881 [Ben Laurie (CHATS)]
6883 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6884 resizing buffers containing secrets, and use where appropriate.
6885 [Ben Laurie (CHATS)]
6887 *) Avoid using fixed size buffers for configuration file location.
6888 [Ben Laurie (CHATS)]
6890 *) Avoid filename truncation for various CA files.
6891 [Ben Laurie (CHATS)]
6893 *) Use sizeof in preference to magic numbers.
6894 [Ben Laurie (CHATS)]
6896 *) Avoid filename truncation in cert requests.
6897 [Ben Laurie (CHATS)]
6899 *) Add assertions to check for (supposedly impossible) buffer
6901 [Ben Laurie (CHATS)]
6903 *) Don't cache truncated DNS entries in the local cache (this could
6904 potentially lead to a spoofing attack).
6905 [Ben Laurie (CHATS)]
6907 *) Fix various buffers to be large enough for hex/decimal
6908 representations in a platform independent manner.
6909 [Ben Laurie (CHATS)]
6911 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6912 resizing buffers containing secrets, and use where appropriate.
6913 [Ben Laurie (CHATS)]
6915 *) Add BIO_indent() to avoid much slightly worrying code to do
6917 [Ben Laurie (CHATS)]
6919 *) Convert sprintf()/BIO_puts() to BIO_printf().
6920 [Ben Laurie (CHATS)]
6922 *) buffer_gets() could terminate with the buffer only half
6924 [Ben Laurie (CHATS)]
6926 *) Add assertions to prevent user-supplied crypto functions from
6927 overflowing internal buffers by having large block sizes, etc.
6928 [Ben Laurie (CHATS)]
6930 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6932 [Ben Laurie (CHATS)]
6934 *) Eliminate unused copy of key in RC4.
6935 [Ben Laurie (CHATS)]
6937 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6938 [Ben Laurie (CHATS)]
6940 *) Fix off-by-one error in EGD path.
6941 [Ben Laurie (CHATS)]
6943 *) If RANDFILE path is too long, ignore instead of truncating.
6944 [Ben Laurie (CHATS)]
6946 *) Eliminate unused and incorrectly sized X.509 structure
6948 [Ben Laurie (CHATS)]
6950 *) Eliminate unused and dangerous function knumber().
6951 [Ben Laurie (CHATS)]
6953 *) Eliminate unused and dangerous structure, KSSL_ERR.
6954 [Ben Laurie (CHATS)]
6956 *) Protect against overlong session ID context length in an encoded
6957 session object. Since these are local, this does not appear to be
6959 [Ben Laurie (CHATS)]
6961 *) Change from security patch (see 0.9.6e below) that did not affect
6962 the 0.9.6 release series:
6964 Remote buffer overflow in SSL3 protocol - an attacker could
6965 supply an oversized master key in Kerberos-enabled versions.
6967 [Ben Laurie (CHATS)]
6969 *) Change the SSL kerb5 codes to match RFC 2712.
6972 *) Make -nameopt work fully for req and add -reqopt switch.
6973 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6975 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6976 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6978 *) Make sure tests can be performed even if the corresponding algorithms
6979 have been removed entirely. This was also the last step to make
6980 OpenSSL compilable with DJGPP under all reasonable conditions.
6981 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6983 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6984 to allow version independent disabling of normally unselected ciphers,
6985 which may be activated as a side-effect of selecting a single cipher.
6987 (E.g., cipher list string "RSA" enables ciphersuites that are left
6988 out of "ALL" because they do not provide symmetric encryption.
6989 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6990 [Lutz Jaenicke, Bodo Moeller]
6992 *) Add appropriate support for separate platform-dependent build
6993 directories. The recommended way to make a platform-dependent
6994 build directory is the following (tested on Linux), maybe with
6997 # Place yourself outside of the OpenSSL source tree. In
6998 # this example, the environment variable OPENSSL_SOURCE
6999 # is assumed to contain the absolute OpenSSL source directory.
7000 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7001 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7002 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7003 mkdir -p `dirname $F`
7004 ln -s $OPENSSL_SOURCE/$F $F
7007 To be absolutely sure not to disturb the source tree, a "make clean"
7008 is a good thing. If it isn't successful, don't worry about it,
7009 it probably means the source directory is very clean.
7012 *) Make sure any ENGINE control commands make local copies of string
7013 pointers passed to them whenever necessary. Otherwise it is possible
7014 the caller may have overwritten (or deallocated) the original string
7015 data when a later ENGINE operation tries to use the stored values.
7016 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7018 *) Improve diagnostics in file reading and command-line digests.
7019 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7021 *) Add AES modes CFB and OFB to the object database. Correct an
7022 error in AES-CFB decryption.
7025 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7026 allows existing EVP_CIPHER_CTX structures to be reused after
7027 calling EVP_*Final(). This behaviour is used by encryption
7028 BIOs and some applications. This has the side effect that
7029 applications must explicitly clean up cipher contexts with
7030 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7033 *) Check the values of dna and dnb in bn_mul_recursive before calling
7034 bn_mul_comba (a non zero value means the a or b arrays do not contain
7035 n2 elements) and fallback to bn_mul_normal if either is not zero.
7038 *) Fix escaping of non-ASCII characters when using the -subj option
7039 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7042 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7043 form for "surname", serialNumber has no short form.
7044 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7045 therefore remove "mail" short name for "internet 7".
7046 The OID for unique identifiers in X509 certificates is
7047 x500UniqueIdentifier, not uniqueIdentifier.
7048 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7051 *) Add an "init" command to the ENGINE config module and auto initialize
7052 ENGINEs. Without any "init" command the ENGINE will be initialized
7053 after all ctrl commands have been executed on it. If init=1 the
7054 ENGINE is initialized at that point (ctrls before that point are run
7055 on the uninitialized ENGINE and after on the initialized one). If
7056 init=0 then the ENGINE will not be initialized at all.
7059 *) Fix the 'app_verify_callback' interface so that the user-defined
7060 argument is actually passed to the callback: In the
7061 SSL_CTX_set_cert_verify_callback() prototype, the callback
7062 declaration has been changed from
7065 int (*cb)(X509_STORE_CTX *,void *);
7066 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7067 i=s->ctx->app_verify_callback(&ctx)
7068 has been changed into
7069 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7071 To update applications using SSL_CTX_set_cert_verify_callback(),
7072 a dummy argument can be added to their callback functions.
7073 [D. K. Smetters <smetters@parc.xerox.com>]
7075 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7076 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7078 *) Add and OPENSSL_LOAD_CONF define which will cause
7079 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7080 This allows older applications to transparently support certain
7081 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7082 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7083 load the config file and OPENSSL_add_all_algorithms_conf() which will
7084 always load it have also been added.
7087 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7088 Adjust NIDs and EVP layer.
7089 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7091 *) Config modules support in openssl utility.
7093 Most commands now load modules from the config file,
7094 though in a few (such as version) this isn't done
7095 because it couldn't be used for anything.
7097 In the case of ca and req the config file used is
7098 the same as the utility itself: that is the -config
7099 command line option can be used to specify an
7103 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7104 use "openssl_conf" if filename is NULL use default openssl config file.
7107 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7108 config section name. Add a new flag to tolerate a missing config file
7109 and move code to CONF_modules_load_file().
7112 *) Support for crypto accelerator cards from Accelerated Encryption
7113 Processing, www.aep.ie. (Use engine 'aep')
7114 The support was copied from 0.9.6c [engine] and adapted/corrected
7115 to work with the new engine framework.
7116 [AEP Inc. and Richard Levitte]
7118 *) Support for SureWare crypto accelerator cards from Baltimore
7119 Technologies. (Use engine 'sureware')
7120 The support was copied from 0.9.6c [engine] and adapted
7121 to work with the new engine framework.
7124 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7125 make the newer ENGINE framework commands for the CHIL engine work.
7126 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7128 *) Make it possible to produce shared libraries on ReliantUNIX.
7129 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7131 *) Add the configuration target debug-linux-ppro.
7132 Make 'openssl rsa' use the general key loading routines
7133 implemented in apps.c, and make those routines able to
7134 handle the key format FORMAT_NETSCAPE and the variant
7136 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7138 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7139 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7141 *) Add -keyform to rsautl, and document -engine.
7142 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7144 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7145 BIO_R_NO_SUCH_FILE error code rather than the generic
7146 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7149 *) Add new functions
7151 ERR_peek_last_error_line
7152 ERR_peek_last_error_line_data.
7153 These are similar to
7156 ERR_peek_error_line_data,
7157 but report on the latest error recorded rather than the first one
7158 still in the error queue.
7159 [Ben Laurie, Bodo Moeller]
7161 *) default_algorithms option in ENGINE config module. This allows things
7163 default_algorithms = ALL
7164 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7167 *) Preliminary ENGINE config module.
7170 *) New experimental application configuration code.
7173 *) Change the AES code to follow the same name structure as all other
7174 symmetric ciphers, and behave the same way. Move everything to
7175 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7176 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7178 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7179 [Ben Laurie and Theo de Raadt]
7181 *) Add option to output public keys in req command.
7182 [Massimiliano Pala madwolf@openca.org]
7184 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7185 (up to about 10% better than before for P-192 and P-224).
7188 *) New functions/macros
7190 SSL_CTX_set_msg_callback(ctx, cb)
7191 SSL_CTX_set_msg_callback_arg(ctx, arg)
7192 SSL_set_msg_callback(ssl, cb)
7193 SSL_set_msg_callback_arg(ssl, arg)
7195 to request calling a callback function
7197 void cb(int write_p, int version, int content_type,
7198 const void *buf, size_t len, SSL *ssl, void *arg)
7200 whenever a protocol message has been completely received
7201 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7202 protocol version according to which the SSL library interprets
7203 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7204 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7205 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7206 specification (change_cipher_spec(20), alert(21), handshake(22)).
7207 'buf' and 'len' point to the actual message, 'ssl' to the
7208 SSL object, and 'arg' is the application-defined value set by
7209 SSL[_CTX]_set_msg_callback_arg().
7211 'openssl s_client' and 'openssl s_server' have new '-msg' options
7212 to enable a callback that displays all protocol messages.
7215 *) Change the shared library support so shared libraries are built as
7216 soon as the corresponding static library is finished, and thereby get
7217 openssl and the test programs linked against the shared library.
7218 This still only happens when the keyword "shard" has been given to
7219 the configuration scripts.
7221 NOTE: shared library support is still an experimental thing, and
7222 backward binary compatibility is still not guaranteed.
7223 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7225 *) Add support for Subject Information Access extension.
7226 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7228 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7229 additional bytes when new memory had to be allocated, not just
7230 when reusing an existing buffer.
7233 *) New command line and configuration option 'utf8' for the req command.
7234 This allows field values to be specified as UTF8 strings.
7237 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7238 runs for the former and machine-readable output for the latter.
7241 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7242 of the e-mail address in the DN (i.e., it will go into a certificate
7243 extension only). The new configuration file option 'email_in_dn = no'
7244 has the same effect.
7245 [Massimiliano Pala madwolf@openca.org]
7247 *) Change all functions with names starting with des_ to be starting
7248 with DES_ instead. Add wrappers that are compatible with libdes,
7249 but are named _ossl_old_des_*. Finally, add macros that map the
7250 des_* symbols to the corresponding _ossl_old_des_* if libdes
7251 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7252 desired, the des_* symbols will be mapped to DES_*, with one
7255 Since we provide two compatibility mappings, the user needs to
7256 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7257 compatibility is desired. The default (i.e., when that macro
7258 isn't defined) is OpenSSL 0.9.6c compatibility.
7260 There are also macros that enable and disable the support of old
7261 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7262 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7263 are defined, the default will apply: to support the old des routines.
7265 In either case, one must include openssl/des.h to get the correct
7266 definitions. Do not try to just include openssl/des_old.h, that
7269 NOTE: This is a major break of an old API into a new one. Software
7270 authors are encouraged to switch to the DES_ style functions. Some
7271 time in the future, des_old.h and the libdes compatibility functions
7272 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7273 default), and then completely removed.
7276 *) Test for certificates which contain unsupported critical extensions.
7277 If such a certificate is found during a verify operation it is
7278 rejected by default: this behaviour can be overridden by either
7279 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7280 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7281 X509_supported_extension() has also been added which returns 1 if a
7282 particular extension is supported.
7285 *) Modify the behaviour of EVP cipher functions in similar way to digests
7286 to retain compatibility with existing code.
7289 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7290 compatibility with existing code. In particular the 'ctx' parameter does
7291 not have to be to be initialized before the call to EVP_DigestInit() and
7292 it is tidied up after a call to EVP_DigestFinal(). New function
7293 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7294 EVP_MD_CTX_copy() changed to not require the destination to be
7295 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7296 requires the destination to be valid.
7298 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7299 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7302 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7303 so that complete 'Handshake' protocol structures are kept in memory
7304 instead of overwriting 'msg_type' and 'length' with 'body' data.
7307 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7308 [Massimo Santin via Richard Levitte]
7310 *) Major restructuring to the underlying ENGINE code. This includes
7311 reduction of linker bloat, separation of pure "ENGINE" manipulation
7312 (initialisation, etc) from functionality dealing with implementations
7313 of specific crypto interfaces. This change also introduces integrated
7314 support for symmetric ciphers and digest implementations - so ENGINEs
7315 can now accelerate these by providing EVP_CIPHER and EVP_MD
7316 implementations of their own. This is detailed in crypto/engine/README
7317 as it couldn't be adequately described here. However, there are a few
7318 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7319 were changed in the original introduction of ENGINE code have now
7320 reverted back - the hooking from this code to ENGINE is now a good
7321 deal more passive and at run-time, operations deal directly with
7322 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7323 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7324 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7325 they were not being used by the framework as there is no concept of a
7326 BIGNUM_METHOD and they could not be generalised to the new
7327 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7328 ENGINE_cpy() has been removed as it cannot be consistently defined in
7332 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7335 *) Change mkdef.pl to sort symbols that get the same entry number,
7336 and make sure the automatically generated functions ERR_load_*
7337 become part of libeay.num as well.
7340 *) New function SSL_renegotiate_pending(). This returns true once
7341 renegotiation has been requested (either SSL_renegotiate() call
7342 or HelloRequest/ClientHello received from the peer) and becomes
7343 false once a handshake has been completed.
7344 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7345 sends a HelloRequest, but does not ensure that a handshake takes
7346 place. SSL_renegotiate_pending() is useful for checking if the
7347 client has followed the request.)
7350 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7351 By default, clients may request session resumption even during
7352 renegotiation (if session ID contexts permit); with this option,
7353 session resumption is possible only in the first handshake.
7355 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7356 more bits available for options that should not be part of
7357 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7360 *) Add some demos for certificate and certificate request creation.
7363 *) Make maximum certificate chain size accepted from the peer application
7364 settable (SSL*_get/set_max_cert_list()), as proposed by
7365 "Douglas E. Engert" <deengert@anl.gov>.
7368 *) Add support for shared libraries for Unixware-7
7369 (Boyd Lynn Gerber <gerberb@zenez.com>).
7372 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7373 be done prior to destruction. Use this to unload error strings from
7374 ENGINEs that load their own error strings. NB: This adds two new API
7375 functions to "get" and "set" this destroy handler in an ENGINE.
7378 *) Alter all existing ENGINE implementations (except "openssl" and
7379 "openbsd") to dynamically instantiate their own error strings. This
7380 makes them more flexible to be built both as statically-linked ENGINEs
7381 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7382 Also, add stub code to each that makes building them as self-contained
7383 shared-libraries easier (see README.ENGINE).
7386 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7387 implementations into applications that are completely implemented in
7388 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7389 commands that can be used to configure what shared-library to load and
7390 to control aspects of the way it is handled. Also, made an update to
7391 the README.ENGINE file that brings its information up-to-date and
7392 provides some information and instructions on the "dynamic" ENGINE
7393 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7396 *) Make it possible to unload ranges of ERR strings with a new
7397 "ERR_unload_strings" function.
7400 *) Add a copy() function to EVP_MD.
7403 *) Make EVP_MD routines take a context pointer instead of just the
7404 md_data void pointer.
7407 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7408 that the digest can only process a single chunk of data
7409 (typically because it is provided by a piece of
7410 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7411 is only going to provide a single chunk of data, and hence the
7412 framework needn't accumulate the data for oneshot drivers.
7415 *) As with "ERR", make it possible to replace the underlying "ex_data"
7416 functions. This change also alters the storage and management of global
7417 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7418 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7419 index counters. The API functions that use this state have been changed
7420 to take a "class_index" rather than pointers to the class's local STACK
7421 and counter, and there is now an API function to dynamically create new
7422 classes. This centralisation allows us to (a) plug a lot of the
7423 thread-safety problems that existed, and (b) makes it possible to clean
7424 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7425 such data would previously have always leaked in application code and
7426 workarounds were in place to make the memory debugging turn a blind eye
7427 to it. Application code that doesn't use this new function will still
7428 leak as before, but their memory debugging output will announce it now
7429 rather than letting it slide.
7431 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7432 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7433 has a return value to indicate success or failure.
7436 *) Make it possible to replace the underlying "ERR" functions such that the
7437 global state (2 LHASH tables and 2 locks) is only used by the "default"
7438 implementation. This change also adds two functions to "get" and "set"
7439 the implementation prior to it being automatically set the first time
7440 any other ERR function takes place. Ie. an application can call "get",
7441 pass the return value to a module it has just loaded, and that module
7442 can call its own "set" function using that value. This means the
7443 module's "ERR" operations will use (and modify) the error state in the
7444 application and not in its own statically linked copy of OpenSSL code.
7447 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7448 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7449 the operation, and provides a more encapsulated way for external code
7450 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7451 to use these functions rather than manually incrementing the counts.
7453 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7456 *) Add EVP test program.
7459 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7462 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7463 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7464 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7465 These allow a CRL to be built without having to access X509_CRL fields
7466 directly. Modify 'ca' application to use new functions.
7469 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7470 bug workarounds. Rollback attack detection is a security feature.
7471 The problem will only arise on OpenSSL servers when TLSv1 is not
7472 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7473 Software authors not wanting to support TLSv1 will have special reasons
7474 for their choice and can explicitly enable this option.
7475 [Bodo Moeller, Lutz Jaenicke]
7477 *) Rationalise EVP so it can be extended: don't include a union of
7478 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7479 (similar to those existing for EVP_CIPHER_CTX).
7484 EVP_MD_CTX_init(&md); /* new function call */
7485 EVP_DigestInit(&md, EVP_sha1());
7486 EVP_DigestUpdate(&md, in, len);
7487 EVP_DigestFinal(&md, out, NULL);
7488 EVP_MD_CTX_cleanup(&md); /* new function call */
7492 *) Make DES key schedule conform to the usual scheme, as well as
7493 correcting its structure. This means that calls to DES functions
7494 now have to pass a pointer to a des_key_schedule instead of a
7495 plain des_key_schedule (which was actually always a pointer
7498 des_key_schedule ks;
7500 des_set_key_checked(..., &ks);
7501 des_ncbc_encrypt(..., &ks, ...);
7503 (Note that a later change renames 'des_...' into 'DES_...'.)
7506 *) Initial reduction of linker bloat: the use of some functions, such as
7507 PEM causes large amounts of unused functions to be linked in due to
7508 poor organisation. For example pem_all.c contains every PEM function
7509 which has a knock on effect of linking in large amounts of (unused)
7510 ASN1 code. Grouping together similar functions and splitting unrelated
7511 functions prevents this.
7514 *) Cleanup of EVP macros.
7517 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7518 correct _ecb suffix.
7521 *) Add initial OCSP responder support to ocsp application. The
7522 revocation information is handled using the text based index
7523 use by the ca application. The responder can either handle
7524 requests generated internally, supplied in files (for example
7525 via a CGI script) or using an internal minimal server.
7528 *) Add configuration choices to get zlib compression for TLS.
7531 *) Changes to Kerberos SSL for RFC 2712 compliance:
7532 1. Implemented real KerberosWrapper, instead of just using
7533 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7534 2. Implemented optional authenticator field of KerberosWrapper.
7536 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7537 and authenticator structs; see crypto/krb5/.
7539 Generalized Kerberos calls to support multiple Kerberos libraries.
7540 [Vern Staats <staatsvr@asc.hpc.mil>,
7541 Jeffrey Altman <jaltman@columbia.edu>
7542 via Richard Levitte]
7544 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7545 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7546 values for each of the key sizes rather than having just
7547 parameters (and 'speed' generating keys each time).
7550 *) Speed up EVP routines.
7553 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7554 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7555 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7556 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7558 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7559 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7560 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7563 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7565 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7568 *) Added the OS2-EMX target.
7569 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7571 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7572 to support NCONF routines in extension code. New function CONF_set_nconf()
7573 to allow functions which take an NCONF to also handle the old LHASH
7574 structure: this means that the old CONF compatible routines can be
7575 retained (in particular wrt extensions) without having to duplicate the
7576 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7579 *) Enhance the general user interface with mechanisms for inner control
7580 and with possibilities to have yes/no kind of prompts.
7583 *) Change all calls to low level digest routines in the library and
7584 applications to use EVP. Add missing calls to HMAC_cleanup() and
7585 don't assume HMAC_CTX can be copied using memcpy().
7586 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7588 *) Add the possibility to control engines through control names but with
7589 arbitrary arguments instead of just a string.
7590 Change the key loaders to take a UI_METHOD instead of a callback
7591 function pointer. NOTE: this breaks binary compatibility with earlier
7592 versions of OpenSSL [engine].
7593 Adapt the nCipher code for these new conditions and add a card insertion
7597 *) Enhance the general user interface with mechanisms to better support
7598 dialog box interfaces, application-defined prompts, the possibility
7599 to use defaults (for example default passwords from somewhere else)
7600 and interrupts/cancellations.
7603 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7604 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7607 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7608 tidy up some unnecessarily weird code in 'sk_new()').
7609 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7611 *) Change the key loading routines for ENGINEs to use the same kind
7612 callback (pem_password_cb) as all other routines that need this
7616 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7617 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7618 than this minimum value is recommended.
7621 *) New random seeder for OpenVMS, using the system process statistics
7622 that are easily reachable.
7625 *) Windows apparently can't transparently handle global
7626 variables defined in DLLs. Initialisations such as:
7628 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7630 won't compile. This is used by the any applications that need to
7631 declare their own ASN1 modules. This was fixed by adding the option
7632 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7633 needed for static libraries under Win32.
7636 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7637 setting of purpose and trust fields. New X509_STORE trust and
7638 purpose functions and tidy up setting in other SSL functions.
7641 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7642 structure. These are inherited by X509_STORE_CTX when it is
7643 initialised. This allows various defaults to be set in the
7644 X509_STORE structure (such as flags for CRL checking and custom
7645 purpose or trust settings) for functions which only use X509_STORE_CTX
7646 internally such as S/MIME.
7648 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7649 trust settings if they are not set in X509_STORE. This allows X509_STORE
7650 purposes and trust (in S/MIME for example) to override any set by default.
7652 Add command line options for CRL checking to smime, s_client and s_server
7656 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7657 are set then the CRL is looked up in the X509_STORE structure and
7658 its validity and signature checked, then if the certificate is found
7659 in the CRL the verify fails with a revoked error.
7661 Various new CRL related callbacks added to X509_STORE_CTX structure.
7663 Command line options added to 'verify' application to support this.
7665 This needs some additional work, such as being able to handle multiple
7666 CRLs with different times, extension based lookup (rather than just
7667 by subject name) and ultimately more complete V2 CRL extension
7671 *) Add a general user interface API (crypto/ui/). This is designed
7672 to replace things like des_read_password and friends (backward
7673 compatibility functions using this new API are provided).
7674 The purpose is to remove prompting functions from the DES code
7675 section as well as provide for prompting through dialog boxes in
7676 a window system and the like.
7679 *) Add "ex_data" support to ENGINE so implementations can add state at a
7680 per-structure level rather than having to store it globally.
7683 *) Make it possible for ENGINE structures to be copied when retrieved by
7684 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7685 This causes the "original" ENGINE structure to act like a template,
7686 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7687 operational state can be localised to each ENGINE structure, despite the
7688 fact they all share the same "methods". New ENGINE structures returned in
7689 this case have no functional references and the return value is the single
7690 structural reference. This matches the single structural reference returned
7691 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7695 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7696 needs to match any other type at all we need to manually clear the
7700 *) Changes to the "openssl engine" utility to include;
7701 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7702 about an ENGINE's available control commands.
7703 - executing control commands from command line arguments using the
7704 '-pre' and '-post' switches. '-post' is only used if '-t' is
7705 specified and the ENGINE is successfully initialised. The syntax for
7706 the individual commands are colon-separated, for example;
7707 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7710 *) New dynamic control command support for ENGINEs. ENGINEs can now
7711 declare their own commands (numbers), names (strings), descriptions,
7712 and input types for run-time discovery by calling applications. A
7713 subset of these commands are implicitly classed as "executable"
7714 depending on their input type, and only these can be invoked through
7715 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7716 can be based on user input, config files, etc). The distinction is
7717 that "executable" commands cannot return anything other than a boolean
7718 result and can only support numeric or string input, whereas some
7719 discoverable commands may only be for direct use through
7720 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7721 pointers, or other custom uses. The "executable" commands are to
7722 support parameterisations of ENGINE behaviour that can be
7723 unambiguously defined by ENGINEs and used consistently across any
7724 OpenSSL-based application. Commands have been added to all the
7725 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7726 control over shared-library paths without source code alterations.
7729 *) Changed all ENGINE implementations to dynamically allocate their
7730 ENGINEs rather than declaring them statically. Apart from this being
7731 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7732 this also allows the implementations to compile without using the
7733 internal engine_int.h header.
7736 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7737 'const' value. Any code that should be able to modify a RAND_METHOD
7738 should already have non-const pointers to it (ie. they should only
7739 modify their own ones).
7742 *) Made a variety of little tweaks to the ENGINE code.
7743 - "atalla" and "ubsec" string definitions were moved from header files
7744 to C code. "nuron" string definitions were placed in variables
7745 rather than hard-coded - allowing parameterisation of these values
7746 later on via ctrl() commands.
7747 - Removed unused "#if 0"'d code.
7748 - Fixed engine list iteration code so it uses ENGINE_free() to release
7749 structural references.
7750 - Constified the RAND_METHOD element of ENGINE structures.
7751 - Constified various get/set functions as appropriate and added
7752 missing functions (including a catch-all ENGINE_cpy that duplicates
7753 all ENGINE values onto a new ENGINE except reference counts/state).
7754 - Removed NULL parameter checks in get/set functions. Setting a method
7755 or function to NULL is a way of cancelling out a previously set
7756 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7757 and doesn't justify the extra error symbols and code.
7758 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7759 flags from engine_int.h to engine.h.
7760 - Changed prototypes for ENGINE handler functions (init(), finish(),
7761 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7764 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7765 to the algorithm using long division. The binary algorithm can be
7766 used only if the modulus is odd. On 32-bit systems, it is faster
7767 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7768 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7769 up to 450 bits. In 64-bit environments, the binary algorithm
7770 appears to be advantageous for much longer moduli; here we use it
7771 for moduli up to 2048 bits.
7774 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7775 could not support the combine flag in choice fields.
7778 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7779 extensions from a certificate request to the certificate.
7782 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7783 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7784 file: this allows the display of the certificate about to be
7785 signed to be customised, to allow certain fields to be included
7786 or excluded and extension details. The old system didn't display
7787 multicharacter strings properly, omitted fields not in the policy
7788 and couldn't display additional details such as extensions.
7791 *) Function EC_POINTs_mul for multiple scalar multiplication
7792 of an arbitrary number of elliptic curve points
7793 \sum scalars[i]*points[i],
7794 optionally including the generator defined for the EC_GROUP:
7795 scalar*generator + \sum scalars[i]*points[i].
7797 EC_POINT_mul is a simple wrapper function for the typical case
7798 that the point list has just one item (besides the optional
7802 *) First EC_METHODs for curves over GF(p):
7804 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7805 operations and provides various method functions that can also
7806 operate with faster implementations of modular arithmetic.
7808 EC_GFp_mont_method() reuses most functions that are part of
7809 EC_GFp_simple_method, but uses Montgomery arithmetic.
7811 [Bodo Moeller; point addition and point doubling
7812 implementation directly derived from source code provided by
7813 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7815 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7816 crypto/ec/ec_lib.c):
7818 Curves are EC_GROUP objects (with an optional group generator)
7819 based on EC_METHODs that are built into the library.
7821 Points are EC_POINT objects based on EC_GROUP objects.
7823 Most of the framework would be able to handle curves over arbitrary
7824 finite fields, but as there are no obvious types for fields other
7825 than GF(p), some functions are limited to that for now.
7828 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7829 that the file contains a complete HTTP response.
7832 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7833 change the def and num file printf format specifier from "%-40sXXX"
7834 to "%-39s XXX". The latter will always guarantee a space after the
7835 field while the former will cause them to run together if the field
7836 is 40 of more characters long.
7839 *) Constify the cipher and digest 'method' functions and structures
7840 and modify related functions to take constant EVP_MD and EVP_CIPHER
7844 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7845 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7848 *) Modify EVP_Digest*() routines so they now return values. Although the
7849 internal software routines can never fail additional hardware versions
7853 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7855 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7856 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7859 ERR_R_NESTED_ASN1_ERROR
7861 ERR_R_MISSING_ASN1_EOS
7862 were 4 .. 9, conflicting with
7863 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7865 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7866 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7868 Add new error code 'ERR_R_INTERNAL_ERROR'.
7871 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7875 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7876 sets the subject name for a new request or supersedes the
7877 subject name in a given request. Formats that can be parsed are
7878 'CN=Some Name, OU=myOU, C=IT'
7880 'CN=Some Name/OU=myOU/C=IT'.
7882 Add options '-batch' and '-verbose' to 'openssl req'.
7883 [Massimiliano Pala <madwolf@hackmasters.net>]
7885 *) Introduce the possibility to access global variables through
7886 functions on platform were that's the best way to handle exporting
7887 global variables in shared libraries. To enable this functionality,
7888 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7889 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7890 is normally done by Configure or something similar).
7892 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7893 in the source file (foo.c) like this:
7895 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7896 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7898 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7899 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7901 OPENSSL_DECLARE_GLOBAL(int,foo);
7902 #define foo OPENSSL_GLOBAL_REF(foo)
7903 OPENSSL_DECLARE_GLOBAL(double,bar);
7904 #define bar OPENSSL_GLOBAL_REF(bar)
7906 The #defines are very important, and therefore so is including the
7907 header file everywhere where the defined globals are used.
7909 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7910 of ASN.1 items, but that structure is a bit different.
7912 The largest change is in util/mkdef.pl which has been enhanced with
7913 better and easier to understand logic to choose which symbols should
7914 go into the Windows .def files as well as a number of fixes and code
7915 cleanup (among others, algorithm keywords are now sorted
7916 lexicographically to avoid constant rewrites).
7919 *) In BN_div() keep a copy of the sign of 'num' before writing the
7920 result to 'rm' because if rm==num the value will be overwritten
7921 and produce the wrong result if 'num' is negative: this caused
7922 problems with BN_mod() and BN_nnmod().
7925 *) Function OCSP_request_verify(). This checks the signature on an
7926 OCSP request and verifies the signer certificate. The signer
7927 certificate is just checked for a generic purpose and OCSP request
7931 *) Add OCSP_check_validity() function to check the validity of OCSP
7932 responses. OCSP responses are prepared in real time and may only
7933 be a few seconds old. Simply checking that the current time lies
7934 between thisUpdate and nextUpdate max reject otherwise valid responses
7935 caused by either OCSP responder or client clock inaccuracy. Instead
7936 we allow thisUpdate and nextUpdate to fall within a certain period of
7937 the current time. The age of the response can also optionally be
7938 checked. Two new options -validity_period and -status_age added to
7942 *) If signature or public key algorithm is unrecognized print out its
7943 OID rather that just UNKNOWN.
7946 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7947 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7948 ID to be generated from the issuer certificate alone which can then be
7949 passed to OCSP_id_issuer_cmp().
7952 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7953 ASN1 modules to export functions returning ASN1_ITEM pointers
7954 instead of the ASN1_ITEM structures themselves. This adds several
7955 new macros which allow the underlying ASN1 function/structure to
7956 be accessed transparently. As a result code should not use ASN1_ITEM
7957 references directly (such as &X509_it) but instead use the relevant
7958 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7959 use of the new ASN1 code on platforms where exporting structures
7960 is problematical (for example in shared libraries) but exporting
7961 functions returning pointers to structures is not.
7964 *) Add support for overriding the generation of SSL/TLS session IDs.
7965 These callbacks can be registered either in an SSL_CTX or per SSL.
7966 The purpose of this is to allow applications to control, if they wish,
7967 the arbitrary values chosen for use as session IDs, particularly as it
7968 can be useful for session caching in multiple-server environments. A
7969 command-line switch for testing this (and any client code that wishes
7970 to use such a feature) has been added to "s_server".
7971 [Geoff Thorpe, Lutz Jaenicke]
7973 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7974 of the form '#if defined(...) || defined(...) || ...' and
7975 '#if !defined(...) && !defined(...) && ...'. This also avoids
7976 the growing number of special cases it was previously handling.
7979 *) Make all configuration macros available for application by making
7980 sure they are available in opensslconf.h, by giving them names starting
7981 with "OPENSSL_" to avoid conflicts with other packages and by making
7982 sure e_os2.h will cover all platform-specific cases together with
7984 Additionally, it is now possible to define configuration/platform-
7985 specific names (called "system identities"). In the C code, these
7986 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7987 macro with the name beginning with "OPENSSL_SYS_", which is determined
7988 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7992 *) New option -set_serial to 'req' and 'x509' this allows the serial
7993 number to use to be specified on the command line. Previously self
7994 signed certificates were hard coded with serial number 0 and the
7995 CA options of 'x509' had to use a serial number in a file which was
7999 *) New options to 'ca' utility to support V2 CRL entry extensions.
8000 Currently CRL reason, invalidity date and hold instruction are
8001 supported. Add new CRL extensions to V3 code and some new objects.
8004 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8005 disable standard block padding (aka PKCS#5 padding) in the EVP
8006 API, which was previously mandatory. This means that the data is
8007 not padded in any way and so the total length much be a multiple
8008 of the block size, otherwise an error occurs.
8011 *) Initial (incomplete) OCSP SSL support.
8014 *) New function OCSP_parse_url(). This splits up a URL into its host,
8015 port and path components: primarily to parse OCSP URLs. New -url
8016 option to ocsp utility.
8019 *) New nonce behavior. The return value of OCSP_check_nonce() now
8020 reflects the various checks performed. Applications can decide
8021 whether to tolerate certain situations such as an absent nonce
8022 in a response when one was present in a request: the ocsp application
8023 just prints out a warning. New function OCSP_add1_basic_nonce()
8024 this is to allow responders to include a nonce in a response even if
8025 the request is nonce-less.
8028 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8029 skipped when using openssl x509 multiple times on a single input file,
8030 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8033 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8034 set string type: to handle setting ASN1_TIME structures. Fix ca
8035 utility to correctly initialize revocation date of CRLs.
8038 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8039 the clients preferred ciphersuites and rather use its own preferences.
8040 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8041 Internet Explorer by ensuring unchanged hash method during stepup.
8042 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8045 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8046 to aes and add a new 'exist' option to print out symbols that don't
8050 *) Additional options to ocsp utility to allow flags to be set and
8051 additional certificates supplied.
8054 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8055 OCSP client a number of certificate to only verify the response
8059 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8060 handle the new API. Currently only ECB, CBC modes supported. Add new
8063 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8064 Encryption Standard (AES) Ciphersuites for Transport Layer
8065 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8066 not enabled by default and were not part of the "ALL" ciphersuite
8067 alias because they were not yet official; they could be
8068 explicitly requested by specifying the "AESdraft" ciphersuite
8069 group alias. In the final release of OpenSSL 0.9.7, the group
8070 alias is called "AES" and is part of "ALL".)
8071 [Ben Laurie, Steve Henson, Bodo Moeller]
8073 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8074 request to response.
8077 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8078 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8079 extract information from a certificate request. OCSP_response_create()
8080 creates a response and optionally adds a basic response structure.
8081 OCSP_basic_add1_status() adds a complete single response to a basic
8082 response and returns the OCSP_SINGLERESP structure just added (to allow
8083 extensions to be included for example). OCSP_basic_add1_cert() adds a
8084 certificate to a basic response and OCSP_basic_sign() signs a basic
8085 response with various flags. New helper functions ASN1_TIME_check()
8086 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8087 (converts ASN1_TIME to GeneralizedTime).
8090 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8091 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8092 structure from a certificate. X509_pubkey_digest() digests the public_key
8093 contents: this is used in various key identifiers.
8096 *) Make sk_sort() tolerate a NULL argument.
8097 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8099 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8100 passed by the function are trusted implicitly. If any of them signed the
8101 response then it is assumed to be valid and is not verified.
8104 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8105 to data. This was previously part of the PKCS7 ASN1 code. This
8106 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8107 [Steve Henson, reported by Kenneth R. Robinette
8108 <support@securenetterm.com>]
8110 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8111 routines: without these tracing memory leaks is very painful.
8112 Fix leaks in PKCS12 and PKCS7 routines.
8115 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8116 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8117 effectively meant GeneralizedTime would never be used. Now it
8118 is initialised to -1 but X509_time_adj() now has to check the value
8119 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8120 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8121 [Steve Henson, reported by Kenneth R. Robinette
8122 <support@securenetterm.com>]
8124 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8125 result in a zero length in the ASN1_INTEGER structure which was
8126 not consistent with the structure when d2i_ASN1_INTEGER() was used
8127 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8128 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8129 where it did not print out a minus for negative ASN1_INTEGER.
8132 *) Add summary printout to ocsp utility. The various functions which
8133 convert status values to strings have been renamed to:
8134 OCSP_response_status_str(), OCSP_cert_status_str() and
8135 OCSP_crl_reason_str() and are no longer static. New options
8136 to verify nonce values and to disable verification. OCSP response
8137 printout format cleaned up.
8140 *) Add additional OCSP certificate checks. These are those specified
8141 in RFC2560. This consists of two separate checks: the CA of the
8142 certificate being checked must either be the OCSP signer certificate
8143 or the issuer of the OCSP signer certificate. In the latter case the
8144 OCSP signer certificate must contain the OCSP signing extended key
8145 usage. This check is performed by attempting to match the OCSP
8146 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8147 in the OCSP_CERTID structures of the response.
8150 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8151 and related routines. This uses the standard OpenSSL certificate
8152 verify routines to perform initial checks (just CA validity) and
8153 to obtain the certificate chain. Then additional checks will be
8154 performed on the chain. Currently the root CA is checked to see
8155 if it is explicitly trusted for OCSP signing. This is used to set
8156 a root CA as a global signing root: that is any certificate that
8157 chains to that CA is an acceptable OCSP signing certificate.
8160 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8161 extensions from a separate configuration file.
8162 As when reading extensions from the main configuration file,
8163 the '-extensions ...' option may be used for specifying the
8165 [Massimiliano Pala <madwolf@comune.modena.it>]
8167 *) New OCSP utility. Allows OCSP requests to be generated or
8168 read. The request can be sent to a responder and the output
8169 parsed, outputed or printed in text form. Not complete yet:
8170 still needs to check the OCSP response validity.
8173 *) New subcommands for 'openssl ca':
8174 'openssl ca -status <serial>' prints the status of the cert with
8175 the given serial number (according to the index file).
8176 'openssl ca -updatedb' updates the expiry status of certificates
8178 [Massimiliano Pala <madwolf@comune.modena.it>]
8180 *) New '-newreq-nodes' command option to CA.pl. This is like
8181 '-newreq', but calls 'openssl req' with the '-nodes' option
8182 so that the resulting key is not encrypted.
8183 [Damien Miller <djm@mindrot.org>]
8185 *) New configuration for the GNU Hurd.
8186 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8188 *) Initial code to implement OCSP basic response verify. This
8189 is currently incomplete. Currently just finds the signer's
8190 certificate and verifies the signature on the response.
8193 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8194 value of OPENSSLDIR. This is available via the new '-d' option
8195 to 'openssl version', and is also included in 'openssl version -a'.
8198 *) Allowing defining memory allocation callbacks that will be given
8199 file name and line number information in additional arguments
8200 (a const char* and an int). The basic functionality remains, as
8201 well as the original possibility to just replace malloc(),
8202 realloc() and free() by functions that do not know about these
8203 additional arguments. To register and find out the current
8204 settings for extended allocation functions, the following
8205 functions are provided:
8207 CRYPTO_set_mem_ex_functions
8208 CRYPTO_set_locked_mem_ex_functions
8209 CRYPTO_get_mem_ex_functions
8210 CRYPTO_get_locked_mem_ex_functions
8212 These work the same way as CRYPTO_set_mem_functions and friends.
8213 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8214 extended allocation function is enabled.
8215 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8216 a conventional allocation function is enabled.
8217 [Richard Levitte, Bodo Moeller]
8219 *) Finish off removing the remaining LHASH function pointer casts.
8220 There should no longer be any prototype-casting required when using
8221 the LHASH abstraction, and any casts that remain are "bugs". See
8222 the callback types and macros at the head of lhash.h for details
8223 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8226 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8227 If /dev/[u]random devices are not available or do not return enough
8228 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8230 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8231 /etc/entropy will be queried once each in this sequence, querying stops
8232 when enough entropy was collected without querying more sockets.
8235 *) Change the Unix RAND_poll() variant to be able to poll several
8236 random devices, as specified by DEVRANDOM, until a sufficient amount
8237 of data has been collected. We spend at most 10 ms on each file
8238 (select timeout) and read in non-blocking mode. DEVRANDOM now
8239 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8240 (previously it was just the string "/dev/urandom"), so on typical
8241 platforms the 10 ms delay will never occur.
8242 Also separate out the Unix variant to its own file, rand_unix.c.
8243 For VMS, there's a currently-empty rand_vms.c.
8246 *) Move OCSP client related routines to ocsp_cl.c. These
8247 provide utility functions which an application needing
8248 to issue a request to an OCSP responder and analyse the
8249 response will typically need: as opposed to those which an
8250 OCSP responder itself would need which will be added later.
8252 OCSP_request_sign() signs an OCSP request with an API similar
8253 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8254 response. OCSP_response_get1_basic() extracts basic response
8255 from response. OCSP_resp_find_status(): finds and extracts status
8256 information from an OCSP_CERTID structure (which will be created
8257 when the request structure is built). These are built from lower
8258 level functions which work on OCSP_SINGLERESP structures but
8259 won't normally be used unless the application wishes to examine
8260 extensions in the OCSP response for example.
8262 Replace nonce routines with a pair of functions.
8263 OCSP_request_add1_nonce() adds a nonce value and optionally
8264 generates a random value. OCSP_check_nonce() checks the
8265 validity of the nonce in an OCSP response.
8268 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8269 This doesn't copy the supplied OCSP_CERTID and avoids the
8270 need to free up the newly created id. Change return type
8271 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8272 This can then be used to add extensions to the request.
8273 Deleted OCSP_request_new(), since most of its functionality
8274 is now in OCSP_REQUEST_new() (and the case insensitive name
8275 clash) apart from the ability to set the request name which
8276 will be added elsewhere.
8279 *) Update OCSP API. Remove obsolete extensions argument from
8280 various functions. Extensions are now handled using the new
8281 OCSP extension code. New simple OCSP HTTP function which
8282 can be used to send requests and parse the response.
8285 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8286 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8287 uses the special reorder version of SET OF to sort the attributes
8288 and reorder them to match the encoded order. This resolves a long
8289 standing problem: a verify on a PKCS7 structure just after signing
8290 it used to fail because the attribute order did not match the
8291 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8292 it uses the received order. This is necessary to tolerate some broken
8293 software that does not order SET OF. This is handled by encoding
8294 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8295 to produce the required SET OF.
8298 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8299 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8300 files to get correct declarations of the ASN.1 item variables.
8303 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8304 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8305 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8306 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8307 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8308 ASN1_ITEM and no wrapper functions.
8311 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8312 replace the old function pointer based I/O routines. Change most of
8313 the *_d2i_bio() and *_d2i_fp() functions to use these.
8316 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8317 lines, recognize more "algorithms" that can be deselected, and make
8318 it complain about algorithm deselection that isn't recognised.
8321 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8322 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8323 to use new functions. Add NO_ASN1_OLD which can be set to remove
8324 some old style ASN1 functions: this can be used to determine if old
8325 code will still work when these eventually go away.
8328 *) New extension functions for OCSP structures, these follow the
8329 same conventions as certificates and CRLs.
8332 *) New function X509V3_add1_i2d(). This automatically encodes and
8333 adds an extension. Its behaviour can be customised with various
8334 flags to append, replace or delete. Various wrappers added for
8335 certificates and CRLs.
8338 *) Fix to avoid calling the underlying ASN1 print routine when
8339 an extension cannot be parsed. Correct a typo in the
8340 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8343 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8344 entries for variables.
8347 *) Add functionality to apps/openssl.c for detecting locking
8348 problems: As the program is single-threaded, all we have
8349 to do is register a locking callback using an array for
8350 storing which locks are currently held by the program.
8353 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8354 SSL_get_ex_data_X509_STORE_idx(), which is used in
8355 ssl_verify_cert_chain() and thus can be called at any time
8356 during TLS/SSL handshakes so that thread-safety is essential.
8357 Unfortunately, the ex_data design is not at all suited
8358 for multi-threaded use, so it probably should be abolished.
8361 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8362 [Broadcom, tweaked and integrated by Geoff Thorpe]
8364 *) Move common extension printing code to new function
8365 X509V3_print_extensions(). Reorganise OCSP print routines and
8366 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8369 *) New function X509_signature_print() to remove duplication in some
8373 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8374 set (this was treated exactly the same as SET OF previously). This
8375 is used to reorder the STACK representing the structure to match the
8376 encoding. This will be used to get round a problem where a PKCS7
8377 structure which was signed could not be verified because the STACK
8378 order did not reflect the encoded order.
8381 *) Reimplement the OCSP ASN1 module using the new code.
8384 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8385 for its ASN1 operations. The old style function pointers still exist
8386 for now but they will eventually go away.
8389 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8390 completely replaces the old ASN1 functionality with a table driven
8391 encoder and decoder which interprets an ASN1_ITEM structure describing
8392 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8393 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8394 has also been converted to the new form.
8397 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8398 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8399 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8400 for negative moduli.
8403 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8404 of not touching the result's sign bit.
8407 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8411 *) Changed the LHASH code to use prototypes for callbacks, and created
8412 macros to declare and implement thin (optionally static) functions
8413 that provide type-safety and avoid function pointer casting for the
8414 type-specific callbacks.
8417 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8419 [Veers Staats <staatsvr@asc.hpc.mil>,
8420 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8422 *) Reformat the FAQ so the different questions and answers can be divided
8423 in sections depending on the subject.
8426 *) Have the zlib compression code load ZLIB.DLL dynamically under
8430 *) New function BN_mod_sqrt for computing square roots modulo a prime
8431 (using the probabilistic Tonelli-Shanks algorithm unless
8432 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8433 be handled deterministically).
8434 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8436 *) Make BN_mod_inverse faster by explicitly handling small quotients
8437 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8438 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8441 *) New function BN_kronecker.
8444 *) Fix BN_gcd so that it works on negative inputs; the result is
8445 positive unless both parameters are zero.
8446 Previously something reasonably close to an infinite loop was
8447 possible because numbers could be growing instead of shrinking
8448 in the implementation of Euclid's algorithm.
8451 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8452 sign of the number in question.
8454 Fix BN_is_word(a,w) to work correctly for w == 0.
8456 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8457 because its test if the absolute value of 'a' equals 'w'.
8458 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8459 it exists mostly for use in the implementations of BN_is_zero(),
8460 BN_is_one(), and BN_is_word().
8463 *) New function BN_swap.
8466 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8467 the exponentiation functions are more likely to produce reasonable
8468 results on negative inputs.
8471 *) Change BN_mod_mul so that the result is always non-negative.
8472 Previously, it could be negative if one of the factors was negative;
8473 I don't think anyone really wanted that behaviour.
8476 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8477 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8478 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8479 and add new functions:
8488 BN_mod_lshift1_quick
8492 These functions always generate non-negative results.
8494 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8495 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8497 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8498 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8499 be reduced modulo m.
8500 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8503 The following entry accidentally appeared in the CHANGES file
8504 distributed with OpenSSL 0.9.7. The modifications described in
8505 it do *not* apply to OpenSSL 0.9.7.
8507 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8508 was actually never needed) and in BN_mul(). The removal in BN_mul()
8509 required a small change in bn_mul_part_recursive() and the addition
8510 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8511 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8512 bn_sub_words() and bn_add_words() except they take arrays with
8517 *) In 'openssl passwd', verify passwords read from the terminal
8518 unless the '-salt' option is used (which usually means that
8519 verification would just waste user's time since the resulting
8520 hash is going to be compared with some given password hash)
8521 or the new '-noverify' option is used.
8523 This is an incompatible change, but it does not affect
8524 non-interactive use of 'openssl passwd' (passwords on the command
8525 line, '-stdin' option, '-in ...' option) and thus should not
8529 *) Remove all references to RSAref, since there's no more need for it.
8532 *) Make DSO load along a path given through an environment variable
8533 (SHLIB_PATH) with shl_load().
8536 *) Constify the ENGINE code as a result of BIGNUM constification.
8537 Also constify the RSA code and most things related to it. In a
8538 few places, most notable in the depth of the ASN.1 code, ugly
8539 casts back to non-const were required (to be solved at a later
8543 *) Make it so the openssl application has all engines loaded by default.
8546 *) Constify the BIGNUM routines a little more.
8549 *) Add the following functions:
8551 ENGINE_load_cswift()
8553 ENGINE_load_atalla()
8555 ENGINE_load_builtin_engines()
8557 That way, an application can itself choose if external engines that
8558 are built-in in OpenSSL shall ever be used or not. The benefit is
8559 that applications won't have to be linked with libdl or other dso
8560 libraries unless it's really needed.
8562 Changed 'openssl engine' to load all engines on demand.
8563 Changed the engine header files to avoid the duplication of some
8564 declarations (they differed!).
8567 *) 'openssl engine' can now list capabilities.
8570 *) Better error reporting in 'openssl engine'.
8573 *) Never call load_dh_param(NULL) in s_server.
8576 *) Add engine application. It can currently list engines by name and
8577 identity, and test if they are actually available.
8580 *) Improve RPM specification file by forcing symbolic linking and making
8581 sure the installed documentation is also owned by root.root.
8582 [Damien Miller <djm@mindrot.org>]
8584 *) Give the OpenSSL applications more possibilities to make use of
8585 keys (public as well as private) handled by engines.
8588 *) Add OCSP code that comes from CertCo.
8591 *) Add VMS support for the Rijndael code.
8594 *) Added untested support for Nuron crypto accelerator.
8597 *) Add support for external cryptographic devices. This code was
8598 previously distributed separately as the "engine" branch.
8599 [Geoff Thorpe, Richard Levitte]
8601 *) Rework the filename-translation in the DSO code. It is now possible to
8602 have far greater control over how a "name" is turned into a filename
8603 depending on the operating environment and any oddities about the
8604 different shared library filenames on each system.
8607 *) Support threads on FreeBSD-elf in Configure.
8610 *) Fix for SHA1 assembly problem with MASM: it produces
8611 warnings about corrupt line number information when assembling
8612 with debugging information. This is caused by the overlapping
8614 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8617 NCONF_get_number() has no error checking at all. As a replacement,
8618 NCONF_get_number_e() is defined (_e for "error checking") and is
8619 promoted strongly. The old NCONF_get_number is kept around for
8620 binary backward compatibility.
8621 Make it possible for methods to load from something other than a BIO,
8622 by providing a function pointer that is given a name instead of a BIO.
8623 For example, this could be used to load configuration data from an
8627 *) Fix for non blocking accept BIOs. Added new I/O special reason
8628 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8629 with non blocking I/O was not possible because no retry code was
8630 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8634 *) Added the beginnings of Rijndael support.
8637 *) Fix for bug in DirectoryString mask setting. Add support for
8638 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8639 to allow certificate printing to more controllable, additional
8640 'certopt' option to 'x509' to allow new printing options to be
8644 *) Clean old EAY MD5 hack from e_os.h.
8647 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8649 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8650 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8651 [Joe Orton, Steve Henson]
8653 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8655 *) Fix additional bug revealed by the NISCC test suite:
8657 Stop bug triggering large recursion when presented with
8658 certain ASN.1 tags (CVE-2003-0851)
8661 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8663 *) Fix various bugs revealed by running the NISCC test suite:
8665 Stop out of bounds reads in the ASN1 code when presented with
8666 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8668 If verify callback ignores invalid public key errors don't try to check
8669 certificate signature with the NULL public key.
8673 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8674 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8678 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8679 extra data after the compression methods not only for TLS 1.0
8680 but also for SSL 3.0 (as required by the specification).
8681 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8683 *) Change X509_certificate_type() to mark the key as exported/exportable
8684 when it's 512 *bits* long, not 512 bytes.
8687 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8689 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8690 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8691 a protocol version number mismatch like a decryption error
8692 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8695 *) Turn on RSA blinding by default in the default implementation
8696 to avoid a timing attack. Applications that don't want it can call
8697 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8698 They would be ill-advised to do so in most cases.
8699 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8701 *) Change RSA blinding code so that it works when the PRNG is not
8702 seeded (in this case, the secret RSA exponent is abused as
8703 an unpredictable seed -- if it is not unpredictable, there
8704 is no point in blinding anyway). Make RSA blinding thread-safe
8705 by remembering the creator's thread ID in rsa->blinding and
8706 having all other threads use local one-time blinding factors
8707 (this requires more computation than sharing rsa->blinding, but
8708 avoids excessive locking; and if an RSA object is not shared
8709 between threads, blinding will still be very fast).
8712 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8714 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8715 via timing by performing a MAC computation even if incorrect
8716 block cipher padding has been found. This is a countermeasure
8717 against active attacks where the attacker has to distinguish
8718 between bad padding and a MAC verification error. (CVE-2003-0078)
8720 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8721 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8722 Martin Vuagnoux (EPFL, Ilion)]
8724 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8726 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8727 memory from it's contents. This is done with a counter that will
8728 place alternating values in each byte. This can be used to solve
8729 two issues: 1) the removal of calls to memset() by highly optimizing
8730 compilers, and 2) cleansing with other values than 0, since those can
8731 be read through on certain media, for example a swap space on disk.
8734 *) Bugfix: client side session caching did not work with external caching,
8735 because the session->cipher setting was not restored when reloading
8736 from the external cache. This problem was masked, when
8737 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8738 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8741 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8742 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8743 [Zeev Lieber <zeev-l@yahoo.com>]
8745 *) Undo an undocumented change introduced in 0.9.6e which caused
8746 repeated calls to OpenSSL_add_all_ciphers() and
8747 OpenSSL_add_all_digests() to be ignored, even after calling
8751 *) Change the default configuration reader to deal with last line not
8752 being properly terminated.
8755 *) Change X509_NAME_cmp() so it applies the special rules on handling
8756 DN values that are of type PrintableString, as well as RDNs of type
8757 emailAddress where the value has the type ia5String.
8758 [stefank@valicert.com via Richard Levitte]
8760 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8761 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8762 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8763 the bitwise-OR of the two for use by the majority of applications
8764 wanting this behaviour, and update the docs. The documented
8765 behaviour and actual behaviour were inconsistent and had been
8766 changing anyway, so this is more a bug-fix than a behavioural
8768 [Geoff Thorpe, diagnosed by Nadav Har'El]
8770 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8771 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8774 *) Fix initialization code race conditions in
8775 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8776 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8777 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8778 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8779 ssl2_get_cipher_by_char(),
8780 ssl3_get_cipher_by_char().
8781 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8783 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8784 the cached sessions are flushed, as the remove_cb() might use ex_data
8785 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8786 (see [openssl.org #212]).
8787 [Geoff Thorpe, Lutz Jaenicke]
8789 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8790 length, instead of the encoding length to d2i_ASN1_OBJECT.
8793 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8795 *) [In 0.9.6g-engine release:]
8796 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8797 [Lynn Gazis <lgazis@rainbow.com>]
8799 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8801 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8802 and get fix the header length calculation.
8803 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8804 Alon Kantor <alonk@checkpoint.com> (and others),
8807 *) Use proper error handling instead of 'assertions' in buffer
8808 overflow checks added in 0.9.6e. This prevents DoS (the
8809 assertions could call abort()).
8810 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8812 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8814 *) Add various sanity checks to asn1_get_length() to reject
8815 the ASN1 length bytes if they exceed sizeof(long), will appear
8816 negative or the content length exceeds the length of the
8818 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8820 *) Fix cipher selection routines: ciphers without encryption had no flags
8821 for the cipher strength set and where therefore not handled correctly
8822 by the selection routines (PR #130).
8825 *) Fix EVP_dsa_sha macro.
8829 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8830 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8831 that was added in OpenSSL 0.9.6d.
8833 As the countermeasure turned out to be incompatible with some
8834 broken SSL implementations, the new option is part of SSL_OP_ALL.
8835 SSL_OP_ALL is usually employed when compatibility with weird SSL
8836 implementations is desired (e.g. '-bugs' option to 's_client' and
8837 's_server'), so the new option is automatically set in many
8841 *) Changes in security patch:
8843 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8844 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8845 Air Force Materiel Command, USAF, under agreement number
8848 *) Add various sanity checks to asn1_get_length() to reject
8849 the ASN1 length bytes if they exceed sizeof(long), will appear
8850 negative or the content length exceeds the length of the
8851 supplied buffer. (CVE-2002-0659)
8852 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8854 *) Assertions for various potential buffer overflows, not known to
8856 [Ben Laurie (CHATS)]
8858 *) Various temporary buffers to hold ASCII versions of integers were
8859 too small for 64 bit platforms. (CVE-2002-0655)
8860 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8862 *) Remote buffer overflow in SSL3 protocol - an attacker could
8863 supply an oversized session ID to a client. (CVE-2002-0656)
8864 [Ben Laurie (CHATS)]
8866 *) Remote buffer overflow in SSL2 protocol - an attacker could
8867 supply an oversized client master key. (CVE-2002-0656)
8868 [Ben Laurie (CHATS)]
8870 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8872 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8873 encoded as NULL) with id-dsa-with-sha1.
8874 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8876 *) Check various X509_...() return values in apps/req.c.
8877 [Nils Larsch <nla@trustcenter.de>]
8879 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8880 an end-of-file condition would erroneously be flagged, when the CRLF
8881 was just at the end of a processed block. The bug was discovered when
8882 processing data through a buffering memory BIO handing the data to a
8883 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8884 <ptsekov@syntrex.com> and Nedelcho Stanev.
8887 *) Implement a countermeasure against a vulnerability recently found
8888 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8889 before application data chunks to avoid the use of known IVs
8890 with data potentially chosen by the attacker.
8893 *) Fix length checks in ssl3_get_client_hello().
8896 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8897 to prevent ssl3_read_internal() from incorrectly assuming that
8898 ssl3_read_bytes() found application data while handshake
8899 processing was enabled when in fact s->s3->in_read_app_data was
8900 merely automatically cleared during the initial handshake.
8901 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8903 *) Fix object definitions for Private and Enterprise: they were not
8904 recognized in their shortname (=lowercase) representation. Extend
8905 obj_dat.pl to issue an error when using undefined keywords instead
8906 of silently ignoring the problem (Svenning Sorensen
8907 <sss@sss.dnsalias.net>).
8910 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8911 generators, i.e. generators other than 2 and 5. (Previously, the
8912 code did not properly initialise the 'add' and 'rem' values to
8913 BN_generate_prime().)
8915 In the new general case, we do not insist that 'generator' is
8916 actually a primitive root: This requirement is rather pointless;
8917 a generator of the order-q subgroup is just as good, if not
8921 *) Map new X509 verification errors to alerts. Discovered and submitted by
8922 Tom Wu <tom@arcot.com>.
8925 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8926 returning non-zero before the data has been completely received
8927 when using non-blocking I/O.
8928 [Bodo Moeller; problem pointed out by John Hughes]
8930 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8931 [Ben Laurie, Lutz Jaenicke]
8933 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8934 Yoram Zahavi <YoramZ@gilian.com>).
8937 *) Add information about CygWin 1.3 and on, and preserve proper
8938 configuration for the versions before that.
8939 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8941 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8942 check whether we deal with a copy of a session and do not delete from
8943 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8944 <izhar@checkpoint.com>.
8947 *) Do not store session data into the internal session cache, if it
8948 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8949 flag is set). Proposed by Aslam <aslam@funk.com>.
8952 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8956 *) [In 0.9.6d-engine release:]
8957 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8958 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8960 *) Add the configuration target linux-s390x.
8961 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8963 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8964 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8965 variable as an indication that a ClientHello message has been
8966 received. As the flag value will be lost between multiple
8967 invocations of ssl3_accept when using non-blocking I/O, the
8968 function may not be aware that a handshake has actually taken
8969 place, thus preventing a new session from being added to the
8972 To avoid this problem, we now set s->new_session to 2 instead of
8973 using a local variable.
8974 [Lutz Jaenicke, Bodo Moeller]
8976 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8977 if the SSL_R_LENGTH_MISMATCH error is detected.
8978 [Geoff Thorpe, Bodo Moeller]
8980 *) New 'shared_ldflag' column in Configure platform table.
8983 *) Fix EVP_CIPHER_mode macro.
8984 ["Dan S. Camper" <dan@bti.net>]
8986 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8987 type, we must throw them away by setting rr->length to 0.
8988 [D P Chang <dpc@qualys.com>]
8990 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8992 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8993 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8994 worked incorrectly for those cases where range = 10..._2 and
8995 3*range is two bits longer than range.)
8998 *) Only add signing time to PKCS7 structures if it is not already
9002 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9003 OBJ_ld_ce should be OBJ_id_ce.
9004 Also some ip-pda OIDs in crypto/objects/objects.txt were
9005 incorrect (cf. RFC 3039).
9006 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9008 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9009 returns early because it has nothing to do.
9010 [Andy Schneider <andy.schneider@bjss.co.uk>]
9012 *) [In 0.9.6c-engine release:]
9013 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9014 [Andy Schneider <andy.schneider@bjss.co.uk>]
9016 *) [In 0.9.6c-engine release:]
9017 Add support for Cryptographic Appliance's keyserver technology.
9018 (Use engine 'keyclient')
9019 [Cryptographic Appliances and Geoff Thorpe]
9021 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9022 is called via tools/c89.sh because arguments have to be
9023 rearranged (all '-L' options must appear before the first object
9025 [Richard Shapiro <rshapiro@abinitio.com>]
9027 *) [In 0.9.6c-engine release:]
9028 Add support for Broadcom crypto accelerator cards, backported
9030 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9032 *) [In 0.9.6c-engine release:]
9033 Add support for SureWare crypto accelerator cards from
9034 Baltimore Technologies. (Use engine 'sureware')
9035 [Baltimore Technologies and Mark Cox]
9037 *) [In 0.9.6c-engine release:]
9038 Add support for crypto accelerator cards from Accelerated
9039 Encryption Processing, www.aep.ie. (Use engine 'aep')
9040 [AEP Inc. and Mark Cox]
9042 *) Add a configuration entry for gcc on UnixWare.
9043 [Gary Benson <gbenson@redhat.com>]
9045 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9046 messages are stored in a single piece (fixed-length part and
9047 variable-length part combined) and fix various bugs found on the way.
9050 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9051 instead. BIO_gethostbyname() does not know what timeouts are
9052 appropriate, so entries would stay in cache even when they have
9054 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9056 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9057 faced with a pathologically small ClientHello fragment that does
9058 not contain client_version: Instead of aborting with an error,
9059 simply choose the highest available protocol version (i.e.,
9060 TLS 1.0 unless it is disabled). In practice, ClientHello
9061 messages are never sent like this, but this change gives us
9062 strictly correct behaviour at least for TLS.
9065 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9066 never resets s->method to s->ctx->method when called from within
9067 one of the SSL handshake functions.
9068 [Bodo Moeller; problem pointed out by Niko Baric]
9070 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9071 (sent using the client's version number) if client_version is
9072 smaller than the protocol version in use. Also change
9073 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9074 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9075 the client will at least see that alert.
9078 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9082 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9083 client receives HelloRequest while in a handshake.
9084 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9086 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9087 should end in 'break', not 'goto end' which circumvents various
9088 cleanups done in state SSL_ST_OK. But session related stuff
9089 must be disabled for SSL_ST_OK in the case that we just sent a
9092 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9093 before just sending a HelloRequest.
9094 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9096 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9097 reveal whether illegal block cipher padding was found or a MAC
9098 verification error occurred. (Neither SSLerr() codes nor alerts
9099 are directly visible to potential attackers, but the information
9100 may leak via logfiles.)
9102 Similar changes are not required for the SSL 2.0 implementation
9103 because the number of padding bytes is sent in clear for SSL 2.0,
9104 and the extra bytes are just ignored. However ssl/s2_pkt.c
9105 failed to verify that the purported number of padding bytes is in
9109 *) Add OpenUNIX-8 support including shared libraries
9110 (Boyd Lynn Gerber <gerberb@zenez.com>).
9113 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9114 'wristwatch attack' using huge encoding parameters (cf.
9115 James H. Manger's CRYPTO 2001 paper). Note that the
9116 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9117 encoding parameters and hence was not vulnerable.
9120 *) BN_sqr() bug fix.
9121 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9123 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9124 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9125 followed by modular reduction.
9126 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9128 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9129 equivalent based on BN_pseudo_rand() instead of BN_rand().
9132 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9133 This function was broken, as the check for a new client hello message
9134 to handle SGC did not allow these large messages.
9135 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9138 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9141 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9142 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9145 *) Rework the configuration and shared library support for Tru64 Unix.
9146 The configuration part makes use of modern compiler features and
9147 still retains old compiler behavior for those that run older versions
9148 of the OS. The shared library support part includes a variant that
9149 uses the RPATH feature, and is available through the special
9150 configuration target "alpha-cc-rpath", which will never be selected
9152 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9154 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9155 with the same message size as in ssl3_get_certificate_request().
9156 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9157 messages might inadvertently be reject as too long.
9158 [Petr Lampa <lampa@fee.vutbr.cz>]
9160 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9163 *) Modified SSL library such that the verify_callback that has been set
9164 specificly for an SSL object with SSL_set_verify() is actually being
9165 used. Before the change, a verify_callback set with this function was
9166 ignored and the verify_callback() set in the SSL_CTX at the time of
9167 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9168 to allow the necessary settings.
9171 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9172 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9173 done automatically (in contradiction to the requirements of the C
9174 standard). This made problems when used from OpenSSH.
9177 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9178 dh->length and always used
9180 BN_rand_range(priv_key, dh->p).
9182 BN_rand_range() is not necessary for Diffie-Hellman, and this
9183 specific range makes Diffie-Hellman unnecessarily inefficient if
9184 dh->length (recommended exponent length) is much smaller than the
9185 length of dh->p. We could use BN_rand_range() if the order of
9186 the subgroup was stored in the DH structure, but we only have
9191 BN_rand(priv_key, l, ...)
9193 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9199 RSA_eay_public_encrypt
9200 RSA_eay_private_decrypt
9201 RSA_eay_private_encrypt (signing)
9202 RSA_eay_public_decrypt (signature verification)
9204 (default implementations for RSA_public_encrypt,
9205 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9206 always reject numbers >= n.
9209 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9210 to synchronize access to 'locking_thread'. This is necessary on
9211 systems where access to 'locking_thread' (an 'unsigned long'
9212 variable) is not atomic.
9215 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9216 *before* setting the 'crypto_lock_rand' flag. The previous code had
9217 a race condition if 0 is a valid thread ID.
9218 [Travis Vitek <vitek@roguewave.com>]
9220 *) Add support for shared libraries under Irix.
9221 [Albert Chin-A-Young <china@thewrittenword.com>]
9223 *) Add configuration option to build on Linux on both big-endian and
9225 [Ralf Baechle <ralf@uni-koblenz.de>]
9227 *) Add the possibility to create shared libraries on HP-UX.
9230 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9232 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9233 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9234 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9235 PRNG state recovery was possible based on the output of
9236 one PRNG request appropriately sized to gain knowledge on
9237 'md' followed by enough consecutive 1-byte PRNG requests
9238 to traverse all of 'state'.
9240 1. When updating 'md_local' (the current thread's copy of 'md')
9241 during PRNG output generation, hash all of the previous
9242 'md_local' value, not just the half used for PRNG output.
9244 2. Make the number of bytes from 'state' included into the hash
9245 independent from the number of PRNG bytes requested.
9247 The first measure alone would be sufficient to avoid
9248 Markku-Juhani's attack. (Actually it had never occurred
9249 to me that the half of 'md_local' used for chaining was the
9250 half from which PRNG output bytes were taken -- I had always
9251 assumed that the secret half would be used.) The second
9252 measure makes sure that additional data from 'state' is never
9253 mixed into 'md_local' in small portions; this heuristically
9254 further strengthens the PRNG.
9257 *) Fix crypto/bn/asm/mips3.s.
9260 *) When only the key is given to "enc", the IV is undefined. Print out
9261 an error message in this case.
9264 *) Handle special case when X509_NAME is empty in X509 printing routines.
9267 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9268 positive and less than q.
9271 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9272 used: it isn't thread safe and the add_lock_callback should handle
9274 [Paul Rose <Paul.Rose@bridge.com>]
9276 *) Verify that incoming data obeys the block size in
9277 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9281 [Ulf Möller, Bodo Möller]
9283 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9284 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9285 when fixing the server behaviour for backwards-compatible 'client
9286 hello' messages. (Note that the attack is impractical against
9287 SSL 3.0 and TLS 1.0 anyway because length and version checking
9288 means that the probability of guessing a valid ciphertext is
9289 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9292 Before 0.9.5, the countermeasure (hide the error by generating a
9293 random 'decryption result') did not work properly because
9294 ERR_clear_error() was missing, meaning that SSL_get_error() would
9295 detect the supposedly ignored error.
9297 Both problems are now fixed.
9300 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9301 (previously it was 1024).
9304 *) Fix for compatibility mode trust settings: ignore trust settings
9305 unless some valid trust or reject settings are present.
9308 *) Fix for blowfish EVP: its a variable length cipher.
9311 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9312 parameters in DSA public key structures and return an error in the
9313 DSA routines if parameters are absent.
9316 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9317 in the current directory if neither $RANDFILE nor $HOME was set.
9318 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9319 caused some confusion to Windows users who haven't defined $HOME.
9320 Thus RAND_file_name() is changed again: e_os.h can define a
9321 DEFAULT_HOME, which will be used if $HOME is not set.
9322 For Windows, we use "C:"; on other platforms, we still require
9323 environment variables.
9325 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9326 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9327 having multiple threads call RAND_poll() concurrently.
9330 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9331 combination of a flag and a thread ID variable.
9332 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9333 flag), *other* threads can enter ssleay_add_bytes without obeying
9334 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9335 that they do not hold after the first thread unsets add_do_not_lock).
9338 *) Change bctest again: '-x' expressions are not available in all
9342 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9344 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9345 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9347 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9348 the default extension for executables, if any. Also, make the perl
9349 scripts that use symlink() to test if it really exists and use "cp"
9350 if it doesn't. All this made OpenSSL compilable and installable in
9354 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9355 If SEQUENCE is length is indefinite just set c->slen to the total
9356 amount of data available.
9357 [Steve Henson, reported by shige@FreeBSD.org]
9358 [This change does not apply to 0.9.7.]
9360 *) Change bctest to avoid here-documents inside command substitution
9361 (workaround for FreeBSD /bin/sh bug).
9362 For compatibility with Ultrix, avoid shell functions (introduced
9363 in the bctest version that searches along $PATH).
9366 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9367 with des_encrypt() defined on some operating systems, like Solaris
9371 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9372 On the Importance of Eliminating Errors in Cryptographic
9373 Computations, J. Cryptology 14 (2001) 2, 101-119,
9374 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9377 *) MIPS assembler BIGNUM division bug fix.
9380 *) Disabled incorrect Alpha assembler code.
9383 *) Fix PKCS#7 decode routines so they correctly update the length
9384 after reading an EOC for the EXPLICIT tag.
9386 [This change does not apply to 0.9.7.]
9388 *) Fix bug in PKCS#12 key generation routines. This was triggered
9389 if a 3DES key was generated with a 0 initial byte. Include
9390 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9391 (but broken) behaviour.
9394 *) Enhance bctest to search for a working bc along $PATH and print
9396 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9398 *) Fix memory leaks in err.c: free err_data string if necessary;
9399 don't write to the wrong index in ERR_set_error_data.
9402 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9406 *) Replace rdtsc with _emit statements for VC++ version 5.
9407 [Jeremy Cooper <jeremy@baymoo.org>]
9409 *) Make it possible to reuse SSLv2 sessions.
9412 *) In copy_email() check for >= 0 as a return value for
9413 X509_NAME_get_index_by_NID() since 0 is a valid index.
9414 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9416 *) Avoid coredump with unsupported or invalid public keys by checking if
9417 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9418 PKCS7_verify() fails with non detached data.
9421 *) Don't use getenv in library functions when run as setuid/setgid.
9422 New function OPENSSL_issetugid().
9425 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9426 due to incorrect handling of multi-threading:
9428 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9430 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9432 3. Count how many times MemCheck_off() has been called so that
9433 nested use can be treated correctly. This also avoids
9434 inband-signalling in the previous code (which relied on the
9435 assumption that thread ID 0 is impossible).
9438 *) Add "-rand" option also to s_client and s_server.
9441 *) Fix CPU detection on Irix 6.x.
9442 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9443 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9445 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9448 [This change does not apply to 0.9.7.]
9450 *) Use the cached encoding of an X509_NAME structure rather than
9451 copying it. This is apparently the reason for the libsafe "errors"
9452 but the code is actually correct.
9455 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9456 Bleichenbacher's DSA attack.
9457 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9458 to be set and top=0 forces the highest bit to be set; top=-1 is new
9459 and leaves the highest bit random.
9460 [Ulf Moeller, Bodo Moeller]
9462 *) In the NCONF_...-based implementations for CONF_... queries
9463 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9464 a temporary CONF structure with the data component set to NULL
9465 (which gives segmentation faults in lh_retrieve).
9466 Instead, use NULL for the CONF pointer in CONF_get_string and
9467 CONF_get_number (which may use environment variables) and directly
9468 return NULL from CONF_get_section.
9471 *) Fix potential buffer overrun for EBCDIC.
9474 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9475 keyUsage if basicConstraints absent for a CA.
9478 *) Make SMIME_write_PKCS7() write mail header values with a format that
9479 is more generally accepted (no spaces before the semicolon), since
9480 some programs can't parse those values properly otherwise. Also make
9481 sure BIO's that break lines after each write do not create invalid
9485 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9486 macros previously used would not encode an empty SEQUENCE OF
9487 and break the signature.
9489 [This change does not apply to 0.9.7.]
9491 *) Zero the premaster secret after deriving the master secret in
9495 *) Add some EVP_add_digest_alias registrations (as found in
9496 OpenSSL_add_all_digests()) to SSL_library_init()
9497 aka OpenSSL_add_ssl_algorithms(). This provides improved
9498 compatibility with peers using X.509 certificates
9499 with unconventional AlgorithmIdentifier OIDs.
9502 *) Fix for Irix with NO_ASM.
9503 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9505 *) ./config script fixes.
9506 [Ulf Moeller, Richard Levitte]
9508 *) Fix 'openssl passwd -1'.
9511 *) Change PKCS12_key_gen_asc() so it can cope with non null
9512 terminated strings whose length is passed in the passlen
9513 parameter, for example from PEM callbacks. This was done
9514 by adding an extra length parameter to asc2uni().
9515 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9517 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9518 call failed, free the DSA structure.
9521 *) Fix to uni2asc() to cope with zero length Unicode strings.
9522 These are present in some PKCS#12 files.
9525 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9526 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9527 when writing a 32767 byte record.
9528 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9530 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9531 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9533 (RSA objects have a reference count access to which is protected
9534 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9535 so they are meant to be shared between threads.)
9536 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9537 "Reddie, Steven" <Steven.Reddie@ca.com>]
9539 *) Fix a deadlock in CRYPTO_mem_leaks().
9542 *) Use better test patterns in bntest.
9545 *) rand_win.c fix for Borland C.
9548 *) BN_rshift bugfix for n == 0.
9551 *) Add a 'bctest' script that checks for some known 'bc' bugs
9552 so that 'make test' does not abort just because 'bc' is broken.
9555 *) Store verify_result within SSL_SESSION also for client side to
9556 avoid potential security hole. (Re-used sessions on the client side
9557 always resulted in verify_result==X509_V_OK, not using the original
9558 result of the server certificate verification.)
9561 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9562 SSL3_RT_APPLICATION_DATA, return 0.
9563 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9567 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9568 releases, have been re-implemented by renaming the previous
9569 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9570 and ssl3_read_internal, respectively, and adding 'peek' parameters
9571 to them. The new ssl[23]_{read,peek} functions are calls to
9572 ssl[23]_read_internal with the 'peek' flag set appropriately.
9573 A 'peek' parameter has also been added to ssl3_read_bytes, which
9574 does the actual work for ssl3_read_internal.
9577 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9578 the method-specific "init()" handler. Also clean up ex_data after
9579 calling the method-specific "finish()" handler. Previously, this was
9580 happening the other way round.
9583 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9584 The previous value, 12, was not always sufficient for BN_mod_exp().
9587 *) Make sure that shared libraries get the internal name engine with
9588 the full version number and not just 0. This should mark the
9589 shared libraries as not backward compatible. Of course, this should
9590 be changed again when we can guarantee backward binary compatibility.
9593 *) Fix typo in get_cert_by_subject() in by_dir.c
9594 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9596 *) Rework the system to generate shared libraries:
9598 - Make note of the expected extension for the shared libraries and
9599 if there is a need for symbolic links from for example libcrypto.so.0
9600 to libcrypto.so.0.9.7. There is extended info in Configure for
9603 - Make as few rebuilds of the shared libraries as possible.
9605 - Still avoid linking the OpenSSL programs with the shared libraries.
9607 - When installing, install the shared libraries separately from the
9611 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9613 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9614 and not in SSL_clear because the latter is also used by the
9615 accept/connect functions; previously, the settings made by
9616 SSL_set_read_ahead would be lost during the handshake.
9617 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9619 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9620 Previously, it would create entries for disabled algorithms no
9624 *) Added several new manual pages for SSL_* function.
9627 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9629 *) In ssl23_get_client_hello, generate an error message when faced
9630 with an initial SSL 3.0/TLS record that is too small to contain the
9631 first two bytes of the ClientHello message, i.e. client_version.
9632 (Note that this is a pathologic case that probably has never happened
9633 in real life.) The previous approach was to use the version number
9634 from the record header as a substitute; but our protocol choice
9635 should not depend on that one because it is not authenticated
9636 by the Finished messages.
9639 *) More robust randomness gathering functions for Windows.
9640 [Jeffrey Altman <jaltman@columbia.edu>]
9642 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9643 not set then we don't setup the error code for issuer check errors
9644 to avoid possibly overwriting other errors which the callback does
9645 handle. If an application does set the flag then we assume it knows
9646 what it is doing and can handle the new informational codes
9650 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9651 a general "ANY" type, as such it should be able to decode anything
9652 including tagged types. However it didn't check the class so it would
9653 wrongly interpret tagged types in the same way as their universal
9654 counterpart and unknown types were just rejected. Changed so that the
9655 tagged and unknown types are handled in the same way as a SEQUENCE:
9656 that is the encoding is stored intact. There is also a new type
9657 "V_ASN1_OTHER" which is used when the class is not universal, in this
9658 case we have no idea what the actual type is so we just lump them all
9662 *) On VMS, stdout may very well lead to a file that is written to
9663 in a record-oriented fashion. That means that every write() will
9664 write a separate record, which will be read separately by the
9665 programs trying to read from it. This can be very confusing.
9667 The solution is to put a BIO filter in the way that will buffer
9668 text until a linefeed is reached, and then write everything a
9669 line at a time, so every record written will be an actual line,
9670 not chunks of lines and not (usually doesn't happen, but I've
9671 seen it once) several lines in one record. BIO_f_linebuffer() is
9674 Currently, it's a VMS-only method, because that's where it has
9675 been tested well enough.
9678 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9679 it can return incorrect results.
9680 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9681 but it was in 0.9.6-beta[12].)
9684 *) Disable the check for content being present when verifying detached
9685 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9686 include zero length content when signing messages.
9689 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9690 BIO_ctrl (for BIO pairs).
9693 *) Add DSO method for VMS.
9696 *) Bug fix: Montgomery multiplication could produce results with the
9700 *) Add RPM specification openssl.spec and modify it to build three
9701 packages. The default package contains applications, application
9702 documentation and run-time libraries. The devel package contains
9703 include files, static libraries and function documentation. The
9704 doc package contains the contents of the doc directory. The original
9705 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9708 *) Add a large number of documentation files for many SSL routines.
9709 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9711 *) Add a configuration entry for Sony News 4.
9712 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9714 *) Don't set the two most significant bits to one when generating a
9715 random number < q in the DSA library.
9718 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9719 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9720 the underlying transport is blocking) if a handshake took place.
9721 (The default behaviour is needed by applications such as s_client
9722 and s_server that use select() to determine when to use SSL_read;
9723 but for applications that know in advance when to expect data, it
9724 just makes things more complicated.)
9727 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9731 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9732 work better on such systems.
9733 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9735 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9736 Update PKCS12_parse() so it copies the friendlyName and the
9737 keyid to the certificates aux info.
9740 *) Fix bug in PKCS7_verify() which caused an infinite loop
9741 if there was more than one signature.
9742 [Sven Uszpelkat <su@celocom.de>]
9744 *) Major change in util/mkdef.pl to include extra information
9745 about each symbol, as well as presenting variables as well
9746 as functions. This change means that there's n more need
9747 to rebuild the .num files when some algorithms are excluded.
9750 *) Allow the verify time to be set by an application,
9751 rather than always using the current time.
9754 *) Phase 2 verify code reorganisation. The certificate
9755 verify code now looks up an issuer certificate by a
9756 number of criteria: subject name, authority key id
9757 and key usage. It also verifies self signed certificates
9758 by the same criteria. The main comparison function is
9759 X509_check_issued() which performs these checks.
9761 Lot of changes were necessary in order to support this
9762 without completely rewriting the lookup code.
9764 Authority and subject key identifier are now cached.
9766 The LHASH 'certs' is X509_STORE has now been replaced
9767 by a STACK_OF(X509_OBJECT). This is mainly because an
9768 LHASH can't store or retrieve multiple objects with
9769 the same hash value.
9771 As a result various functions (which were all internal
9772 use only) have changed to handle the new X509_STORE
9773 structure. This will break anything that messed round
9774 with X509_STORE internally.
9776 The functions X509_STORE_add_cert() now checks for an
9777 exact match, rather than just subject name.
9779 The X509_STORE API doesn't directly support the retrieval
9780 of multiple certificates matching a given criteria, however
9781 this can be worked round by performing a lookup first
9782 (which will fill the cache with candidate certificates)
9783 and then examining the cache for matches. This is probably
9784 the best we can do without throwing out X509_LOOKUP
9785 entirely (maybe later...).
9787 The X509_VERIFY_CTX structure has been enhanced considerably.
9789 All certificate lookup operations now go via a get_issuer()
9790 callback. Although this currently uses an X509_STORE it
9791 can be replaced by custom lookups. This is a simple way
9792 to bypass the X509_STORE hackery necessary to make this
9793 work and makes it possible to use more efficient techniques
9794 in future. A very simple version which uses a simple
9795 STACK for its trusted certificate store is also provided
9796 using X509_STORE_CTX_trusted_stack().
9798 The verify_cb() and verify() callbacks now have equivalents
9799 in the X509_STORE_CTX structure.
9801 X509_STORE_CTX also has a 'flags' field which can be used
9802 to customise the verify behaviour.
9805 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9806 excludes S/MIME capabilities.
9809 *) When a certificate request is read in keep a copy of the
9810 original encoding of the signed data and use it when outputting
9811 again. Signatures then use the original encoding rather than
9812 a decoded, encoded version which may cause problems if the
9813 request is improperly encoded.
9816 *) For consistency with other BIO_puts implementations, call
9817 buffer_write(b, ...) directly in buffer_puts instead of calling
9820 In BIO_puts, increment b->num_write as in BIO_write.
9821 [Peter.Sylvester@EdelWeb.fr]
9823 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9824 BN_zero, we may not return a BIGNUM with an array consisting of
9828 *) Avoid calling abort() from within the library when problems are
9829 detected, except if preprocessor symbols have been defined
9830 (such as REF_CHECK, BN_DEBUG etc.).
9833 *) New openssl application 'rsautl'. This utility can be
9834 used for low level RSA operations. DER public key
9835 BIO/fp routines also added.
9838 *) New Configure entry and patches for compiling on QNX 4.
9839 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9841 *) A demo state-machine implementation was sponsored by
9842 Nuron (http://www.nuron.com/) and is now available in
9843 demos/state_machine.
9846 *) New options added to the 'dgst' utility for signature
9847 generation and verification.
9850 *) Unrecognized PKCS#7 content types are now handled via a
9851 catch all ASN1_TYPE structure. This allows unsupported
9852 types to be stored as a "blob" and an application can
9853 encode and decode it manually.
9856 *) Fix various signed/unsigned issues to make a_strex.c
9858 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9860 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9861 length if passed a buffer. ASN1_INTEGER_to_BN failed
9862 if passed a NULL BN and its argument was negative.
9863 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9865 *) Modification to PKCS#7 encoding routines to output definite
9866 length encoding. Since currently the whole structures are in
9867 memory there's not real point in using indefinite length
9868 constructed encoding. However if OpenSSL is compiled with
9869 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9872 *) Added BIO_vprintf() and BIO_vsnprintf().
9875 *) Added more prefixes to parse for in the strings written
9876 through a logging bio, to cover all the levels that are available
9877 through syslog. The prefixes are now:
9879 PANIC, EMERG, EMR => LOG_EMERG
9880 ALERT, ALR => LOG_ALERT
9881 CRIT, CRI => LOG_CRIT
9882 ERROR, ERR => LOG_ERR
9883 WARNING, WARN, WAR => LOG_WARNING
9884 NOTICE, NOTE, NOT => LOG_NOTICE
9885 INFO, INF => LOG_INFO
9886 DEBUG, DBG => LOG_DEBUG
9888 and as before, if none of those prefixes are present at the
9889 beginning of the string, LOG_ERR is chosen.
9891 On Win32, the LOG_* levels are mapped according to this:
9893 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9894 LOG_WARNING => EVENTLOG_WARNING_TYPE
9895 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9899 *) Made it possible to reconfigure with just the configuration
9900 argument "reconf" or "reconfigure". The command line arguments
9901 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9902 and are retrieved from there when reconfiguring.
9906 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9908 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9911 *) The obj_dat.pl script was messing up the sorting of object
9912 names. The reason was that it compared the quoted version
9913 of strings as a result "OCSP" > "OCSP Signing" because
9914 " > SPACE. Changed script to store unquoted versions of
9915 names and add quotes on output. It was also omitting some
9916 names from the lookup table if they were given a default
9917 value (that is if SN is missing it is given the same
9918 value as LN and vice versa), these are now added on the
9919 grounds that if an object has a name we should be able to
9920 look it up. Finally added warning output when duplicate
9921 short or long names are found.
9924 *) Changes needed for Tandem NSK.
9925 [Scott Uroff <scott@xypro.com>]
9927 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9928 RSA_padding_check_SSLv23(), special padding was never detected
9929 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9930 version rollback attacks was not effective.
9932 In s23_clnt.c, don't use special rollback-attack detection padding
9933 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9934 client; similarly, in s23_srvr.c, don't do the rollback check if
9935 SSL 2.0 is the only protocol enabled in the server.
9938 *) Make it possible to get hexdumps of unprintable data with 'openssl
9939 asn1parse'. By implication, the functions ASN1_parse_dump() and
9940 BIO_dump_indent() are added.
9943 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9944 these print out strings and name structures based on various
9945 flags including RFC2253 support and proper handling of
9946 multibyte characters. Added options to the 'x509' utility
9947 to allow the various flags to be set.
9950 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9951 Also change the functions X509_cmp_current_time() and
9952 X509_gmtime_adj() work with an ASN1_TIME structure,
9953 this will enable certificates using GeneralizedTime in validity
9954 dates to be checked.
9957 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9958 negative public key encodings) on by default,
9959 NO_NEG_PUBKEY_BUG can be set to disable it.
9962 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9963 content octets. An i2c_ASN1_OBJECT is unnecessary because
9964 the encoding can be trivially obtained from the structure.
9967 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9968 not read locks (CRYPTO_r_[un]lock).
9971 *) A first attempt at creating official support for shared
9972 libraries through configuration. I've kept it so the
9973 default is static libraries only, and the OpenSSL programs
9974 are always statically linked for now, but there are
9975 preparations for dynamic linking in place.
9976 This has been tested on Linux and Tru64.
9979 *) Randomness polling function for Win9x, as described in:
9980 Peter Gutmann, Software Generation of Practically Strong
9984 *) Fix so PRNG is seeded in req if using an already existing
9988 *) New options to smime application. -inform and -outform
9989 allow alternative formats for the S/MIME message including
9990 PEM and DER. The -content option allows the content to be
9991 specified separately. This should allow things like Netscape
9992 form signing output easier to verify.
9995 *) Fix the ASN1 encoding of tags using the 'long form'.
9998 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9999 STRING types. These convert content octets to and from the
10000 underlying type. The actual tag and length octets are
10001 already assumed to have been read in and checked. These
10002 are needed because all other string types have virtually
10003 identical handling apart from the tag. By having versions
10004 of the ASN1 functions that just operate on content octets
10005 IMPLICIT tagging can be handled properly. It also allows
10006 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10007 and ASN1_INTEGER are identical apart from the tag.
10010 *) Change the handling of OID objects as follows:
10012 - New object identifiers are inserted in objects.txt, following
10013 the syntax given in objects.README.
10014 - objects.pl is used to process obj_mac.num and create a new
10016 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10019 This is currently kind of a hack, and the perl code in objects.pl
10020 isn't very elegant, but it works as I intended. The simplest way
10021 to check that it worked correctly is to look in obj_dat.h and
10022 check the array nid_objs and make sure the objects haven't moved
10023 around (this is important!). Additions are OK, as well as
10024 consistent name changes.
10027 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10030 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10031 The given file adds to whatever has already been seeded into the
10032 random pool through the RANDFILE configuration file option or
10033 environment variable, or the default random state file.
10036 *) mkstack.pl now sorts each macro group into lexical order.
10037 Previously the output order depended on the order the files
10038 appeared in the directory, resulting in needless rewriting
10042 *) Patches to make OpenSSL compile under Win32 again. Mostly
10043 work arounds for the VC++ problem that it treats func() as
10044 func(void). Also stripped out the parts of mkdef.pl that
10045 added extra typesafe functions: these no longer exist.
10048 *) Reorganisation of the stack code. The macros are now all
10049 collected in safestack.h . Each macro is defined in terms of
10050 a "stack macro" of the form SKM_<name>(type, a, b). The
10051 DEBUG_SAFESTACK is now handled in terms of function casts,
10052 this has the advantage of retaining type safety without the
10053 use of additional functions. If DEBUG_SAFESTACK is not defined
10054 then the non typesafe macros are used instead. Also modified the
10055 mkstack.pl script to handle the new form. Needs testing to see
10056 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10057 the default if no major problems. Similar behaviour for ASN1_SET_OF
10058 and PKCS12_STACK_OF.
10061 *) When some versions of IIS use the 'NET' form of private key the
10062 key derivation algorithm is different. Normally MD5(password) is
10063 used as a 128 bit RC4 key. In the modified case
10064 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10065 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10066 as the old Netscape_RSA functions except they have an additional
10067 'sgckey' parameter which uses the modified algorithm. Also added
10068 an -sgckey command line option to the rsa utility. Thanks to
10069 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10070 algorithm to openssl-dev.
10073 *) The evp_local.h macros were using 'c.##kname' which resulted in
10074 invalid expansion on some systems (SCO 5.0.5 for example).
10075 Corrected to 'c.kname'.
10076 [Phillip Porch <root@theporch.com>]
10078 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10079 a STACK of email addresses from a certificate or request, these look
10080 in the subject name and the subject alternative name extensions and
10081 omit any duplicate addresses.
10084 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10085 This makes DSA verification about 2 % faster.
10088 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10089 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10090 plus overhead for 1024 bit moduli).
10091 This makes exponentiations about 0.5 % faster for 1024 bit
10092 exponents (as measured by "openssl speed rsa2048").
10095 *) Rename memory handling macros to avoid conflicts with other
10097 Malloc => OPENSSL_malloc
10098 Malloc_locked => OPENSSL_malloc_locked
10099 Realloc => OPENSSL_realloc
10100 Free => OPENSSL_free
10103 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10104 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10107 *) CygWin32 support.
10108 [John Jarvie <jjarvie@newsguy.com>]
10110 *) The type-safe stack code has been rejigged. It is now only compiled
10111 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10112 by default all type-specific stack functions are "#define"d back to
10113 standard stack functions. This results in more streamlined output
10114 but retains the type-safety checking possibilities of the original
10118 *) The STACK code has been cleaned up, and certain type declarations
10119 that didn't make a lot of sense have been brought in line. This has
10120 also involved a cleanup of sorts in safestack.h to more correctly
10121 map type-safe stack functions onto their plain stack counterparts.
10122 This work has also resulted in a variety of "const"ifications of
10123 lots of the code, especially "_cmp" operations which should normally
10124 be prototyped with "const" parameters anyway.
10127 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10128 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10129 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10130 where all of 'md' is used each time the PRNG is used, but 'state'
10131 is used only indexed by a cyclic counter. As entropy may not be
10132 well distributed from the beginning, 'md' is important as a
10133 chaining variable. However, the output function chains only half
10134 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10135 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10136 in all of 'state' being rewritten, with the new values depending
10137 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10140 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10141 the handshake is continued after ssl_verify_cert_chain();
10142 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10143 can lead to 'unexplainable' connection aborts later.
10144 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10146 *) Major EVP API cipher revision.
10147 Add hooks for extra EVP features. This allows various cipher
10148 parameters to be set in the EVP interface. Support added for variable
10149 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10150 setting of RC2 and RC5 parameters.
10152 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10155 Remove lots of duplicated code from the EVP library. For example *every*
10156 cipher init() function handles the 'iv' in the same way according to the
10157 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10158 for CFB and OFB modes they zero ctx->num.
10160 New functionality allows removal of S/MIME code RC2 hack.
10162 Most of the routines have the same form and so can be declared in terms
10165 By shifting this to the top level EVP_CipherInit() it can be removed from
10166 all individual ciphers. If the cipher wants to handle IVs or keys
10167 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10170 Change lots of functions like EVP_EncryptUpdate() to now return a
10171 value: although software versions of the algorithms cannot fail
10172 any installed hardware versions can.
10175 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10176 this option is set, tolerate broken clients that send the negotiated
10177 protocol version number instead of the requested protocol version
10181 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10182 i.e. non-zero for export ciphersuites, zero otherwise.
10183 Previous versions had this flag inverted, inconsistent with
10184 rsa_tmp_cb (..._TMP_RSA_CB).
10185 [Bodo Moeller; problem reported by Amit Chopra]
10187 *) Add missing DSA library text string. Work around for some IIS
10188 key files with invalid SEQUENCE encoding.
10191 *) Add a document (doc/standards.txt) that list all kinds of standards
10192 and so on that are implemented in OpenSSL.
10195 *) Enhance c_rehash script. Old version would mishandle certificates
10196 with the same subject name hash and wouldn't handle CRLs at all.
10197 Added -fingerprint option to crl utility, to support new c_rehash
10201 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10204 *) Fix for SSL server purpose checking. Server checking was
10205 rejecting certificates which had extended key usage present
10206 but no ssl client purpose.
10207 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10209 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10210 is a little unclear about how a blank password is handled.
10211 Since the password in encoded as a BMPString with terminating
10212 double NULL a zero length password would end up as just the
10213 double NULL. However no password at all is different and is
10214 handled differently in the PKCS#12 key generation code. NS
10215 treats a blank password as zero length. MSIE treats it as no
10216 password on export: but it will try both on import. We now do
10217 the same: PKCS12_parse() tries zero length and no password if
10218 the password is set to "" or NULL (NULL is now a valid password:
10219 it wasn't before) as does the pkcs12 application.
10222 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10223 perror when PEM_read_bio_X509_REQ fails, the error message must
10224 be obtained from the error queue.
10227 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10228 it in ERR_remove_state if appropriate, and change ERR_get_state
10229 accordingly to avoid race conditions (this is necessary because
10230 thread_hash is no longer constant once set).
10233 *) Bugfix for linux-elf makefile.one.
10236 *) RSA_get_default_method() will now cause a default
10237 RSA_METHOD to be chosen if one doesn't exist already.
10238 Previously this was only set during a call to RSA_new()
10239 or RSA_new_method(NULL) meaning it was possible for
10240 RSA_get_default_method() to return NULL.
10243 *) Added native name translation to the existing DSO code
10244 that will convert (if the flag to do so is set) filenames
10245 that are sufficiently small and have no path information
10246 into a canonical native form. Eg. "blah" converted to
10247 "libblah.so" or "blah.dll" etc.
10250 *) New function ERR_error_string_n(e, buf, len) which is like
10251 ERR_error_string(e, buf), but writes at most 'len' bytes
10252 including the 0 terminator. For ERR_error_string_n, 'buf'
10254 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10256 *) CONF library reworked to become more general. A new CONF
10257 configuration file reader "class" is implemented as well as a
10258 new functions (NCONF_*, for "New CONF") to handle it. The now
10259 old CONF_* functions are still there, but are reimplemented to
10260 work in terms of the new functions. Also, a set of functions
10261 to handle the internal storage of the configuration data is
10262 provided to make it easier to write new configuration file
10263 reader "classes" (I can definitely see something reading a
10264 configuration file in XML format, for example), called _CONF_*,
10265 or "the configuration storage API"...
10267 The new configuration file reading functions are:
10269 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10270 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10272 NCONF_default, NCONF_WIN32
10274 NCONF_dump_fp, NCONF_dump_bio
10276 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10277 NCONF_new creates a new CONF object. This works in the same way
10278 as other interfaces in OpenSSL, like the BIO interface.
10279 NCONF_dump_* dump the internal storage of the configuration file,
10280 which is useful for debugging. All other functions take the same
10281 arguments as the old CONF_* functions wth the exception of the
10282 first that must be a `CONF *' instead of a `LHASH *'.
10284 To make it easer to use the new classes with the old CONF_* functions,
10285 the function CONF_set_default_method is provided.
10288 *) Add '-tls1' option to 'openssl ciphers', which was already
10289 mentioned in the documentation but had not been implemented.
10290 (This option is not yet really useful because even the additional
10291 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10294 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10295 OpenSSL-based applications) load shared libraries and bind to
10296 them in a portable way.
10297 [Geoff Thorpe, with contributions from Richard Levitte]
10299 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10301 *) Make sure _lrotl and _lrotr are only used with MSVC.
10303 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10304 (the default implementation of RAND_status).
10306 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10307 to '-clrext' (= clear extensions), as intended and documented.
10308 [Bodo Moeller; inconsistency pointed out by Michael Attili
10309 <attili@amaxo.com>]
10311 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10312 was larger than the MD block size.
10313 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10315 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10316 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10317 using the passed key: if the passed key was a private key the result
10318 of X509_print(), for example, would be to print out all the private key
10322 *) des_quad_cksum() byte order bug fix.
10323 [Ulf Möller, using the problem description in krb4-0.9.7, where
10324 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10326 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10328 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10330 *) For easily testing in shell scripts whether some command
10331 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10332 returns with exit code 0 iff no command of the given name is available.
10333 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10334 the output goes to stdout and nothing is printed to stderr.
10335 Additional arguments are always ignored.
10337 Since for each cipher there is a command of the same name,
10338 the 'no-cipher' compilation switches can be tested this way.
10340 ('openssl no-XXX' is not able to detect pseudo-commands such
10341 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10344 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10347 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10348 is set; it will be thrown away anyway because each handshake creates
10350 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10351 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10352 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10353 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10356 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10357 'Q' and 'R' lose their special meanings (quit/renegotiate).
10358 This is part of what -quiet does; unlike -quiet, -ign_eof
10359 does not suppress any output.
10362 *) Add compatibility options to the purpose and trust code. The
10363 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10364 accepts a certificate or CA, this was the previous behaviour,
10365 with all the associated security issues.
10367 X509_TRUST_COMPAT is the old trust behaviour: only and
10368 automatically trust self signed roots in certificate store. A
10369 new trust setting X509_TRUST_DEFAULT is used to specify that
10370 a purpose has no associated trust setting and it should instead
10371 use the value in the default purpose.
10374 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10375 and fix a memory leak.
10378 *) In util/mkerr.pl (which implements 'make errors'), preserve
10379 reason strings from the previous version of the .c file, as
10380 the default to have only downcase letters (and digits) in
10381 automatically generated reasons codes is not always appropriate.
10384 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10385 using strerror. Previously, ERR_reason_error_string() returned
10386 library names as reason strings for SYSerr; but SYSerr is a special
10387 case where small numbers are errno values, not library numbers.
10390 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10391 converts DSA parameters into DH parameters. (When creating parameters,
10392 DSA_generate_parameters is used.)
10395 *) Include 'length' (recommended exponent length) in C code generated
10396 by 'openssl dhparam -C'.
10399 *) The second argument to set_label in perlasm was already being used
10400 so couldn't be used as a "file scope" flag. Moved to third argument
10404 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10405 instead of RAND_bytes for encryption IVs and salts.
10408 *) Include RAND_status() into RAND_METHOD instead of implementing
10409 it only for md_rand.c Otherwise replacing the PRNG by calling
10410 RAND_set_rand_method would be impossible.
10413 *) Don't let DSA_generate_key() enter an infinite loop if the random
10414 number generation fails.
10417 *) New 'rand' application for creating pseudo-random output.
10420 *) Added configuration support for Linux/IA64
10421 [Rolf Haberrecker <rolf@suse.de>]
10423 *) Assembler module support for Mingw32.
10426 *) Shared library support for HPUX (in shlib/).
10427 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10429 *) Shared library support for Solaris gcc.
10430 [Lutz Behnke <behnke@trustcenter.de>]
10432 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10434 *) PKCS7_encrypt() was adding text MIME headers twice because they
10435 were added manually and by SMIME_crlf_copy().
10438 *) In bntest.c don't call BN_rand with zero bits argument.
10439 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10441 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10442 case was implemented. This caused BN_div_recp() to fail occasionally.
10445 *) Add an optional second argument to the set_label() in the perl
10446 assembly language builder. If this argument exists and is set
10447 to 1 it signals that the assembler should use a symbol whose
10448 scope is the entire file, not just the current function. This
10449 is needed with MASM which uses the format label:: for this scope.
10450 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10452 *) Change the ASN1 types so they are typedefs by default. Before
10453 almost all types were #define'd to ASN1_STRING which was causing
10454 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10458 *) Change names of new functions to the new get1/get0 naming
10459 convention: After 'get1', the caller owns a reference count
10460 and has to call ..._free; 'get0' returns a pointer to some
10461 data structure without incrementing reference counters.
10462 (Some of the existing 'get' functions increment a reference
10463 counter, some don't.)
10464 Similarly, 'set1' and 'add1' functions increase reference
10465 counters or duplicate objects.
10468 *) Allow for the possibility of temp RSA key generation failure:
10469 the code used to assume it always worked and crashed on failure.
10472 *) Fix potential buffer overrun problem in BIO_printf().
10473 [Ulf Möller, using public domain code by Patrick Powell; problem
10474 pointed out by David Sacerdote <das33@cornell.edu>]
10476 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10477 RAND_egd() and RAND_status(). In the command line application,
10478 the EGD socket can be specified like a seed file using RANDFILE
10482 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10483 Some CAs (e.g. Verisign) distribute certificates in this form.
10486 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10487 list to exclude them. This means that no special compilation option
10488 is needed to use anonymous DH: it just needs to be included in the
10492 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10493 EVP_MD_type. The old functionality is available in a new macro called
10494 EVP_MD_md(). Change code that uses it and update docs.
10497 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10498 where the 'void *' argument is replaced by a function pointer argument.
10499 Previously 'void *' was abused to point to functions, which works on
10500 many platforms, but is not correct. As these functions are usually
10501 called by macros defined in OpenSSL header files, most source code
10502 should work without changes.
10505 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10506 sections with information on -D... compiler switches used for
10507 compiling the library so that applications can see them. To enable
10508 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10509 must be defined. E.g.,
10510 #define OPENSSL_ALGORITHM_DEFINES
10511 #include <openssl/opensslconf.h>
10512 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10513 [Richard Levitte, Ulf and Bodo Möller]
10515 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10519 *) Change the 'other' type in certificate aux info to a STACK_OF
10520 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10521 the required ASN1 format: arbitrary types determined by an OID.
10524 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10525 argument to 'req'. This is not because the function is newer or
10526 better than others it just uses the work 'NEW' in the certificate
10527 request header lines. Some software needs this.
10530 *) Reorganise password command line arguments: now passwords can be
10531 obtained from various sources. Delete the PEM_cb function and make
10532 it the default behaviour: i.e. if the callback is NULL and the
10533 usrdata argument is not NULL interpret it as a null terminated pass
10534 phrase. If usrdata and the callback are NULL then the pass phrase
10535 is prompted for as usual.
10538 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10539 the support is automatically enabled. The resulting binaries will
10540 autodetect the card and use it if present.
10541 [Ben Laurie and Compaq Inc.]
10543 *) Work around for Netscape hang bug. This sends certificate request
10544 and server done in one record. Since this is perfectly legal in the
10545 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10546 the bugs/SSLv3 entry for more info.
10549 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10552 *) Add -rand argument to smime and pkcs12 applications and read/write
10556 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10559 *) Add command line password options to the remaining applications.
10562 *) Bug fix for BN_div_recp() for numerators with an even number of
10566 *) More tests in bntest.c, and changed test_bn output.
10569 *) ./config recognizes MacOS X now.
10572 *) Bug fix for BN_div() when the first words of num and divisor are
10573 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10576 *) Add support for various broken PKCS#8 formats, and command line
10577 options to produce them.
10580 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10581 get temporary BIGNUMs from a BN_CTX.
10584 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10588 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10589 include a #define from the old name to the new. The original intent
10590 was that statically linked binaries could for example just call
10591 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10592 link with digests. This never worked because SSLeay_add_all_digests()
10593 and SSLeay_add_all_ciphers() were in the same source file so calling
10594 one would link with the other. They are now in separate source files.
10597 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10600 *) Use a less unusual form of the Miller-Rabin primality test (it used
10601 a binary algorithm for exponentiation integrated into the Miller-Rabin
10602 loop, our standard modexp algorithms are faster).
10605 *) Support for the EBCDIC character set completed.
10606 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10608 *) Source code cleanups: use const where appropriate, eliminate casts,
10609 use void * instead of char * in lhash.
10612 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10613 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10614 this the server could overwrite ephemeral keys that the client
10618 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10619 using 50 iterations of the Rabin-Miller test.
10621 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10622 iterations of the Rabin-Miller test as required by the appendix
10623 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10624 As BN_is_prime_fasttest includes trial division, DSA parameter
10625 generation becomes much faster.
10627 This implies a change for the callback functions in DSA_is_prime
10628 and DSA_generate_parameters: The callback function is called once
10629 for each positive witness in the Rabin-Miller test, not just
10630 occasionally in the inner loop; and the parameters to the
10631 callback function now provide an iteration count for the outer
10632 loop rather than for the current invocation of the inner loop.
10633 DSA_generate_parameters additionally can call the callback
10634 function with an 'iteration count' of -1, meaning that a
10635 candidate has passed the trial division test (when q is generated
10636 from an application-provided seed, trial division is skipped).
10639 *) New function BN_is_prime_fasttest that optionally does trial
10640 division before starting the Rabin-Miller test and has
10641 an additional BN_CTX * argument (whereas BN_is_prime always
10642 has to allocate at least one BN_CTX).
10643 'callback(1, -1, cb_arg)' is called when a number has passed the
10644 trial division stage.
10647 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10651 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10654 *) New function BN_pseudo_rand().
10657 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10658 bignum version of BN_from_montgomery() with the working code from
10659 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10663 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10664 made it impossible to use the same SSL_SESSION data structure in
10665 SSL2 clients in multiple threads.
10668 *) The return value of RAND_load_file() no longer counts bytes obtained
10669 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10670 to seed the PRNG (previously an explicit byte count was required).
10671 [Ulf Möller, Bodo Möller]
10673 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10674 used (char *) instead of (void *) and had casts all over the place.
10677 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10680 *) Retain source code compatibility for BN_prime_checks macro:
10681 BN_is_prime(..., BN_prime_checks, ...) now uses
10682 BN_prime_checks_for_size to determine the appropriate number of
10683 Rabin-Miller iterations.
10686 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10687 DH_CHECK_P_NOT_SAFE_PRIME.
10688 (Check if this is true? OpenPGP calls them "strong".)
10691 *) Merge the functionality of "dh" and "gendh" programs into a new program
10692 "dhparam". The old programs are retained for now but will handle DH keys
10693 (instead of parameters) in future.
10696 *) Make the ciphers, s_server and s_client programs check the return values
10697 when a new cipher list is set.
10700 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10701 ciphers. Before when the 56bit ciphers were enabled the sorting was
10704 The syntax for the cipher sorting has been extended to support sorting by
10705 cipher-strength (using the strength_bits hard coded in the tables).
10706 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10708 Fix a bug in the cipher-command parser: when supplying a cipher command
10709 string with an "undefined" symbol (neither command nor alphanumeric
10710 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10711 an error is flagged.
10713 Due to the strength-sorting extension, the code of the
10714 ssl_create_cipher_list() function was completely rearranged. I hope that
10715 the readability was also increased :-)
10716 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10718 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10719 for the first serial number and places 2 in the serial number file. This
10720 avoids problems when the root CA is created with serial number zero and
10721 the first user certificate has the same issuer name and serial number
10725 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10726 the new code. Add documentation for this stuff.
10729 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10730 X509_*() to X509at_*() on the grounds that they don't handle X509
10731 structures and behave in an analogous way to the X509v3 functions:
10732 they shouldn't be called directly but wrapper functions should be used
10735 So we also now have some wrapper functions that call the X509at functions
10736 when passed certificate requests. (TO DO: similar things can be done with
10737 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10738 things. Some of these need some d2i or i2d and print functionality
10739 because they handle more complex structures.)
10742 *) Add missing #ifndefs that caused missing symbols when building libssl
10743 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10744 NO_RSA in ssl/s2*.c.
10745 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10747 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10748 has a return value which indicates the quality of the random data
10749 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10750 error queue. New function RAND_pseudo_bytes() generates output that is
10751 guaranteed to be unique but not unpredictable. RAND_add is like
10752 RAND_seed, but takes an extra argument for an entropy estimate
10753 (RAND_seed always assumes full entropy).
10756 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10757 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10758 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10759 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10760 false-positive rate of at most 2^-80 for random input.
10763 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10766 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10767 in the 0.9.5 release), this returns the chain
10768 from an X509_CTX structure with a dup of the stack and all
10769 the X509 reference counts upped: so the stack will exist
10770 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10773 Also make SSL_SESSION_print() print out the verify return
10777 *) Add manpage for the pkcs12 command. Also change the default
10778 behaviour so MAC iteration counts are used unless the new
10779 -nomaciter option is used. This improves file security and
10780 only older versions of MSIE (4.0 for example) need it.
10783 *) Honor the no-xxx Configure options when creating .DEF files.
10786 *) Add PKCS#10 attributes to field table: challengePassword,
10787 unstructuredName and unstructuredAddress. These are taken from
10788 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10789 international characters are used.
10791 More changes to X509_ATTRIBUTE code: allow the setting of types
10792 based on strings. Remove the 'loc' parameter when adding
10793 attributes because these will be a SET OF encoding which is sorted
10797 *) Initial changes to the 'req' utility to allow request generation
10798 automation. This will allow an application to just generate a template
10799 file containing all the field values and have req construct the
10802 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10803 used all over the place including certificate requests and PKCS#7
10804 structures. They are currently handled manually where necessary with
10805 some primitive wrappers for PKCS#7. The new functions behave in a
10806 manner analogous to the X509 extension functions: they allow
10807 attributes to be looked up by NID and added.
10809 Later something similar to the X509V3 code would be desirable to
10810 automatically handle the encoding, decoding and printing of the
10811 more complex types. The string types like challengePassword can
10812 be handled by the string table functions.
10814 Also modified the multi byte string table handling. Now there is
10815 a 'global mask' which masks out certain types. The table itself
10816 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10817 is useful when for example there is only one permissible type
10818 (as in countryName) and using the mask might result in no valid
10822 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10823 SSL_get_peer_finished to allow applications to obtain the latest
10824 Finished messages sent to the peer or expected from the peer,
10825 respectively. (SSL_get_peer_finished is usually the Finished message
10826 actually received from the peer, otherwise the protocol will be aborted.)
10828 As the Finished message are message digests of the complete handshake
10829 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10830 be used for external authentication procedures when the authentication
10831 provided by SSL/TLS is not desired or is not enough.
10834 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10835 the host supports BWX extension and if Compaq C is present on the
10836 $PATH. Just exploiting of the BWX extension results in 20-30%
10837 performance kick for some algorithms, e.g. DES and RC4 to mention
10838 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10842 *) Add support for MS "fast SGC". This is arguably a violation of the
10843 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10844 weak crypto and after checking the certificate is SGC a second one
10845 with strong crypto. MS SGC stops the first handshake after receiving
10846 the server certificate message and sends a second client hello. Since
10847 a server will typically do all the time consuming operations before
10848 expecting any further messages from the client (server key exchange
10849 is the most expensive) there is little difference between the two.
10851 To get OpenSSL to support MS SGC we have to permit a second client
10852 hello message after we have sent server done. In addition we have to
10853 reset the MAC if we do get this second client hello.
10856 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10857 if a DER encoded private key is RSA or DSA traditional format. Changed
10858 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10859 format DER encoded private key. Newer code should use PKCS#8 format which
10860 has the key type encoded in the ASN1 structure. Added DER private key
10861 support to pkcs8 application.
10864 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10865 ciphersuites has been selected (as required by the SSL 3/TLS 1
10866 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10867 is set, we interpret this as a request to violate the specification
10868 (the worst that can happen is a handshake failure, and 'correct'
10869 behaviour would result in a handshake failure anyway).
10872 *) In SSL_CTX_add_session, take into account that there might be multiple
10873 SSL_SESSION structures with the same session ID (e.g. when two threads
10874 concurrently obtain them from an external cache).
10875 The internal cache can handle only one SSL_SESSION with a given ID,
10876 so if there's a conflict, we now throw out the old one to achieve
10880 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10881 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10882 some routines that use cipher OIDs: some ciphers do not have OIDs
10883 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10887 *) Simplify the trust setting structure and code. Now we just have
10888 two sequences of OIDs for trusted and rejected settings. These will
10889 typically have values the same as the extended key usage extension
10890 and any application specific purposes.
10892 The trust checking code now has a default behaviour: it will just
10893 check for an object with the same NID as the passed id. Functions can
10894 be provided to override either the default behaviour or the behaviour
10895 for a given id. SSL client, server and email already have functions
10896 in place for compatibility: they check the NID and also return "trusted"
10897 if the certificate is self signed.
10900 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10901 traditional format into an EVP_PKEY structure.
10904 *) Add a password callback function PEM_cb() which either prompts for
10905 a password if usr_data is NULL or otherwise assumes it is a null
10906 terminated password. Allow passwords to be passed on command line
10907 environment or config files in a few more utilities.
10910 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10911 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10912 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10913 Update documentation.
10916 *) Support for ASN1 "NULL" type. This could be handled before by using
10917 ASN1_TYPE but there wasn't any function that would try to read a NULL
10918 and produce an error if it couldn't. For compatibility we also have
10919 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10920 don't allocate anything because they don't need to.
10923 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10925 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10927 *) Rebuild of the memory allocation routines used by OpenSSL code and
10928 possibly others as well. The purpose is to make an interface that
10929 provide hooks so anyone can build a separate set of allocation and
10930 deallocation routines to be used by OpenSSL, for example memory
10931 pool implementations, or something else, which was previously hard
10932 since Malloc(), Realloc() and Free() were defined as macros having
10933 the values malloc, realloc and free, respectively (except for Win32
10934 compilations). The same is provided for memory debugging code.
10935 OpenSSL already comes with functionality to find memory leaks, but
10936 this gives people a chance to debug other memory problems.
10938 With these changes, a new set of functions and macros have appeared:
10940 CRYPTO_set_mem_debug_functions() [F]
10941 CRYPTO_get_mem_debug_functions() [F]
10942 CRYPTO_dbg_set_options() [F]
10943 CRYPTO_dbg_get_options() [F]
10944 CRYPTO_malloc_debug_init() [M]
10946 The memory debug functions are NULL by default, unless the library
10947 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10948 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10949 gives the standard debugging functions that come with OpenSSL) or
10950 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10951 provided by the library user) must be used. When the standard
10952 debugging functions are used, CRYPTO_dbg_set_options can be used to
10953 request additional information:
10954 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10955 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10957 Also, things like CRYPTO_set_mem_functions will always give the
10958 expected result (the new set of functions is used for allocation
10959 and deallocation) at all times, regardless of platform and compiler
10962 To finish it up, some functions that were never use in any other
10963 way than through macros have a new API and new semantic:
10965 CRYPTO_dbg_malloc()
10966 CRYPTO_dbg_realloc()
10969 All macros of value have retained their old syntax.
10970 [Richard Levitte and Bodo Moeller]
10972 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10973 ordering of SMIMECapabilities wasn't in "strength order" and there
10974 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10978 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10979 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10980 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10982 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10983 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10984 functionality to handle multipart/signed properly) and a utility
10985 called 'smime' to call all this stuff. This is based on code I
10986 originally wrote for Celo who have kindly allowed it to be
10987 included in OpenSSL.
10990 *) Add variants des_set_key_checked and des_set_key_unchecked of
10991 des_set_key (aka des_key_sched). Global variable des_check_key
10992 decides which of these is called by des_set_key; this way
10993 des_check_key behaves as it always did, but applications and
10994 the library itself, which was buggy for des_check_key == 1,
10995 have a cleaner way to pick the version they need.
10998 *) New function PKCS12_newpass() which changes the password of a
11002 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11003 dynamic mix. In both cases the ids can be used as an index into the
11004 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11005 functions so they accept a list of the field values and the
11006 application doesn't need to directly manipulate the X509_TRUST
11010 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11014 *) Modify the way the V3 extension code looks up extensions. This now
11015 works in a similar way to the object code: we have some "standard"
11016 extensions in a static table which is searched with OBJ_bsearch()
11017 and the application can add dynamic ones if needed. The file
11018 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11019 updated whenever a new extension is added to the core code and kept
11020 in ext_nid order. There is a simple program 'tabtest.c' which checks
11021 this. New extensions are not added too often so this file can readily
11022 be maintained manually.
11024 There are two big advantages in doing things this way. The extensions
11025 can be looked up immediately and no longer need to be "added" using
11026 X509V3_add_standard_extensions(): this function now does nothing.
11027 [Side note: I get *lots* of email saying the extension code doesn't
11028 work because people forget to call this function]
11029 Also no dynamic allocation is done unless new extensions are added:
11030 so if we don't add custom extensions there is no need to call
11031 X509V3_EXT_cleanup().
11034 *) Modify enc utility's salting as follows: make salting the default. Add a
11035 magic header, so unsalted files fail gracefully instead of just decrypting
11036 to garbage. This is because not salting is a big security hole, so people
11037 should be discouraged from doing it.
11040 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11041 digest to be passed on the command line but it only used this
11042 parameter when signing a certificate. Modified so all relevant
11043 operations are affected by the digest parameter including the
11044 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11045 DSA key was used because it didn't fix the digest.
11048 *) Initial certificate chain verify code. Currently tests the untrusted
11049 certificates for consistency with the verify purpose (which is set
11050 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11052 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11053 this is because it will reject chains with invalid extensions whereas
11054 every previous version of OpenSSL and SSLeay made no checks at all.
11056 Trust code: checks the root CA for the relevant trust settings. Trust
11057 settings have an initial value consistent with the verify purpose: e.g.
11058 if the verify purpose is for SSL client use it expects the CA to be
11059 trusted for SSL client use. However the default value can be changed to
11060 permit custom trust settings: one example of this would be to only trust
11061 certificates from a specific "secure" set of CAs.
11063 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11064 which should be used for version portability: especially since the
11065 verify structure is likely to change more often now.
11067 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11068 to set them. If not set then assume SSL clients will verify SSL servers
11071 Two new options to the verify program: -untrusted allows a set of
11072 untrusted certificates to be passed in and -purpose which sets the
11073 intended purpose of the certificate. If a purpose is set then the
11074 new chain verify code is used to check extension consistency.
11077 *) Support for the authority information access extension.
11080 *) Modify RSA and DSA PEM read routines to transparently handle
11081 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11082 public keys in a format compatible with certificate
11083 SubjectPublicKeyInfo structures. Unfortunately there were already
11084 functions called *_PublicKey_* which used various odd formats so
11085 these are retained for compatibility: however the DSA variants were
11086 never in a public release so they have been deleted. Changed dsa/rsa
11087 utilities to handle the new format: note no releases ever handled public
11088 keys so we should be OK.
11090 The primary motivation for this change is to avoid the same fiasco
11091 that dogs private keys: there are several incompatible private key
11092 formats some of which are standard and some OpenSSL specific and
11093 require various evil hacks to allow partial transparent handling and
11094 even then it doesn't work with DER formats. Given the option anything
11095 other than PKCS#8 should be dumped: but the other formats have to
11096 stay in the name of compatibility.
11098 With public keys and the benefit of hindsight one standard format
11099 is used which works with EVP_PKEY, RSA or DSA structures: though
11100 it clearly returns an error if you try to read the wrong kind of key.
11102 Added a -pubkey option to the 'x509' utility to output the public key.
11103 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11104 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11105 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11106 that do the same as the EVP_PKEY_assign_*() except they up the
11107 reference count of the added key (they don't "swallow" the
11111 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11112 CRLs would fail if the file contained no certificates or no CRLs:
11113 added a new function to read in both types and return the number
11114 read: this means that if none are read it will be an error. The
11115 DER versions of the certificate and CRL reader would always fail
11116 because it isn't possible to mix certificates and CRLs in DER format
11117 without choking one or the other routine. Changed this to just read
11118 a certificate: this is the best we can do. Also modified the code
11119 in apps/verify.c to take notice of return codes: it was previously
11120 attempting to read in certificates from NULL pointers and ignoring
11121 any errors: this is one reason why the cert and CRL reader seemed
11122 to work. It doesn't check return codes from the default certificate
11123 routines: these may well fail if the certificates aren't installed.
11126 *) Code to support otherName option in GeneralName.
11129 *) First update to verify code. Change the verify utility
11130 so it warns if it is passed a self signed certificate:
11131 for consistency with the normal behaviour. X509_verify
11132 has been modified to it will now verify a self signed
11133 certificate if *exactly* the same certificate appears
11134 in the store: it was previously impossible to trust a
11135 single self signed certificate. This means that:
11136 openssl verify ss.pem
11137 now gives a warning about a self signed certificate but
11138 openssl verify -CAfile ss.pem ss.pem
11142 *) For servers, store verify_result in SSL_SESSION data structure
11143 (and add it to external session representation).
11144 This is needed when client certificate verifications fails,
11145 but an application-provided verification callback (set by
11146 SSL_CTX_set_cert_verify_callback) allows accepting the session
11147 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11148 but returns 1): When the session is reused, we have to set
11149 ssl->verify_result to the appropriate error code to avoid
11151 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11153 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11154 case in PKCS7_dataInit() where the signed PKCS7 structure
11155 didn't contain any existing data because it was being created.
11156 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11158 *) Add a salt to the key derivation routines in enc.c. This
11159 forms the first 8 bytes of the encrypted file. Also add a
11160 -S option to allow a salt to be input on the command line.
11163 *) New function X509_cmp(). Oddly enough there wasn't a function
11164 to compare two certificates. We do this by working out the SHA1
11165 hash and comparing that. X509_cmp() will be needed by the trust
11169 *) SSL_get1_session() is like SSL_get_session(), but increments
11170 the reference count in the SSL_SESSION returned.
11171 [Geoff Thorpe <geoff@eu.c2.net>]
11173 *) Fix for 'req': it was adding a null to request attributes.
11174 Also change the X509_LOOKUP and X509_INFO code to handle
11175 certificate auxiliary information.
11178 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11182 *) Add the possibility to add extra information to the memory leak
11183 detecting output, to form tracebacks, showing from where each
11184 allocation was originated: CRYPTO_push_info("constant string") adds
11185 the string plus current file name and line number to a per-thread
11186 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11187 is like calling CYRPTO_pop_info() until the stack is empty.
11188 Also updated memory leak detection code to be multi-thread-safe.
11191 *) Add options -text and -noout to pkcs7 utility and delete the
11192 encryption options which never did anything. Update docs.
11195 *) Add options to some of the utilities to allow the pass phrase
11196 to be included on either the command line (not recommended on
11197 OSes like Unix) or read from the environment. Update the
11198 manpages and fix a few bugs.
11201 *) Add a few manpages for some of the openssl commands.
11204 *) Fix the -revoke option in ca. It was freeing up memory twice,
11205 leaking and not finding already revoked certificates.
11208 *) Extensive changes to support certificate auxiliary information.
11209 This involves the use of X509_CERT_AUX structure and X509_AUX
11210 functions. An X509_AUX function such as PEM_read_X509_AUX()
11211 can still read in a certificate file in the usual way but it
11212 will also read in any additional "auxiliary information". By
11213 doing things this way a fair degree of compatibility can be
11214 retained: existing certificates can have this information added
11215 using the new 'x509' options.
11217 Current auxiliary information includes an "alias" and some trust
11218 settings. The trust settings will ultimately be used in enhanced
11219 certificate chain verification routines: currently a certificate
11220 can only be trusted if it is self signed and then it is trusted
11224 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11225 The problem was that one of the replacement routines had not been working
11226 since SSLeay releases. For now the offending routine has been replaced
11227 with non-optimised assembler. Even so, this now gives around 95%
11228 performance improvement for 1024 bit RSA signs.
11231 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11232 handling. Most clients have the effective key size in bits equal to
11233 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11234 A few however don't do this and instead use the size of the decrypted key
11235 to determine the RC2 key length and the AlgorithmIdentifier to determine
11236 the effective key length. In this case the effective key length can still
11237 be 40 bits but the key length can be 168 bits for example. This is fixed
11238 by manually forcing an RC2 key into the EVP_PKEY structure because the
11239 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11240 the key length and effective key length are equal.
11243 *) Add a bunch of functions that should simplify the creation of
11244 X509_NAME structures. Now you should be able to do:
11245 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11246 and have it automatically work out the correct field type and fill in
11247 the structures. The more adventurous can try:
11248 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11249 and it will (hopefully) work out the correct multibyte encoding.
11252 *) Change the 'req' utility to use the new field handling and multibyte
11253 copy routines. Before the DN field creation was handled in an ad hoc
11254 way in req, ca, and x509 which was rather broken and didn't support
11255 BMPStrings or UTF8Strings. Since some software doesn't implement
11256 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11257 using the dirstring_type option. See the new comment in the default
11258 openssl.cnf for more info.
11261 *) Make crypto/rand/md_rand.c more robust:
11262 - Assure unique random numbers after fork().
11263 - Make sure that concurrent threads access the global counter and
11264 md serializably so that we never lose entropy in them
11265 or use exactly the same state in multiple threads.
11266 Access to the large state is not always serializable because
11267 the additional locking could be a performance killer, and
11268 md should be large enough anyway.
11271 *) New file apps/app_rand.c with commonly needed functionality
11272 for handling the random seed file.
11274 Use the random seed file in some applications that previously did not:
11276 dsaparam -genkey (which also ignored its '-rand' option),
11279 x509 (when signing).
11280 Except on systems with /dev/urandom, it is crucial to have a random
11281 seed file at least for key creation, DSA signing, and for DH exchanges;
11282 for RSA signatures we could do without one.
11284 gendh and gendsa (unlike genrsa) used to read only the first byte
11285 of each file listed in the '-rand' option. The function as previously
11286 found in genrsa is now in app_rand.c and is used by all programs
11287 that support '-rand'.
11290 *) In RAND_write_file, use mode 0600 for creating files;
11291 don't just chmod when it may be too late.
11294 *) Report an error from X509_STORE_load_locations
11295 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11298 *) New function ASN1_mbstring_copy() this copies a string in either
11299 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11300 into an ASN1_STRING type. A mask of permissible types is passed
11301 and it chooses the "minimal" type to use or an error if not type
11305 *) Add function equivalents to the various macros in asn1.h. The old
11306 macros are retained with an M_ prefix. Code inside the library can
11307 use the M_ macros. External code (including the openssl utility)
11308 should *NOT* in order to be "shared library friendly".
11311 *) Add various functions that can check a certificate's extensions
11312 to see if it usable for various purposes such as SSL client,
11313 server or S/MIME and CAs of these types. This is currently
11314 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11315 verification. Also added a -purpose flag to x509 utility to
11316 print out all the purposes.
11319 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11323 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11324 for, obtain and decode and extension and obtain its critical flag.
11325 This allows all the necessary extension code to be handled in a
11326 single function call.
11329 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11330 platforms. See crypto/rc4/rc4_enc.c for further details.
11333 *) New -noout option to asn1parse. This causes no output to be produced
11334 its main use is when combined with -strparse and -out to extract data
11335 from a file (which may not be in ASN.1 format).
11338 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11339 when producing the local key id.
11340 [Richard Levitte <levitte@stacken.kth.se>]
11342 *) New option -dhparam in s_server. This allows a DH parameter file to be
11343 stated explicitly. If it is not stated then it tries the first server
11344 certificate file. The previous behaviour hard coded the filename
11348 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11349 a public key to be input or output. For example:
11350 openssl rsa -in key.pem -pubout -out pubkey.pem
11351 Also added necessary DSA public key functions to handle this.
11354 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11355 in the message. This was handled by allowing
11356 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11357 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11359 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11360 to the end of the strings whereas this didn't. This would cause problems
11361 if strings read with d2i_ASN1_bytes() were later modified.
11362 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11364 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11365 data and it contains EOF it will end up returning an error. This is
11366 caused by input 46 bytes long. The cause is due to the way base64
11367 BIOs find the start of base64 encoded data. They do this by trying a
11368 trial decode on each line until they find one that works. When they
11369 do a flag is set and it starts again knowing it can pass all the
11370 data directly through the decoder. Unfortunately it doesn't reset
11371 the context it uses. This means that if EOF is reached an attempt
11372 is made to pass two EOFs through the context and this causes the
11373 resulting error. This can also cause other problems as well. As is
11374 usual with these problems it takes *ages* to find and the fix is
11375 trivial: move one line.
11376 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11378 *) Ugly workaround to get s_client and s_server working under Windows. The
11379 old code wouldn't work because it needed to select() on sockets and the
11380 tty (for keypresses and to see if data could be written). Win32 only
11381 supports select() on sockets so we select() with a 1s timeout on the
11382 sockets and then see if any characters are waiting to be read, if none
11383 are present then we retry, we also assume we can always write data to
11384 the tty. This isn't nice because the code then blocks until we've
11385 received a complete line of data and it is effectively polling the
11386 keyboard at 1s intervals: however it's quite a bit better than not
11387 working at all :-) A dedicated Windows application might handle this
11388 with an event loop for example.
11391 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11392 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11393 will be called when RSA_sign() and RSA_verify() are used. This is useful
11394 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11395 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11396 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11397 This necessitated the support of an extra signature type NID_md5_sha1
11398 for SSL signatures and modifications to the SSL library to use it instead
11399 of calling RSA_public_decrypt() and RSA_private_encrypt().
11402 *) Add new -verify -CAfile and -CApath options to the crl program, these
11403 will lookup a CRL issuers certificate and verify the signature in a
11404 similar way to the verify program. Tidy up the crl program so it
11405 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11406 less strict. It will now permit CRL extensions even if it is not
11407 a V2 CRL: this will allow it to tolerate some broken CRLs.
11410 *) Initialize all non-automatic variables each time one of the openssl
11411 sub-programs is started (this is necessary as they may be started
11412 multiple times from the "OpenSSL>" prompt).
11413 [Lennart Bang, Bodo Moeller]
11415 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11416 removing all other RSA functionality (this is what NO_RSA does). This
11417 is so (for example) those in the US can disable those operations covered
11418 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11422 *) Non-copying interface to BIO pairs.
11423 (still largely untested)
11426 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11427 ASCII string. This was handled independently in various places before.
11430 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11431 UTF8 strings a character at a time.
11434 *) Use client_version from client hello to select the protocol
11435 (s23_srvr.c) and for RSA client key exchange verification
11436 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11439 *) Add various utility functions to handle SPKACs, these were previously
11440 handled by poking round in the structure internals. Added new function
11441 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11442 print, verify and generate SPKACs. Based on an original idea from
11443 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11446 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11449 *) Allow the config file extension section to be overwritten on the
11450 command line. Based on an original idea from Massimiliano Pala
11451 <madwolf@comune.modena.it>. The new option is called -extensions
11452 and can be applied to ca, req and x509. Also -reqexts to override
11453 the request extensions in req and -crlexts to override the crl extensions
11457 *) Add new feature to the SPKAC handling in ca. Now you can include
11458 the same field multiple times by preceding it by "XXXX." for example:
11461 this is the same syntax as used in the req config file.
11464 *) Allow certificate extensions to be added to certificate requests. These
11465 are specified in a 'req_extensions' option of the req section of the
11466 config file. They can be printed out with the -text option to req but
11467 are otherwise ignored at present.
11470 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11471 data read consists of only the final block it would not decrypted because
11472 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11473 A misplaced 'break' also meant the decrypted final block might not be
11474 copied until the next read.
11477 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11478 a few extra parameters to the DH structure: these will be useful if
11479 for example we want the value of 'q' or implement X9.42 DH.
11482 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11483 provides hooks that allow the default DSA functions or functions on a
11484 "per key" basis to be replaced. This allows hardware acceleration and
11485 hardware key storage to be handled without major modification to the
11486 library. Also added low level modexp hooks and CRYPTO_EX structure and
11487 associated functions.
11490 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11491 as "read only": it can't be written to and the buffer it points to will
11492 not be freed. Reading from a read only BIO is much more efficient than
11493 a normal memory BIO. This was added because there are several times when
11494 an area of memory needs to be read from a BIO. The previous method was
11495 to create a memory BIO and write the data to it, this results in two
11496 copies of the data and an O(n^2) reading algorithm. There is a new
11497 function BIO_new_mem_buf() which creates a read only memory BIO from
11498 an area of memory. Also modified the PKCS#7 routines to use read only
11502 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11503 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11504 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11505 but a retry condition occurred while trying to read the rest.
11508 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11509 NID_pkcs7_encrypted by default: this was wrong since this should almost
11510 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11511 the encrypted data type: this is a more sensible place to put it and it
11512 allows the PKCS#12 code to be tidied up that duplicated this
11516 *) Changed obj_dat.pl script so it takes its input and output files on
11517 the command line. This should avoid shell escape redirection problems
11521 *) Initial support for certificate extension requests, these are included
11522 in things like Xenroll certificate requests. Included functions to allow
11523 extensions to be obtained and added.
11526 *) -crlf option to s_client and s_server for sending newlines as
11527 CRLF (as required by many protocols).
11530 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11532 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11533 [Ralf S. Engelschall]
11535 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11536 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11538 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11542 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11543 DH parameters/keys (q is lost during that conversion, but the resulting
11544 DH parameters contain its length).
11546 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11547 much faster than DH_generate_parameters (which creates parameters
11548 where p = 2*q + 1), and also the smaller q makes DH computations
11549 much more efficient (160-bit exponentiation instead of 1024-bit
11550 exponentiation); so this provides a convenient way to support DHE
11551 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11552 utter importance to use
11553 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11555 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11556 when such DH parameters are used, because otherwise small subgroup
11557 attacks may become possible!
11560 *) Avoid memory leak in i2d_DHparams.
11563 *) Allow the -k option to be used more than once in the enc program:
11564 this allows the same encrypted message to be read by multiple recipients.
11567 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11568 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11569 it will always use the numerical form of the OID, even if it has a short
11573 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11574 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11575 otherwise bn_mod_exp was called. In the case of hardware keys for example
11576 no private key components need be present and it might store extra data
11577 in the RSA structure, which cannot be accessed from bn_mod_exp.
11578 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11579 private key operations.
11582 *) Added support for SPARC Linux.
11585 *) pem_password_cb function type incompatibly changed from
11586 typedef int pem_password_cb(char *buf, int size, int rwflag);
11588 ....(char *buf, int size, int rwflag, void *userdata);
11589 so that applications can pass data to their callbacks:
11590 The PEM[_ASN1]_{read,write}... functions and macros now take an
11591 additional void * argument, which is just handed through whenever
11592 the password callback is called.
11593 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11595 New function SSL_CTX_set_default_passwd_cb_userdata.
11597 Compatibility note: As many C implementations push function arguments
11598 onto the stack in reverse order, the new library version is likely to
11599 interoperate with programs that have been compiled with the old
11600 pem_password_cb definition (PEM_whatever takes some data that
11601 happens to be on the stack as its last argument, and the callback
11602 just ignores this garbage); but there is no guarantee whatsoever that
11605 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11606 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11607 problems not only on Windows, but also on some Unix platforms.
11608 To avoid problematic command lines, these definitions are now in an
11609 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11610 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11613 *) MIPS III/IV assembler module is reimplemented.
11616 *) More DES library cleanups: remove references to srand/rand and
11617 delete an unused file.
11620 *) Add support for the free Netwide assembler (NASM) under Win32,
11621 since not many people have MASM (ml) and it can be hard to obtain.
11622 This is currently experimental but it seems to work OK and pass all
11623 the tests. Check out INSTALL.W32 for info.
11626 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11627 without temporary keys kept an extra copy of the server key,
11628 and connections with temporary keys did not free everything in case
11632 *) New function RSA_check_key and new openssl rsa option -check
11633 for verifying the consistency of RSA keys.
11634 [Ulf Moeller, Bodo Moeller]
11636 *) Various changes to make Win32 compile work:
11637 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11638 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11639 comparison" warnings.
11640 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11643 *) Add a debugging option to PKCS#5 v2 key generation function: when
11644 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11645 derived keys are printed to stderr.
11648 *) Copy the flags in ASN1_STRING_dup().
11649 [Roman E. Pavlov <pre@mo.msk.ru>]
11651 *) The x509 application mishandled signing requests containing DSA
11652 keys when the signing key was also DSA and the parameters didn't match.
11654 It was supposed to omit the parameters when they matched the signing key:
11655 the verifying software was then supposed to automatically use the CA's
11656 parameters if they were absent from the end user certificate.
11658 Omitting parameters is no longer recommended. The test was also
11659 the wrong way round! This was probably due to unusual behaviour in
11660 EVP_cmp_parameters() which returns 1 if the parameters match.
11661 This meant that parameters were omitted when they *didn't* match and
11662 the certificate was useless. Certificates signed with 'ca' didn't have
11664 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11666 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11667 The interface is as follows:
11668 Applications can use
11669 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11670 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11671 "off" is now the default.
11672 The library internally uses
11673 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11674 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11675 to disable memory-checking temporarily.
11677 Some inconsistent states that previously were possible (and were
11678 even the default) are now avoided.
11680 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11681 with each memory chunk allocated; this is occasionally more helpful
11682 than just having a counter.
11684 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11686 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11690 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11691 which largely parallels "options", but is for changing API behaviour,
11692 whereas "options" are about protocol behaviour.
11693 Initial "mode" flags are:
11695 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11696 a single record has been written.
11697 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11698 retries use the same buffer location.
11699 (But all of the contents must be
11703 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11706 *) Fix problems with no-hmac etc.
11707 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11709 *) New functions RSA_get_default_method(), RSA_set_method() and
11710 RSA_get_method(). These allows replacement of RSA_METHODs without having
11711 to mess around with the internals of an RSA structure.
11714 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11715 Also really enable memory leak checks in openssl.c and in some
11717 [Chad C. Mulligan, Bodo Moeller]
11719 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11720 up the length of negative integers. This has now been simplified to just
11721 store the length when it is first determined and use it later, rather
11722 than trying to keep track of where data is copied and updating it to
11724 [Steve Henson, reported by Brien Wheeler
11725 <bwheeler@authentica-security.com>]
11727 *) Add a new function PKCS7_signatureVerify. This allows the verification
11728 of a PKCS#7 signature but with the signing certificate passed to the
11729 function itself. This contrasts with PKCS7_dataVerify which assumes the
11730 certificate is present in the PKCS#7 structure. This isn't always the
11731 case: certificates can be omitted from a PKCS#7 structure and be
11732 distributed by "out of band" means (such as a certificate database).
11735 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11736 function prototypes in pem.h, also change util/mkdef.pl to add the
11737 necessary function names.
11740 *) mk1mf.pl (used by Windows builds) did not properly read the
11741 options set by Configure in the top level Makefile, and Configure
11742 was not even able to write more than one option correctly.
11743 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11746 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11747 file to be loaded from a BIO or FILE pointer. The BIO version will
11748 for example allow memory BIOs to contain config info.
11751 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11752 Whoever hopes to achieve shared-library compatibility across versions
11753 must use this, not the compile-time macro.
11754 (Exercise 0.9.4: Which is the minimum library version required by
11756 Note: All this applies only to multi-threaded programs, others don't
11760 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11761 through a BIO pair triggered the default case, i.e.
11762 SSLerr(...,SSL_R_UNKNOWN_STATE).
11765 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11766 can use the SSL library even if none of the specific BIOs is
11770 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11771 for the encoded length.
11772 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11774 *) Add initial documentation of the X509V3 functions.
11777 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11778 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11779 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11780 secure PKCS#8 private key format with a high iteration count.
11783 *) Fix determination of Perl interpreter: A perl or perl5
11784 _directory_ in $PATH was also accepted as the interpreter.
11785 [Ralf S. Engelschall]
11787 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11788 wrong with it but it was very old and did things like calling
11789 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11790 unusual formatting.
11793 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11794 to use the new extension code.
11797 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11798 with macros. This should make it easier to change their form, add extra
11799 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11803 *) Add to configuration table a new entry that can specify an alternative
11804 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11805 according to Mark Crispin <MRC@Panda.COM>.
11809 *) DES CBC did not update the IV. Weird.
11812 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11813 Changing the behaviour of the former might break existing programs --
11814 where IV updating is needed, des_ncbc_encrypt can be used.
11817 *) When bntest is run from "make test" it drives bc to check its
11818 calculations, as well as internally checking them. If an internal check
11819 fails, it needs to cause bc to give a non-zero result or make test carries
11820 on without noticing the failure. Fixed.
11823 *) DES library cleanups.
11826 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11827 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11828 ciphers. NOTE: although the key derivation function has been verified
11829 against some published test vectors it has not been extensively tested
11830 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11834 *) Instead of "mkdir -p", which is not fully portable, use new
11835 Perl script "util/mkdir-p.pl".
11838 *) Rewrite the way password based encryption (PBE) is handled. It used to
11839 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11840 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11841 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11842 the 'parameter' field of the AlgorithmIdentifier is passed to the
11843 underlying key generation function so it must do its own ASN1 parsing.
11844 This has also changed the EVP_PBE_CipherInit() function which now has a
11845 'parameter' argument instead of literal salt and iteration count values
11846 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11849 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11850 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11851 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11852 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11853 value was just used as a "magic string" and not used directly its
11854 value doesn't matter.
11857 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11861 *) "linux-sparc64" configuration (ultrapenguin).
11862 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11863 "linux-sparc" configuration.
11864 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11866 *) config now generates no-xxx options for missing ciphers.
11869 *) Support the EBCDIC character set (work in progress).
11870 File ebcdic.c not yet included because it has a different license.
11871 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11873 *) Support BS2000/OSD-POSIX.
11874 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11876 *) Make callbacks for key generation use void * instead of char *.
11879 *) Make S/MIME samples compile (not yet tested).
11882 *) Additional typesafe stacks.
11885 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11889 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11891 *) New configuration variant "sco5-gcc".
11893 *) Updated some demos.
11894 [Sean O Riordain, Wade Scholine]
11896 *) Add missing BIO_free at exit of pkcs12 application.
11899 *) Fix memory leak in conf.c.
11902 *) Updates for Win32 to assembler version of MD5.
11905 *) Set #! path to perl in apps/der_chop to where we found it
11906 instead of using a fixed path.
11909 *) SHA library changes for irix64-mips4-cc.
11912 *) Improvements for VMS support.
11916 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11918 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11919 This also avoids the problems with SC4.2 and unpatched SC5.
11920 [Andy Polyakov <appro@fy.chalmers.se>]
11922 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11923 These are required because of the typesafe stack would otherwise break
11924 existing code. If old code used a structure member which used to be STACK
11925 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11926 sk_num or sk_value it would produce an error because the num, data members
11927 are not present in STACK_OF. Now it just produces a warning. sk_set
11928 replaces the old method of assigning a value to sk_value
11929 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11930 that does this will no longer work (and should use sk_set instead) but
11931 this could be regarded as a "questionable" behaviour anyway.
11934 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11935 correctly handle encrypted S/MIME data.
11938 *) Change type of various DES function arguments from des_cblock
11939 (which means, in function argument declarations, pointer to char)
11940 to des_cblock * (meaning pointer to array with 8 char elements),
11941 which allows the compiler to do more typechecking; it was like
11942 that back in SSLeay, but with lots of ugly casts.
11944 Introduce new type const_des_cblock.
11947 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11948 problems: find RecipientInfo structure that matches recipient certificate
11949 and initialise the ASN1 structures properly based on passed cipher.
11952 *) Belatedly make the BN tests actually check the results.
11955 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11956 to and from BNs: it was completely broken. New compilation option
11957 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11958 key elements as negative integers.
11961 *) Reorganize and speed up MD5.
11962 [Andy Polyakov <appro@fy.chalmers.se>]
11965 [Richard Levitte <richard@levitte.org>]
11967 *) New option -out to asn1parse to allow the parsed structure to be
11968 output to a file. This is most useful when combined with the -strparse
11969 option to examine the output of things like OCTET STRINGS.
11972 *) Make SSL library a little more fool-proof by not requiring any longer
11973 that SSL_set_{accept,connect}_state be called before
11974 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11975 in many applications because usually everything *appeared* to work as
11976 intended anyway -- now it really works as intended).
11979 *) Move openssl.cnf out of lib/.
11982 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11983 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11984 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11985 [Ralf S. Engelschall]
11987 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11988 handle PKCS#7 enveloped data properly.
11989 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11991 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11992 copying pointers. The cert_st handling is changed by this in
11993 various ways (and thus what used to be known as ctx->default_cert
11994 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11995 any longer when s->cert does not give us what we need).
11996 ssl_cert_instantiate becomes obsolete by this change.
11997 As soon as we've got the new code right (possibly it already is?),
11998 we have solved a couple of bugs of the earlier code where s->cert
11999 was used as if it could not have been shared with other SSL structures.
12001 Note that using the SSL API in certain dirty ways now will result
12002 in different behaviour than observed with earlier library versions:
12003 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12004 does not influence s as it used to.
12006 In order to clean up things more thoroughly, inside SSL_SESSION
12007 we don't use CERT any longer, but a new structure SESS_CERT
12008 that holds per-session data (if available); currently, this is
12009 the peer's certificate chain and, for clients, the server's certificate
12010 and temporary key. CERT holds only those values that can have
12011 meaningful defaults in an SSL_CTX.
12014 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12015 from the internal representation. Various PKCS#7 fixes: remove some
12016 evil casts and set the enc_dig_alg field properly based on the signing
12020 *) Allow PKCS#12 password to be set from the command line or the
12021 environment. Let 'ca' get its config file name from the environment
12022 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12026 *) Allow certificate policies extension to use an IA5STRING for the
12027 organization field. This is contrary to the PKIX definition but
12028 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12032 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12033 without disallowing inline assembler and the like for non-pedantic builds.
12036 *) Support Borland C++ builder.
12037 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12039 *) Support Mingw32.
12042 *) SHA-1 cleanups and performance enhancements.
12043 [Andy Polyakov <appro@fy.chalmers.se>]
12045 *) Sparc v8plus assembler for the bignum library.
12046 [Andy Polyakov <appro@fy.chalmers.se>]
12048 *) Accept any -xxx and +xxx compiler options in Configure.
12051 *) Update HPUX configuration.
12054 *) Add missing sk_<type>_unshift() function to safestack.h
12055 [Ralf S. Engelschall]
12057 *) New function SSL_CTX_use_certificate_chain_file that sets the
12058 "extra_cert"s in addition to the certificate. (This makes sense
12059 only for "PEM" format files, as chains as a whole are not
12063 *) Support verify_depth from the SSL API.
12064 x509_vfy.c had what can be considered an off-by-one-error:
12065 Its depth (which was not part of the external interface)
12066 was actually counting the number of certificates in a chain;
12067 now it really counts the depth.
12070 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12071 instead of X509err, which often resulted in confusing error
12072 messages since the error codes are not globally unique
12073 (e.g. an alleged error in ssl3_accept when a certificate
12074 didn't match the private key).
12076 *) New function SSL_CTX_set_session_id_context that allows to set a default
12077 value (so that you don't need SSL_set_session_id_context for each
12078 connection using the SSL_CTX).
12081 *) OAEP decoding bug fix.
12084 *) Support INSTALL_PREFIX for package builders, as proposed by
12088 *) New Configure options "threads" and "no-threads". For systems
12089 where the proper compiler options are known (currently Solaris
12090 and Linux), "threads" is the default.
12093 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12096 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12097 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12098 such as /usr/local/bin.
12101 *) "make linux-shared" to build shared libraries.
12102 [Niels Poppe <niels@netbox.org>]
12104 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12107 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12108 extension adding in x509 utility.
12111 *) Remove NOPROTO sections and error code comments.
12114 *) Partial rewrite of the DEF file generator to now parse the ANSI
12118 *) New Configure options --prefix=DIR and --openssldir=DIR.
12121 *) Complete rewrite of the error code script(s). It is all now handled
12122 by one script at the top level which handles error code gathering,
12123 header rewriting and C source file generation. It should be much better
12124 than the old method: it now uses a modified version of Ulf's parser to
12125 read the ANSI prototypes in all header files (thus the old K&R definitions
12126 aren't needed for error creation any more) and do a better job of
12127 translating function codes into names. The old 'ASN1 error code imbedded
12128 in a comment' is no longer necessary and it doesn't use .err files which
12129 have now been deleted. Also the error code call doesn't have to appear all
12130 on one line (which resulted in some large lines...).
12133 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12136 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12137 0 (which usually indicates a closed connection), but continue reading.
12140 *) Fix some race conditions.
12143 *) Add support for CRL distribution points extension. Add Certificate
12144 Policies and CRL distribution points documentation.
12147 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12150 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12151 8 of keying material. Merlin has also confirmed interop with this fix
12152 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12153 [Merlin Hughes <merlin@baltimore.ie>]
12155 *) Fix lots of warnings.
12156 [Richard Levitte <levitte@stacken.kth.se>]
12158 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12159 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12160 [Richard Levitte <levitte@stacken.kth.se>]
12162 *) Fix problems with sizeof(long) == 8.
12163 [Andy Polyakov <appro@fy.chalmers.se>]
12165 *) Change functions to ANSI C.
12168 *) Fix typos in error codes.
12169 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12171 *) Remove defunct assembler files from Configure.
12174 *) SPARC v8 assembler BIGNUM implementation.
12175 [Andy Polyakov <appro@fy.chalmers.se>]
12177 *) Support for Certificate Policies extension: both print and set.
12178 Various additions to support the r2i method this uses.
12181 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12182 return a const string when you are expecting an allocated buffer.
12185 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12186 types DirectoryString and DisplayText.
12189 *) Add code to allow r2i extensions to access the configuration database,
12190 add an LHASH database driver and add several ctx helper functions.
12193 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12194 fail when they extended the size of a BIGNUM.
12197 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12198 support typesafe stack.
12201 *) Fix typo in SSL_[gs]et_options().
12202 [Nils Frostberg <nils@medcom.se>]
12204 *) Delete various functions and files that belonged to the (now obsolete)
12205 old X509V3 handling code.
12208 *) New Configure option "rsaref".
12211 *) Don't auto-generate pem.h.
12214 *) Introduce type-safe ASN.1 SETs.
12217 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12218 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12220 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12221 that links with OpenSSL (well at least cause lots of warnings), but fear
12222 not: the conversion is trivial, and it eliminates loads of evil casts. A
12223 few STACKed things have been converted already. Feel free to convert more.
12224 In the fullness of time, I'll do away with the STACK type altogether.
12227 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12228 specified in <certfile> by updating the entry in the index.txt file.
12229 This way one no longer has to edit the index.txt file manually for
12230 revoking a certificate. The -revoke option does the gory details now.
12231 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12233 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12234 `-text' option at all and this way the `-noout -text' combination was
12235 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12236 [Ralf S. Engelschall]
12238 *) Make sure a corresponding plain text error message exists for the
12239 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12240 verify callback function determined that a certificate was revoked.
12241 [Ralf S. Engelschall]
12243 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12244 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12245 all available ciphers including rc5, which was forgotten until now.
12246 In order to let the testing shell script know which algorithms
12247 are available, a new (up to now undocumented) command
12248 "openssl list-cipher-commands" is used.
12251 *) Bugfix: s_client occasionally would sleep in select() when
12252 it should have checked SSL_pending() first.
12255 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12256 the raw DSA values prior to ASN.1 encoding.
12259 *) Tweaks to Configure
12260 [Niels Poppe <niels@netbox.org>]
12262 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12266 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12269 *) New config option to avoid instructions that are illegal on the 80386.
12270 The default code is faster, but requires at least a 486.
12273 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12274 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12275 same as SSL2_VERSION anyway.
12278 *) New "-showcerts" option for s_client.
12281 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12282 application. Various cleanups and fixes.
12285 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12286 modify error routines to work internally. Add error codes and PBE init
12287 to library startup routines.
12290 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12291 packing functions to asn1 and evp. Changed function names and error
12292 codes along the way.
12295 *) PKCS12 integration: and so it begins... First of several patches to
12296 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12297 objects to objects.h
12300 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12301 and display support for Thawte strong extranet extension.
12304 *) Add LinuxPPC support.
12305 [Jeff Dubrule <igor@pobox.org>]
12307 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12308 bn_div_words in alpha.s.
12309 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12311 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12312 OAEP isn't supported when OpenSSL is built with RSAref.
12313 [Ulf Moeller <ulf@fitug.de>]
12315 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12316 so they no longer are missing under -DNOPROTO.
12317 [Soren S. Jorvang <soren@t.dk>]
12320 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12322 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12323 doesn't work when the session is reused. Coming soon!
12326 *) Fix a security hole, that allows sessions to be reused in the wrong
12327 context thus bypassing client cert protection! All software that uses
12328 client certs and session caches in multiple contexts NEEDS PATCHING to
12329 allow session reuse! A fuller solution is in the works.
12330 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12332 *) Some more source tree cleanups (removed obsolete files
12333 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12334 permission on "config" script to be executable) and a fix for the INSTALL
12336 [Ulf Moeller <ulf@fitug.de>]
12338 *) Remove some legacy and erroneous uses of malloc, free instead of
12340 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12342 *) Make rsa_oaep_test return non-zero on error.
12343 [Ulf Moeller <ulf@fitug.de>]
12345 *) Add support for native Solaris shared libraries. Configure
12346 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12347 if someone would make that last step automatic.
12348 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12350 *) ctx_size was not built with the right compiler during "make links". Fixed.
12353 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12354 except NULL ciphers". This means the default cipher list will no longer
12355 enable NULL ciphers. They need to be specifically enabled e.g. with
12356 the string "DEFAULT:eNULL".
12359 *) Fix to RSA private encryption routines: if p < q then it would
12360 occasionally produce an invalid result. This will only happen with
12361 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12364 *) Be less restrictive and allow also `perl util/perlpath.pl
12365 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12366 because this way one can also use an interpreter named `perl5' (which is
12367 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12368 installed as `perl').
12369 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12371 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12372 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12374 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12375 advapi32.lib to Win32 build and change the pem test comparison
12376 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12377 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12378 and crypto/des/ede_cbcm_enc.c.
12381 *) DES quad checksum was broken on big-endian architectures. Fixed.
12384 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12385 Win32 test batch file so it (might) work again. The Win32 test batch file
12386 is horrible: I feel ill....
12389 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12390 in e_os.h. Audit of header files to check ANSI and non ANSI
12391 sections: 10 functions were absent from non ANSI section and not exported
12392 from Windows DLLs. Fixed up libeay.num for new functions.
12395 *) Make `openssl version' output lines consistent.
12396 [Ralf S. Engelschall]
12398 *) Fix Win32 symbol export lists for BIO functions: Added
12399 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12400 to ms/libeay{16,32}.def.
12401 [Ralf S. Engelschall]
12403 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12404 fine under Unix and passes some trivial tests I've now added. But the
12405 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12406 added to make sure no one expects that this stuff really works in the
12407 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12408 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12410 [Ralf S. Engelschall]
12412 *) Fix the generation of two part addresses in perl.
12413 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12415 *) Add config entry for Linux on MIPS.
12416 [John Tobey <jtobey@channel1.com>]
12418 *) Make links whenever Configure is run, unless we are on Windoze.
12421 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12422 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12426 *) Add a useful kludge to allow package maintainers to specify compiler and
12427 other platforms details on the command line without having to patch the
12428 Configure script everytime: One now can use ``perl Configure
12429 <id>:<details>'', i.e. platform ids are allowed to have details appended
12430 to them (separated by colons). This is treated as there would be a static
12431 pre-configured entry in Configure's %table under key <id> with value
12432 <details> and ``perl Configure <id>'' is called. So, when you want to
12433 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12434 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12435 now, which overrides the FreeBSD-elf entry on-the-fly.
12436 [Ralf S. Engelschall]
12438 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12441 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12442 on the `perl Configure ...' command line. This way one can compile
12443 OpenSSL libraries with Position Independent Code (PIC) which is needed
12444 for linking it into DSOs.
12445 [Ralf S. Engelschall]
12447 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12451 *) Cleaned up the LICENSE document: The official contact for any license
12452 questions now is the OpenSSL core team under openssl-core@openssl.org.
12453 And add a paragraph about the dual-license situation to make sure people
12454 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12455 to the OpenSSL toolkit.
12456 [Ralf S. Engelschall]
12458 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12459 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12460 Additionally cleaned up the `make links' target: Remove unnecessary
12461 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12462 to speed processing and no longer clutter the display with confusing
12463 stuff. Instead only the actually done links are displayed.
12464 [Ralf S. Engelschall]
12466 *) Permit null encryption ciphersuites, used for authentication only. It used
12467 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12468 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12472 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12473 signed attributes when verifying signatures (this would break them),
12474 the detached data encoding was wrong and public keys obtained using
12475 X509_get_pubkey() weren't freed.
12478 *) Add text documentation for the BUFFER functions. Also added a work around
12479 to a Win95 console bug. This was triggered by the password read stuff: the
12480 last character typed gets carried over to the next fread(). If you were
12481 generating a new cert request using 'req' for example then the last
12482 character of the passphrase would be CR which would then enter the first
12486 *) Added the new `Includes OpenSSL Cryptography Software' button as
12487 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12488 button and can be used by applications based on OpenSSL to show the
12489 relationship to the OpenSSL project.
12490 [Ralf S. Engelschall]
12492 *) Remove confusing variables in function signatures in files
12493 ssl/ssl_lib.c and ssl/ssl.h.
12494 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12496 *) Don't install bss_file.c under PREFIX/include/
12497 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12499 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12500 functions that return function pointers and has support for NT specific
12501 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12502 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12503 unsigned to signed types: this was killing the Win32 compile.
12506 *) Add new certificate file to stack functions,
12507 SSL_add_dir_cert_subjects_to_stack() and
12508 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12509 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12510 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12511 This means that Apache-SSL and similar packages don't have to mess around
12512 to add as many CAs as they want to the preferred list.
12515 *) Experiment with doxygen documentation. Currently only partially applied to
12517 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12518 openssl.doxy as the configuration file.
12521 *) Get rid of remaining C++-style comments which strict C compilers hate.
12522 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12524 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12525 compiled in by default: it has problems with large keys.
12528 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12529 DH private keys and/or callback functions which directly correspond to
12530 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12531 is needed for applications which have to configure certificates on a
12532 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12534 For the RSA certificate situation is makes no difference, but
12535 for the DSA certificate situation this fixes the "no shared cipher"
12536 problem where the OpenSSL cipher selection procedure failed because the
12537 temporary keys were not overtaken from the context and the API provided
12538 no way to reconfigure them.
12539 The new functions now let applications reconfigure the stuff and they
12540 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12541 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12542 non-public-API function ssl_cert_instantiate() is used as a helper
12543 function and also to reduce code redundancy inside ssl_rsa.c.
12544 [Ralf S. Engelschall]
12546 *) Move s_server -dcert and -dkey options out of the undocumented feature
12547 area because they are useful for the DSA situation and should be
12548 recognized by the users.
12549 [Ralf S. Engelschall]
12551 *) Fix the cipher decision scheme for export ciphers: the export bits are
12552 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12553 SSL_EXP_MASK. So, the original variable has to be used instead of the
12554 already masked variable.
12555 [Richard Levitte <levitte@stacken.kth.se>]
12557 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12558 [Richard Levitte <levitte@stacken.kth.se>]
12560 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12561 from `int' to `unsigned int' because it's a length and initialized by
12562 EVP_DigestFinal() which expects an `unsigned int *'.
12563 [Richard Levitte <levitte@stacken.kth.se>]
12565 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12566 script. Instead use the usual Shell->Perl transition trick.
12567 [Ralf S. Engelschall]
12569 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12570 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12571 -noout -modulus' as it's already the case for `openssl rsa -noout
12572 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12573 currently the public key is printed (a decision which was already done by
12574 `openssl dsa -modulus' in the past) which serves a similar purpose.
12575 Additionally the NO_RSA no longer completely removes the whole -modulus
12576 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12578 [Ralf S. Engelschall]
12580 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12581 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12582 [Arne Ansper <arne@ats.cyber.ee>]
12584 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12585 to be added. Now both 'req' and 'ca' can use new objects defined in the
12589 *) Add cool BIO that does syslog (or event log on NT).
12590 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12592 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12593 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12594 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12595 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12598 *) Add preliminary config info for new extension code.
12601 *) Make RSA_NO_PADDING really use no padding.
12602 [Ulf Moeller <ulf@fitug.de>]
12604 *) Generate errors when private/public key check is done.
12607 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12608 for some CRL extensions and new objects added.
12611 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12612 key usage extension and fuller support for authority key id.
12615 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12616 padding method for RSA, which is recommended for new applications in PKCS
12617 #1 v2.0 (RFC 2437, October 1998).
12618 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12619 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12620 against Bleichbacher's attack on RSA.
12621 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12624 *) Updates to the new SSL compression code
12625 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12627 *) Fix so that the version number in the master secret, when passed
12628 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12629 (because the server will not accept higher), that the version number
12630 is 0x03,0x01, not 0x03,0x00
12631 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12633 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12634 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12635 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12638 *) Support for RAW extensions where an arbitrary extension can be
12639 created by including its DER encoding. See apps/openssl.cnf for
12643 *) Make sure latest Perl versions don't interpret some generated C array
12644 code as Perl array code in the crypto/err/err_genc.pl script.
12645 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12647 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12648 not many people have the assembler. Various Win32 compilation fixes and
12649 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12650 build instructions.
12653 *) Modify configure script 'Configure' to automatically create crypto/date.h
12654 file under Win32 and also build pem.h from pem.org. New script
12655 util/mkfiles.pl to create the MINFO file on environments that can't do a
12656 'make files': perl util/mkfiles.pl >MINFO should work.
12659 *) Major rework of DES function declarations, in the pursuit of correctness
12660 and purity. As a result, many evil casts evaporated, and some weirdness,
12661 too. You may find this causes warnings in your code. Zapping your evil
12662 casts will probably fix them. Mostly.
12665 *) Fix for a typo in asn1.h. Bug fix to object creation script
12666 obj_dat.pl. It considered a zero in an object definition to mean
12667 "end of object": none of the objects in objects.h have any zeros
12668 so it wasn't spotted.
12669 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12671 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12672 Masking (CBCM). In the absence of test vectors, the best I have been able
12673 to do is check that the decrypt undoes the encrypt, so far. Send me test
12674 vectors if you have them.
12677 *) Correct calculation of key length for export ciphers (too much space was
12678 allocated for null ciphers). This has not been tested!
12681 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12682 message is now correct (it understands "crypto" and "ssl" on its
12683 command line). There is also now an "update" option. This will update
12684 the util/ssleay.num and util/libeay.num files with any new functions.
12686 perl util/mkdef.pl crypto ssl update
12687 it will update them.
12690 *) Overhauled the Perl interface (perl/*):
12691 - ported BN stuff to OpenSSL's different BN library
12692 - made the perl/ source tree CVS-aware
12693 - renamed the package from SSLeay to OpenSSL (the files still contain
12694 their history because I've copied them in the repository)
12695 - removed obsolete files (the test scripts will be replaced
12696 by better Test::Harness variants in the future)
12697 [Ralf S. Engelschall]
12699 *) First cut for a very conservative source tree cleanup:
12700 1. merge various obsolete readme texts into doc/ssleay.txt
12701 where we collect the old documents and readme texts.
12702 2. remove the first part of files where I'm already sure that we no
12703 longer need them because of three reasons: either they are just temporary
12704 files which were left by Eric or they are preserved original files where
12705 I've verified that the diff is also available in the CVS via "cvs diff
12706 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12707 the crypto/md/ stuff).
12708 [Ralf S. Engelschall]
12710 *) More extension code. Incomplete support for subject and issuer alt
12711 name, issuer and authority key id. Change the i2v function parameters
12712 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12713 what that's for :-) Fix to ASN1 macro which messed up
12714 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12717 *) Preliminary support for ENUMERATED type. This is largely copied from the
12721 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12722 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12724 *) Make sure `make rehash' target really finds the `openssl' program.
12725 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12727 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12728 like to hear about it if this slows down other processors.
12731 *) Add CygWin32 platform information to Configure script.
12732 [Alan Batie <batie@aahz.jf.intel.com>]
12734 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12735 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12737 *) New program nseq to manipulate netscape certificate sequences
12740 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12744 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12745 but the BN code had some problems that would cause failures when
12746 doing certificate verification and some other functions.
12747 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12749 *) Add ASN1 and PEM code to support netscape certificate sequences.
12752 *) Add ASN1 and PEM code to support netscape certificate sequences.
12755 *) Add several PKIX and private extended key usage OIDs.
12758 *) Modify the 'ca' program to handle the new extension code. Modify
12759 openssl.cnf for new extension format, add comments.
12762 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12763 and add a sample to openssl.cnf so req -x509 now adds appropriate
12767 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12768 error code, add initial support to X509_print() and x509 application.
12771 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12772 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12773 stuff is currently isolated and isn't even compiled yet.
12776 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12777 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12778 Removed the versions check from X509 routines when loading extensions:
12779 this allows certain broken certificates that don't set the version
12780 properly to be processed.
12783 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12784 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12785 can still be regenerated with "make depend".
12788 *) Spelling mistake in C version of CAST-128.
12789 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12791 *) Changes to the error generation code. The perl script err-code.pl
12792 now reads in the old error codes and retains the old numbers, only
12793 adding new ones if necessary. It also only changes the .err files if new
12794 codes are added. The makefiles have been modified to only insert errors
12795 when needed (to avoid needlessly modifying header files). This is done
12796 by only inserting errors if the .err file is newer than the auto generated
12797 C file. To rebuild all the error codes from scratch (the old behaviour)
12798 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12799 or delete all the .err files.
12802 *) CAST-128 was incorrectly implemented for short keys. The C version has
12803 been fixed, but is untested. The assembler versions are also fixed, but
12804 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12805 to regenerate it if needed.
12806 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12807 Hagino <itojun@kame.net>]
12809 *) File was opened incorrectly in randfile.c.
12810 [Ulf Möller <ulf@fitug.de>]
12812 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12813 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12814 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12815 al: it's just almost always a UTCTime. Note this patch adds new error
12816 codes so do a "make errors" if there are problems.
12819 *) Correct Linux 1 recognition in config.
12820 [Ulf Möller <ulf@fitug.de>]
12822 *) Remove pointless MD5 hash when using DSA keys in ca.
12823 [Anonymous <nobody@replay.com>]
12825 *) Generate an error if given an empty string as a cert directory. Also
12826 generate an error if handed NULL (previously returned 0 to indicate an
12827 error, but didn't set one).
12828 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12830 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12833 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12834 parameters. This was causing a warning which killed off the Win32 compile.
12837 *) Remove C++ style comments from crypto/bn/bn_local.h.
12838 [Neil Costigan <neil.costigan@celocom.com>]
12840 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12841 based on a text string, looking up short and long names and finally
12842 "dot" format. The "dot" format stuff didn't work. Added new function
12843 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12844 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12845 OID is not part of the table.
12848 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12849 X509_LOOKUP_by_alias().
12852 *) Sort openssl functions by name.
12855 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12856 encryption from sample DSA keys (in case anyone is interested the password
12860 *) Make _all_ *_free functions accept a NULL pointer.
12861 [Frans Heymans <fheymans@isaserver.be>]
12863 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12865 [Anonymous <nobody@replay.com>]
12867 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12868 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12870 *) Don't blow it for numeric -newkey arguments to apps/req.
12871 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12873 *) Temp key "for export" tests were wrong in s3_srvr.c.
12874 [Anonymous <nobody@replay.com>]
12876 *) Add prototype for temp key callback functions
12877 SSL_CTX_set_tmp_{rsa,dh}_callback().
12880 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12881 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12884 *) X509_name_add_entry() freed the wrong thing after an error.
12885 [Arne Ansper <arne@ats.cyber.ee>]
12887 *) rsa_eay.c would attempt to free a NULL context.
12888 [Arne Ansper <arne@ats.cyber.ee>]
12890 *) BIO_s_socket() had a broken should_retry() on Windoze.
12891 [Arne Ansper <arne@ats.cyber.ee>]
12893 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12894 [Arne Ansper <arne@ats.cyber.ee>]
12896 *) Make sure the already existing X509_STORE->depth variable is initialized
12897 in X509_STORE_new(), but document the fact that this variable is still
12898 unused in the certificate verification process.
12899 [Ralf S. Engelschall]
12901 *) Fix the various library and apps files to free up pkeys obtained from
12902 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12905 *) Fix reference counting in X509_PUBKEY_get(). This makes
12906 demos/maurice/example2.c work, amongst others, probably.
12907 [Steve Henson and Ben Laurie]
12909 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12910 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12911 are no longer created. This way we have a single and consistent command
12912 line interface `openssl <command>', similar to `cvs <command>'.
12913 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12915 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12916 BIT STRING wrapper always have zero unused bits.
12919 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12922 *) Make the top-level INSTALL documentation easier to understand.
12925 *) Makefiles updated to exit if an error occurs in a sub-directory
12926 make (including if user presses ^C) [Paul Sutton]
12928 *) Make Montgomery context stuff explicit in RSA data structure.
12931 *) Fix build order of pem and err to allow for generated pem.h.
12934 *) Fix renumbering bug in X509_NAME_delete_entry().
12937 *) Enhanced the err-ins.pl script so it makes the error library number
12938 global and can add a library name. This is needed for external ASN1 and
12939 other error libraries.
12942 *) Fixed sk_insert which never worked properly.
12945 *) Fix ASN1 macros so they can handle indefinite length constructed
12946 EXPLICIT tags. Some non standard certificates use these: they can now
12950 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12951 into a single doc/ssleay.txt bundle. This way the information is still
12952 preserved but no longer messes up this directory. Now it's new room for
12953 the new set of documentation files.
12954 [Ralf S. Engelschall]
12956 *) SETs were incorrectly DER encoded. This was a major pain, because they
12957 shared code with SEQUENCEs, which aren't coded the same. This means that
12958 almost everything to do with SETs or SEQUENCEs has either changed name or
12959 number of arguments.
12960 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12962 *) Fix test data to work with the above.
12965 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12966 was already fixed by Eric for 0.9.1 it seems.
12967 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12969 *) Autodetect FreeBSD3.
12972 *) Fix various bugs in Configure. This affects the following platforms:
12976 unixware-2.0-pentium
12980 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12981 before they are needed.
12984 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12988 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12990 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12991 changed SSLeay to OpenSSL in version strings.
12992 [Ralf S. Engelschall]
12994 *) Some fixups to the top-level documents.
12997 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12998 because the symlink to include/ was missing.
12999 [Ralf S. Engelschall]
13001 *) Incorporated the popular no-RSA/DSA-only patches
13002 which allow to compile a RSA-free SSLeay.
13003 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13005 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13006 when "ssleay" is still not found.
13007 [Ralf S. Engelschall]
13009 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13010 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13012 *) Updated the README file.
13013 [Ralf S. Engelschall]
13015 *) Added various .cvsignore files in the CVS repository subdirs
13016 to make a "cvs update" really silent.
13017 [Ralf S. Engelschall]
13019 *) Recompiled the error-definition header files and added
13020 missing symbols to the Win32 linker tables.
13021 [Ralf S. Engelschall]
13023 *) Cleaned up the top-level documents;
13024 o new files: CHANGES and LICENSE
13025 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13026 o merged COPYRIGHT into LICENSE
13027 o removed obsolete TODO file
13028 o renamed MICROSOFT to INSTALL.W32
13029 [Ralf S. Engelschall]
13031 *) Removed dummy files from the 0.9.1b source tree:
13032 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13033 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13034 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13035 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13036 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13037 [Ralf S. Engelschall]
13039 *) Added various platform portability fixes.
13042 *) The Genesis of the OpenSSL rpject:
13043 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13044 Young and Tim J. Hudson created while they were working for C2Net until
13046 [The OpenSSL Project]
13049 Changes between 0.9.0b and 0.9.1b [not released]
13051 *) Updated a few CA certificates under certs/
13054 *) Changed some BIGNUM api stuff.
13057 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13058 DGUX x86, Linux Alpha, etc.
13061 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13062 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13066 *) Add -strparse option to asn1pars program which parses nested
13068 [Dr Stephen Henson <shenson@bigfoot.com>]
13070 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13073 *) DSA fix for "ca" program.
13076 *) Added "-genkey" option to "dsaparam" program.
13079 *) Added RIPE MD160 (rmd160) message digest.
13082 *) Added -a (all) option to "ssleay version" command.
13085 *) Added PLATFORM define which is the id given to Configure.
13088 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13091 *) Extended the ASN.1 parser routines.
13094 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13097 *) Added a BN_CTX to the BN library.
13100 *) Fixed the weak key values in DES library
13103 *) Changed API in EVP library for cipher aliases.
13106 *) Added support for RC2/64bit cipher.
13109 *) Converted the lhash library to the crypto/mem.c functions.
13112 *) Added more recognized ASN.1 object ids.
13115 *) Added more RSA padding checks for SSL/TLS.
13118 *) Added BIO proxy/filter functionality.
13121 *) Added extra_certs to SSL_CTX which can be used
13122 send extra CA certificates to the client in the CA cert chain sending
13123 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13126 *) Now Fortezza is denied in the authentication phase because
13127 this is key exchange mechanism is not supported by SSLeay at all.
13130 *) Additional PKCS1 checks.
13133 *) Support the string "TLSv1" for all TLS v1 ciphers.
13136 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13137 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13140 *) Fixed a few memory leaks.
13143 *) Fixed various code and comment typos.
13146 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13147 bytes sent in the client random.
13148 [Edward Bishop <ebishop@spyglass.com>]