5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.1 and 1.1.2 [xx XXX xxxx]
12 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
13 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
14 are retained for backwards compatibility.
17 *) AES-XTS mode now enforces that its two keys are different to mitigate
18 the attacked described in "Efficient Instantiations of Tweakable
19 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
20 Details of this attack can be obtained from:
21 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
24 *) Rename the object files, i.e. give them other names than in previous
25 versions. Their names now include the name of the final product, as
26 well as its type mnemonic (bin, lib, shlib).
29 *) Added new option for 'openssl list', '-objects', which will display the
30 list of built in objects, i.e. OIDs with names.
33 Changes between 1.1.1 and 1.1.1a [xx XXX xxxx]
35 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
36 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
37 of two gigabytes and the error handling improved.
39 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
40 categorized as a normal bug, not a security issue, because the DRBG reseeds
41 automatically and is fully functional even without additional randomness
42 provided by the application.
44 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
46 *) Add a new ClientHello callback. Provides a callback interface that gives
47 the application the ability to adjust the nascent SSL object at the
48 earliest stage of ClientHello processing, immediately after extensions have
49 been collected but before they have been processed. In particular, this
50 callback can adjust the supported TLS versions in response to the contents
54 *) Add SM2 base algorithm support.
57 *) s390x assembly pack: add (improved) hardware-support for the following
58 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
59 aes-cfb/cfb8, aes-ecb.
62 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
63 parameter is no longer accepted, as it leads to a corrupt table. NULL
64 pem_str is reserved for alias entries only.
67 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
68 step for prime curves. The new implementation is based on formulae from
69 differential addition-and-doubling in homogeneous projective coordinates
70 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
71 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
72 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
73 to work in projective coordinates.
74 [Billy Bob Brumley, Nicola Tuveri]
76 *) Change generating and checking of primes so that the error rate of not
77 being prime depends on the intended use based on the size of the input.
78 For larger primes this will result in more rounds of Miller-Rabin.
79 The maximal error rate for primes with more than 1080 bits is lowered
81 [Kurt Roeckx, Annie Yousar]
83 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
86 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
87 moving between systems, and to avoid confusion when a Windows build is
88 done with mingw vs with MSVC. For POSIX installs, there's still a
89 symlink or copy named 'tsget' to avoid that confusion as well.
92 *) Revert blinding in ECDSA sign and instead make problematic addition
93 length-invariant. Switch even to fixed-length Montgomery multiplication.
96 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
97 step for binary curves. The new implementation is based on formulae from
98 differential addition-and-doubling in mixed Lopez-Dahab projective
99 coordinates, modified to independently blind the operands.
100 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
102 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
103 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
104 EC_METHODs to implement their own specialized "ladder step", to take
105 advantage of more favorable coordinate systems or more efficient
106 differential addition-and-doubling algorithms.
107 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
109 *) Modified the random device based seed sources to keep the relevant
110 file descriptors open rather than reopening them on each access.
111 This allows such sources to operate in a chroot() jail without
112 the associated device nodes being available. This behaviour can be
113 controlled using RAND_keep_random_devices_open().
116 *) Numerous side-channel attack mitigations have been applied. This may have
117 performance impacts for some algorithms for the benefit of improved
118 security. Specific changes are noted in this change log by their respective
122 *) AIX shared library support overhaul. Switch to AIX "natural" way of
123 handling shared libraries, which means collecting shared objects of
124 different versions and bitnesses in one common archive. This allows to
125 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
126 doesn't affect the way 3rd party applications are linked, only how
127 multi-version installation is managed.
130 *) Make ec_group_do_inverse_ord() more robust and available to other
131 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
132 mitigations are applied to the fallback BN_mod_inverse().
133 When using this function rather than BN_mod_inverse() directly, new
134 EC cryptosystem implementations are then safer-by-default.
137 *) Add coordinate blinding for EC_POINT and implement projective
138 coordinate blinding for generic prime curves as a countermeasure to
139 chosen point SCA attacks.
140 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
142 *) Add blinding to ECDSA and DSA signatures to protect against side channel
143 attacks discovered by Keegan Ryan (NCC Group).
146 *) Enforce checking in the pkeyutl command line app to ensure that the input
147 length does not exceed the maximum supported digest length when performing
148 a sign, verify or verifyrecover operation.
151 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
152 I/O in combination with something like select() or poll() will hang. This
153 can be turned off again using SSL_CTX_clear_mode().
154 Many applications do not properly handle non-application data records, and
155 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
156 around the problems in those applications, but can also break some.
157 It's recommended to read the manpages about SSL_read(), SSL_write(),
158 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
159 SSL_CTX_set_read_ahead() again.
162 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
163 now allow empty (zero character) pass phrases.
166 *) Apply blinding to binary field modular inversion and remove patent
167 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
170 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
171 binary and prime elliptic curves.
174 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
175 constant time fixed point multiplication.
178 *) Revise elliptic curve scalar multiplication with timing attack
179 defenses: ec_wNAF_mul redirects to a constant time implementation
180 when computing fixed point and variable point multiplication (which
181 in OpenSSL are mostly used with secret scalars in keygen, sign,
182 ECDH derive operations).
183 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
186 *) Updated CONTRIBUTING
189 *) Updated DRBG / RAND to request nonce and additional low entropy
190 randomness from the system.
191 [Matthias St. Pierre]
193 *) Updated 'openssl rehash' to use OpenSSL consistent default.
196 *) Moved the load of the ssl_conf module to libcrypto, which helps
197 loading engines that libssl uses before libssl is initialised.
200 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
203 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
204 [Ingo Schwarze, Rich Salz]
206 *) Added output of accepting IP address and port for 'openssl s_server'
209 *) Added a new API for TLSv1.3 ciphersuites:
210 SSL_CTX_set_ciphersuites()
211 SSL_set_ciphersuites()
214 *) Memory allocation failures consistenly add an error to the error
218 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
219 in libcrypto when run as setuid/setgid.
222 *) Load any config file by default when libssl is used.
225 *) Added new public header file <openssl/rand_drbg.h> and documentation
226 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
227 [Matthias St. Pierre]
229 *) QNX support removed (cannot find contributors to get their approval
230 for the license change).
233 *) TLSv1.3 replay protection for early data has been implemented. See the
234 SSL_read_early_data() man page for further details.
237 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
238 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
239 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
240 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
241 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
242 configuration has been separated out. See the ciphers man page or the
243 SSL_CTX_set_ciphersuites() man page for more information.
246 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
247 in responder mode now supports the new "-multi" option, which
248 spawns the specified number of child processes to handle OCSP
249 requests. The "-timeout" option now also limits the OCSP
250 responder's patience to wait to receive the full client request
251 on a newly accepted connection. Child processes are respawned
252 as needed, and the CA index file is automatically reloaded
253 when changed. This makes it possible to run the "ocsp" responder
254 as a long-running service, making the OpenSSL CA somewhat more
255 feature-complete. In this mode, most diagnostic messages logged
256 after entering the event loop are logged via syslog(3) rather than
260 *) Added support for X448 and Ed448. Heavily based on original work by
264 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
265 objects loaded. This adds the functions OSSL_STORE_expect() and
266 OSSL_STORE_find() as well as needed tools to construct searches and
267 get the search data out of them.
270 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
271 version of OpenSSL should review their configuration settings to ensure
272 that they are still appropriate for TLSv1.3. For further information see:
273 https://wiki.openssl.org/index.php/TLS1.3
276 *) Grand redesign of the OpenSSL random generator
278 The default RAND method now utilizes an AES-CTR DRBG according to
279 NIST standard SP 800-90Ar1. The new random generator is essentially
280 a port of the default random generator from the OpenSSL FIPS 2.0
281 object module. It is a hybrid deterministic random bit generator
282 using an AES-CTR bit stream and which seeds and reseeds itself
283 automatically using trusted system entropy sources.
285 Some of its new features are:
286 o Support for multiple DRBG instances with seed chaining.
287 o The default RAND method makes use of a DRBG.
288 o There is a public and private DRBG instance.
289 o The DRBG instances are fork-safe.
290 o Keep all global DRBG instances on the secure heap if it is enabled.
291 o The public and private DRBG instance are per thread for lock free
293 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
295 *) Changed Configure so it only says what it does and doesn't dump
296 so much data. Instead, ./configdata.pm should be used as a script
297 to display all sorts of configuration data.
300 *) Added processing of "make variables" to Configure.
303 *) Added SHA512/224 and SHA512/256 algorithm support.
306 *) The last traces of Netware support, first removed in 1.1.0, have
310 *) Get rid of Makefile.shared, and in the process, make the processing
311 of certain files (rc.obj, or the .def/.map/.opt files produced from
312 the ordinal files) more visible and hopefully easier to trace and
313 debug (or make silent).
316 *) Make it possible to have environment variable assignments as
317 arguments to config / Configure.
320 *) Add multi-prime RSA (RFC 8017) support.
323 *) Add SM3 implemented according to GB/T 32905-2016
324 [ Jack Lloyd <jack.lloyd@ribose.com>,
325 Ronald Tse <ronald.tse@ribose.com>,
326 Erick Borsboom <erick.borsboom@ribose.com> ]
328 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
329 as documented in RFC6066.
330 Based on a patch from Tomasz Moń
331 [Filipe Raimundo da Silva]
333 *) Add SM4 implemented according to GB/T 32907-2016.
334 [ Jack Lloyd <jack.lloyd@ribose.com>,
335 Ronald Tse <ronald.tse@ribose.com>,
336 Erick Borsboom <erick.borsboom@ribose.com> ]
338 *) Reimplement -newreq-nodes and ERR_error_string_n; the
339 original author does not agree with the license change.
342 *) Add ARIA AEAD TLS support.
345 *) Some macro definitions to support VS6 have been removed. Visual
346 Studio 6 has not worked since 1.1.0
349 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
350 without clearing the errors.
353 *) Add "atfork" functions. If building on a system that without
354 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
355 requirements. The RAND facility now uses/requires this.
361 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
362 not possible to disable entirely. However, it's still possible to
363 disable the console reading UI method, UI_OpenSSL() (use UI_null()
366 To disable, configure with 'no-ui-console'. 'no-ui' is still
367 possible to use as an alias. Check at compile time with the
368 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
369 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
372 *) Add a STORE module, which implements a uniform and URI based reader of
373 stores that can contain keys, certificates, CRLs and numerous other
374 objects. The main API is loosely based on a few stdio functions,
375 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
376 OSSL_STORE_error and OSSL_STORE_close.
377 The implementation uses backends called "loaders" to implement arbitrary
378 URI schemes. There is one built in "loader" for the 'file' scheme.
381 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
382 then adjusted to work on FreeBSD 8.4 as well.
383 Enable by configuring with 'enable-devcryptoeng'. This is done by default
384 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
387 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
388 util/mkerr.pl, which is adapted to allow those prefixes, leading to
389 error code calls like this:
391 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
393 With this change, we claim the namespaces OSSL and OPENSSL in a manner
394 that can be encoded in C. For the foreseeable future, this will only
396 [Richard Levitte and Tim Hudson]
398 *) Removed BSD cryptodev engine.
401 *) Add a build target 'build_all_generated', to build all generated files
402 and only that. This can be used to prepare everything that requires
403 things like perl for a system that lacks perl and then move everything
404 to that system and do the rest of the build there.
407 *) In the UI interface, make it possible to duplicate the user data. This
408 can be used by engines that need to retain the data for a longer time
409 than just the call where this user data is passed.
412 *) Ignore the '-named_curve auto' value for compatibility of applications
414 [Tomas Mraz <tmraz@fedoraproject.org>]
416 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
417 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
418 alerts across multiple records (some of which could be empty). In practice
419 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
420 prohibts this altogether and other libraries (BoringSSL, NSS) do not
421 support this at all. Supporting it adds significant complexity to the
422 record layer, and its removal is unlikely to cause inter-operability
426 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
427 with Z. These are meant to replace LONG and ZLONG and to be size safe.
428 The use of LONG and ZLONG is discouraged and scheduled for deprecation
432 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
433 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
434 [Richard Levitte, Andy Polyakov]
436 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
440 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
441 platform rather than 'mingw'.
444 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
445 success if they are asked to add an object which already exists
446 in the store. This change cascades to other functions which load
447 certificates and CRLs.
450 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
451 facilitate stack unwinding even from assembly subroutines.
454 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
455 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
458 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
459 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
460 which is the minimum version we support.
463 *) Certificate time validation (X509_cmp_time) enforces stricter
464 compliance with RFC 5280. Fractional seconds and timezone offsets
465 are no longer allowed.
468 *) Add support for ARIA
471 *) s_client will now send the Server Name Indication (SNI) extension by
472 default unless the new "-noservername" option is used. The server name is
473 based on the host provided to the "-connect" option unless overridden by
477 *) Add support for SipHash
480 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
481 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
482 prevent issues where no progress is being made and the peer continually
483 sends unrecognised record types, using up resources processing them.
486 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
487 using the algorithm defined in
488 https://www.akkadia.org/drepper/SHA-crypt.txt
491 *) Heartbeat support has been removed; the ABI is changed for now.
492 [Richard Levitte, Rich Salz]
494 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
497 *) The RSA "null" method, which was partially supported to avoid patent
498 issues, has been replaced to always returns NULL.
502 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
504 *) Client DoS due to large DH parameter
506 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
507 malicious server can send a very large prime value to the client. This will
508 cause the client to spend an unreasonably long period of time generating a
509 key for this prime resulting in a hang until the client has finished. This
510 could be exploited in a Denial Of Service attack.
512 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
516 *) Cache timing vulnerability in RSA Key Generation
518 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
519 a cache timing side channel attack. An attacker with sufficient access to
520 mount cache timing attacks during the RSA key generation process could
521 recover the private key.
523 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
524 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
528 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
529 parameter is no longer accepted, as it leads to a corrupt table. NULL
530 pem_str is reserved for alias entries only.
533 *) Revert blinding in ECDSA sign and instead make problematic addition
534 length-invariant. Switch even to fixed-length Montgomery multiplication.
537 *) Change generating and checking of primes so that the error rate of not
538 being prime depends on the intended use based on the size of the input.
539 For larger primes this will result in more rounds of Miller-Rabin.
540 The maximal error rate for primes with more than 1080 bits is lowered
542 [Kurt Roeckx, Annie Yousar]
544 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
547 *) Add blinding to ECDSA and DSA signatures to protect against side channel
548 attacks discovered by Keegan Ryan (NCC Group).
551 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
552 now allow empty (zero character) pass phrases.
555 *) Certificate time validation (X509_cmp_time) enforces stricter
556 compliance with RFC 5280. Fractional seconds and timezone offsets
557 are no longer allowed.
560 *) Fixed a text canonicalisation bug in CMS
562 Where a CMS detached signature is used with text content the text goes
563 through a canonicalisation process first prior to signing or verifying a
564 signature. This process strips trailing space at the end of lines, converts
565 line terminators to CRLF and removes additional trailing line terminators
566 at the end of a file. A bug in the canonicalisation process meant that
567 some characters, such as form-feed, were incorrectly treated as whitespace
568 and removed. This is contrary to the specification (RFC5485). This fix
569 could mean that detached text data signed with an earlier version of
570 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
571 signed with a fixed OpenSSL may fail to verify with an earlier version of
572 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
573 and use the "-binary" flag (for the "cms" command line application) or set
574 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
577 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
579 *) Constructed ASN.1 types with a recursive definition could exceed the stack
581 Constructed ASN.1 types with a recursive definition (such as can be found
582 in PKCS7) could eventually exceed the stack given malicious input with
583 excessive recursion. This could result in a Denial Of Service attack. There
584 are no such structures used within SSL/TLS that come from untrusted sources
585 so this is considered safe.
587 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
592 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
594 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
595 effectively reduced to only comparing the least significant bit of each
596 byte. This allows an attacker to forge messages that would be considered as
597 authenticated in an amount of tries lower than that guaranteed by the
598 security claims of the scheme. The module can only be compiled by the
599 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
601 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
606 *) Add a build target 'build_all_generated', to build all generated files
607 and only that. This can be used to prepare everything that requires
608 things like perl for a system that lacks perl and then move everything
609 to that system and do the rest of the build there.
612 *) Backport SSL_OP_NO_RENGOTIATION
614 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
615 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
616 changes this is no longer possible in 1.1.0. Therefore the new
617 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
618 1.1.0 to provide equivalent functionality.
620 Note that if an application built against 1.1.0h headers (or above) is run
621 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
622 accepted but nothing will happen, i.e. renegotiation will not be prevented.
625 *) Removed the OS390-Unix config target. It relied on a script that doesn't
629 *) rsaz_1024_mul_avx2 overflow bug on x86_64
631 There is an overflow bug in the AVX2 Montgomery multiplication procedure
632 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
633 Analysis suggests that attacks against RSA and DSA as a result of this
634 defect would be very difficult to perform and are not believed likely.
635 Attacks against DH1024 are considered just feasible, because most of the
636 work necessary to deduce information about a private key may be performed
637 offline. The amount of resources required for such an attack would be
638 significant. However, for an attack on TLS to be meaningful, the server
639 would have to share the DH1024 private key among multiple clients, which is
640 no longer an option since CVE-2016-0701.
642 This only affects processors that support the AVX2 but not ADX extensions
643 like Intel Haswell (4th generation).
645 This issue was reported to OpenSSL by David Benjamin (Google). The issue
646 was originally found via the OSS-Fuzz project.
650 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
652 *) bn_sqrx8x_internal carry bug on x86_64
654 There is a carry propagating bug in the x86_64 Montgomery squaring
655 procedure. No EC algorithms are affected. Analysis suggests that attacks
656 against RSA and DSA as a result of this defect would be very difficult to
657 perform and are not believed likely. Attacks against DH are considered just
658 feasible (although very difficult) because most of the work necessary to
659 deduce information about a private key may be performed offline. The amount
660 of resources required for such an attack would be very significant and
661 likely only accessible to a limited number of attackers. An attacker would
662 additionally need online access to an unpatched system using the target
663 private key in a scenario with persistent DH parameters and a private
664 key that is shared between multiple clients.
666 This only affects processors that support the BMI1, BMI2 and ADX extensions
667 like Intel Broadwell (5th generation) and later or AMD Ryzen.
669 This issue was reported to OpenSSL by the OSS-Fuzz project.
673 *) Malformed X.509 IPAddressFamily could cause OOB read
675 If an X.509 certificate has a malformed IPAddressFamily extension,
676 OpenSSL could do a one-byte buffer overread. The most likely result
677 would be an erroneous display of the certificate in text format.
679 This issue was reported to OpenSSL by the OSS-Fuzz project.
683 Changes between 1.1.0e and 1.1.0f [25 May 2017]
685 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
686 platform rather than 'mingw'.
689 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
690 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
691 which is the minimum version we support.
694 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
696 *) Encrypt-Then-Mac renegotiation crash
698 During a renegotiation handshake if the Encrypt-Then-Mac extension is
699 negotiated where it was not in the original handshake (or vice-versa) then
700 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
701 and servers are affected.
703 This issue was reported to OpenSSL by Joe Orton (Red Hat).
707 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
709 *) Truncated packet could crash via OOB read
711 If one side of an SSL/TLS path is running on a 32-bit host and a specific
712 cipher is being used, then a truncated packet can cause that host to
713 perform an out-of-bounds read, usually resulting in a crash.
715 This issue was reported to OpenSSL by Robert Święcki of Google.
719 *) Bad (EC)DHE parameters cause a client crash
721 If a malicious server supplies bad parameters for a DHE or ECDHE key
722 exchange then this can result in the client attempting to dereference a
723 NULL pointer leading to a client crash. This could be exploited in a Denial
726 This issue was reported to OpenSSL by Guido Vranken.
730 *) BN_mod_exp may produce incorrect results on x86_64
732 There is a carry propagating bug in the x86_64 Montgomery squaring
733 procedure. No EC algorithms are affected. Analysis suggests that attacks
734 against RSA and DSA as a result of this defect would be very difficult to
735 perform and are not believed likely. Attacks against DH are considered just
736 feasible (although very difficult) because most of the work necessary to
737 deduce information about a private key may be performed offline. The amount
738 of resources required for such an attack would be very significant and
739 likely only accessible to a limited number of attackers. An attacker would
740 additionally need online access to an unpatched system using the target
741 private key in a scenario with persistent DH parameters and a private
742 key that is shared between multiple clients. For example this can occur by
743 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
744 similar to CVE-2015-3193 but must be treated as a separate problem.
746 This issue was reported to OpenSSL by the OSS-Fuzz project.
750 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
752 *) ChaCha20/Poly1305 heap-buffer-overflow
754 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
755 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
756 crash. This issue is not considered to be exploitable beyond a DoS.
758 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
762 *) CMS Null dereference
764 Applications parsing invalid CMS structures can crash with a NULL pointer
765 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
766 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
767 structure callback if an attempt is made to free certain invalid encodings.
768 Only CHOICE structures using a callback which do not handle NULL value are
771 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
775 *) Montgomery multiplication may produce incorrect results
777 There is a carry propagating bug in the Broadwell-specific Montgomery
778 multiplication procedure that handles input lengths divisible by, but
779 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
780 and DH private keys are impossible. This is because the subroutine in
781 question is not used in operations with the private key itself and an input
782 of the attacker's direct choice. Otherwise the bug can manifest itself as
783 transient authentication and key negotiation failures or reproducible
784 erroneous outcome of public-key operations with specially crafted input.
785 Among EC algorithms only Brainpool P-512 curves are affected and one
786 presumably can attack ECDH key negotiation. Impact was not analyzed in
787 detail, because pre-requisites for attack are considered unlikely. Namely
788 multiple clients have to choose the curve in question and the server has to
789 share the private key among them, neither of which is default behaviour.
790 Even then only clients that chose the curve will be affected.
792 This issue was publicly reported as transient failures and was not
793 initially recognized as a security issue. Thanks to Richard Morgan for
794 providing reproducible case.
798 *) Removed automatic addition of RPATH in shared libraries and executables,
799 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
802 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
804 *) Fix Use After Free for large message sizes
806 The patch applied to address CVE-2016-6307 resulted in an issue where if a
807 message larger than approx 16k is received then the underlying buffer to
808 store the incoming message is reallocated and moved. Unfortunately a
809 dangling pointer to the old location is left which results in an attempt to
810 write to the previously freed location. This is likely to result in a
811 crash, however it could potentially lead to execution of arbitrary code.
813 This issue only affects OpenSSL 1.1.0a.
815 This issue was reported to OpenSSL by Robert Święcki.
819 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
821 *) OCSP Status Request extension unbounded memory growth
823 A malicious client can send an excessively large OCSP Status Request
824 extension. If that client continually requests renegotiation, sending a
825 large OCSP Status Request extension each time, then there will be unbounded
826 memory growth on the server. This will eventually lead to a Denial Of
827 Service attack through memory exhaustion. Servers with a default
828 configuration are vulnerable even if they do not support OCSP. Builds using
829 the "no-ocsp" build time option are not affected.
831 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
835 *) SSL_peek() hang on empty record
837 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
838 sends an empty record. This could be exploited by a malicious peer in a
839 Denial Of Service attack.
841 This issue was reported to OpenSSL by Alex Gaynor.
845 *) Excessive allocation of memory in tls_get_message_header() and
846 dtls1_preprocess_fragment()
848 A (D)TLS message includes 3 bytes for its length in the header for the
849 message. This would allow for messages up to 16Mb in length. Messages of
850 this length are excessive and OpenSSL includes a check to ensure that a
851 peer is sending reasonably sized messages in order to avoid too much memory
852 being consumed to service a connection. A flaw in the logic of version
853 1.1.0 means that memory for the message is allocated too early, prior to
854 the excessive message length check. Due to way memory is allocated in
855 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
856 to service a connection. This could lead to a Denial of Service through
857 memory exhaustion. However, the excessive message length check still takes
858 place, and this would cause the connection to immediately fail. Assuming
859 that the application calls SSL_free() on the failed connection in a timely
860 manner then the 21Mb of allocated memory will then be immediately freed
861 again. Therefore the excessive memory allocation will be transitory in
862 nature. This then means that there is only a security impact if:
864 1) The application does not call SSL_free() in a timely manner in the event
865 that the connection fails
867 2) The application is working in a constrained environment where there is
868 very little free memory
870 3) The attacker initiates multiple connection attempts such that there are
871 multiple connections in a state where memory has been allocated for the
872 connection; SSL_free() has not yet been called; and there is insufficient
873 memory to service the multiple requests.
875 Except in the instance of (1) above any Denial Of Service is likely to be
876 transitory because as soon as the connection fails the memory is
877 subsequently freed again in the SSL_free() call. However there is an
878 increased risk during this period of application crashes due to the lack of
879 memory - which would then mean a more serious Denial of Service.
881 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
882 (CVE-2016-6307 and CVE-2016-6308)
885 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
886 had to be removed. Primary reason is that vendor assembler can't
887 assemble our modules with -KPIC flag. As result it, assembly
888 support, was not even available as option. But its lack means
889 lack of side-channel resistant code, which is incompatible with
890 security by todays standards. Fortunately gcc is readily available
891 prepackaged option, which we firmly point at...
894 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
896 *) Windows command-line tool supports UTF-8 opt-in option for arguments
897 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
898 (to any value) allows Windows user to access PKCS#12 file generated
899 with Windows CryptoAPI and protected with non-ASCII password, as well
900 as files generated under UTF-8 locale on Linux also protected with
904 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
905 have been disabled by default and removed from DEFAULT, just like RC4.
906 See the RC4 item below to re-enable both.
909 *) The method for finding the storage location for the Windows RAND seed file
910 has changed. First we check %RANDFILE%. If that is not set then we check
911 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
912 all else fails we fall back to C:\.
915 *) The EVP_EncryptUpdate() function has had its return type changed from void
916 to int. A return of 0 indicates and error while a return of 1 indicates
920 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
921 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
922 off the constant time implementation for RSA, DSA and DH have been made
923 no-ops and deprecated.
926 *) Windows RAND implementation was simplified to only get entropy by
927 calling CryptGenRandom(). Various other RAND-related tickets
929 [Joseph Wylie Yandle, Rich Salz]
931 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
932 and OPENSSL_LH_, respectively. The old names are available
933 with API compatibility. They new names are now completely documented.
936 *) Unify TYPE_up_ref(obj) methods signature.
937 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
938 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
939 int (instead of void) like all others TYPE_up_ref() methods.
940 So now these methods also check the return value of CRYPTO_atomic_add(),
941 and the validity of object reference counter.
942 [fdasilvayy@gmail.com]
944 *) With Windows Visual Studio builds, the .pdb files are installed
945 alongside the installed libraries and executables. For a static
946 library installation, ossl_static.pdb is the associate compiler
947 generated .pdb file to be used when linking programs.
950 *) Remove openssl.spec. Packaging files belong with the packagers.
953 *) Automatic Darwin/OSX configuration has had a refresh, it will now
954 recognise x86_64 architectures automatically. You can still decide
955 to build for a different bitness with the environment variable
956 KERNEL_BITS (can be 32 or 64), for example:
958 KERNEL_BITS=32 ./config
962 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
963 256 bit AES and HMAC with SHA256.
966 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
969 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
972 *) To enable users to have their own config files and build file templates,
973 Configure looks in the directory indicated by the environment variable
974 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
975 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
976 name and is used as is.
979 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
980 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
981 X509_CERT_FILE_CTX was removed.
984 *) "shared" builds are now the default. To create only static libraries use
985 the "no-shared" Configure option.
988 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
989 All of these option have not worked for some while and are fundamental
993 *) Make various cleanup routines no-ops and mark them as deprecated. Most
994 global cleanup functions are no longer required because they are handled
995 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
996 Explicitly de-initing can cause problems (e.g. where a library that uses
997 OpenSSL de-inits, but an application is still using it). The affected
998 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
999 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1000 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1001 COMP_zlib_cleanup().
1004 *) --strict-warnings no longer enables runtime debugging options
1005 such as REF_DEBUG. Instead, debug options are automatically
1006 enabled with '--debug' builds.
1007 [Andy Polyakov, Emilia Käsper]
1009 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1010 have been moved out of the public header files. New functions for managing
1011 these have been added.
1014 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1015 objects have been moved out of the public header files. New
1016 functions for managing these have been added.
1019 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1020 have been moved out of the public header files. New functions for managing
1021 these have been added.
1024 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1025 moved out of the public header files. New functions for managing these
1029 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1032 *) Removed the mk1mf build scripts.
1035 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1036 it is always safe to #include a header now.
1039 *) Removed the aged BC-32 config and all its supporting scripts
1042 *) Removed support for Ultrix, Netware, and OS/2.
1045 *) Add support for HKDF.
1046 [Alessandro Ghedini]
1048 *) Add support for blake2b and blake2s
1051 *) Added support for "pipelining". Ciphers that have the
1052 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1053 encryptions/decryptions simultaneously. There are currently no built-in
1054 ciphers with this property but the expectation is that engines will be able
1055 to offer it to significantly improve throughput. Support has been extended
1056 into libssl so that multiple records for a single connection can be
1057 processed in one go (for >=TLS 1.1).
1060 *) Added the AFALG engine. This is an async capable engine which is able to
1061 offload work to the Linux kernel. In this initial version it only supports
1062 AES128-CBC. The kernel must be version 4.1.0 or greater.
1065 *) OpenSSL now uses a new threading API. It is no longer necessary to
1066 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1067 are two supported threading models: pthreads and windows threads. It is
1068 also possible to configure OpenSSL at compile time for "no-threads". The
1069 old threading API should no longer be used. The functions have been
1070 replaced with "no-op" compatibility macros.
1071 [Alessandro Ghedini, Matt Caswell]
1073 *) Modify behavior of ALPN to invoke callback after SNI/servername
1074 callback, such that updates to the SSL_CTX affect ALPN.
1077 *) Add SSL_CIPHER queries for authentication and key-exchange.
1080 *) Changes to the DEFAULT cipherlist:
1081 - Prefer (EC)DHE handshakes over plain RSA.
1082 - Prefer AEAD ciphers over legacy ciphers.
1083 - Prefer ECDSA over RSA when both certificates are available.
1084 - Prefer TLSv1.2 ciphers/PRF.
1085 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1089 *) Change the ECC default curve list to be this, in order: x25519,
1090 secp256r1, secp521r1, secp384r1.
1093 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1094 disabled by default. They can be re-enabled using the
1095 enable-weak-ssl-ciphers option to Configure.
1098 *) If the server has ALPN configured, but supports no protocols that the
1099 client advertises, send a fatal "no_application_protocol" alert.
1100 This behaviour is SHALL in RFC 7301, though it isn't universally
1101 implemented by other servers.
1104 *) Add X25519 support.
1105 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1106 for public and private key encoding using the format documented in
1107 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1108 key generation and key derivation.
1110 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1114 *) Deprecate SRP_VBASE_get_by_user.
1115 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1116 In order to fix an unavoidable memory leak (CVE-2016-0798),
1117 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1118 seed, even if the seed is configured.
1120 Users should use SRP_VBASE_get1_by_user instead. Note that in
1121 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1122 also that even though configuring the SRP seed attempts to hide
1123 invalid usernames by continuing the handshake with fake
1124 credentials, this behaviour is not constant time and no strong
1125 guarantees are made that the handshake is indistinguishable from
1126 that of a valid user.
1129 *) Configuration change; it's now possible to build dynamic engines
1130 without having to build shared libraries and vice versa. This
1131 only applies to the engines in engines/, those in crypto/engine/
1132 will always be built into libcrypto (i.e. "static").
1134 Building dynamic engines is enabled by default; to disable, use
1135 the configuration option "disable-dynamic-engine".
1137 The only requirements for building dynamic engines are the
1138 presence of the DSO module and building with position independent
1139 code, so they will also automatically be disabled if configuring
1140 with "disable-dso" or "disable-pic".
1142 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1143 are also taken away from openssl/opensslconf.h, as they are
1147 *) Configuration change; if there is a known flag to compile
1148 position independent code, it will always be applied on the
1149 libcrypto and libssl object files, and never on the application
1150 object files. This means other libraries that use routines from
1151 libcrypto / libssl can be made into shared libraries regardless
1152 of how OpenSSL was configured.
1154 If this isn't desirable, the configuration options "disable-pic"
1155 or "no-pic" can be used to disable the use of PIC. This will
1156 also disable building shared libraries and dynamic engines.
1159 *) Removed JPAKE code. It was experimental and has no wide use.
1162 *) The INSTALL_PREFIX Makefile variable has been renamed to
1163 DESTDIR. That makes for less confusion on what this variable
1164 is for. Also, the configuration option --install_prefix is
1168 *) Heartbeat for TLS has been removed and is disabled by default
1169 for DTLS; configure with enable-heartbeats. Code that uses the
1170 old #define's might need to be updated.
1171 [Emilia Käsper, Rich Salz]
1173 *) Rename REF_CHECK to REF_DEBUG.
1176 *) New "unified" build system
1178 The "unified" build system is aimed to be a common system for all
1179 platforms we support. With it comes new support for VMS.
1181 This system builds supports building in a different directory tree
1182 than the source tree. It produces one Makefile (for unix family
1183 or lookalikes), or one descrip.mms (for VMS).
1185 The source of information to make the Makefile / descrip.mms is
1186 small files called 'build.info', holding the necessary
1187 information for each directory with source to compile, and a
1188 template in Configurations, like unix-Makefile.tmpl or
1191 With this change, the library names were also renamed on Windows
1192 and on VMS. They now have names that are closer to the standard
1193 on Unix, and include the major version number, and in certain
1194 cases, the architecture they are built for. See "Notes on shared
1195 libraries" in INSTALL.
1197 We rely heavily on the perl module Text::Template.
1200 *) Added support for auto-initialisation and de-initialisation of the library.
1201 OpenSSL no longer requires explicit init or deinit routines to be called,
1202 except in certain circumstances. See the OPENSSL_init_crypto() and
1203 OPENSSL_init_ssl() man pages for further information.
1206 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1207 "peer" argument is now expected to be a BIO_ADDR object.
1209 *) Rewrite of BIO networking library. The BIO library lacked consistent
1210 support of IPv6, and adding it required some more extensive
1211 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1212 which hold all types of addresses and chains of address information.
1213 It also introduces a new API, with functions like BIO_socket,
1214 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1215 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1216 have been adapted accordingly.
1219 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1223 *) CRIME protection: disable compression by default, even if OpenSSL is
1224 compiled with zlib enabled. Applications can still enable compression
1225 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1226 using the SSL_CONF library to configure compression.
1229 *) The signature of the session callback configured with
1230 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1231 was explicitly marked as 'const unsigned char*' instead of
1235 *) Always DPURIFY. Remove the use of uninitialized memory in the
1236 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1239 *) Removed many obsolete configuration items, including
1240 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1241 MD2_CHAR, MD2_INT, MD2_LONG
1243 IDEA_SHORT, IDEA_LONG
1244 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1245 [Rich Salz, with advice from Andy Polyakov]
1247 *) Many BN internals have been moved to an internal header file.
1248 [Rich Salz with help from Andy Polyakov]
1250 *) Configuration and writing out the results from it has changed.
1251 Files such as Makefile include/openssl/opensslconf.h and are now
1252 produced through general templates, such as Makefile.in and
1253 crypto/opensslconf.h.in and some help from the perl module
1256 Also, the center of configuration information is no longer
1257 Makefile. Instead, Configure produces a perl module in
1258 configdata.pm which holds most of the config data (in the hash
1259 table %config), the target data that comes from the target
1260 configuration in one of the Configurations/*.conf files (in
1264 *) To clarify their intended purposes, the Configure options
1265 --prefix and --openssldir change their semantics, and become more
1266 straightforward and less interdependent.
1268 --prefix shall be used exclusively to give the location INSTALLTOP
1269 where programs, scripts, libraries, include files and manuals are
1270 going to be installed. The default is now /usr/local.
1272 --openssldir shall be used exclusively to give the default
1273 location OPENSSLDIR where certificates, private keys, CRLs are
1274 managed. This is also where the default openssl.cnf gets
1276 If the directory given with this option is a relative path, the
1277 values of both the --prefix value and the --openssldir value will
1278 be combined to become OPENSSLDIR.
1279 The default for --openssldir is INSTALLTOP/ssl.
1281 Anyone who uses --openssldir to specify where OpenSSL is to be
1282 installed MUST change to use --prefix instead.
1285 *) The GOST engine was out of date and therefore it has been removed. An up
1286 to date GOST engine is now being maintained in an external repository.
1287 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1288 support for GOST ciphersuites (these are only activated if a GOST engine
1292 *) EGD is no longer supported by default; use enable-egd when
1294 [Ben Kaduk and Rich Salz]
1296 *) The distribution now has Makefile.in files, which are used to
1297 create Makefile's when Configure is run. *Configure must be run
1298 before trying to build now.*
1301 *) The return value for SSL_CIPHER_description() for error conditions
1305 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1307 Obtaining and performing DNSSEC validation of TLSA records is
1308 the application's responsibility. The application provides
1309 the TLSA records of its choice to OpenSSL, and these are then
1310 used to authenticate the peer.
1312 The TLSA records need not even come from DNS. They can, for
1313 example, be used to implement local end-entity certificate or
1314 trust-anchor "pinning", where the "pin" data takes the form
1315 of TLSA records, which can augment or replace verification
1316 based on the usual WebPKI public certification authorities.
1319 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1320 continues to support deprecated interfaces in default builds.
1321 However, applications are strongly advised to compile their
1322 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1323 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1324 or the 1.1.0 releases.
1326 In environments in which all applications have been ported to
1327 not use any deprecated interfaces OpenSSL's Configure script
1328 should be used with the --api=1.1.0 option to entirely remove
1329 support for the deprecated features from the library and
1330 unconditionally disable them in the installed headers.
1331 Essentially the same effect can be achieved with the "no-deprecated"
1332 argument to Configure, except that this will always restrict
1333 the build to just the latest API, rather than a fixed API
1336 As applications are ported to future revisions of the API,
1337 they should update their compile-time OPENSSL_API_COMPAT define
1338 accordingly, but in most cases should be able to continue to
1339 compile with later releases.
1341 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1342 0x10000000L and 0x00908000L, respectively. However those
1343 versions did not support the OPENSSL_API_COMPAT feature, and
1344 so applications are not typically tested for explicit support
1345 of just the undeprecated features of either release.
1348 *) Add support for setting the minimum and maximum supported protocol.
1349 It can bet set via the SSL_set_min_proto_version() and
1350 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1351 MaxProtocol. It's recommended to use the new APIs to disable
1352 protocols instead of disabling individual protocols using
1353 SSL_set_options() or SSL_CONF's Protocol. This change also
1354 removes support for disabling TLS 1.2 in the OpenSSL TLS
1355 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1358 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1361 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1362 and integrates ECDSA and ECDH functionality into EC. Implementations can
1363 now redirect key generation and no longer need to convert to or from
1366 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1367 include the ec.h header file instead.
1370 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1371 ciphers who are no longer supported and drops support the ephemeral RSA key
1372 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1375 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1376 opaque. For HMAC_CTX, the following constructors and destructors
1379 HMAC_CTX *HMAC_CTX_new(void);
1380 void HMAC_CTX_free(HMAC_CTX *ctx);
1382 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1383 destroy such methods has been added. See EVP_MD_meth_new(3) and
1384 EVP_CIPHER_meth_new(3) for documentation.
1387 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1388 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1389 EVP_MD_CTX_reset() should be called instead to reinitialise
1390 an already created structure.
1391 2) For consistency with the majority of our object creators and
1392 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1393 EVP_MD_CTX_(new|free). The old names are retained as macros
1394 for deprecated builds.
1397 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1398 cryptographic operations to be performed asynchronously as long as an
1399 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1400 further details. Libssl has also had this capability integrated with the
1401 introduction of the new mode SSL_MODE_ASYNC and associated error
1402 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1403 pages. This work was developed in partnership with Intel Corp.
1406 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1407 always enabled now. If you want to disable the support you should
1408 exclude it using the list of supported ciphers. This also means that the
1409 "-no_ecdhe" option has been removed from s_server.
1412 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1413 SSL_{CTX_}set1_curves() which can set a list.
1416 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1417 curve you want to support using SSL_{CTX_}set1_curves().
1420 *) State machine rewrite. The state machine code has been significantly
1421 refactored in order to remove much duplication of code and solve issues
1422 with the old code (see ssl/statem/README for further details). This change
1423 does have some associated API changes. Notably the SSL_state() function
1424 has been removed and replaced by SSL_get_state which now returns an
1425 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1426 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1430 *) All instances of the string "ssleay" in the public API were replaced
1431 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1432 Some error codes related to internal RSA_eay API's were renamed.
1435 *) The demo files in crypto/threads were moved to demo/threads.
1438 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1440 [Matt Caswell, Rich Salz]
1442 *) New ASN.1 embed macro.
1444 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1445 structure is not allocated: it is part of the parent. That is instead of
1453 This reduces memory fragmentation and make it impossible to accidentally
1454 set a mandatory field to NULL.
1456 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1457 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1458 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1462 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1465 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1466 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1467 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1468 DES and RC4 ciphersuites.
1471 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1472 This changes the decoding behaviour for some invalid messages,
1473 though the change is mostly in the more lenient direction, and
1474 legacy behaviour is preserved as much as possible.
1477 *) Fix no-stdio build.
1478 [ David Woodhouse <David.Woodhouse@intel.com> and also
1479 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1481 *) New testing framework
1482 The testing framework has been largely rewritten and is now using
1483 perl and the perl modules Test::Harness and an extended variant of
1484 Test::More called OpenSSL::Test to do its work. All test scripts in
1485 test/ have been rewritten into test recipes, and all direct calls to
1486 executables in test/Makefile have become individual recipes using the
1487 simplified testing OpenSSL::Test::Simple.
1489 For documentation on our testing modules, do:
1491 perldoc test/testlib/OpenSSL/Test/Simple.pm
1492 perldoc test/testlib/OpenSSL/Test.pm
1496 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1497 are used; the latter aborts on memory leaks (usually checked on exit).
1498 Some undocumented "set malloc, etc., hooks" functions were removed
1499 and others were changed. All are now documented.
1502 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1504 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1506 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1507 from RFC4279, RFC4785, RFC5487, RFC5489.
1509 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1510 original RSA_PSK patch.
1513 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1514 era flag was never set throughout the codebase (only read). Also removed
1515 SSL3_FLAGS_POP_BUFFER which was only used if
1516 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1519 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1520 to be "oneline" instead of "compat".
1523 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1524 not aware of clients that still exhibit this bug, and the workaround
1525 hasn't been working properly for a while.
1528 *) The return type of BIO_number_read() and BIO_number_written() as well as
1529 the corresponding num_read and num_write members in the BIO structure has
1530 changed from unsigned long to uint64_t. On platforms where an unsigned
1531 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1535 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1536 OpenSSL without support for them. It also means that maintaining
1537 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1538 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1541 *) Removed support for the two export grade static DH ciphersuites
1542 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1543 were newly added (along with a number of other static DH ciphersuites) to
1544 1.0.2. However the two export ones have *never* worked since they were
1545 introduced. It seems strange in any case to be adding new export
1546 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1549 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1550 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1551 and turned into macros which simply call the new preferred function names
1552 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1553 should use the new names instead. Also as part of this change the ssl23.h
1554 header file has been removed.
1557 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1558 code and the associated standard is no longer considered fit-for-purpose.
1561 *) RT2547 was closed. When generating a private key, try to make the
1562 output file readable only by the owner. This behavior change might
1563 be noticeable when interacting with other software.
1565 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1569 *) Added HTTP GET support to the ocsp command.
1572 *) Changed default digest for the dgst and enc commands from MD5 to
1576 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1579 *) Added support for TLS extended master secret from
1580 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1581 initial patch which was a great help during development.
1584 *) All libssl internal structures have been removed from the public header
1585 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1586 now redundant). Users should not attempt to access internal structures
1587 directly. Instead they should use the provided API functions.
1590 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1591 Access to deprecated functions can be re-enabled by running config with
1592 "enable-deprecated". In addition applications wishing to use deprecated
1593 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1594 will, by default, disable some transitive includes that previously existed
1595 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1598 *) Added support for OCB mode. OpenSSL has been granted a patent license
1599 compatible with the OpenSSL license for use of OCB. Details are available
1600 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1601 for OCB can be removed by calling config with no-ocb.
1604 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1605 compatible client hello.
1608 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1609 done while fixing the error code for the key-too-small case.
1610 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1612 *) CA.sh has been removed; use CA.pl instead.
1615 *) Removed old DES API.
1618 *) Remove various unsupported platforms:
1624 Sinix/ReliantUNIX RM400
1629 16-bit platforms such as WIN16
1632 *) Clean up OPENSSL_NO_xxx #define's
1633 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1634 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1635 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1636 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1637 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1638 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1639 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1640 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1641 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1642 Remove MS_STATIC; it's a relic from platforms <32 bits.
1645 *) Cleaned up dead code
1646 Remove all but one '#ifdef undef' which is to be looked at.
1649 *) Clean up calling of xxx_free routines.
1650 Just like free(), fix most of the xxx_free routines to accept
1651 NULL. Remove the non-null checks from callers. Save much code.
1654 *) Add secure heap for storage of private keys (when possible).
1655 Add BIO_s_secmem(), CBIGNUM, etc.
1656 Contributed by Akamai Technologies under our Corporate CLA.
1659 *) Experimental support for a new, fast, unbiased prime candidate generator,
1660 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1661 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1663 *) New output format NSS in the sess_id command line tool. This allows
1664 exporting the session id and the master key in NSS keylog format.
1665 [Martin Kaiser <martin@kaiser.cx>]
1667 *) Harmonize version and its documentation. -f flag is used to display
1669 [mancha <mancha1@zoho.com>]
1671 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1672 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1673 [mancha <mancha1@zoho.com>]
1675 *) Fix some double frees. These are not thought to be exploitable.
1676 [mancha <mancha1@zoho.com>]
1678 *) A missing bounds check in the handling of the TLS heartbeat extension
1679 can be used to reveal up to 64k of memory to a connected client or
1682 Thanks for Neel Mehta of Google Security for discovering this bug and to
1683 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1684 preparing the fix (CVE-2014-0160)
1685 [Adam Langley, Bodo Moeller]
1687 *) Fix for the attack described in the paper "Recovering OpenSSL
1688 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1689 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1690 http://eprint.iacr.org/2014/140
1692 Thanks to Yuval Yarom and Naomi Benger for discovering this
1693 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1694 [Yuval Yarom and Naomi Benger]
1696 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1697 this fixes a limitation in previous versions of OpenSSL.
1700 *) Experimental encrypt-then-mac support.
1702 Experimental support for encrypt then mac from
1703 draft-gutmann-tls-encrypt-then-mac-02.txt
1705 To enable it set the appropriate extension number (0x42 for the test
1706 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1708 For non-compliant peers (i.e. just about everything) this should have no
1711 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1715 *) Add EVP support for key wrapping algorithms, to avoid problems with
1716 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1717 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1718 algorithms and include tests cases.
1721 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1725 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1726 MGF1 digest and OAEP label.
1729 *) Make openssl verify return errors.
1730 [Chris Palmer <palmer@google.com> and Ben Laurie]
1732 *) New function ASN1_TIME_diff to calculate the difference between two
1733 ASN1_TIME structures or one structure and the current time.
1736 *) Update fips_test_suite to support multiple command line options. New
1737 test to induce all self test errors in sequence and check expected
1741 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1742 sign or verify all in one operation.
1745 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1746 test programs and fips_test_suite. Includes functionality to parse
1747 the minimal script output of fipsalgest.pl directly.
1750 *) Add authorisation parameter to FIPS_module_mode_set().
1753 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1756 *) Use separate DRBG fields for internal and external flags. New function
1757 FIPS_drbg_health_check() to perform on demand health checking. Add
1758 generation tests to fips_test_suite with reduced health check interval to
1759 demonstrate periodic health checking. Add "nodh" option to
1760 fips_test_suite to skip very slow DH test.
1763 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1767 *) More extensive health check for DRBG checking many more failure modes.
1768 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1769 combination: call this in fips_test_suite.
1772 *) Add support for canonical generation of DSA parameter 'g'. See
1775 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1776 POST to handle HMAC cases.
1779 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1780 to return numerical and string versions of the FIPS module number.
1783 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1784 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1785 outside the validated module in the FIPS capable OpenSSL.
1788 *) Minor change to DRBG entropy callback semantics. In some cases
1789 there is no multiple of the block length between min_len and
1790 max_len. Allow the callback to return more than max_len bytes
1791 of entropy but discard any extra: it is the callback's responsibility
1792 to ensure that the extra data discarded does not impact the
1793 requested amount of entropy.
1796 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1797 information in FIPS186-3, SP800-57 and SP800-131A.
1800 *) CCM support via EVP. Interface is very similar to GCM case except we
1801 must supply all data in one chunk (i.e. no update, final) and the
1802 message length must be supplied if AAD is used. Add algorithm test
1806 *) Initial version of POST overhaul. Add POST callback to allow the status
1807 of POST to be monitored and/or failures induced. Modify fips_test_suite
1808 to use callback. Always run all selftests even if one fails.
1811 *) XTS support including algorithm test driver in the fips_gcmtest program.
1812 Note: this does increase the maximum key length from 32 to 64 bytes but
1813 there should be no binary compatibility issues as existing applications
1814 will never use XTS mode.
1817 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1818 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1819 performs algorithm blocking for unapproved PRNG types. Also do not
1820 set PRNG type in FIPS_mode_set(): leave this to the application.
1821 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1822 the standard OpenSSL PRNG: set additional data to a date time vector.
1825 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1826 This shouldn't present any incompatibility problems because applications
1827 shouldn't be using these directly and any that are will need to rethink
1828 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1831 *) Extensive self tests and health checking required by SP800-90 DRBG.
1832 Remove strength parameter from FIPS_drbg_instantiate and always
1833 instantiate at maximum supported strength.
1836 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1839 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1842 *) New function DH_compute_key_padded() to compute a DH key and pad with
1843 leading zeroes if needed: this complies with SP800-56A et al.
1846 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1847 anything, incomplete, subject to change and largely untested at present.
1850 *) Modify fipscanisteronly build option to only build the necessary object
1851 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1854 *) Add experimental option FIPSSYMS to give all symbols in
1855 fipscanister.o and FIPS or fips prefix. This will avoid
1856 conflicts with future versions of OpenSSL. Add perl script
1857 util/fipsas.pl to preprocess assembly language source files
1858 and rename any affected symbols.
1861 *) Add selftest checks and algorithm block of non-fips algorithms in
1862 FIPS mode. Remove DES2 from selftests.
1865 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1866 return internal method without any ENGINE dependencies. Add new
1867 tiny fips sign and verify functions.
1870 *) New build option no-ec2m to disable characteristic 2 code.
1873 *) New build option "fipscanisteronly". This only builds fipscanister.o
1874 and (currently) associated fips utilities. Uses the file Makefile.fips
1875 instead of Makefile.org as the prototype.
1878 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1879 Update fips_gcmtest to use IV generator.
1882 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1883 setting output buffer to NULL. The *Final function must be
1884 called although it will not retrieve any additional data. The tag
1885 can be set or retrieved with a ctrl. The IV length is by default 12
1886 bytes (96 bits) but can be set to an alternative value. If the IV
1887 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1891 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1892 underlying do_cipher function handles all cipher semantics itself
1893 including padding and finalisation. This is useful if (for example)
1894 an ENGINE cipher handles block padding itself. The behaviour of
1895 do_cipher is subtly changed if this flag is set: the return value
1896 is the number of characters written to the output buffer (zero is
1897 no longer an error code) or a negative error code. Also if the
1898 input buffer is NULL and length 0 finalisation should be performed.
1901 *) If a candidate issuer certificate is already part of the constructed
1902 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1905 *) Improve forward-security support: add functions
1907 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1908 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1910 for use by SSL/TLS servers; the callback function will be called whenever a
1911 new session is created, and gets to decide whether the session may be
1912 cached to make it resumable (return 0) or not (return 1). (As by the
1913 SSL/TLS protocol specifications, the session_id sent by the server will be
1914 empty to indicate that the session is not resumable; also, the server will
1915 not generate RFC 4507 (RFC 5077) session tickets.)
1917 A simple reasonable callback implementation is to return is_forward_secure.
1918 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1919 by the SSL/TLS server library, indicating whether it can provide forward
1921 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1923 *) New -verify_name option in command line utilities to set verification
1927 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1928 Add CMAC pkey methods.
1931 *) Experimental renegotiation in s_server -www mode. If the client
1932 browses /reneg connection is renegotiated. If /renegcert it is
1933 renegotiated requesting a certificate.
1936 *) Add an "external" session cache for debugging purposes to s_server. This
1937 should help trace issues which normally are only apparent in deployed
1938 multi-process servers.
1941 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1942 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1943 BIO_set_cipher() and some obscure PEM functions were changed so they
1944 can now return an error. The RAND changes required a change to the
1945 RAND_METHOD structure.
1948 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1949 a gcc attribute to warn if the result of a function is ignored. This
1950 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1951 whose return value is often ignored.
1954 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1955 These allow SCTs (signed certificate timestamps) to be requested and
1956 validated when establishing a connection.
1957 [Rob Percival <robpercival@google.com>]
1959 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1961 *) Prevent padding oracle in AES-NI CBC MAC check
1963 A MITM attacker can use a padding oracle attack to decrypt traffic
1964 when the connection uses an AES CBC cipher and the server support
1967 This issue was introduced as part of the fix for Lucky 13 padding
1968 attack (CVE-2013-0169). The padding check was rewritten to be in
1969 constant time by making sure that always the same bytes are read and
1970 compared against either the MAC or padding bytes. But it no longer
1971 checked that there was enough data to have both the MAC and padding
1974 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1978 *) Fix EVP_EncodeUpdate overflow
1980 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1981 Base64 encoding of binary data. If an attacker is able to supply very large
1982 amounts of input data then a length check can overflow resulting in a heap
1985 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1986 the PEM_write_bio* family of functions. These are mainly used within the
1987 OpenSSL command line applications, so any application which processes data
1988 from an untrusted source and outputs it as a PEM file should be considered
1989 vulnerable to this issue. User applications that call these APIs directly
1990 with large amounts of untrusted data may also be vulnerable.
1992 This issue was reported by Guido Vranken.
1996 *) Fix EVP_EncryptUpdate overflow
1998 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1999 is able to supply very large amounts of input data after a previous call to
2000 EVP_EncryptUpdate() with a partial block then a length check can overflow
2001 resulting in a heap corruption. Following an analysis of all OpenSSL
2002 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2003 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2004 the first called function after an EVP_EncryptInit(), and therefore that
2005 specific call must be safe. The second form is where the length passed to
2006 EVP_EncryptUpdate() can be seen from the code to be some small value and
2007 therefore there is no possibility of an overflow. Since all instances are
2008 one of these two forms, it is believed that there can be no overflows in
2009 internal code due to this problem. It should be noted that
2010 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2011 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2012 of these calls have also been analysed too and it is believed there are no
2013 instances in internal usage where an overflow could occur.
2015 This issue was reported by Guido Vranken.
2019 *) Prevent ASN.1 BIO excessive memory allocation
2021 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2022 a short invalid encoding can cause allocation of large amounts of memory
2023 potentially consuming excessive resources or exhausting memory.
2025 Any application parsing untrusted data through d2i BIO functions is
2026 affected. The memory based functions such as d2i_X509() are *not* affected.
2027 Since the memory based functions are used by the TLS library, TLS
2028 applications are not affected.
2030 This issue was reported by Brian Carpenter.
2036 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2037 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2038 in arbitrary stack data being returned in the buffer.
2040 This issue was reported by Guido Vranken.
2044 *) Modify behavior of ALPN to invoke callback after SNI/servername
2045 callback, such that updates to the SSL_CTX affect ALPN.
2048 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2052 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2053 methods are enabled and ssl2 is disabled the methods return NULL.
2056 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2058 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2059 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2060 provide any "EXPORT" or "LOW" strength ciphers.
2063 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2064 is by default disabled at build-time. Builds that are not configured with
2065 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2066 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2067 will need to explicitly call either of:
2069 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2071 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2073 as appropriate. Even if either of those is used, or the application
2074 explicitly uses the version-specific SSLv2_method() or its client and
2075 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2076 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2077 ciphers, and SSLv2 56-bit DES are no longer available.
2081 *) Fix a double-free in DSA code
2083 A double free bug was discovered when OpenSSL parses malformed DSA private
2084 keys and could lead to a DoS attack or memory corruption for applications
2085 that receive DSA private keys from untrusted sources. This scenario is
2088 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2093 *) Disable SRP fake user seed to address a server memory leak.
2095 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2097 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2098 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2099 was changed to ignore the "fake user" SRP seed, even if the seed
2102 Users should use SRP_VBASE_get1_by_user instead. Note that in
2103 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2104 also that even though configuring the SRP seed attempts to hide
2105 invalid usernames by continuing the handshake with fake
2106 credentials, this behaviour is not constant time and no strong
2107 guarantees are made that the handshake is indistinguishable from
2108 that of a valid user.
2112 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2114 In the BN_hex2bn function the number of hex digits is calculated using an
2115 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2116 large values of |i| this can result in |bn_expand| not allocating any
2117 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2118 field as NULL leading to a subsequent NULL ptr deref. For very large values
2119 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2120 In this case memory is allocated to the internal BIGNUM data field, but it
2121 is insufficiently sized leading to heap corruption. A similar issue exists
2122 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2123 is ever called by user applications with very large untrusted hex/dec data.
2124 This is anticipated to be a rare occurrence.
2126 All OpenSSL internal usage of these functions use data that is not expected
2127 to be untrusted, e.g. config file data or application command line
2128 arguments. If user developed applications generate config file data based
2129 on untrusted data then it is possible that this could also lead to security
2130 consequences. This is also anticipated to be rare.
2132 This issue was reported to OpenSSL by Guido Vranken.
2136 *) Fix memory issues in BIO_*printf functions
2138 The internal |fmtstr| function used in processing a "%s" format string in
2139 the BIO_*printf functions could overflow while calculating the length of a
2140 string and cause an OOB read when printing very long strings.
2142 Additionally the internal |doapr_outch| function can attempt to write to an
2143 OOB memory location (at an offset from the NULL pointer) in the event of a
2144 memory allocation failure. In 1.0.2 and below this could be caused where
2145 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2146 could be in processing a very long "%s" format string. Memory leaks can
2149 The first issue may mask the second issue dependent on compiler behaviour.
2150 These problems could enable attacks where large amounts of untrusted data
2151 is passed to the BIO_*printf functions. If applications use these functions
2152 in this way then they could be vulnerable. OpenSSL itself uses these
2153 functions when printing out human-readable dumps of ASN.1 data. Therefore
2154 applications that print this data could be vulnerable if the data is from
2155 untrusted sources. OpenSSL command line applications could also be
2156 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2157 as command line arguments.
2159 Libssl is not considered directly vulnerable. Additionally certificates etc
2160 received via remote connections via libssl are also unlikely to be able to
2161 trigger these issues because of message size limits enforced within libssl.
2163 This issue was reported to OpenSSL Guido Vranken.
2167 *) Side channel attack on modular exponentiation
2169 A side-channel attack was found which makes use of cache-bank conflicts on
2170 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2171 of RSA keys. The ability to exploit this issue is limited as it relies on
2172 an attacker who has control of code in a thread running on the same
2173 hyper-threaded core as the victim thread which is performing decryptions.
2175 This issue was reported to OpenSSL by Yuval Yarom, The University of
2176 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2177 Nadia Heninger, University of Pennsylvania with more information at
2178 http://cachebleed.info.
2182 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2183 if no keysize is specified with default_bits. This fixes an
2184 omission in an earlier change that changed all RSA/DSA key generation
2185 apps to use 2048 bits by default.
2188 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2189 *) DH small subgroups
2191 Historically OpenSSL only ever generated DH parameters based on "safe"
2192 primes. More recently (in version 1.0.2) support was provided for
2193 generating X9.42 style parameter files such as those required for RFC 5114
2194 support. The primes used in such files may not be "safe". Where an
2195 application is using DH configured with parameters based on primes that are
2196 not "safe" then an attacker could use this fact to find a peer's private
2197 DH exponent. This attack requires that the attacker complete multiple
2198 handshakes in which the peer uses the same private DH exponent. For example
2199 this could be used to discover a TLS server's private DH exponent if it's
2200 reusing the private DH exponent or it's using a static DH ciphersuite.
2202 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2203 TLS. It is not on by default. If the option is not set then the server
2204 reuses the same private DH exponent for the life of the server process and
2205 would be vulnerable to this attack. It is believed that many popular
2206 applications do set this option and would therefore not be at risk.
2208 The fix for this issue adds an additional check where a "q" parameter is
2209 available (as is the case in X9.42 based parameters). This detects the
2210 only known attack, and is the only possible defense for static DH
2211 ciphersuites. This could have some performance impact.
2213 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2214 default and cannot be disabled. This could have some performance impact.
2216 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2220 *) SSLv2 doesn't block disabled ciphers
2222 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2223 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2224 been disabled, provided that the SSLv2 protocol was not also disabled via
2227 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2228 and Sebastian Schinzel.
2232 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2234 *) BN_mod_exp may produce incorrect results on x86_64
2236 There is a carry propagating bug in the x86_64 Montgomery squaring
2237 procedure. No EC algorithms are affected. Analysis suggests that attacks
2238 against RSA and DSA as a result of this defect would be very difficult to
2239 perform and are not believed likely. Attacks against DH are considered just
2240 feasible (although very difficult) because most of the work necessary to
2241 deduce information about a private key may be performed offline. The amount
2242 of resources required for such an attack would be very significant and
2243 likely only accessible to a limited number of attackers. An attacker would
2244 additionally need online access to an unpatched system using the target
2245 private key in a scenario with persistent DH parameters and a private
2246 key that is shared between multiple clients. For example this can occur by
2247 default in OpenSSL DHE based SSL/TLS ciphersuites.
2249 This issue was reported to OpenSSL by Hanno Böck.
2253 *) Certificate verify crash with missing PSS parameter
2255 The signature verification routines will crash with a NULL pointer
2256 dereference if presented with an ASN.1 signature using the RSA PSS
2257 algorithm and absent mask generation function parameter. Since these
2258 routines are used to verify certificate signature algorithms this can be
2259 used to crash any certificate verification operation and exploited in a
2260 DoS attack. Any application which performs certificate verification is
2261 vulnerable including OpenSSL clients and servers which enable client
2264 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2268 *) X509_ATTRIBUTE memory leak
2270 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2271 memory. This structure is used by the PKCS#7 and CMS routines so any
2272 application which reads PKCS#7 or CMS data from untrusted sources is
2273 affected. SSL/TLS is not affected.
2275 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2280 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2281 This changes the decoding behaviour for some invalid messages,
2282 though the change is mostly in the more lenient direction, and
2283 legacy behaviour is preserved as much as possible.
2286 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2288 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2290 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2292 *) Alternate chains certificate forgery
2294 During certificate verification, OpenSSL will attempt to find an
2295 alternative certificate chain if the first attempt to build such a chain
2296 fails. An error in the implementation of this logic can mean that an
2297 attacker could cause certain checks on untrusted certificates to be
2298 bypassed, such as the CA flag, enabling them to use a valid leaf
2299 certificate to act as a CA and "issue" an invalid certificate.
2301 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2305 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2307 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2308 incompatibility in the handling of HMAC. The previous ABI has now been
2312 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2314 *) Malformed ECParameters causes infinite loop
2316 When processing an ECParameters structure OpenSSL enters an infinite loop
2317 if the curve specified is over a specially malformed binary polynomial
2320 This can be used to perform denial of service against any
2321 system which processes public keys, certificate requests or
2322 certificates. This includes TLS clients and TLS servers with
2323 client authentication enabled.
2325 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2329 *) Exploitable out-of-bounds read in X509_cmp_time
2331 X509_cmp_time does not properly check the length of the ASN1_TIME
2332 string and can read a few bytes out of bounds. In addition,
2333 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2336 An attacker can use this to craft malformed certificates and CRLs of
2337 various sizes and potentially cause a segmentation fault, resulting in
2338 a DoS on applications that verify certificates or CRLs. TLS clients
2339 that verify CRLs are affected. TLS clients and servers with client
2340 authentication enabled may be affected if they use custom verification
2343 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2344 independently by Hanno Böck.
2348 *) PKCS7 crash with missing EnvelopedContent
2350 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2351 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2352 with missing content and trigger a NULL pointer dereference on parsing.
2354 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2355 structures from untrusted sources are affected. OpenSSL clients and
2356 servers are not affected.
2358 This issue was reported to OpenSSL by Michal Zalewski (Google).
2362 *) CMS verify infinite loop with unknown hash function
2364 When verifying a signedData message the CMS code can enter an infinite loop
2365 if presented with an unknown hash function OID. This can be used to perform
2366 denial of service against any system which verifies signedData messages using
2368 This issue was reported to OpenSSL by Johannes Bauer.
2372 *) Race condition handling NewSessionTicket
2374 If a NewSessionTicket is received by a multi-threaded client when attempting to
2375 reuse a previous ticket then a race condition can occur potentially leading to
2376 a double free of the ticket data.
2380 *) Only support 256-bit or stronger elliptic curves with the
2381 'ecdh_auto' setting (server) or by default (client). Of supported
2382 curves, prefer P-256 (both).
2385 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2387 *) ClientHello sigalgs DoS fix
2389 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2390 invalid signature algorithms extension a NULL pointer dereference will
2391 occur. This can be exploited in a DoS attack against the server.
2393 This issue was was reported to OpenSSL by David Ramos of Stanford
2396 [Stephen Henson and Matt Caswell]
2398 *) Multiblock corrupted pointer fix
2400 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2401 feature only applies on 64 bit x86 architecture platforms that support AES
2402 NI instructions. A defect in the implementation of "multiblock" can cause
2403 OpenSSL's internal write buffer to become incorrectly set to NULL when
2404 using non-blocking IO. Typically, when the user application is using a
2405 socket BIO for writing, this will only result in a failed connection.
2406 However if some other BIO is used then it is likely that a segmentation
2407 fault will be triggered, thus enabling a potential DoS attack.
2409 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2413 *) Segmentation fault in DTLSv1_listen fix
2415 The DTLSv1_listen function is intended to be stateless and processes the
2416 initial ClientHello from many peers. It is common for user code to loop
2417 over the call to DTLSv1_listen until a valid ClientHello is received with
2418 an associated cookie. A defect in the implementation of DTLSv1_listen means
2419 that state is preserved in the SSL object from one invocation to the next
2420 that can lead to a segmentation fault. Errors processing the initial
2421 ClientHello can trigger this scenario. An example of such an error could be
2422 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2425 This issue was reported to OpenSSL by Per Allansson.
2429 *) Segmentation fault in ASN1_TYPE_cmp fix
2431 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2432 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2433 certificate signature algorithm consistency this can be used to crash any
2434 certificate verification operation and exploited in a DoS attack. Any
2435 application which performs certificate verification is vulnerable including
2436 OpenSSL clients and servers which enable client authentication.
2440 *) Segmentation fault for invalid PSS parameters fix
2442 The signature verification routines will crash with a NULL pointer
2443 dereference if presented with an ASN.1 signature using the RSA PSS
2444 algorithm and invalid parameters. Since these routines are used to verify
2445 certificate signature algorithms this can be used to crash any
2446 certificate verification operation and exploited in a DoS attack. Any
2447 application which performs certificate verification is vulnerable including
2448 OpenSSL clients and servers which enable client authentication.
2450 This issue was was reported to OpenSSL by Brian Carpenter.
2454 *) ASN.1 structure reuse memory corruption fix
2456 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2457 memory corruption via an invalid write. Such reuse is and has been
2458 strongly discouraged and is believed to be rare.
2460 Applications that parse structures containing CHOICE or ANY DEFINED BY
2461 components may be affected. Certificate parsing (d2i_X509 and related
2462 functions) are however not affected. OpenSSL clients and servers are
2467 *) PKCS7 NULL pointer dereferences fix
2469 The PKCS#7 parsing code does not handle missing outer ContentInfo
2470 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2471 missing content and trigger a NULL pointer dereference on parsing.
2473 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2474 otherwise parse PKCS#7 structures from untrusted sources are
2475 affected. OpenSSL clients and servers are not affected.
2477 This issue was reported to OpenSSL by Michal Zalewski (Google).
2481 *) DoS via reachable assert in SSLv2 servers fix
2483 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2484 servers that both support SSLv2 and enable export cipher suites by sending
2485 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2487 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2488 (OpenSSL development team).
2492 *) Empty CKE with client auth and DHE fix
2494 If client auth is used then a server can seg fault in the event of a DHE
2495 ciphersuite being selected and a zero length ClientKeyExchange message
2496 being sent by the client. This could be exploited in a DoS attack.
2500 *) Handshake with unseeded PRNG fix
2502 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2503 with an unseeded PRNG. The conditions are:
2504 - The client is on a platform where the PRNG has not been seeded
2505 automatically, and the user has not seeded manually
2506 - A protocol specific client method version has been used (i.e. not
2507 SSL_client_methodv23)
2508 - A ciphersuite is used that does not require additional random data from
2509 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2511 If the handshake succeeds then the client random that has been used will
2512 have been generated from a PRNG with insufficient entropy and therefore the
2513 output may be predictable.
2515 For example using the following command with an unseeded openssl will
2516 succeed on an unpatched platform:
2518 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2522 *) Use After Free following d2i_ECPrivatekey error fix
2524 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2525 could cause a use after free condition. This, in turn, could cause a double
2526 free in several private key parsing functions (such as d2i_PrivateKey
2527 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2528 for applications that receive EC private keys from untrusted
2529 sources. This scenario is considered rare.
2531 This issue was discovered by the BoringSSL project and fixed in their
2536 *) X509_to_X509_REQ NULL pointer deref fix
2538 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2539 the certificate key is invalid. This function is rarely used in practice.
2541 This issue was discovered by Brian Carpenter.
2545 *) Removed the export ciphers from the DEFAULT ciphers
2548 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2550 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2551 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2552 So far those who have to target multiple platforms would compromise
2553 and argue that binary targeting say ARMv5 would still execute on
2554 ARMv8. "Universal" build resolves this compromise by providing
2555 near-optimal performance even on newer platforms.
2558 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2559 (other platforms pending).
2560 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2562 *) Add support for the SignedCertificateTimestampList certificate and
2563 OCSP response extensions from RFC6962.
2566 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2567 for corner cases. (Certain input points at infinity could lead to
2568 bogus results, with non-infinity inputs mapped to infinity too.)
2571 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2572 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2573 common cases are optimized and there still is room for further
2574 improvements. Vector Permutation AES for Altivec is also added.
2577 *) Add support for little-endian ppc64 Linux target.
2578 [Marcelo Cerri (IBM)]
2580 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2581 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2582 are optimized and there still is room for further improvements.
2583 Both 32- and 64-bit modes are supported.
2584 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2586 *) Improved ARMv7 NEON support.
2589 *) Support for SPARC Architecture 2011 crypto extensions, first
2590 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2591 SHA256/512, MD5, GHASH and modular exponentiation.
2592 [Andy Polyakov, David Miller]
2594 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2596 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2598 *) Support for new and upcoming Intel processors, including AVX2,
2599 BMI and SHA ISA extensions. This includes additional "stitched"
2600 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2603 This work was sponsored by Intel Corp.
2606 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2607 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2608 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2611 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2612 this fixes a limitation in previous versions of OpenSSL.
2615 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2616 MGF1 digest and OAEP label.
2619 *) Add EVP support for key wrapping algorithms, to avoid problems with
2620 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2621 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2622 algorithms and include tests cases.
2625 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2627 [Douglas E. Engert, Steve Henson]
2629 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2630 difference in days and seconds between two tm or ASN1_TIME structures.
2633 *) Add -rev test option to s_server to just reverse order of characters
2634 received by client and send back to server. Also prints an abbreviated
2635 summary of the connection parameters.
2638 *) New option -brief for s_client and s_server to print out a brief summary
2639 of connection parameters.
2642 *) Add callbacks for arbitrary TLS extensions.
2643 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2645 *) New option -crl_download in several openssl utilities to download CRLs
2646 from CRLDP extension in certificates.
2649 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2652 *) New function X509_CRL_diff to generate a delta CRL from the difference
2653 of two full CRLs. Add support to "crl" utility.
2656 *) New functions to set lookup_crls function and to retrieve
2657 X509_STORE from X509_STORE_CTX.
2660 *) Print out deprecated issuer and subject unique ID fields in
2664 *) Extend OCSP I/O functions so they can be used for simple general purpose
2665 HTTP as well as OCSP. New wrapper function which can be used to download
2666 CRLs using the OCSP API.
2669 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2672 *) SSL_CONF* functions. These provide a common framework for application
2673 configuration using configuration files or command lines.
2676 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2677 message callback and prints the results. Needs compile time option
2678 "enable-ssl-trace". New options to s_client and s_server to enable
2682 *) New ctrl and macro to retrieve supported points extensions.
2683 Print out extension in s_server and s_client.
2686 *) New functions to retrieve certificate signature and signature
2690 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2694 *) New Suite B modes for TLS code. These use and enforce the requirements
2695 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2696 only use Suite B curves. The Suite B modes can be set by using the
2697 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2700 *) New chain verification flags for Suite B levels of security. Check
2701 algorithms are acceptable when flags are set in X509_verify_cert.
2704 *) Make tls1_check_chain return a set of flags indicating checks passed
2705 by a certificate chain. Add additional tests to handle client
2706 certificates: checks for matching certificate type and issuer name
2710 *) If an attempt is made to use a signature algorithm not in the peer
2711 preference list abort the handshake. If client has no suitable
2712 signature algorithms in response to a certificate request do not
2713 use the certificate.
2716 *) If server EC tmp key is not in client preference list abort handshake.
2719 *) Add support for certificate stores in CERT structure. This makes it
2720 possible to have different stores per SSL structure or one store in
2721 the parent SSL_CTX. Include distinct stores for certificate chain
2722 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2723 to build and store a certificate chain in CERT structure: returning
2724 an error if the chain cannot be built: this will allow applications
2725 to test if a chain is correctly configured.
2727 Note: if the CERT based stores are not set then the parent SSL_CTX
2728 store is used to retain compatibility with existing behaviour.
2732 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2733 mask based on the current session, check mask when sending client
2734 hello and checking the requested ciphersuite.
2737 *) New ctrls to retrieve and set certificate types in a certificate
2738 request message. Print out received values in s_client. If certificate
2739 types is not set with custom values set sensible values based on
2740 supported signature algorithms.
2743 *) Support for distinct client and server supported signature algorithms.
2746 *) Add certificate callback. If set this is called whenever a certificate
2747 is required by client or server. An application can decide which
2748 certificate chain to present based on arbitrary criteria: for example
2749 supported signature algorithms. Add very simple example to s_server.
2750 This fixes many of the problems and restrictions of the existing client
2751 certificate callback: for example you can now clear an existing
2752 certificate and specify the whole chain.
2755 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2756 the certificate can be used for (if anything). Set valid_flags field
2757 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2758 to have similar checks in it.
2760 Add new "cert_flags" field to CERT structure and include a "strict mode".
2761 This enforces some TLS certificate requirements (such as only permitting
2762 certificate signature algorithms contained in the supported algorithms
2763 extension) which some implementations ignore: this option should be used
2764 with caution as it could cause interoperability issues.
2767 *) Update and tidy signature algorithm extension processing. Work out
2768 shared signature algorithms based on preferences and peer algorithms
2769 and print them out in s_client and s_server. Abort handshake if no
2770 shared signature algorithms.
2773 *) Add new functions to allow customised supported signature algorithms
2774 for SSL and SSL_CTX structures. Add options to s_client and s_server
2778 *) New function SSL_certs_clear() to delete all references to certificates
2779 from an SSL structure. Before this once a certificate had been added
2780 it couldn't be removed.
2783 *) Integrate hostname, email address and IP address checking with certificate
2784 verification. New verify options supporting checking in openssl utility.
2787 *) Fixes and wildcard matching support to hostname and email checking
2788 functions. Add manual page.
2789 [Florian Weimer (Red Hat Product Security Team)]
2791 *) New functions to check a hostname email or IP address against a
2792 certificate. Add options x509 utility to print results of checks against
2796 *) Fix OCSP checking.
2797 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2799 *) Initial experimental support for explicitly trusted non-root CAs.
2800 OpenSSL still tries to build a complete chain to a root but if an
2801 intermediate CA has a trust setting included that is used. The first
2802 setting is used: whether to trust (e.g., -addtrust option to the x509
2806 *) Add -trusted_first option which attempts to find certificates in the
2807 trusted store even if an untrusted chain is also supplied.
2810 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2811 platform support for Linux and Android.
2814 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2817 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2818 When in FIPS mode the approved implementations are used as normal,
2819 when not in FIPS mode the internal unapproved versions are used instead.
2820 This means that the FIPS capable OpenSSL isn't forced to use the
2821 (often lower performance) FIPS implementations outside FIPS mode.
2824 *) Transparently support X9.42 DH parameters when calling
2825 PEM_read_bio_DHparameters. This means existing applications can handle
2826 the new parameter format automatically.
2829 *) Initial experimental support for X9.42 DH parameter format: mainly
2830 to support use of 'q' parameter for RFC5114 parameters.
2833 *) Add DH parameters from RFC5114 including test data to dhtest.
2836 *) Support for automatic EC temporary key parameter selection. If enabled
2837 the most preferred EC parameters are automatically used instead of
2838 hardcoded fixed parameters. Now a server just has to call:
2839 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2840 support ECDH and use the most appropriate parameters.
2843 *) Enhance and tidy EC curve and point format TLS extension code. Use
2844 static structures instead of allocation if default values are used.
2845 New ctrls to set curves we wish to support and to retrieve shared curves.
2846 Print out shared curves in s_server. New options to s_server and s_client
2847 to set list of supported curves.
2850 *) New ctrls to retrieve supported signature algorithms and
2851 supported curve values as an array of NIDs. Extend openssl utility
2852 to print out received values.
2855 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2856 between NIDs and the more common NIST names such as "P-256". Enhance
2857 ecparam utility and ECC method to recognise the NIST names for curves.
2860 *) Enhance SSL/TLS certificate chain handling to support different
2861 chains for each certificate instead of one chain in the parent SSL_CTX.
2864 *) Support for fixed DH ciphersuite client authentication: where both
2865 server and client use DH certificates with common parameters.
2868 *) Support for fixed DH ciphersuites: those requiring DH server
2872 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2874 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2875 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2876 X509_CINF_get_signature were reverted post internal team review.
2878 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2880 *) Build fixes for the Windows and OpenVMS platforms
2881 [Matt Caswell and Richard Levitte]
2883 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2885 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2886 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2887 dereference. This could lead to a Denial Of Service attack. Thanks to
2888 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2892 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2893 dtls1_buffer_record function under certain conditions. In particular this
2894 could occur if an attacker sent repeated DTLS records with the same
2895 sequence number but for the next epoch. The memory leak could be exploited
2896 by an attacker in a Denial of Service attack through memory exhaustion.
2897 Thanks to Chris Mueller for reporting this issue.
2901 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2902 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2903 method would be set to NULL which could later result in a NULL pointer
2904 dereference. Thanks to Frank Schmirler for reporting this issue.
2908 *) Abort handshake if server key exchange message is omitted for ephemeral
2911 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2912 reporting this issue.
2916 *) Remove non-export ephemeral RSA code on client and server. This code
2917 violated the TLS standard by allowing the use of temporary RSA keys in
2918 non-export ciphersuites and could be used by a server to effectively
2919 downgrade the RSA key length used to a value smaller than the server
2920 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2921 INRIA or reporting this issue.
2925 *) Fixed issue where DH client certificates are accepted without verification.
2926 An OpenSSL server will accept a DH certificate for client authentication
2927 without the certificate verify message. This effectively allows a client to
2928 authenticate without the use of a private key. This only affects servers
2929 which trust a client certificate authority which issues certificates
2930 containing DH keys: these are extremely rare and hardly ever encountered.
2931 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2936 *) Ensure that the session ID context of an SSL is updated when its
2937 SSL_CTX is updated via SSL_set_SSL_CTX.
2939 The session ID context is typically set from the parent SSL_CTX,
2940 and can vary with the CTX.
2943 *) Fix various certificate fingerprint issues.
2945 By using non-DER or invalid encodings outside the signed portion of a
2946 certificate the fingerprint can be changed without breaking the signature.
2947 Although no details of the signed portion of the certificate can be changed
2948 this can cause problems with some applications: e.g. those using the
2949 certificate fingerprint for blacklists.
2951 1. Reject signatures with non zero unused bits.
2953 If the BIT STRING containing the signature has non zero unused bits reject
2954 the signature. All current signature algorithms require zero unused bits.
2956 2. Check certificate algorithm consistency.
2958 Check the AlgorithmIdentifier inside TBS matches the one in the
2959 certificate signature. NB: this will result in signature failure
2960 errors for some broken certificates.
2962 Thanks to Konrad Kraszewski from Google for reporting this issue.
2964 3. Check DSA/ECDSA signatures use DER.
2966 Re-encode DSA/ECDSA signatures and compare with the original received
2967 signature. Return an error if there is a mismatch.
2969 This will reject various cases including garbage after signature
2970 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2971 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2972 (negative or with leading zeroes).
2974 Further analysis was conducted and fixes were developed by Stephen Henson
2975 of the OpenSSL core team.
2980 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2981 results on some platforms, including x86_64. This bug occurs at random
2982 with a very low probability, and is not known to be exploitable in any
2983 way, though its exact impact is difficult to determine. Thanks to Pieter
2984 Wuille (Blockstream) who reported this issue and also suggested an initial
2985 fix. Further analysis was conducted by the OpenSSL development team and
2986 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2987 the OpenSSL core team.
2991 *) Do not resume sessions on the server if the negotiated protocol
2992 version does not match the session's version. Resuming with a different
2993 version, while not strictly forbidden by the RFC, is of questionable
2994 sanity and breaks all known clients.
2995 [David Benjamin, Emilia Käsper]
2997 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2998 early CCS messages during renegotiation. (Note that because
2999 renegotiation is encrypted, this early CCS was not exploitable.)
3002 *) Tighten client-side session ticket handling during renegotiation:
3003 ensure that the client only accepts a session ticket if the server sends
3004 the extension anew in the ServerHello. Previously, a TLS client would
3005 reuse the old extension state and thus accept a session ticket if one was
3006 announced in the initial ServerHello.
3008 Similarly, ensure that the client requires a session ticket if one
3009 was advertised in the ServerHello. Previously, a TLS client would
3010 ignore a missing NewSessionTicket message.
3013 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3015 *) SRTP Memory Leak.
3017 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3018 sends a carefully crafted handshake message, to cause OpenSSL to fail
3019 to free up to 64k of memory causing a memory leak. This could be
3020 exploited in a Denial Of Service attack. This issue affects OpenSSL
3021 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3022 whether SRTP is used or configured. Implementations of OpenSSL that
3023 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3025 The fix was developed by the OpenSSL team.
3029 *) Session Ticket Memory Leak.
3031 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3032 integrity of that ticket is first verified. In the event of a session
3033 ticket integrity check failing, OpenSSL will fail to free memory
3034 causing a memory leak. By sending a large number of invalid session
3035 tickets an attacker could exploit this issue in a Denial Of Service
3040 *) Build option no-ssl3 is incomplete.
3042 When OpenSSL is configured with "no-ssl3" as a build option, servers
3043 could accept and complete a SSL 3.0 handshake, and clients could be
3044 configured to send them.
3046 [Akamai and the OpenSSL team]
3048 *) Add support for TLS_FALLBACK_SCSV.
3049 Client applications doing fallback retries should call
3050 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3052 [Adam Langley, Bodo Moeller]
3054 *) Add additional DigestInfo checks.
3056 Re-encode DigestInto in DER and check against the original when
3057 verifying RSA signature: this will reject any improperly encoded
3058 DigestInfo structures.
3060 Note: this is a precautionary measure and no attacks are currently known.
3064 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3066 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3067 SRP code can be overrun an internal buffer. Add sanity check that
3068 g, A, B < N to SRP code.
3070 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3071 Group for discovering this issue.
3075 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3076 TLS 1.0 instead of higher protocol versions when the ClientHello message
3077 is badly fragmented. This allows a man-in-the-middle attacker to force a
3078 downgrade to TLS 1.0 even if both the server and the client support a
3079 higher protocol version, by modifying the client's TLS records.
3081 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3082 researching this issue.
3086 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3087 to a denial of service attack. A malicious server can crash the client
3088 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3089 ciphersuite and sending carefully crafted handshake messages.
3091 Thanks to Felix Gröbert (Google) for discovering and researching this
3096 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3097 to leak memory. This can be exploited through a Denial of Service attack.
3098 Thanks to Adam Langley for discovering and researching this issue.
3102 *) An attacker can force openssl to consume large amounts of memory whilst
3103 processing DTLS handshake messages. This can be exploited through a
3104 Denial of Service attack.
3105 Thanks to Adam Langley for discovering and researching this issue.
3109 *) An attacker can force an error condition which causes openssl to crash
3110 whilst processing DTLS packets due to memory being freed twice. This
3111 can be exploited through a Denial of Service attack.
3112 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3117 *) If a multithreaded client connects to a malicious server using a resumed
3118 session and the server sends an ec point format extension it could write
3119 up to 255 bytes to freed memory.
3121 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3126 *) A malicious server can crash an OpenSSL client with a null pointer
3127 dereference (read) by specifying an SRP ciphersuite even though it was not
3128 properly negotiated with the client. This can be exploited through a
3129 Denial of Service attack.
3131 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3132 discovering and researching this issue.
3136 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3137 X509_name_oneline, X509_name_print_ex et al. to leak some information
3138 from the stack. Applications may be affected if they echo pretty printing
3139 output to the attacker.
3141 Thanks to Ivan Fratric (Google) for discovering this issue.
3143 [Emilia Käsper, and Steve Henson]
3145 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3146 for corner cases. (Certain input points at infinity could lead to
3147 bogus results, with non-infinity inputs mapped to infinity too.)
3150 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3152 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3153 handshake can force the use of weak keying material in OpenSSL
3154 SSL/TLS clients and servers.
3156 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3157 researching this issue. (CVE-2014-0224)
3158 [KIKUCHI Masashi, Steve Henson]
3160 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3161 OpenSSL DTLS client the code can be made to recurse eventually crashing
3164 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3166 [Imre Rad, Steve Henson]
3168 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3169 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3170 client or server. This is potentially exploitable to run arbitrary
3171 code on a vulnerable client or server.
3173 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3174 [Jüri Aedla, Steve Henson]
3176 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3177 are subject to a denial of service attack.
3179 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3180 this issue. (CVE-2014-3470)
3181 [Felix Gröbert, Ivan Fratric, Steve Henson]
3183 *) Harmonize version and its documentation. -f flag is used to display
3185 [mancha <mancha1@zoho.com>]
3187 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3188 in i2d_ECPrivateKey.
3189 [mancha <mancha1@zoho.com>]
3191 *) Fix some double frees. These are not thought to be exploitable.
3192 [mancha <mancha1@zoho.com>]
3194 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3196 *) A missing bounds check in the handling of the TLS heartbeat extension
3197 can be used to reveal up to 64k of memory to a connected client or
3200 Thanks for Neel Mehta of Google Security for discovering this bug and to
3201 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3202 preparing the fix (CVE-2014-0160)
3203 [Adam Langley, Bodo Moeller]
3205 *) Fix for the attack described in the paper "Recovering OpenSSL
3206 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3207 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3208 http://eprint.iacr.org/2014/140
3210 Thanks to Yuval Yarom and Naomi Benger for discovering this
3211 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3212 [Yuval Yarom and Naomi Benger]
3214 *) TLS pad extension: draft-agl-tls-padding-03
3216 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3217 TLS client Hello record length value would otherwise be > 255 and
3218 less that 512 pad with a dummy extension containing zeroes so it
3219 is at least 512 bytes long.
3221 [Adam Langley, Steve Henson]
3223 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3225 *) Fix for TLS record tampering bug. A carefully crafted invalid
3226 handshake could crash OpenSSL with a NULL pointer exception.
3227 Thanks to Anton Johansson for reporting this issues.
3230 *) Keep original DTLS digest and encryption contexts in retransmission
3231 structures so we can use the previous session parameters if they need
3232 to be resent. (CVE-2013-6450)
3235 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3236 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3237 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3238 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3239 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3240 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3241 [Rob Stradling, Adam Langley]
3243 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3245 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3246 supporting platforms or when small records were transferred.
3247 [Andy Polyakov, Steve Henson]
3249 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3251 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3253 This addresses the flaw in CBC record processing discovered by
3254 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3255 at: http://www.isg.rhul.ac.uk/tls/
3257 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3258 Security Group at Royal Holloway, University of London
3259 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3260 Emilia Käsper for the initial patch.
3262 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3264 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3265 ciphersuites which can be exploited in a denial of service attack.
3266 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3267 and detecting this bug and to Wolfgang Ettlinger
3268 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3272 *) Return an error when checking OCSP signatures when key is NULL.
3273 This fixes a DoS attack. (CVE-2013-0166)
3276 *) Make openssl verify return errors.
3277 [Chris Palmer <palmer@google.com> and Ben Laurie]
3279 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3280 the right response is stapled. Also change SSL_get_certificate()
3281 so it returns the certificate actually sent.
3282 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3283 [Rob Stradling <rob.stradling@comodo.com>]
3285 *) Fix possible deadlock when decoding public keys.
3288 *) Don't use TLS 1.0 record version number in initial client hello
3292 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3294 *) Sanity check record length before skipping explicit IV in TLS
3295 1.2, 1.1 and DTLS to fix DoS attack.
3297 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3298 fuzzing as a service testing platform.
3302 *) Initialise tkeylen properly when encrypting CMS messages.
3303 Thanks to Solar Designer of Openwall for reporting this issue.
3306 *) In FIPS mode don't try to use composite ciphers as they are not
3310 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3312 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3313 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3314 mean any application compiled against OpenSSL 1.0.0 headers setting
3315 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3316 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3317 0x10000000L Any application which was previously compiled against
3318 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3319 will need to be recompiled as a result. Letting be results in
3320 inability to disable specifically TLS 1.1 and in client context,
3321 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3324 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3325 disable just protocol X, but all protocols above X *if* there are
3326 protocols *below* X still enabled. In more practical terms it means
3327 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3328 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3329 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3333 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3335 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3336 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3337 in CRYPTO_realloc_clean.
3339 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3340 issue and to Adam Langley <agl@chromium.org> for fixing it.
3342 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3344 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3347 *) Workarounds for some broken servers that "hang" if a client hello
3348 record length exceeds 255 bytes.
3350 1. Do not use record version number > TLS 1.0 in initial client
3351 hello: some (but not all) hanging servers will now work.
3352 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3353 the number of ciphers sent in the client hello. This should be
3354 set to an even number, such as 50, for example by passing:
3355 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3356 Most broken servers should now work.
3357 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3358 TLS 1.2 client support entirely.
3361 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3364 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3366 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3367 STRING form instead of a DigestInfo.
3370 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3371 and the RSA_sign/RSA_verify functions. This was made more apparent when
3372 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3373 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3374 the correct format in RSA_verify so both forms transparently work.
3377 *) Some servers which support TLS 1.0 can choke if we initially indicate
3378 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3379 encrypted premaster secret. As a workaround use the maximum permitted
3380 client version in client hello, this should keep such servers happy
3381 and still work with previous versions of OpenSSL.
3384 *) Add support for TLS/DTLS heartbeats.
3385 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3387 *) Add support for SCTP.
3388 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3390 *) Improved PRNG seeding for VOS.
3391 [Paul Green <Paul.Green@stratus.com>]
3393 *) Extensive assembler packs updates, most notably:
3395 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3396 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3397 - x86_64: bit-sliced AES implementation;
3398 - ARM: NEON support, contemporary platforms optimizations;
3399 - s390x: z196 support;
3400 - *: GHASH and GF(2^m) multiplication implementations;
3404 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3405 (removal of unnecessary code)
3406 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3408 *) Add TLS key material exporter from RFC 5705.
3411 *) Add DTLS-SRTP negotiation from RFC 5764.
3414 *) Add Next Protocol Negotiation,
3415 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3416 disabled with a no-npn flag to config or Configure. Code donated
3418 [Adam Langley <agl@google.com> and Ben Laurie]
3420 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3421 NIST-P256, NIST-P521, with constant-time single point multiplication on
3422 typical inputs. Compiler support for the nonstandard type __uint128_t is
3423 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3424 Code made available under Apache License version 2.0.
3426 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3427 line to include this in your build of OpenSSL, and run "make depend" (or
3428 "make update"). This enables the following EC_METHODs:
3430 EC_GFp_nistp224_method()
3431 EC_GFp_nistp256_method()
3432 EC_GFp_nistp521_method()
3434 EC_GROUP_new_by_curve_name() will automatically use these (while
3435 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3437 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3439 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3440 all platforms. Move ssize_t definition from e_os.h to the public
3441 header file e_os2.h as it now appears in public header file cms.h
3444 *) New -sigopt option to the ca, req and x509 utilities. Additional
3445 signature parameters can be passed using this option and in
3449 *) Add RSA PSS signing function. This will generate and set the
3450 appropriate AlgorithmIdentifiers for PSS based on those in the
3451 corresponding EVP_MD_CTX structure. No application support yet.
3454 *) Support for companion algorithm specific ASN1 signing routines.
3455 New function ASN1_item_sign_ctx() signs a pre-initialised
3456 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3457 the appropriate parameters.
3460 *) Add new algorithm specific ASN1 verification initialisation function
3461 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3462 handling will be the same no matter what EVP_PKEY_METHOD is used.
3463 Add a PSS handler to support verification of PSS signatures: checked
3464 against a number of sample certificates.
3467 *) Add signature printing for PSS. Add PSS OIDs.
3468 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3470 *) Add algorithm specific signature printing. An individual ASN1 method
3471 can now print out signatures instead of the standard hex dump.
3473 More complex signatures (e.g. PSS) can print out more meaningful
3474 information. Include DSA version that prints out the signature
3478 *) Password based recipient info support for CMS library: implementing
3482 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3483 neatly separates the code into cipher and PBE sections and is required
3484 for some algorithms that split PBES2 into separate pieces (such as
3485 password based CMS).
3488 *) Session-handling fixes:
3489 - Fix handling of connections that are resuming with a session ID,
3490 but also support Session Tickets.
3491 - Fix a bug that suppressed issuing of a new ticket if the client
3492 presented a ticket with an expired session.
3493 - Try to set the ticket lifetime hint to something reasonable.
3494 - Make tickets shorter by excluding irrelevant information.
3495 - On the client side, don't ignore renewed tickets.
3496 [Adam Langley, Bodo Moeller (Google)]
3498 *) Fix PSK session representation.
3501 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3503 This work was sponsored by Intel.
3506 *) Add GCM support to TLS library. Some custom code is needed to split
3507 the IV between the fixed (from PRF) and explicit (from TLS record)
3508 portions. This adds all GCM ciphersuites supported by RFC5288 and
3509 RFC5289. Generalise some AES* cipherstrings to include GCM and
3510 add a special AESGCM string for GCM only.
3513 *) Expand range of ctrls for AES GCM. Permit setting invocation
3514 field on decrypt and retrieval of invocation field only on encrypt.
3517 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3518 As required by RFC5289 these ciphersuites cannot be used if for
3519 versions of TLS earlier than 1.2.
3522 *) For FIPS capable OpenSSL interpret a NULL default public key method
3523 as unset and return the appropriate default but do *not* set the default.
3524 This means we can return the appropriate method in applications that
3525 switch between FIPS and non-FIPS modes.
3528 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3529 ENGINE is used then we cannot handle that in the FIPS module so we
3530 keep original code iff non-FIPS operations are allowed.
3533 *) Add -attime option to openssl utilities.
3534 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3536 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3539 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3540 FIPS EC methods unconditionally for now.
3543 *) New build option no-ec2m to disable characteristic 2 code.
3546 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3547 all cases can be covered as some introduce binary incompatibilities.
3550 *) Redirect RSA operations to FIPS module including keygen,
3551 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3554 *) Add similar low level API blocking to ciphers.
3557 *) Low level digest APIs are not approved in FIPS mode: any attempt
3558 to use these will cause a fatal error. Applications that *really* want
3559 to use them can use the private_* version instead.
3562 *) Redirect cipher operations to FIPS module for FIPS builds.
3565 *) Redirect digest operations to FIPS module for FIPS builds.
3568 *) Update build system to add "fips" flag which will link in fipscanister.o
3569 for static and shared library builds embedding a signature if needed.
3572 *) Output TLS supported curves in preference order instead of numerical
3573 order. This is currently hardcoded for the highest order curves first.
3574 This should be configurable so applications can judge speed vs strength.
3577 *) Add TLS v1.2 server support for client authentication.
3580 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3584 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3585 FIPS modules versions.
3588 *) Add TLS v1.2 client side support for client authentication. Keep cache
3589 of handshake records longer as we don't know the hash algorithm to use
3590 until after the certificate request message is received.
3593 *) Initial TLS v1.2 client support. Add a default signature algorithms
3594 extension including all the algorithms we support. Parse new signature
3595 format in client key exchange. Relax some ECC signing restrictions for
3596 TLS v1.2 as indicated in RFC5246.
3599 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3600 to new signature format when needed using client digest preference.
3601 All server ciphersuites should now work correctly in TLS v1.2. No client
3602 support yet and no support for client certificates.
3605 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3606 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3607 ciphersuites. At present only RSA key exchange ciphersuites work with
3608 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3609 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3610 and version checking.
3613 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3614 with this defined it will not be affected by any changes to ssl internal
3615 structures. Add several utility functions to allow openssl application
3616 to work with OPENSSL_NO_SSL_INTERN defined.
3619 *) A long standing patch to add support for SRP from EdelWeb (Peter
3620 Sylvester and Christophe Renou) was integrated.
3621 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3622 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3625 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3628 *) Permit abbreviated handshakes when renegotiating using the function
3629 SSL_renegotiate_abbreviated().
3630 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3632 *) Add call to ENGINE_register_all_complete() to
3633 ENGINE_load_builtin_engines(), so some implementations get used
3634 automatically instead of needing explicit application support.
3637 *) Add support for TLS key exporter as described in RFC5705.
3638 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3640 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3641 a few changes are required:
3643 Add SSL_OP_NO_TLSv1_1 flag.
3644 Add TLSv1_1 methods.
3645 Update version checking logic to handle version 1.1.
3646 Add explicit IV handling (ported from DTLS code).
3647 Add command line options to s_client/s_server.
3650 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3652 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3653 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3654 content decryption and always return the same error. Note: this attack
3655 needs on average 2^20 messages so it only affects automated senders. The
3656 old behaviour can be re-enabled in the CMS code by setting the
3657 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3658 an MMA defence is not necessary.
3659 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3660 this issue. (CVE-2012-0884)
3663 *) Fix CVE-2011-4619: make sure we really are receiving a
3664 client hello before rejecting multiple SGC restarts. Thanks to
3665 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3668 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3670 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3671 Thanks to Antonio Martin, Enterprise Secure Access Research and
3672 Development, Cisco Systems, Inc. for discovering this bug and
3673 preparing a fix. (CVE-2012-0050)
3676 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3678 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3679 of the Vaudenay padding oracle attack on CBC mode encryption
3680 which enables an efficient plaintext recovery attack against
3681 the OpenSSL implementation of DTLS. Their attack exploits timing
3682 differences arising during decryption processing. A research
3683 paper describing this attack can be found at:
3684 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3685 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3686 Security Group at Royal Holloway, University of London
3687 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3688 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3689 for preparing the fix. (CVE-2011-4108)
3690 [Robin Seggelmann, Michael Tuexen]
3692 *) Clear bytes used for block padding of SSL 3.0 records.
3694 [Adam Langley (Google)]
3696 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3697 Kadianakis <desnacked@gmail.com> for discovering this issue and
3698 Adam Langley for preparing the fix. (CVE-2011-4619)
3699 [Adam Langley (Google)]
3701 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3702 [Andrey Kulikov <amdeich@gmail.com>]
3704 *) Prevent malformed RFC3779 data triggering an assertion failure.
3705 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3706 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3707 [Rob Austein <sra@hactrn.net>]
3709 *) Improved PRNG seeding for VOS.
3710 [Paul Green <Paul.Green@stratus.com>]
3712 *) Fix ssl_ciph.c set-up race.
3713 [Adam Langley (Google)]
3715 *) Fix spurious failures in ecdsatest.c.
3716 [Emilia Käsper (Google)]
3718 *) Fix the BIO_f_buffer() implementation (which was mixing different
3719 interpretations of the '..._len' fields).
3720 [Adam Langley (Google)]
3722 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3723 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3724 threads won't reuse the same blinding coefficients.
3726 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3727 lock to call BN_BLINDING_invert_ex, and avoids one use of
3728 BN_BLINDING_update for each BN_BLINDING structure (previously,
3729 the last update always remained unused).
3730 [Emilia Käsper (Google)]
3732 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3733 [Bob Buckholz (Google)]
3735 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3737 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3738 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3739 [Kaspar Brand <ossl@velox.ch>]
3741 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3742 for multi-threaded use of ECDH. (CVE-2011-3210)
3743 [Adam Langley (Google)]
3745 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3748 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3749 signature public key algorithm by using OID xref utilities instead.
3750 Before this you could only use some ECC ciphersuites with SHA1 only.
3753 *) Add protection against ECDSA timing attacks as mentioned in the paper
3754 by Billy Bob Brumley and Nicola Tuveri, see:
3756 http://eprint.iacr.org/2011/232.pdf
3758 [Billy Bob Brumley and Nicola Tuveri]
3760 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3762 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3763 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3765 *) Fix bug in string printing code: if *any* escaping is enabled we must
3766 escape the escape character (backslash) or the resulting string is
3770 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3772 *) Disable code workaround for ancient and obsolete Netscape browsers
3773 and servers: an attacker can use it in a ciphersuite downgrade attack.
3774 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3777 *) Fixed J-PAKE implementation error, originally discovered by
3778 Sebastien Martini, further info and confirmation from Stefan
3779 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3782 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3784 *) Fix extension code to avoid race conditions which can result in a buffer
3785 overrun vulnerability: resumed sessions must not be modified as they can
3786 be shared by multiple threads. CVE-2010-3864
3789 *) Fix WIN32 build system to correctly link an ENGINE directory into
3793 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3795 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3797 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3799 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3801 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3802 context. The operation can be customised via the ctrl mechanism in
3803 case ENGINEs want to include additional functionality.
3806 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3809 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3810 output hashes compatible with older versions of OpenSSL.
3811 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3813 *) Fix compression algorithm handling: if resuming a session use the
3814 compression algorithm of the resumed session instead of determining
3815 it from client hello again. Don't allow server to change algorithm.
3818 *) Add load_crls() function to apps tidying load_certs() too. Add option
3819 to verify utility to allow additional CRLs to be included.
3822 *) Update OCSP request code to permit adding custom headers to the request:
3823 some responders need this.
3826 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3828 [Julia Lawall <julia@diku.dk>]
3830 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3831 needlessly dereferenced structures, used obsolete functions and
3832 didn't handle all updated verify codes correctly.
3835 *) Disable MD2 in the default configuration.
3838 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3839 indicate the initial BIO being pushed or popped. This makes it possible
3840 to determine whether the BIO is the one explicitly called or as a result
3841 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3842 it handles reference counts correctly and doesn't zero out the I/O bio
3843 when it is not being explicitly popped. WARNING: applications which
3844 included workarounds for the old buggy behaviour will need to be modified
3845 or they could free up already freed BIOs.
3848 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3849 renaming to all platforms (within the 0.9.8 branch, this was
3850 done conditionally on Netware platforms to avoid a name clash).
3851 [Guenter <lists@gknw.net>]
3853 *) Add ECDHE and PSK support to DTLS.
3854 [Michael Tuexen <tuexen@fh-muenster.de>]
3856 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3860 *) Add "missing" function EVP_MD_flags() (without this the only way to
3861 retrieve a digest flags is by accessing the structure directly. Update
3862 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3863 or cipher is registered as in the "from" argument. Print out all
3864 registered digests in the dgst usage message instead of manually
3865 attempting to work them out.
3868 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3869 this allows the use of compression and extensions. Change default cipher
3870 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3871 by default unless an application cipher string requests it.
3874 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3875 key ids to find matching certificates and keys but some PKCS#12 files
3876 don't follow the (somewhat unwritten) rules and this strategy fails.
3877 Now just gather all certificates together and the first private key
3878 then look for the first certificate that matches the key.
3881 *) Support use of registered digest and cipher names for dgst and cipher
3882 commands instead of having to add each one as a special case. So now
3889 openssl dgst -sha256 foo
3891 and this works for ENGINE based algorithms too.
3895 *) Update Gost ENGINE to support parameter files.
3896 [Victor B. Wagner <vitus@cryptocom.ru>]
3898 *) Support GeneralizedTime in ca utility.
3899 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3901 *) Enhance the hash format used for certificate directory links. The new
3902 form uses the canonical encoding (meaning equivalent names will work
3903 even if they aren't identical) and uses SHA1 instead of MD5. This form
3904 is incompatible with the older format and as a result c_rehash should
3905 be used to rebuild symbolic links.
3908 *) Make PKCS#8 the default write format for private keys, replacing the
3909 traditional format. This form is standardised, more secure and doesn't
3910 include an implicit MD5 dependency.
3913 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3914 committed to OpenSSL should pass this lot as a minimum.
3917 *) Add session ticket override functionality for use by EAP-FAST.
3918 [Jouni Malinen <j@w1.fi>]
3920 *) Modify HMAC functions to return a value. Since these can be implemented
3921 in an ENGINE errors can occur.
3924 *) Type-checked OBJ_bsearch_ex.
3927 *) Type-checked OBJ_bsearch. Also some constification necessitated
3928 by type-checking. Still to come: TXT_DB, bsearch(?),
3929 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3933 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3934 seconds to a tm structure directly, instead of going through OS
3935 specific date routines. This avoids any issues with OS routines such
3936 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3937 and X509_time_adj_ex() to cover the extended range. The existing
3938 X509_time_adj() is still usable and will no longer have any date issues.
3941 *) Delta CRL support. New use deltas option which will attempt to locate
3942 and search any appropriate delta CRLs available.
3944 This work was sponsored by Google.
3947 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3948 code and add additional score elements. Validate alternate CRL paths
3949 as part of the CRL checking and indicate a new error "CRL path validation
3950 error" in this case. Applications wanting additional details can use
3951 the verify callback and check the new "parent" field. If this is not
3952 NULL CRL path validation is taking place. Existing applications won't
3953 see this because it requires extended CRL support which is off by
3956 This work was sponsored by Google.
3959 *) Support for freshest CRL extension.
3961 This work was sponsored by Google.
3964 *) Initial indirect CRL support. Currently only supported in the CRLs
3965 passed directly and not via lookup. Process certificate issuer
3966 CRL entry extension and lookup CRL entries by bother issuer name
3967 and serial number. Check and process CRL issuer entry in IDP extension.
3969 This work was sponsored by Google.
3972 *) Add support for distinct certificate and CRL paths. The CRL issuer
3973 certificate is validated separately in this case. Only enabled if
3974 an extended CRL support flag is set: this flag will enable additional
3975 CRL functionality in future.
3977 This work was sponsored by Google.
3980 *) Add support for policy mappings extension.
3982 This work was sponsored by Google.
3985 *) Fixes to pathlength constraint, self issued certificate handling,
3986 policy processing to align with RFC3280 and PKITS tests.
3988 This work was sponsored by Google.
3991 *) Support for name constraints certificate extension. DN, email, DNS
3992 and URI types are currently supported.
3994 This work was sponsored by Google.
3997 *) To cater for systems that provide a pointer-based thread ID rather
3998 than numeric, deprecate the current numeric thread ID mechanism and
3999 replace it with a structure and associated callback type. This
4000 mechanism allows a numeric "hash" to be extracted from a thread ID in
4001 either case, and on platforms where pointers are larger than 'long',
4002 mixing is done to help ensure the numeric 'hash' is usable even if it
4003 can't be guaranteed unique. The default mechanism is to use "&errno"
4004 as a pointer-based thread ID to distinguish between threads.
4006 Applications that want to provide their own thread IDs should now use
4007 CRYPTO_THREADID_set_callback() to register a callback that will call
4008 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4010 Note that ERR_remove_state() is now deprecated, because it is tied
4011 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4012 to free the current thread's error state should be replaced by
4013 ERR_remove_thread_state(NULL).
4015 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4016 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4017 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4018 application was previously providing a numeric thread callback that
4019 was inappropriate for distinguishing threads, then uniqueness might
4020 have been obtained with &errno that happened immediately in the
4021 intermediate development versions of OpenSSL; this is no longer the
4022 case, the numeric thread callback will now override the automatic use
4024 [Geoff Thorpe, with help from Bodo Moeller]
4026 *) Initial support for different CRL issuing certificates. This covers a
4027 simple case where the self issued certificates in the chain exist and
4028 the real CRL issuer is higher in the existing chain.
4030 This work was sponsored by Google.
4033 *) Removed effectively defunct crypto/store from the build.
4036 *) Revamp of STACK to provide stronger type-checking. Still to come:
4037 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4038 ASN1_STRING, CONF_VALUE.
4041 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4042 RAM on SSL connections. This option can save about 34k per idle SSL.
4045 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4046 STACK, TXT_DB, bsearch, qsort.
4049 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4050 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4051 support for data, signedData, compressedData, digestedData and
4052 encryptedData, envelopedData types included. Scripts to check against
4053 RFC4134 examples draft and interop and consistency checks of many
4054 content types and variants.
4057 *) Add options to enc utility to support use of zlib compression BIO.
4060 *) Extend mk1mf to support importing of options and assembly language
4061 files from Configure script, currently only included in VC-WIN32.
4062 The assembly language rules can now optionally generate the source
4063 files from the associated perl scripts.
4066 *) Implement remaining functionality needed to support GOST ciphersuites.
4067 Interop testing has been performed using CryptoPro implementations.
4068 [Victor B. Wagner <vitus@cryptocom.ru>]
4070 *) s390x assembler pack.
4073 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4077 *) Implement Opaque PRF Input TLS extension as specified in
4078 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4079 official specification yet and no extension type assignment by
4080 IANA exists, this extension (for now) will have to be explicitly
4081 enabled when building OpenSSL by providing the extension number
4082 to use. For example, specify an option
4084 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4086 to the "config" or "Configure" script to enable the extension,
4087 assuming extension number 0x9527 (which is a completely arbitrary
4088 and unofficial assignment based on the MD5 hash of the Internet
4089 Draft). Note that by doing so, you potentially lose
4090 interoperability with other TLS implementations since these might
4091 be using the same extension number for other purposes.
4093 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4094 opaque PRF input value to use in the handshake. This will create
4095 an internal copy of the length-'len' string at 'src', and will
4096 return non-zero for success.
4098 To get more control and flexibility, provide a callback function
4101 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4102 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4106 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4109 Callback function 'cb' will be called in handshakes, and is
4110 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4111 Argument 'arg' is for application purposes (the value as given to
4112 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4113 be provided to the callback function). The callback function
4114 has to return non-zero to report success: usually 1 to use opaque
4115 PRF input just if possible, or 2 to enforce use of the opaque PRF
4116 input. In the latter case, the library will abort the handshake
4117 if opaque PRF input is not successfully negotiated.
4119 Arguments 'peerinput' and 'len' given to the callback function
4120 will always be NULL and 0 in the case of a client. A server will
4121 see the client's opaque PRF input through these variables if
4122 available (NULL and 0 otherwise). Note that if the server
4123 provides an opaque PRF input, the length must be the same as the
4124 length of the client's opaque PRF input.
4126 Note that the callback function will only be called when creating
4127 a new session (session resumption can resume whatever was
4128 previously negotiated), and will not be called in SSL 2.0
4129 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4130 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4131 for applications that need to enforce opaque PRF input.
4135 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4138 [Victor B. Wagner <vitus@cryptocom.ru>]
4140 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4141 RFC4507bis. The encrypted ticket format is an encrypted encoded
4142 SSL_SESSION structure, that way new session features are automatically
4145 If a client application caches session in an SSL_SESSION structure
4146 support is transparent because tickets are now stored in the encoded
4149 The SSL_CTX structure automatically generates keys for ticket
4150 protection in servers so again support should be possible
4151 with no application modification.
4153 If a client or server wishes to disable RFC4507 support then the option
4154 SSL_OP_NO_TICKET can be set.
4156 Add a TLS extension debugging callback to allow the contents of any client
4157 or server extensions to be examined.
4159 This work was sponsored by Google.
4162 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4163 OpenSSL should now compile cleanly on gcc 4.2
4164 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4166 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4167 support including streaming MAC support: this is required for GOST
4168 ciphersuite support.
4169 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4171 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4172 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4173 to output in BER and PEM format.
4176 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4177 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4178 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4179 ENGINE support for HMAC keys which are unextractable. New -mac and
4180 -macopt options to dgst utility.
4183 *) New option -sigopt to dgst utility. Update dgst to use
4184 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4185 alternative signing parameters such as X9.31 or PSS in the dgst
4189 *) Change ssl_cipher_apply_rule(), the internal function that does
4190 the work each time a ciphersuite string requests enabling
4191 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4192 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4193 the order of disabled ciphersuites such that those ciphersuites
4194 that most recently went from enabled to disabled not only stay
4195 in order with respect to each other, but also have higher priority
4196 than other disabled ciphersuites the next time ciphersuites are
4199 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4200 the same ciphersuites as with "HIGH" alone, but in a specific
4201 order where the PSK ciphersuites come first (since they are the
4202 most recently disabled ciphersuites when "HIGH" is parsed).
4204 Also, change ssl_create_cipher_list() (using this new
4205 functionality) such that between otherwise identical
4206 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4210 *) Change ssl_create_cipher_list() so that it automatically
4211 arranges the ciphersuites in reasonable order before starting
4212 to process the rule string. Thus, the definition for "DEFAULT"
4213 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4214 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4215 This makes it much easier to arrive at a reasonable default order
4216 in applications for which anonymous ciphers are OK (meaning
4217 that you can't actually use DEFAULT).
4218 [Bodo Moeller; suggested by Victor Duchovni]
4220 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4221 processing) into multiple integers instead of setting
4222 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4223 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4224 (These masks as well as the individual bit definitions are hidden
4225 away into the non-exported interface ssl/ssl_locl.h, so this
4226 change to the definition of the SSL_CIPHER structure shouldn't
4227 affect applications.) This give us more bits for each of these
4228 categories, so there is no longer a need to coagulate AES128 and
4229 AES256 into a single algorithm bit, and to coagulate Camellia128
4230 and Camellia256 into a single algorithm bit, which has led to all
4233 Thus, among other things, the kludge introduced in 0.9.7m and
4234 0.9.8e for masking out AES256 independently of AES128 or masking
4235 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4237 With the change, we also introduce new ciphersuite aliases that
4238 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4242 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4243 Use the leftmost N bytes of the signature input if the input is
4244 larger than the prime q (with N being the size in bytes of q).
4247 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4248 it yet and it is largely untested.
4251 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4254 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4255 some compilers (gcc 4.2 and later) reject their use. Safestack is
4256 reimplemented. Update ASN1 to avoid use of legacy functions.
4259 *) Win32/64 targets are linked with Winsock2.
4262 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4263 to external functions. This can be used to increase CRL handling
4264 efficiency especially when CRLs are very large by (for example) storing
4265 the CRL revoked certificates in a database.
4268 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4269 new CRLs added to a directory can be used. New command line option
4270 -verify_return_error to s_client and s_server. This causes real errors
4271 to be returned by the verify callback instead of carrying on no matter
4272 what. This reflects the way a "real world" verify callback would behave.
4275 *) GOST engine, supporting several GOST algorithms and public key formats.
4276 Kindly donated by Cryptocom.
4279 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4280 partitioned by DP are handled but no indirect CRL or reason partitioning
4281 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4282 selected via a scoring technique which handles IDP and AKID in CRLs.
4285 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4286 will ultimately be used for all verify operations: this will remove the
4287 X509_STORE dependency on certificate verification and allow alternative
4288 lookup methods. X509_STORE based implementations of these two callbacks.
4291 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4292 Modify get_crl() to find a valid (unexpired) CRL if possible.
4295 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4296 this would be called X509_CRL_cmp() but that name is already used by
4297 a function that just compares CRL issuer names. Cache several CRL
4298 extensions in X509_CRL structure and cache CRLDP in X509.
4301 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4302 this maps equivalent X509_NAME structures into a consistent structure.
4303 Name comparison can then be performed rapidly using memcmp().
4306 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4310 *) Allow digests to supply their own micalg string for S/MIME type using
4311 the ctrl EVP_MD_CTRL_MICALG.
4314 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4315 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4316 ctrl. It can then customise the structure before and/or after signing
4320 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4321 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4322 to free up any added signature OIDs.
4325 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4326 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4327 digest and cipher tables. New options added to openssl utility:
4328 list-message-digest-algorithms and list-cipher-algorithms.
4331 *) Change the array representation of binary polynomials: the list
4332 of degrees of non-zero coefficients is now terminated with -1.
4333 Previously it was terminated with 0, which was also part of the
4334 value; thus, the array representation was not applicable to
4335 polynomials where t^0 has coefficient zero. This change makes
4336 the array representation useful in a more general context.
4339 *) Various modifications and fixes to SSL/TLS cipher string
4340 handling. For ECC, the code now distinguishes between fixed ECDH
4341 with RSA certificates on the one hand and with ECDSA certificates
4342 on the other hand, since these are separate ciphersuites. The
4343 unused code for Fortezza ciphersuites has been removed.
4345 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4346 (not "ECDHE"). For consistency with the code for DH
4347 certificates, use of ECDH certificates is now considered ECDH
4348 authentication, not RSA or ECDSA authentication (the latter is
4349 merely the CA's signing algorithm and not actively used in the
4352 The temporary ciphersuite alias "ECCdraft" is no longer
4353 available, and ECC ciphersuites are no longer excluded from "ALL"
4354 and "DEFAULT". The following aliases now exist for RFC 4492
4355 ciphersuites, most of these by analogy with the DH case:
4357 kECDHr - ECDH cert, signed with RSA
4358 kECDHe - ECDH cert, signed with ECDSA
4359 kECDH - ECDH cert (signed with either RSA or ECDSA)
4360 kEECDH - ephemeral ECDH
4361 ECDH - ECDH cert or ephemeral ECDH
4367 AECDH - anonymous ECDH
4368 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4372 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4373 Use correct micalg parameters depending on digest(s) in signed message.
4376 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4377 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4380 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4381 an engine to register a method. Add ENGINE lookups for methods and
4382 functional reference processing.
4385 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4386 EVP_{Sign,Verify}* which allow an application to customise the signature
4390 *) New -resign option to smime utility. This adds one or more signers
4391 to an existing PKCS#7 signedData structure. Also -md option to use an
4392 alternative message digest algorithm for signing.
4395 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4396 create PKCS7 structures containing multiple signers. Update smime
4397 application to support multiple signers.
4400 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4404 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4405 Reorganize PBE internals to lookup from a static table using NIDs,
4406 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4407 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4408 PRF which will be automatically used with PBES2.
4411 *) Replace the algorithm specific calls to generate keys in "req" with the
4415 *) Update PKCS#7 enveloped data routines to use new API. This is now
4416 supported by any public key method supporting the encrypt operation. A
4417 ctrl is added to allow the public key algorithm to examine or modify
4418 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4422 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4423 a default digest type to use. In most cases this will be SHA1 but some
4424 algorithms (such as GOST) need to specify an alternative digest. The
4425 return value indicates how strong the preference is 1 means optional and
4426 2 is mandatory (that is it is the only supported type). Modify
4427 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4428 use the default md. Update openssl utilities to use the default digest
4429 type for signing if it is not explicitly indicated.
4432 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4433 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4434 signing method from the key type. This effectively removes the link
4435 between digests and public key types.
4438 *) Add an OID cross reference table and utility functions. Its purpose is to
4439 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4440 rsaEncryption. This will allow some of the algorithm specific hackery
4441 needed to use the correct OID to be removed.
4444 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4445 structures for PKCS7_sign(). They are now set up by the relevant public
4449 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4452 *) Add support for key derivation (agreement) in the API, DH method and
4456 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4457 public and private key formats. As a side effect these add additional
4458 command line functionality not previously available: DSA signatures can be
4459 generated and verified using pkeyutl and DH key support and generation in
4464 [Oliver Tappe <zooey@hirschkaefer.de>]
4466 *) New make target "install_html_docs" installs HTML renditions of the
4468 [Oliver Tappe <zooey@hirschkaefer.de>]
4470 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4471 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4472 support key and parameter generation and add initial key generation
4473 functionality for RSA.
4476 *) Add functions for main EVP_PKEY_method operations. The undocumented
4477 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4478 EVP_PKEY_{encrypt,decrypt}_old.
4481 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4482 key API, doesn't do much yet.
4485 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4486 public key algorithms. New option to openssl utility:
4487 "list-public-key-algorithms" to print out info.
4490 *) Implement the Supported Elliptic Curves Extension for
4491 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4494 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4495 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4498 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4499 utilities such as rsa, dsa, dsaparam etc except they process any key
4503 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4504 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4505 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4509 *) Initial support for pluggable public key ASN1.
4510 De-spaghettify the public key ASN1 handling. Move public and private
4511 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4512 algorithm specific handling to a single module within the relevant
4513 algorithm directory. Add functions to allow (near) opaque processing
4514 of public and private key structures.
4517 *) Implement the Supported Point Formats Extension for
4518 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4521 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4522 for the psk identity [hint] and the psk callback functions to the
4523 SSL_SESSION, SSL and SSL_CTX structure.
4526 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4530 SSL_CTX_use_psk_identity_hint
4531 SSL_get_psk_identity_hint
4532 SSL_get_psk_identity
4533 SSL_use_psk_identity_hint
4535 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4537 *) Add RFC 3161 compliant time stamp request creation, response generation
4538 and response verification functionality.
4539 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4541 *) Add initial support for TLS extensions, specifically for the server_name
4542 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4543 have new members for a host name. The SSL data structure has an
4544 additional member SSL_CTX *initial_ctx so that new sessions can be
4545 stored in that context to allow for session resumption, even after the
4546 SSL has been switched to a new SSL_CTX in reaction to a client's
4547 server_name extension.
4549 New functions (subject to change):
4551 SSL_get_servername()
4552 SSL_get_servername_type()
4555 New CTRL codes and macros (subject to change):
4557 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4558 - SSL_CTX_set_tlsext_servername_callback()
4559 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4560 - SSL_CTX_set_tlsext_servername_arg()
4561 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4563 openssl s_client has a new '-servername ...' option.
4565 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4566 '-key2 ...', '-servername_fatal' (subject to change). This allows
4567 testing the HostName extension for a specific single host name ('-cert'
4568 and '-key' remain fallbacks for handshakes without HostName
4569 negotiation). If the unrecognized_name alert has to be sent, this by
4570 default is a warning; it becomes fatal with the '-servername_fatal'
4573 [Peter Sylvester, Remy Allais, Christophe Renou]
4575 *) Whirlpool hash implementation is added.
4578 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4579 bn(64,32). Because of instruction set limitations it doesn't have
4580 any negative impact on performance. This was done mostly in order
4581 to make it possible to share assembler modules, such as bn_mul_mont
4582 implementations, between 32- and 64-bit builds without hassle.
4585 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4586 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4590 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4591 dedicated Montgomery multiplication procedure, is introduced.
4592 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4593 "64-bit" performance on certain 32-bit targets.
4596 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4597 in SSL structures. New SSL ctrl to set maximum send fragment size.
4598 Save memory by setting the I/O buffer sizes dynamically instead of
4599 using the maximum available value.
4602 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4603 in addition to the text details.
4606 *) Very, very preliminary EXPERIMENTAL support for printing of general
4607 ASN1 structures. This currently produces rather ugly output and doesn't
4608 handle several customised structures at all.
4611 *) Integrated support for PVK file format and some related formats such
4612 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4613 these in the 'rsa' and 'dsa' utilities.
4616 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4619 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4620 place for the (very old) "NETSCAPE" format certificates which are now
4621 handled using new ASN1 code equivalents.
4624 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4625 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4626 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4629 *) Modify CRL distribution points extension code to print out previously
4630 unsupported fields. Enhance extension setting code to allow setting of
4634 *) Add print and set support for Issuing Distribution Point CRL extension.
4637 *) Change 'Configure' script to enable Camellia by default.
4640 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4642 *) When rejecting SSL/TLS records due to an incorrect version number, never
4643 update s->server with a new major version number. As of
4644 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4645 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4646 the previous behavior could result in a read attempt at NULL when
4647 receiving specific incorrect SSL/TLS records once record payload
4648 protection is active. (CVE-2010-0740)
4649 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4651 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4652 could be crashed if the relevant tables were not present (e.g. chrooted).
4653 [Tomas Hoger <thoger@redhat.com>]
4655 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4657 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4658 [Martin Olsson, Neel Mehta]
4660 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4661 accommodate for stack sorting, always a write lock!).
4664 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4665 excessive delays in the RAND_poll(): over a minute. As a workaround
4666 include a time check in the inner Heap32Next loop too.
4669 *) The code that handled flushing of data in SSL/TLS originally used the
4670 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4671 the problem outlined in PR#1949. The fix suggested there however can
4672 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4673 of Apache). So instead simplify the code to flush unconditionally.
4674 This should be fine since flushing with no data to flush is a no op.
4677 *) Handle TLS versions 2.0 and later properly and correctly use the
4678 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4679 off ancient servers have a habit of sticking around for a while...
4682 *) Modify compression code so it frees up structures without using the
4683 ex_data callbacks. This works around a problem where some applications
4684 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4685 restarting) then use compression (e.g. SSL with compression) later.
4686 This results in significant per-connection memory leaks and
4687 has caused some security issues including CVE-2008-1678 and
4691 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4692 change when encrypting or decrypting.
4695 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4696 connect and renegotiate with servers which do not support RI.
4697 Until RI is more widely deployed this option is enabled by default.
4700 *) Add "missing" ssl ctrls to clear options and mode.
4703 *) If client attempts to renegotiate and doesn't support RI respond with
4704 a no_renegotiation alert as required by RFC5746. Some renegotiating
4705 TLS clients will continue a connection gracefully when they receive
4706 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4707 waiting for a server hello which it will never receive. Now we treat a
4708 received no_renegotiation alert as a fatal error. This is because
4709 applications requesting a renegotiation might well expect it to succeed
4710 and would have no code in place to handle the server denying it so the
4711 only safe thing to do is to terminate the connection.
4714 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4715 peer supports secure renegotiation and 0 otherwise. Print out peer
4716 renegotiation support in s_client/s_server.
4719 *) Replace the highly broken and deprecated SPKAC certification method with
4720 the updated NID creation version. This should correctly handle UTF8.
4723 *) Implement RFC5746. Re-enable renegotiation but require the extension
4724 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4725 turns out to be a bad idea. It has been replaced by
4726 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4727 SSL_CTX_set_options(). This is really not recommended unless you
4728 know what you are doing.
4729 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4731 *) Fixes to stateless session resumption handling. Use initial_ctx when
4732 issuing and attempting to decrypt tickets in case it has changed during
4733 servername handling. Use a non-zero length session ID when attempting
4734 stateless session resumption: this makes it possible to determine if
4735 a resumption has occurred immediately after receiving server hello
4736 (several places in OpenSSL subtly assume this) instead of later in
4740 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4741 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4742 fixes for a few places where the return code is not checked
4744 [Julia Lawall <julia@diku.dk>]
4746 *) Add --strict-warnings option to Configure script to include devteam
4747 warnings in other configurations.
4750 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4751 makes it possible to install openssl libraries in locations which
4752 have names other than "lib", for example "/usr/lib64" which some
4754 [Steve Henson, based on patch from Jeremy Utley]
4756 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4757 X690 8.9.12 and can produce some misleading textual output of OIDs.
4758 [Steve Henson, reported by Dan Kaminsky]
4760 *) Delete MD2 from algorithm tables. This follows the recommendation in
4761 several standards that it is not used in new applications due to
4762 several cryptographic weaknesses. For binary compatibility reasons
4763 the MD2 API is still compiled in by default.
4766 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4770 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4771 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4773 [Guenter <lists@gknw.net>]
4775 *) Fix the server certificate chain building code to use X509_verify_cert(),
4776 it used to have an ad-hoc builder which was unable to cope with anything
4777 other than a simple chain.
4778 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4780 *) Don't check self signed certificate signatures in X509_verify_cert()
4781 by default (a flag can override this): it just wastes time without
4782 adding any security. As a useful side effect self signed root CAs
4783 with non-FIPS digests are now usable in FIPS mode.
4786 *) In dtls1_process_out_of_seq_message() the check if the current message
4787 is already buffered was missing. For every new message was memory
4788 allocated, allowing an attacker to perform an denial of service attack
4789 with sending out of seq handshake messages until there is no memory
4790 left. Additionally every future message was buffered, even if the
4791 sequence number made no sense and would be part of another handshake.
4792 So only messages with sequence numbers less than 10 in advance will be
4793 buffered. (CVE-2009-1378)
4794 [Robin Seggelmann, discovered by Daniel Mentz]
4796 *) Records are buffered if they arrive with a future epoch to be
4797 processed after finishing the corresponding handshake. There is
4798 currently no limitation to this buffer allowing an attacker to perform
4799 a DOS attack with sending records with future epochs until there is no
4800 memory left. This patch adds the pqueue_size() function to determine
4801 the size of a buffer and limits the record buffer to 100 entries.
4803 [Robin Seggelmann, discovered by Daniel Mentz]
4805 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4806 parent structure is freed. (CVE-2009-1379)
4809 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4810 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4813 [Ilya O. <vrghost@gmail.com>]
4815 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4817 *) Disable renegotiation completely - this fixes a severe security
4818 problem (CVE-2009-3555) at the cost of breaking all
4819 renegotiation. Renegotiation can be re-enabled by setting
4820 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4821 run-time. This is really not recommended unless you know what
4825 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4827 *) Don't set val to NULL when freeing up structures, it is freed up by
4828 underlying code. If sizeof(void *) > sizeof(long) this can result in
4829 zeroing past the valid field. (CVE-2009-0789)
4830 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4832 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4833 checked correctly. This would allow some invalid signed attributes to
4834 appear to verify correctly. (CVE-2009-0591)
4835 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4837 *) Reject UniversalString and BMPString types with invalid lengths. This
4838 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4839 a legal length. (CVE-2009-0590)
4842 *) Set S/MIME signing as the default purpose rather than setting it
4843 unconditionally. This allows applications to override it at the store
4847 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4848 to handle some structures.
4851 *) Improve efficiency of mem_gets: don't search whole buffer each time
4853 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4855 *) New -hex option for openssl rand.
4858 *) Print out UTF8String and NumericString when parsing ASN1.
4861 *) Support NumericString type for name components.
4864 *) Allow CC in the environment to override the automatically chosen
4865 compiler. Note that nothing is done to ensure flags work with the
4869 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4871 *) Properly check EVP_VerifyFinal() and similar return values
4873 [Ben Laurie, Bodo Moeller, Google Security Team]
4875 *) Enable TLS extensions by default.
4878 *) Allow the CHIL engine to be loaded, whether the application is
4879 multithreaded or not. (This does not release the developer from the
4880 obligation to set up the dynamic locking callbacks.)
4881 [Sander Temme <sander@temme.net>]
4883 *) Use correct exit code if there is an error in dgst command.
4884 [Steve Henson; problem pointed out by Roland Dirlewanger]
4886 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4887 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4890 *) Add experimental JPAKE support, including demo authentication in
4891 s_client and s_server.
4894 *) Set the comparison function in v3_addr_canonize().
4895 [Rob Austein <sra@hactrn.net>]
4897 *) Add support for XMPP STARTTLS in s_client.
4898 [Philip Paeps <philip@freebsd.org>]
4900 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4901 to ensure that even with this option, only ciphersuites in the
4902 server's preference list will be accepted. (Note that the option
4903 applies only when resuming a session, so the earlier behavior was
4904 just about the algorithm choice for symmetric cryptography.)
4907 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4909 *) Fix NULL pointer dereference if a DTLS server received
4910 ChangeCipherSpec as first record (CVE-2009-1386).
4913 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4914 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4917 *) The fix in 0.9.8c that supposedly got rid of unsafe
4918 double-checked locking was incomplete for RSA blinding,
4919 addressing just one layer of what turns out to have been
4920 doubly unsafe triple-checked locking.
4922 So now fix this for real by retiring the MONT_HELPER macro
4923 in crypto/rsa/rsa_eay.c.
4925 [Bodo Moeller; problem pointed out by Marius Schilder]
4927 *) Various precautionary measures:
4929 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4931 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4932 (NB: This would require knowledge of the secret session ticket key
4933 to exploit, in which case you'd be SOL either way.)
4935 - Change bn_nist.c so that it will properly handle input BIGNUMs
4936 outside the expected range.
4938 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4941 [Neel Mehta, Bodo Moeller]
4943 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4944 the load fails. Useful for distros.
4945 [Ben Laurie and the FreeBSD team]
4947 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4950 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4953 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4955 This work was sponsored by Logica.
4958 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4959 keystores. Support for SSL/TLS client authentication too.
4960 Not compiled unless enable-capieng specified to Configure.
4962 This work was sponsored by Logica.
4965 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4966 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4967 attribute creation routines such as certificate requests and PKCS#12
4971 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4973 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4974 handshake which could lead to a client crash as found using the
4975 Codenomicon TLS test suite (CVE-2008-1672)
4976 [Steve Henson, Mark Cox]
4978 *) Fix double free in TLS server name extensions which could lead to
4979 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4982 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4984 Clear the error queue to ensure that error entries left from
4985 older function calls do not interfere with the correct operation.
4986 [Lutz Jaenicke, Erik de Castro Lopo]
4988 *) Remove root CA certificates of commercial CAs:
4990 The OpenSSL project does not recommend any specific CA and does not
4991 have any policy with respect to including or excluding any CA.
4992 Therefore it does not make any sense to ship an arbitrary selection
4993 of root CA certificates with the OpenSSL software.
4996 *) RSA OAEP patches to fix two separate invalid memory reads.
4997 The first one involves inputs when 'lzero' is greater than
4998 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4999 before the beginning of from). The second one involves inputs where
5000 the 'db' section contains nothing but zeroes (there is a one-byte
5001 invalid read after the end of 'db').
5002 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5004 *) Partial backport from 0.9.9-dev:
5006 Introduce bn_mul_mont (dedicated Montgomery multiplication
5007 procedure) as a candidate for BIGNUM assembler implementation.
5008 While 0.9.9-dev uses assembler for various architectures, only
5009 x86_64 is available by default here in the 0.9.8 branch, and
5010 32-bit x86 is available through a compile-time setting.
5012 To try the 32-bit x86 assembler implementation, use Configure
5013 option "enable-montasm" (which exists only for this backport).
5015 As "enable-montasm" for 32-bit x86 disclaims code stability
5016 anyway, in this constellation we activate additional code
5017 backported from 0.9.9-dev for further performance improvements,
5018 namely BN_from_montgomery_word. (To enable this otherwise,
5019 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5021 [Andy Polyakov (backport partially by Bodo Moeller)]
5023 *) Add TLS session ticket callback. This allows an application to set
5024 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5025 values. This is useful for key rollover for example where several key
5026 sets may exist with different names.
5029 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5030 This was broken until now in 0.9.8 releases, such that the only way
5031 a registered ENGINE could be used (assuming it initialises
5032 successfully on the host) was to explicitly set it as the default
5033 for the relevant algorithms. This is in contradiction with 0.9.7
5034 behaviour and the documentation. With this fix, when an ENGINE is
5035 registered into a given algorithm's table of implementations, the
5036 'uptodate' flag is reset so that auto-discovery will be used next
5037 time a new context for that algorithm attempts to select an
5039 [Ian Lister (tweaked by Geoff Thorpe)]
5041 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5042 implementation in the following ways:
5044 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5047 Lack of BER streaming support means one pass streaming processing is
5048 only supported if data is detached: setting the streaming flag is
5049 ignored for embedded content.
5051 CMS support is disabled by default and must be explicitly enabled
5052 with the enable-cms configuration option.
5055 *) Update the GMP engine glue to do direct copies between BIGNUM and
5056 mpz_t when openssl and GMP use the same limb size. Otherwise the
5057 existing "conversion via a text string export" trick is still used.
5058 [Paul Sheer <paulsheer@gmail.com>]
5060 *) Zlib compression BIO. This is a filter BIO which compressed and
5061 uncompresses any data passed through it.
5064 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5065 RFC3394 compatible AES key wrapping.
5068 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5069 sets string data without copying. X509_ALGOR_set0() and
5070 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5071 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5072 from an X509_ATTRIBUTE structure optionally checking it occurs only
5073 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5077 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5078 to get the expected BN_FLG_CONSTTIME behavior.
5079 [Bodo Moeller (Google)]
5083 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5084 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5085 - added some more tests to do_tests.pl
5086 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5087 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5088 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5089 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5090 - various changes to netware.pl to enable gcc-cross builds on Win32
5092 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5093 - various changes to fix missing prototype warnings
5094 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5095 - added AES, WHIRLPOOL and CPUID assembler code to build files
5096 - added missing AES assembler make rules to mk1mf.pl
5097 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5098 [Guenter Knauf <eflash@gmx.net>]
5100 *) Implement certificate status request TLS extension defined in RFC3546.
5101 A client can set the appropriate parameters and receive the encoded
5102 OCSP response via a callback. A server can query the supplied parameters
5103 and set the encoded OCSP response in the callback. Add simplified examples
5104 to s_client and s_server.
5107 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5109 *) Fix various bugs:
5110 + Binary incompatibility of ssl_ctx_st structure
5111 + DTLS interoperation with non-compliant servers
5112 + Don't call get_session_cb() without proposed session
5113 + Fix ia64 assembler code
5114 [Andy Polyakov, Steve Henson]
5116 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5118 *) DTLS Handshake overhaul. There were longstanding issues with
5119 OpenSSL DTLS implementation, which were making it impossible for
5120 RFC 4347 compliant client to communicate with OpenSSL server.
5121 Unfortunately just fixing these incompatibilities would "cut off"
5122 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5123 server keeps tolerating non RFC compliant syntax. The opposite is
5124 not true, 0.9.8f client can not communicate with earlier server.
5125 This update even addresses CVE-2007-4995.
5128 *) Changes to avoid need for function casts in OpenSSL: some compilers
5129 (gcc 4.2 and later) reject their use.
5130 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5133 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5134 RFC4507bis. The encrypted ticket format is an encrypted encoded
5135 SSL_SESSION structure, that way new session features are automatically
5138 If a client application caches session in an SSL_SESSION structure
5139 support is transparent because tickets are now stored in the encoded
5142 The SSL_CTX structure automatically generates keys for ticket
5143 protection in servers so again support should be possible
5144 with no application modification.
5146 If a client or server wishes to disable RFC4507 support then the option
5147 SSL_OP_NO_TICKET can be set.
5149 Add a TLS extension debugging callback to allow the contents of any client
5150 or server extensions to be examined.
5152 This work was sponsored by Google.
5155 *) Add initial support for TLS extensions, specifically for the server_name
5156 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5157 have new members for a host name. The SSL data structure has an
5158 additional member SSL_CTX *initial_ctx so that new sessions can be
5159 stored in that context to allow for session resumption, even after the
5160 SSL has been switched to a new SSL_CTX in reaction to a client's
5161 server_name extension.
5163 New functions (subject to change):
5165 SSL_get_servername()
5166 SSL_get_servername_type()
5169 New CTRL codes and macros (subject to change):
5171 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5172 - SSL_CTX_set_tlsext_servername_callback()
5173 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5174 - SSL_CTX_set_tlsext_servername_arg()
5175 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5177 openssl s_client has a new '-servername ...' option.
5179 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5180 '-key2 ...', '-servername_fatal' (subject to change). This allows
5181 testing the HostName extension for a specific single host name ('-cert'
5182 and '-key' remain fallbacks for handshakes without HostName
5183 negotiation). If the unrecognized_name alert has to be sent, this by
5184 default is a warning; it becomes fatal with the '-servername_fatal'
5187 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5189 *) Add AES and SSE2 assembly language support to VC++ build.
5192 *) Mitigate attack on final subtraction in Montgomery reduction.
5195 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5196 (which previously caused an internal error).
5199 *) Squeeze another 10% out of IGE mode when in != out.
5202 *) AES IGE mode speedup.
5203 [Dean Gaudet (Google)]
5205 *) Add the Korean symmetric 128-bit cipher SEED (see
5206 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5207 add SEED ciphersuites from RFC 4162:
5209 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5210 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5211 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5212 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5214 To minimize changes between patchlevels in the OpenSSL 0.9.8
5215 series, SEED remains excluded from compilation unless OpenSSL
5216 is configured with 'enable-seed'.
5217 [KISA, Bodo Moeller]
5219 *) Mitigate branch prediction attacks, which can be practical if a
5220 single processor is shared, allowing a spy process to extract
5221 information. For detailed background information, see
5222 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5223 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5224 and Necessary Software Countermeasures"). The core of the change
5225 are new versions BN_div_no_branch() and
5226 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5227 respectively, which are slower, but avoid the security-relevant
5228 conditional branches. These are automatically called by BN_div()
5229 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5230 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5231 remove a conditional branch.
5233 BN_FLG_CONSTTIME is the new name for the previous
5234 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5235 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5236 in the exponent causes BN_mod_exp_mont() to use the alternative
5237 implementation in BN_mod_exp_mont_consttime().) The old name
5238 remains as a deprecated alias.
5240 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5241 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5242 constant-time implementations for more than just exponentiation.
5243 Here too the old name is kept as a deprecated alias.
5245 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5246 the BN_BLINDING structure gets an independent copy of the
5247 modulus. This means that the previous "BIGNUM *m" argument to
5248 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5249 essentially becomes "const BIGNUM *m", although we can't actually
5250 change this in the header file before 0.9.9. It allows
5251 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5252 enable BN_FLG_CONSTTIME.
5254 [Matthew D Wood (Intel Corp)]
5256 *) In the SSL/TLS server implementation, be strict about session ID
5257 context matching (which matters if an application uses a single
5258 external cache for different purposes). Previously,
5259 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5260 set. This did ensure strict client verification, but meant that,
5261 with applications using a single external cache for quite
5262 different requirements, clients could circumvent ciphersuite
5263 restrictions for a given session ID context by starting a session
5264 in a different context.
5267 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5268 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5269 authentication-only ciphersuites.
5272 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5273 not complete and could lead to a possible single byte overflow
5274 (CVE-2007-5135) [Ben Laurie]
5276 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5278 *) Since AES128 and AES256 (and similarly Camellia128 and
5279 Camellia256) share a single mask bit in the logic of
5280 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5281 kludge to work properly if AES128 is available and AES256 isn't
5282 (or if Camellia128 is available and Camellia256 isn't).
5285 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5286 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5287 When a point or a seed is encoded in a BIT STRING, we need to
5288 prevent the removal of trailing zero bits to get the proper DER
5289 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5290 of a NamedBitList, for which trailing 0 bits need to be removed.)
5293 *) Have SSL/TLS server implementation tolerate "mismatched" record
5294 protocol version while receiving ClientHello even if the
5295 ClientHello is fragmented. (The server can't insist on the
5296 particular protocol version it has chosen before the ServerHello
5297 message has informed the client about his choice.)
5300 *) Add RFC 3779 support.
5301 [Rob Austein for ARIN, Ben Laurie]
5303 *) Load error codes if they are not already present instead of using a
5304 static variable. This allows them to be cleanly unloaded and reloaded.
5305 Improve header file function name parsing.
5308 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5309 or CAPABILITY handshake as required by RFCs.
5312 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5314 *) Introduce limits to prevent malicious keys being able to
5315 cause a denial of service. (CVE-2006-2940)
5316 [Steve Henson, Bodo Moeller]
5318 *) Fix ASN.1 parsing of certain invalid structures that can result
5319 in a denial of service. (CVE-2006-2937) [Steve Henson]
5321 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5322 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5324 *) Fix SSL client code which could crash if connecting to a
5325 malicious SSLv2 server. (CVE-2006-4343)
5326 [Tavis Ormandy and Will Drewry, Google Security Team]
5328 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5329 match only those. Before that, "AES256-SHA" would be interpreted
5330 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5331 the same strength classification in 0.9.7h) as we currently only
5332 have a single AES bit in the ciphersuite description bitmap.
5333 That change, however, also applied to ciphersuite strings such as
5334 "RC4-MD5" that intentionally matched multiple ciphersuites --
5335 namely, SSL 2.0 ciphersuites in addition to the more common ones
5336 from SSL 3.0/TLS 1.0.
5338 So we change the selection algorithm again: Naming an explicit
5339 ciphersuite selects this one ciphersuite, and any other similar
5340 ciphersuite (same bitmap) from *other* protocol versions.
5341 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5342 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5344 Since SSL 2.0 does not have any ciphersuites for which the
5345 128/256 bit distinction would be relevant, this works for now.
5346 The proper fix will be to use different bits for AES128 and
5347 AES256, which would have avoided the problems from the beginning;
5348 however, bits are scarce, so we can only do this in a new release
5349 (not just a patchlevel) when we can change the SSL_CIPHER
5350 definition to split the single 'unsigned long mask' bitmap into
5351 multiple values to extend the available space.
5355 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5357 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5358 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5360 *) Add AES IGE and biIGE modes.
5363 *) Change the Unix randomness entropy gathering to use poll() when
5364 possible instead of select(), since the latter has some
5365 undesirable limitations.
5366 [Darryl Miles via Richard Levitte and Bodo Moeller]
5368 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5369 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5370 cannot be implicitly activated as part of, e.g., the "AES" alias.
5371 However, please upgrade to OpenSSL 0.9.9[-dev] for
5372 non-experimental use of the ECC ciphersuites to get TLS extension
5373 support, which is required for curve and point format negotiation
5374 to avoid potential handshake problems.
5377 *) Disable rogue ciphersuites:
5379 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5380 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5381 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5383 The latter two were purportedly from
5384 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5387 Also deactivate the remaining ciphersuites from
5388 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5389 unofficial, and the ID has long expired.
5392 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5393 dual-core machines) and other potential thread-safety issues.
5396 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5397 versions), which is now available for royalty-free use
5398 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5399 Also, add Camellia TLS ciphersuites from RFC 4132.
5401 To minimize changes between patchlevels in the OpenSSL 0.9.8
5402 series, Camellia remains excluded from compilation unless OpenSSL
5403 is configured with 'enable-camellia'.
5406 *) Disable the padding bug check when compression is in use. The padding
5407 bug check assumes the first packet is of even length, this is not
5408 necessarily true if compression is enabled and can result in false
5409 positives causing handshake failure. The actual bug test is ancient
5410 code so it is hoped that implementations will either have fixed it by
5411 now or any which still have the bug do not support compression.
5414 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5416 *) When applying a cipher rule check to see if string match is an explicit
5417 cipher suite and only match that one cipher suite if it is.
5420 *) Link in manifests for VC++ if needed.
5421 [Austin Ziegler <halostatue@gmail.com>]
5423 *) Update support for ECC-based TLS ciphersuites according to
5424 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5425 TLS extensions, which are supported starting with the 0.9.9
5426 branch, not in the OpenSSL 0.9.8 branch).
5429 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5430 opaque EVP_CIPHER_CTX handling.
5433 *) Fixes and enhancements to zlib compression code. We now only use
5434 "zlib1.dll" and use the default __cdecl calling convention on Win32
5435 to conform with the standards mentioned here:
5436 http://www.zlib.net/DLL_FAQ.txt
5437 Static zlib linking now works on Windows and the new --with-zlib-include
5438 --with-zlib-lib options to Configure can be used to supply the location
5439 of the headers and library. Gracefully handle case where zlib library
5443 *) Several fixes and enhancements to the OID generation code. The old code
5444 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5445 handle numbers larger than ULONG_MAX, truncated printing and had a
5446 non standard OBJ_obj2txt() behaviour.
5449 *) Add support for building of engines under engine/ as shared libraries
5450 under VC++ build system.
5453 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5454 Hopefully, we will not see any false combination of paths any more.
5457 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5459 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5460 (part of SSL_OP_ALL). This option used to disable the
5461 countermeasure against man-in-the-middle protocol-version
5462 rollback in the SSL 2.0 server implementation, which is a bad
5463 idea. (CVE-2005-2969)
5465 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5466 for Information Security, National Institute of Advanced Industrial
5467 Science and Technology [AIST], Japan)]
5469 *) Add two function to clear and return the verify parameter flags.
5472 *) Keep cipherlists sorted in the source instead of sorting them at
5473 runtime, thus removing the need for a lock.
5476 *) Avoid some small subgroup attacks in Diffie-Hellman.
5477 [Nick Mathewson and Ben Laurie]
5479 *) Add functions for well-known primes.
5482 *) Extended Windows CE support.
5483 [Satoshi Nakamura and Andy Polyakov]
5485 *) Initialize SSL_METHOD structures at compile time instead of during
5486 runtime, thus removing the need for a lock.
5489 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5490 attempting to decrypt each encrypted key in turn. Add support to
5494 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5496 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5499 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5502 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5503 key into the same file any more.
5506 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5509 *) Add -utf8 command line and config file option to 'ca'.
5510 [Stefan <stf@udoma.org]
5512 *) Removed the macro des_crypt(), as it seems to conflict with some
5513 libraries. Use DES_crypt().
5516 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5517 involves renaming the source and generated shared-libs for
5518 both. The engines will accept the corrected or legacy ids
5519 ('ncipher' and '4758_cca' respectively) when binding. NB,
5520 this only applies when building 'shared'.
5521 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5523 *) Add attribute functions to EVP_PKEY structure. Modify
5524 PKCS12_create() to recognize a CSP name attribute and
5525 use it. Make -CSP option work again in pkcs12 utility.
5528 *) Add new functionality to the bn blinding code:
5529 - automatic re-creation of the BN_BLINDING parameters after
5530 a fixed number of uses (currently 32)
5531 - add new function for parameter creation
5532 - introduce flags to control the update behaviour of the
5533 BN_BLINDING parameters
5534 - hide BN_BLINDING structure
5535 Add a second BN_BLINDING slot to the RSA structure to improve
5536 performance when a single RSA object is shared among several
5540 *) Add support for DTLS.
5541 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5543 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5544 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5547 *) Remove buggy and incomplete DH cert support from
5548 ssl/ssl_rsa.c and ssl/s3_both.c
5551 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5552 the apps/openssl applications.
5555 *) Compile clean with "-Wall -Wmissing-prototypes
5556 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5557 DEBUG_SAFESTACK must also be set.
5560 *) Change ./Configure so that certain algorithms can be disabled by default.
5561 The new counterpiece to "no-xxx" is "enable-xxx".
5563 The patented RC5 and MDC2 algorithms will now be disabled unless
5564 "enable-rc5" and "enable-mdc2", respectively, are specified.
5566 (IDEA remains enabled despite being patented. This is because IDEA
5567 is frequently required for interoperability, and there is no license
5568 fee for non-commercial use. As before, "no-idea" can be used to
5569 avoid this algorithm.)
5573 *) Add processing of proxy certificates (see RFC 3820). This work was
5574 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5575 EGEE (Enabling Grids for E-science in Europe).
5578 *) RC4 performance overhaul on modern architectures/implementations, such
5579 as Intel P4, IA-64 and AMD64.
5582 *) New utility extract-section.pl. This can be used specify an alternative
5583 section number in a pod file instead of having to treat each file as
5584 a separate case in Makefile. This can be done by adding two lines to the
5587 =for comment openssl_section:XXX
5589 The blank line is mandatory.
5593 *) New arguments -certform, -keyform and -pass for s_client and s_server
5594 to allow alternative format key and certificate files and passphrase
5598 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5599 update associated structures and add various utility functions.
5601 Add new policy related verify parameters, include policy checking in
5602 standard verify code. Enhance 'smime' application with extra parameters
5603 to support policy checking and print out.
5606 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5607 Nehemiah processors. These extensions support AES encryption in hardware
5608 as well as RNG (though RNG support is currently disabled).
5609 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5611 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5614 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5615 [Andy Polyakov and a number of other people]
5617 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5618 implementation contributed by IBM.
5619 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5621 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5622 exponent rather than 'unsigned long'. There is a corresponding change to
5623 the new 'rsa_keygen' element of the RSA_METHOD structure.
5624 [Jelte Jansen, Geoff Thorpe]
5626 *) Functionality for creating the initial serial number file is now
5627 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5629 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5630 number file to 1, which is bound to cause problems. To avoid
5631 the problems while respecting compatibility between different 0.9.7
5632 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5633 CA.pl for serial number initialization. With the new release 0.9.8,
5634 we can fix the problem directly in the 'ca' utility.)
5637 *) Reduced header interdependencies by declaring more opaque objects in
5638 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5639 give fewer recursive includes, which could break lazy source code - so
5640 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5641 developers should define this symbol when building and using openssl to
5642 ensure they track the recommended behaviour, interfaces, [etc], but
5643 backwards-compatible behaviour prevails when this isn't defined.
5646 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5649 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5650 This will generate a random key of the appropriate length based on the
5651 cipher context. The EVP_CIPHER can provide its own random key generation
5652 routine to support keys of a specific form. This is used in the des and
5653 3des routines to generate a key of the correct parity. Update S/MIME
5654 code to use new functions and hence generate correct parity DES keys.
5655 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5656 valid (weak or incorrect parity).
5659 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5660 as looking them up. This is useful when the verified structure may contain
5661 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5662 present unless the new PKCS7_NO_CRL flag is asserted.
5665 *) Extend ASN1 oid configuration module. It now additionally accepts the
5668 shortName = some long name, 1.2.3.4
5671 *) Reimplemented the BN_CTX implementation. There is now no more static
5672 limitation on the number of variables it can handle nor the depth of the
5673 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5674 information can now expand as required, and rather than having a single
5675 static array of bignums, BN_CTX now uses a linked-list of such arrays
5676 allowing it to expand on demand whilst maintaining the usefulness of
5677 BN_CTX's "bundling".
5680 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5681 to allow all RSA operations to function using a single BN_CTX.
5684 *) Preliminary support for certificate policy evaluation and checking. This
5685 is initially intended to pass the tests outlined in "Conformance Testing
5686 of Relying Party Client Certificate Path Processing Logic" v1.07.
5689 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5690 remained unused and not that useful. A variety of other little bignum
5691 tweaks and fixes have also been made continuing on from the audit (see
5695 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5696 associated ASN1, EVP and SSL functions and old ASN1 macros.
5699 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5700 and this should never fail. So the return value from the use of
5701 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5702 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5705 *) BN_CTX_get() should return zero-valued bignums, providing the same
5706 initialised value as BN_new().
5707 [Geoff Thorpe, suggested by Ulf Möller]
5709 *) Support for inhibitAnyPolicy certificate extension.
5712 *) An audit of the BIGNUM code is underway, for which debugging code is
5713 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5714 is considered valid when processing BIGNUMs, and causes execution to
5715 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5716 further steps are taken to deliberately pollute unused data in BIGNUM
5717 structures to try and expose faulty code further on. For now, openssl will
5718 (in its default mode of operation) continue to tolerate the inconsistent
5719 forms that it has tolerated in the past, but authors and packagers should
5720 consider trying openssl and their own applications when compiled with
5721 these debugging symbols defined. It will help highlight potential bugs in
5722 their own code, and will improve the test coverage for OpenSSL itself. At
5723 some point, these tighter rules will become openssl's default to improve
5724 maintainability, though the assert()s and other overheads will remain only
5725 in debugging configurations. See bn.h for more details.
5726 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5728 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5729 that can only be obtained through BN_CTX_new() (which implicitly
5730 initialises it). The presence of this function only made it possible
5731 to overwrite an existing structure (and cause memory leaks).
5734 *) Because of the callback-based approach for implementing LHASH as a
5735 template type, lh_insert() adds opaque objects to hash-tables and
5736 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5737 to clean up those corresponding objects before destroying the hash table
5738 (and losing the object pointers). So some over-zealous constifications in
5739 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5740 objects as "const" and the lh_doall[_arg] callback wrappers are not
5741 prototyped to have "const" restrictions on the object pointers they are
5742 given (and so aren't required to cast them away any more).
5745 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5746 (speed) prefers to use its own implementation. The two implementations
5747 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5748 its object type properly exposed (MS_TM) instead of casting to/from "char
5749 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5750 aren't necessarily the greatest nomenclatures - but this is what was used
5751 internally to the implementation so I've used that for now.
5754 *) Ensure that deprecated functions do not get compiled when
5755 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5756 the self-tests were still using deprecated key-generation functions so
5757 these have been updated also.
5760 *) Reorganise PKCS#7 code to separate the digest location functionality
5761 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5762 New function PKCS7_set_digest() to set the digest type for PKCS#7
5763 digestedData type. Add additional code to correctly generate the
5764 digestedData type and add support for this type in PKCS7 initialization
5768 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5769 structure of type "other".
5772 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5773 sure the loop does correctly stop and breaking ("division by zero")
5774 modulus operations are not performed. The (pre-generated) prime
5775 table crypto/bn/bn_prime.h was already correct, but it could not be
5776 re-generated on some platforms because of the "division by zero"
5777 situation in the script.
5778 [Ralf S. Engelschall]
5780 *) Update support for ECC-based TLS ciphersuites according to
5781 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5782 SHA-1 now is only used for "small" curves (where the
5783 representation of a field element takes up to 24 bytes); for
5784 larger curves, the field element resulting from ECDH is directly
5785 used as premaster secret.
5786 [Douglas Stebila (Sun Microsystems Laboratories)]
5788 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5789 curve secp160r1 to the tests.
5790 [Douglas Stebila (Sun Microsystems Laboratories)]
5792 *) Add the possibility to load symbols globally with DSO.
5793 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5795 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5796 control of the error stack.
5799 *) Add support for STORE in ENGINE.
5802 *) Add the STORE type. The intention is to provide a common interface
5803 to certificate and key stores, be they simple file-based stores, or
5804 HSM-type store, or LDAP stores, or...
5805 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5808 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5809 pass a list of arguments to any function as well as provide a way
5810 for a function to pass data back to the caller.
5813 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5814 works like BUF_strdup() but can be used to duplicate a portion of
5815 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5819 *) Add the function sk_find_ex() which works like sk_find(), but will
5820 return an index to an element even if an exact match couldn't be
5821 found. The index is guaranteed to point at the element where the
5822 searched-for key would be inserted to preserve sorting order.
5825 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5826 takes an extra flags argument for optional functionality. Currently,
5827 the following flags are defined:
5829 OBJ_BSEARCH_VALUE_ON_NOMATCH
5830 This one gets OBJ_bsearch_ex() to return a pointer to the first
5831 element where the comparing function returns a negative or zero
5834 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5835 This one gets OBJ_bsearch_ex() to return a pointer to the first
5836 element where the comparing function returns zero. This is useful
5837 if there are more than one element where the comparing function
5841 *) Make it possible to create self-signed certificates with 'openssl ca'
5842 in such a way that the self-signed certificate becomes part of the
5843 CA database and uses the same mechanisms for serial number generation
5844 as all other certificate signing. The new flag '-selfsign' enables
5845 this functionality. Adapt CA.sh and CA.pl.in.
5848 *) Add functionality to check the public key of a certificate request
5849 against a given private. This is useful to check that a certificate
5850 request can be signed by that key (self-signing).
5853 *) Make it possible to have multiple active certificates with the same
5854 subject in the CA index file. This is done only if the keyword
5855 'unique_subject' is set to 'no' in the main CA section (default
5856 if 'CA_default') of the configuration file. The value is saved
5857 with the database itself in a separate index attribute file,
5858 named like the index file with '.attr' appended to the name.
5861 *) Generate multi-valued AVAs using '+' notation in config files for
5865 *) Support for nameConstraints certificate extension.
5868 *) Support for policyConstraints certificate extension.
5871 *) Support for policyMappings certificate extension.
5874 *) Make sure the default DSA_METHOD implementation only uses its
5875 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5876 and change its own handlers to be NULL so as to remove unnecessary
5877 indirection. This lets alternative implementations fallback to the
5878 default implementation more easily.
5881 *) Support for directoryName in GeneralName related extensions
5885 *) Make it possible to link applications using Makefile.shared.
5886 Make that possible even when linking against static libraries!
5889 *) Support for single pass processing for S/MIME signing. This now
5890 means that S/MIME signing can be done from a pipe, in addition
5891 cleartext signing (multipart/signed type) is effectively streaming
5892 and the signed data does not need to be all held in memory.
5894 This is done with a new flag PKCS7_STREAM. When this flag is set
5895 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5896 is done after the data is output (and digests calculated) in
5897 SMIME_write_PKCS7().
5900 *) Add full support for -rpath/-R, both in shared libraries and
5901 applications, at least on the platforms where it's known how
5905 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5906 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5907 will now compute a table of multiples of the generator that
5908 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5909 faster (notably in the case of a single point multiplication,
5910 scalar * generator).
5911 [Nils Larsch, Bodo Moeller]
5913 *) IPv6 support for certificate extensions. The various extensions
5914 which use the IP:a.b.c.d can now take IPv6 addresses using the
5915 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5919 *) Added an ENGINE that implements RSA by performing private key
5920 exponentiations with the GMP library. The conversions to and from
5921 GMP's mpz_t format aren't optimised nor are any montgomery forms
5922 cached, and on x86 it appears OpenSSL's own performance has caught up.
5923 However there are likely to be other architectures where GMP could
5924 provide a boost. This ENGINE is not built in by default, but it can be
5925 specified at Configure time and should be accompanied by the necessary
5926 linker additions, eg;
5927 ./config -DOPENSSL_USE_GMP -lgmp
5930 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5931 testing availability of engines with "-t" - the old behaviour is
5932 produced by increasing the feature's verbosity with "-tt".
5935 *) ECDSA routines: under certain error conditions uninitialized BN objects
5936 could be freed. Solution: make sure initialization is performed early
5937 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5941 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5942 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5943 software implementations. For DSA and DH, parameter generation can
5944 also be overridden by providing the appropriate method callbacks.
5947 *) Change the "progress" mechanism used in key-generation and
5948 primality testing to functions that take a new BN_GENCB pointer in
5949 place of callback/argument pairs. The new API functions have "_ex"
5950 postfixes and the older functions are reimplemented as wrappers for
5951 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5952 declarations of the old functions to help (graceful) attempts to
5953 migrate to the new functions. Also, the new key-generation API
5954 functions operate on a caller-supplied key-structure and return
5955 success/failure rather than returning a key or NULL - this is to
5956 help make "keygen" another member function of RSA_METHOD etc.
5958 Example for using the new callback interface:
5960 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5964 BN_GENCB_set(&my_cb, my_callback, my_arg);
5966 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5967 /* For the meaning of a, b in calls to my_callback(), see the
5968 * documentation of the function that calls the callback.
5969 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5970 * my_callback should return 1 if it wants BN_is_prime_ex()
5971 * to continue, or 0 to stop.
5976 *) Change the ZLIB compression method to be stateful, and make it
5977 available to TLS with the number defined in
5978 draft-ietf-tls-compression-04.txt.
5981 *) Add the ASN.1 structures and functions for CertificatePair, which
5982 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5984 CertificatePair ::= SEQUENCE {
5985 forward [0] Certificate OPTIONAL,
5986 reverse [1] Certificate OPTIONAL,
5987 -- at least one of the pair shall be present -- }
5989 Also implement the PEM functions to read and write certificate
5990 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5992 This needed to be defined, mostly for the sake of the LDAP
5993 attribute crossCertificatePair, but may prove useful elsewhere as
5997 *) Make it possible to inhibit symlinking of shared libraries in
5998 Makefile.shared, for Cygwin's sake.
6001 *) Extend the BIGNUM API by creating a function
6002 void BN_set_negative(BIGNUM *a, int neg);
6003 and a macro that behave like
6004 int BN_is_negative(const BIGNUM *a);
6006 to avoid the need to access 'a->neg' directly in applications.
6009 *) Implement fast modular reduction for pseudo-Mersenne primes
6010 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6011 EC_GROUP_new_curve_GFp() will now automatically use this
6013 [Nils Larsch <nla@trustcenter.de>]
6015 *) Add new lock type (CRYPTO_LOCK_BN).
6018 *) Change the ENGINE framework to automatically load engines
6019 dynamically from specific directories unless they could be
6020 found to already be built in or loaded. Move all the
6021 current engines except for the cryptodev one to a new
6023 The engines in engines/ are built as shared libraries if
6024 the "shared" options was given to ./Configure or ./config.
6025 Otherwise, they are inserted in libcrypto.a.
6026 /usr/local/ssl/engines is the default directory for dynamic
6027 engines, but that can be overridden at configure time through
6028 the usual use of --prefix and/or --openssldir, and at run
6029 time with the environment variable OPENSSL_ENGINES.
6030 [Geoff Thorpe and Richard Levitte]
6032 *) Add Makefile.shared, a helper makefile to build shared
6033 libraries. Adapt Makefile.org.
6036 *) Add version info to Win32 DLLs.
6037 [Peter 'Luna' Runestig" <peter@runestig.com>]
6039 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6040 can be added using this API to created arbitrary PKCS#12
6041 files while avoiding the low level API.
6043 New options to PKCS12_create(), key or cert can be NULL and
6044 will then be omitted from the output file. The encryption
6045 algorithm NIDs can be set to -1 for no encryption, the mac
6046 iteration count can be set to 0 to omit the mac.
6048 Enhance pkcs12 utility by making the -nokeys and -nocerts
6049 options work when creating a PKCS#12 file. New option -nomac
6050 to omit the mac, NONE can be set for an encryption algorithm.
6051 New code is modified to use the enhanced PKCS12_create()
6052 instead of the low level API.
6055 *) Extend ASN1 encoder to support indefinite length constructed
6056 encoding. This can output sequences tags and octet strings in
6057 this form. Modify pk7_asn1.c to support indefinite length
6058 encoding. This is experimental and needs additional code to
6059 be useful, such as an ASN1 bio and some enhanced streaming
6062 Extend template encode functionality so that tagging is passed
6063 down to the template encoder.
6066 *) Let 'openssl req' fail if an argument to '-newkey' is not
6067 recognized instead of using RSA as a default.
6070 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6071 As these are not official, they are not included in "ALL";
6072 the "ECCdraft" ciphersuite group alias can be used to select them.
6073 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6075 *) Add ECDH engine support.
6076 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6078 *) Add ECDH in new directory crypto/ecdh/.
6079 [Douglas Stebila (Sun Microsystems Laboratories)]
6081 *) Let BN_rand_range() abort with an error after 100 iterations
6082 without success (which indicates a broken PRNG).
6085 *) Change BN_mod_sqrt() so that it verifies that the input value
6086 is really the square of the return value. (Previously,
6087 BN_mod_sqrt would show GIGO behaviour.)
6090 *) Add named elliptic curves over binary fields from X9.62, SECG,
6091 and WAP/WTLS; add OIDs that were still missing.
6093 [Sheueling Chang Shantz and Douglas Stebila
6094 (Sun Microsystems Laboratories)]
6096 *) Extend the EC library for elliptic curves over binary fields
6097 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6100 EC_GF2m_simple_method
6104 EC_GROUP_new_curve_GF2m
6105 EC_GROUP_set_curve_GF2m
6106 EC_GROUP_get_curve_GF2m
6107 EC_POINT_set_affine_coordinates_GF2m
6108 EC_POINT_get_affine_coordinates_GF2m
6109 EC_POINT_set_compressed_coordinates_GF2m
6111 Point compression for binary fields is disabled by default for
6112 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6115 As binary polynomials are represented as BIGNUMs, various members
6116 of the EC_GROUP and EC_POINT data structures can be shared
6117 between the implementations for prime fields and binary fields;
6118 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6119 are essentially identical to their ..._GFp counterparts.
6120 (For simplicity, the '..._GFp' prefix has been dropped from
6121 various internal method names.)
6123 An internal 'field_div' method (similar to 'field_mul' and
6124 'field_sqr') has been added; this is used only for binary fields.
6126 [Sheueling Chang Shantz and Douglas Stebila
6127 (Sun Microsystems Laboratories)]
6129 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6130 through methods ('mul', 'precompute_mult').
6132 The generic implementations (now internally called 'ec_wNAF_mul'
6133 and 'ec_wNAF_precomputed_mult') remain the default if these
6134 methods are undefined.
6136 [Sheueling Chang Shantz and Douglas Stebila
6137 (Sun Microsystems Laboratories)]
6139 *) New function EC_GROUP_get_degree, which is defined through
6140 EC_METHOD. For curves over prime fields, this returns the bit
6141 length of the modulus.
6143 [Sheueling Chang Shantz and Douglas Stebila
6144 (Sun Microsystems Laboratories)]
6146 *) New functions EC_GROUP_dup, EC_POINT_dup.
6147 (These simply call ..._new and ..._copy).
6149 [Sheueling Chang Shantz and Douglas Stebila
6150 (Sun Microsystems Laboratories)]
6152 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6153 Polynomials are represented as BIGNUMs (where the sign bit is not
6154 used) in the following functions [macros]:
6157 BN_GF2m_sub [= BN_GF2m_add]
6158 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6159 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6160 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6162 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6163 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6164 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6165 BN_GF2m_cmp [= BN_ucmp]
6167 (Note that only the 'mod' functions are actually for fields GF(2^m).
6168 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6170 For some functions, an the irreducible polynomial defining a
6171 field can be given as an 'unsigned int[]' with strictly
6172 decreasing elements giving the indices of those bits that are set;
6173 i.e., p[] represents the polynomial
6174 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6176 p[0] > p[1] > ... > p[k] = 0.
6177 This applies to the following functions:
6182 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6183 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6185 BN_GF2m_mod_sqrt_arr
6186 BN_GF2m_mod_solve_quad_arr
6190 Conversion can be performed by the following functions:
6195 bntest.c has additional tests for binary polynomial arithmetic.
6197 Two implementations for BN_GF2m_mod_div() are available.
6198 The default algorithm simply uses BN_GF2m_mod_inv() and
6199 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6200 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6201 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6203 [Sheueling Chang Shantz and Douglas Stebila
6204 (Sun Microsystems Laboratories)]
6206 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6207 functionality is disabled at compile-time.
6208 [Douglas Stebila <douglas.stebila@sun.com>]
6210 *) Change default behaviour of 'openssl asn1parse' so that more
6211 information is visible when viewing, e.g., a certificate:
6213 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6214 mode the content of non-printable OCTET STRINGs is output in a
6215 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6216 avoid the appearance of a printable string.
6217 [Nils Larsch <nla@trustcenter.de>]
6219 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6221 EC_GROUP_set_asn1_flag()
6222 EC_GROUP_get_asn1_flag()
6223 EC_GROUP_set_point_conversion_form()
6224 EC_GROUP_get_point_conversion_form()
6225 These control ASN1 encoding details:
6226 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6227 has been set to OPENSSL_EC_NAMED_CURVE.
6228 - Points are encoded in uncompressed form by default; options for
6229 asn1_for are as for point2oct, namely
6230 POINT_CONVERSION_COMPRESSED
6231 POINT_CONVERSION_UNCOMPRESSED
6232 POINT_CONVERSION_HYBRID
6234 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6237 EC_GROUP_get0_seed()
6238 EC_GROUP_get_seed_len()
6239 This is used only for ASN1 purposes (so far).
6240 [Nils Larsch <nla@trustcenter.de>]
6242 *) Add 'field_type' member to EC_METHOD, which holds the NID
6243 of the appropriate field type OID. The new function
6244 EC_METHOD_get_field_type() returns this value.
6245 [Nils Larsch <nla@trustcenter.de>]
6250 EC_POINT_point2hex()
6251 EC_POINT_hex2point()
6252 providing useful interfaces to EC_POINT_point2oct() and
6253 EC_POINT_oct2point().
6254 [Nils Larsch <nla@trustcenter.de>]
6256 *) Change internals of the EC library so that the functions
6257 EC_GROUP_set_generator()
6258 EC_GROUP_get_generator()
6259 EC_GROUP_get_order()
6260 EC_GROUP_get_cofactor()
6261 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6262 to methods, which would lead to unnecessary code duplication when
6263 adding different types of curves.
6264 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6266 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6267 arithmetic, and such that modified wNAFs are generated
6268 (which avoid length expansion in many cases).
6271 *) Add a function EC_GROUP_check_discriminant() (defined via
6272 EC_METHOD) that verifies that the curve discriminant is non-zero.
6274 Add a function EC_GROUP_check() that makes some sanity tests
6275 on a EC_GROUP, its generator and order. This includes
6276 EC_GROUP_check_discriminant().
6277 [Nils Larsch <nla@trustcenter.de>]
6279 *) Add ECDSA in new directory crypto/ecdsa/.
6281 Add applications 'openssl ecparam' and 'openssl ecdsa'
6282 (these are based on 'openssl dsaparam' and 'openssl dsa').
6284 ECDSA support is also included in various other files across the
6285 library. Most notably,
6286 - 'openssl req' now has a '-newkey ecdsa:file' option;
6287 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6288 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6289 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6290 them suitable for ECDSA where domain parameters must be
6291 extracted before the specific public key;
6292 - ECDSA engine support has been added.
6293 [Nils Larsch <nla@trustcenter.de>]
6295 *) Include some named elliptic curves, and add OIDs from X9.62,
6296 SECG, and WAP/WTLS. Each curve can be obtained from the new
6298 EC_GROUP_new_by_curve_name(),
6299 and the list of available named curves can be obtained with
6300 EC_get_builtin_curves().
6301 Also add a 'curve_name' member to EC_GROUP objects, which can be
6303 EC_GROUP_set_curve_name()
6304 EC_GROUP_get_curve_name()
6305 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6307 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6308 was actually never needed) and in BN_mul(). The removal in BN_mul()
6309 required a small change in bn_mul_part_recursive() and the addition
6310 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6311 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6312 bn_sub_words() and bn_add_words() except they take arrays with
6316 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6318 *) Cleanse PEM buffers before freeing them since they may contain
6320 [Benjamin Bennett <ben@psc.edu>]
6322 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6323 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6324 authentication-only ciphersuites.
6327 *) Since AES128 and AES256 share a single mask bit in the logic of
6328 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6329 kludge to work properly if AES128 is available and AES256 isn't.
6332 *) Expand security boundary to match 1.1.1 module.
6335 *) Remove redundant features: hash file source, editing of test vectors
6336 modify fipsld to use external fips_premain.c signature.
6339 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6340 run algorithm test programs.
6343 *) Make algorithm test programs more tolerant of whitespace.
6346 *) Have SSL/TLS server implementation tolerate "mismatched" record
6347 protocol version while receiving ClientHello even if the
6348 ClientHello is fragmented. (The server can't insist on the
6349 particular protocol version it has chosen before the ServerHello
6350 message has informed the client about his choice.)
6353 *) Load error codes if they are not already present instead of using a
6354 static variable. This allows them to be cleanly unloaded and reloaded.
6357 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6359 *) Introduce limits to prevent malicious keys being able to
6360 cause a denial of service. (CVE-2006-2940)
6361 [Steve Henson, Bodo Moeller]
6363 *) Fix ASN.1 parsing of certain invalid structures that can result
6364 in a denial of service. (CVE-2006-2937) [Steve Henson]
6366 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6367 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6369 *) Fix SSL client code which could crash if connecting to a
6370 malicious SSLv2 server. (CVE-2006-4343)
6371 [Tavis Ormandy and Will Drewry, Google Security Team]
6373 *) Change ciphersuite string processing so that an explicit
6374 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6375 will no longer include "AES128-SHA"), and any other similar
6376 ciphersuite (same bitmap) from *other* protocol versions (so that
6377 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6378 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6379 changes from 0.9.8b and 0.9.8d.
6382 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6384 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6385 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6387 *) Change the Unix randomness entropy gathering to use poll() when
6388 possible instead of select(), since the latter has some
6389 undesirable limitations.
6390 [Darryl Miles via Richard Levitte and Bodo Moeller]
6392 *) Disable rogue ciphersuites:
6394 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6395 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6396 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6398 The latter two were purportedly from
6399 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6402 Also deactivate the remaining ciphersuites from
6403 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6404 unofficial, and the ID has long expired.
6407 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6408 dual-core machines) and other potential thread-safety issues.
6411 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6413 *) Adapt fipsld and the build system to link against the validated FIPS
6414 module in FIPS mode.
6417 *) Fixes for VC++ 2005 build under Windows.
6420 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6421 from a Windows bash shell such as MSYS. It is autodetected from the
6422 "config" script when run from a VC++ environment. Modify standard VC++
6423 build to use fipscanister.o from the GNU make build.
6426 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6428 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6429 The value now differs depending on if you build for FIPS or not.
6430 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6431 safely run with a non-FIPSed libcrypto, as it may crash because of
6432 the difference induced by this change.
6435 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6437 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6438 (part of SSL_OP_ALL). This option used to disable the
6439 countermeasure against man-in-the-middle protocol-version
6440 rollback in the SSL 2.0 server implementation, which is a bad
6441 idea. (CVE-2005-2969)
6443 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6444 for Information Security, National Institute of Advanced Industrial
6445 Science and Technology [AIST], Japan)]
6447 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6448 mainly for FIPS compliance and not fully integrated at this stage.
6451 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6452 the exponentiation using a fixed-length exponent. (Otherwise,
6453 the information leaked through timing could expose the secret key
6454 after many signatures; cf. Bleichenbacher's attack on DSA with
6458 *) Make a new fixed-window mod_exp implementation the default for
6459 RSA, DSA, and DH private-key operations so that the sequence of
6460 squares and multiplies and the memory access pattern are
6461 independent of the particular secret key. This will mitigate
6462 cache-timing and potential related attacks.
6464 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6465 and this is automatically used by BN_mod_exp_mont() if the new flag
6466 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6467 will use this BN flag for private exponents unless the flag
6468 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6469 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6471 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6473 *) Change the client implementation for SSLv23_method() and
6474 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6475 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6476 (Previously, the SSL 2.0 backwards compatible Client Hello
6477 message format would be used even with SSL_OP_NO_SSLv2.)
6480 *) Add support for smime-type MIME parameter in S/MIME messages which some
6484 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6485 a threadsafe manner. Modify rsa code to use new function and add calls
6486 to dsa and dh code (which had race conditions before).
6489 *) Include the fixed error library code in the C error file definitions
6490 instead of fixing them up at runtime. This keeps the error code
6491 structures constant.
6494 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6496 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6499 *) Fixes for newer kerberos headers. NB: the casts are needed because
6500 the 'length' field is signed on one version and unsigned on another
6501 with no (?) obvious way to tell the difference, without these VC++
6502 complains. Also the "definition" of FAR (blank) is no longer included
6503 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6504 some needed definitions.
6507 *) Undo Cygwin change.
6510 *) Added support for proxy certificates according to RFC 3820.
6511 Because they may be a security thread to unaware applications,
6512 they must be explicitly allowed in run-time. See
6513 docs/HOWTO/proxy_certificates.txt for further information.
6516 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6518 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6519 server and client random values. Previously
6520 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6521 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6523 This change has negligible security impact because:
6525 1. Server and client random values still have 24 bytes of pseudo random
6528 2. Server and client random values are sent in the clear in the initial
6531 3. The master secret is derived using the premaster secret (48 bytes in
6532 size for static RSA ciphersuites) as well as client server and random
6535 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6538 [Stephen Henson, reported by UK NISCC]
6540 *) Use Windows randomness collection on Cygwin.
6543 *) Fix hang in EGD/PRNGD query when communication socket is closed
6544 prematurely by EGD/PRNGD.
6545 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6547 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6550 *) Back-port of selected performance improvements from development
6551 branch, as well as improved support for PowerPC platforms.
6554 *) Add lots of checks for memory allocation failure, error codes to indicate
6555 failure and freeing up memory if a failure occurs.
6556 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6558 *) Add new -passin argument to dgst.
6561 *) Perform some character comparisons of different types in X509_NAME_cmp:
6562 this is needed for some certificates that re-encode DNs into UTF8Strings
6563 (in violation of RFC3280) and can't or won't issue name rollover
6567 *) Make an explicit check during certificate validation to see that
6568 the CA setting in each certificate on the chain is correct. As a
6569 side effect always do the following basic checks on extensions,
6570 not just when there's an associated purpose to the check:
6572 - if there is an unhandled critical extension (unless the user
6573 has chosen to ignore this fault)
6574 - if the path length has been exceeded (if one is set at all)
6575 - that certain extensions fit the associated purpose (if one has
6579 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6581 *) Avoid a race condition when CRLs are checked in a multi threaded
6582 environment. This would happen due to the reordering of the revoked
6583 entries during signature checking and serial number lookup. Now the
6584 encoding is cached and the serial number sort performed under a lock.
6585 Add new STACK function sk_is_sorted().
6588 *) Add Delta CRL to the extension code.
6591 *) Various fixes to s3_pkt.c so alerts are sent properly.
6592 [David Holmes <d.holmes@f5.com>]
6594 *) Reduce the chances of duplicate issuer name and serial numbers (in
6595 violation of RFC3280) using the OpenSSL certificate creation utilities.
6596 This is done by creating a random 64 bit value for the initial serial
6597 number when a serial number file is created or when a self signed
6598 certificate is created using 'openssl req -x509'. The initial serial
6599 number file is created using 'openssl x509 -next_serial' in CA.pl
6600 rather than being initialized to 1.
6603 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6605 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6606 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6607 [Joe Orton, Steve Henson]
6609 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6611 [Joe Orton, Steve Henson]
6613 *) Make it possible to have multiple active certificates with the same
6614 subject in the CA index file. This is done only if the keyword
6615 'unique_subject' is set to 'no' in the main CA section (default
6616 if 'CA_default') of the configuration file. The value is saved
6617 with the database itself in a separate index attribute file,
6618 named like the index file with '.attr' appended to the name.
6621 *) X509 verify fixes. Disable broken certificate workarounds when
6622 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6623 keyUsage extension present. Don't accept CRLs with unhandled critical
6624 extensions: since verify currently doesn't process CRL extensions this
6625 rejects a CRL with *any* critical extensions. Add new verify error codes
6629 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6630 A clarification of RFC2560 will require the use of OCTET STRINGs and
6631 some implementations cannot handle the current raw format. Since OpenSSL
6632 copies and compares OCSP nonces as opaque blobs without any attempt at
6633 parsing them this should not create any compatibility issues.
6636 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6637 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6638 this HMAC (and other) operations are several times slower than OpenSSL
6642 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6643 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6645 *) Use the correct content when signing type "other".
6648 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6650 *) Fix various bugs revealed by running the NISCC test suite:
6652 Stop out of bounds reads in the ASN1 code when presented with
6653 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6655 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6657 If verify callback ignores invalid public key errors don't try to check
6658 certificate signature with the NULL public key.
6662 *) New -ignore_err option in ocsp application to stop the server
6663 exiting on the first error in a request.
6666 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6667 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6671 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6672 extra data after the compression methods not only for TLS 1.0
6673 but also for SSL 3.0 (as required by the specification).
6674 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6676 *) Change X509_certificate_type() to mark the key as exported/exportable
6677 when it's 512 *bits* long, not 512 bytes.
6680 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6681 blocks during encryption.
6684 *) Various fixes to base64 BIO and non blocking I/O. On write
6685 flushes were not handled properly if the BIO retried. On read
6686 data was not being buffered properly and had various logic bugs.
6687 This also affects blocking I/O when the data being decoded is a
6691 *) Various S/MIME bugfixes and compatibility changes:
6692 output correct application/pkcs7 MIME type if
6693 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6694 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6695 of files as .eml work). Correctly handle very long lines in MIME
6699 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6701 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6702 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6703 a protocol version number mismatch like a decryption error
6704 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6707 *) Turn on RSA blinding by default in the default implementation
6708 to avoid a timing attack. Applications that don't want it can call
6709 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6710 They would be ill-advised to do so in most cases.
6711 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6713 *) Change RSA blinding code so that it works when the PRNG is not
6714 seeded (in this case, the secret RSA exponent is abused as
6715 an unpredictable seed -- if it is not unpredictable, there
6716 is no point in blinding anyway). Make RSA blinding thread-safe
6717 by remembering the creator's thread ID in rsa->blinding and
6718 having all other threads use local one-time blinding factors
6719 (this requires more computation than sharing rsa->blinding, but
6720 avoids excessive locking; and if an RSA object is not shared
6721 between threads, blinding will still be very fast).
6724 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6725 ENGINE as defaults for all supported algorithms irrespective of
6726 the 'flags' parameter. 'flags' is now honoured, so applications
6727 should make sure they are passing it correctly.
6730 *) Target "mingw" now allows native Windows code to be generated in
6731 the Cygwin environment as well as with the MinGW compiler.
6734 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6736 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6737 via timing by performing a MAC computation even if incorrect
6738 block cipher padding has been found. This is a countermeasure
6739 against active attacks where the attacker has to distinguish
6740 between bad padding and a MAC verification error. (CVE-2003-0078)
6742 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6743 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6744 Martin Vuagnoux (EPFL, Ilion)]
6746 *) Make the no-err option work as intended. The intention with no-err
6747 is not to have the whole error stack handling routines removed from
6748 libcrypto, it's only intended to remove all the function name and
6749 reason texts, thereby removing some of the footprint that may not
6750 be interesting if those errors aren't displayed anyway.
6752 NOTE: it's still possible for any application or module to have it's
6753 own set of error texts inserted. The routines are there, just not
6754 used by default when no-err is given.
6757 *) Add support for FreeBSD on IA64.
6758 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6760 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6761 Kerberos function mit_des_cbc_cksum(). Before this change,
6762 the value returned by DES_cbc_cksum() was like the one from
6763 mit_des_cbc_cksum(), except the bytes were swapped.
6764 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6766 *) Allow an application to disable the automatic SSL chain building.
6767 Before this a rather primitive chain build was always performed in
6768 ssl3_output_cert_chain(): an application had no way to send the
6769 correct chain if the automatic operation produced an incorrect result.
6771 Now the chain builder is disabled if either:
6773 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6775 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6777 The reasoning behind this is that an application would not want the
6778 auto chain building to take place if extra chain certificates are
6779 present and it might also want a means of sending no additional
6780 certificates (for example the chain has two certificates and the
6784 *) Add the possibility to build without the ENGINE framework.
6785 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6787 *) Under Win32 gmtime() can return NULL: check return value in
6788 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6791 *) DSA routines: under certain error conditions uninitialized BN objects
6792 could be freed. Solution: make sure initialization is performed early
6793 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6794 Nils Larsch <nla@trustcenter.de> via PR#459)
6797 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6798 checked on reconnect on the client side, therefore session resumption
6799 could still fail with a "ssl session id is different" error. This
6800 behaviour is masked when SSL_OP_ALL is used due to
6801 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6802 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6803 followup to PR #377.
6806 *) IA-32 assembler support enhancements: unified ELF targets, support
6807 for SCO/Caldera platforms, fix for Cygwin shared build.
6810 *) Add support for FreeBSD on sparc64. As a consequence, support for
6811 FreeBSD on non-x86 processors is separate from x86 processors on
6812 the config script, much like the NetBSD support.
6813 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6815 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6817 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6820 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6821 code (06) was taken as the first octet of the session ID and the last
6822 octet was ignored consequently. As a result SSLv2 client side session
6823 caching could not have worked due to the session ID mismatch between
6825 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6829 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6830 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6834 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6835 seems that in spite of existing for more than a year, many application
6836 author have done nothing to provide the necessary callbacks, which
6837 means that this particular engine will not work properly anywhere.
6838 This is a very unfortunate situation which forces us, in the name
6839 of usability, to give the hw_ncipher.c a static lock, which is part
6841 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6842 appear in 0.9.8 or later. We EXPECT application authors to have
6843 dealt properly with this when 0.9.8 is released (unless we actually
6844 make such changes in the libcrypto locking code that changes will
6845 have to be made anyway).
6848 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6849 octets have been read, EOF or an error occurs. Without this change
6850 some truncated ASN1 structures will not produce an error.
6853 *) Disable Heimdal support, since it hasn't been fully implemented.
6854 Still give the possibility to force the use of Heimdal, but with
6855 warnings and a request that patches get sent to openssl-dev.
6858 *) Add the VC-CE target, introduce the WINCE sysname, and add
6859 INSTALL.WCE and appropriate conditionals to make it build.
6860 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6862 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6863 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6864 edit numbers of the version.
6865 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6867 *) Introduce safe string copy and catenation functions
6868 (BUF_strlcpy() and BUF_strlcat()).
6869 [Ben Laurie (CHATS) and Richard Levitte]
6871 *) Avoid using fixed-size buffers for one-line DNs.
6872 [Ben Laurie (CHATS)]
6874 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6875 resizing buffers containing secrets, and use where appropriate.
6876 [Ben Laurie (CHATS)]
6878 *) Avoid using fixed size buffers for configuration file location.
6879 [Ben Laurie (CHATS)]
6881 *) Avoid filename truncation for various CA files.
6882 [Ben Laurie (CHATS)]
6884 *) Use sizeof in preference to magic numbers.
6885 [Ben Laurie (CHATS)]
6887 *) Avoid filename truncation in cert requests.
6888 [Ben Laurie (CHATS)]
6890 *) Add assertions to check for (supposedly impossible) buffer
6892 [Ben Laurie (CHATS)]
6894 *) Don't cache truncated DNS entries in the local cache (this could
6895 potentially lead to a spoofing attack).
6896 [Ben Laurie (CHATS)]
6898 *) Fix various buffers to be large enough for hex/decimal
6899 representations in a platform independent manner.
6900 [Ben Laurie (CHATS)]
6902 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6903 resizing buffers containing secrets, and use where appropriate.
6904 [Ben Laurie (CHATS)]
6906 *) Add BIO_indent() to avoid much slightly worrying code to do
6908 [Ben Laurie (CHATS)]
6910 *) Convert sprintf()/BIO_puts() to BIO_printf().
6911 [Ben Laurie (CHATS)]
6913 *) buffer_gets() could terminate with the buffer only half
6915 [Ben Laurie (CHATS)]
6917 *) Add assertions to prevent user-supplied crypto functions from
6918 overflowing internal buffers by having large block sizes, etc.
6919 [Ben Laurie (CHATS)]
6921 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6923 [Ben Laurie (CHATS)]
6925 *) Eliminate unused copy of key in RC4.
6926 [Ben Laurie (CHATS)]
6928 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6929 [Ben Laurie (CHATS)]
6931 *) Fix off-by-one error in EGD path.
6932 [Ben Laurie (CHATS)]
6934 *) If RANDFILE path is too long, ignore instead of truncating.
6935 [Ben Laurie (CHATS)]
6937 *) Eliminate unused and incorrectly sized X.509 structure
6939 [Ben Laurie (CHATS)]
6941 *) Eliminate unused and dangerous function knumber().
6942 [Ben Laurie (CHATS)]
6944 *) Eliminate unused and dangerous structure, KSSL_ERR.
6945 [Ben Laurie (CHATS)]
6947 *) Protect against overlong session ID context length in an encoded
6948 session object. Since these are local, this does not appear to be
6950 [Ben Laurie (CHATS)]
6952 *) Change from security patch (see 0.9.6e below) that did not affect
6953 the 0.9.6 release series:
6955 Remote buffer overflow in SSL3 protocol - an attacker could
6956 supply an oversized master key in Kerberos-enabled versions.
6958 [Ben Laurie (CHATS)]
6960 *) Change the SSL kerb5 codes to match RFC 2712.
6963 *) Make -nameopt work fully for req and add -reqopt switch.
6964 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6966 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6967 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6969 *) Make sure tests can be performed even if the corresponding algorithms
6970 have been removed entirely. This was also the last step to make
6971 OpenSSL compilable with DJGPP under all reasonable conditions.
6972 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6974 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6975 to allow version independent disabling of normally unselected ciphers,
6976 which may be activated as a side-effect of selecting a single cipher.
6978 (E.g., cipher list string "RSA" enables ciphersuites that are left
6979 out of "ALL" because they do not provide symmetric encryption.
6980 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6981 [Lutz Jaenicke, Bodo Moeller]
6983 *) Add appropriate support for separate platform-dependent build
6984 directories. The recommended way to make a platform-dependent
6985 build directory is the following (tested on Linux), maybe with
6988 # Place yourself outside of the OpenSSL source tree. In
6989 # this example, the environment variable OPENSSL_SOURCE
6990 # is assumed to contain the absolute OpenSSL source directory.
6991 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6992 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6993 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6994 mkdir -p `dirname $F`
6995 ln -s $OPENSSL_SOURCE/$F $F
6998 To be absolutely sure not to disturb the source tree, a "make clean"
6999 is a good thing. If it isn't successful, don't worry about it,
7000 it probably means the source directory is very clean.
7003 *) Make sure any ENGINE control commands make local copies of string
7004 pointers passed to them whenever necessary. Otherwise it is possible
7005 the caller may have overwritten (or deallocated) the original string
7006 data when a later ENGINE operation tries to use the stored values.
7007 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7009 *) Improve diagnostics in file reading and command-line digests.
7010 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7012 *) Add AES modes CFB and OFB to the object database. Correct an
7013 error in AES-CFB decryption.
7016 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7017 allows existing EVP_CIPHER_CTX structures to be reused after
7018 calling EVP_*Final(). This behaviour is used by encryption
7019 BIOs and some applications. This has the side effect that
7020 applications must explicitly clean up cipher contexts with
7021 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7024 *) Check the values of dna and dnb in bn_mul_recursive before calling
7025 bn_mul_comba (a non zero value means the a or b arrays do not contain
7026 n2 elements) and fallback to bn_mul_normal if either is not zero.
7029 *) Fix escaping of non-ASCII characters when using the -subj option
7030 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7033 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7034 form for "surname", serialNumber has no short form.
7035 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7036 therefore remove "mail" short name for "internet 7".
7037 The OID for unique identifiers in X509 certificates is
7038 x500UniqueIdentifier, not uniqueIdentifier.
7039 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7042 *) Add an "init" command to the ENGINE config module and auto initialize
7043 ENGINEs. Without any "init" command the ENGINE will be initialized
7044 after all ctrl commands have been executed on it. If init=1 the
7045 ENGINE is initialized at that point (ctrls before that point are run
7046 on the uninitialized ENGINE and after on the initialized one). If
7047 init=0 then the ENGINE will not be initialized at all.
7050 *) Fix the 'app_verify_callback' interface so that the user-defined
7051 argument is actually passed to the callback: In the
7052 SSL_CTX_set_cert_verify_callback() prototype, the callback
7053 declaration has been changed from
7056 int (*cb)(X509_STORE_CTX *,void *);
7057 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7058 i=s->ctx->app_verify_callback(&ctx)
7059 has been changed into
7060 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7062 To update applications using SSL_CTX_set_cert_verify_callback(),
7063 a dummy argument can be added to their callback functions.
7064 [D. K. Smetters <smetters@parc.xerox.com>]
7066 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7067 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7069 *) Add and OPENSSL_LOAD_CONF define which will cause
7070 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7071 This allows older applications to transparently support certain
7072 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7073 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7074 load the config file and OPENSSL_add_all_algorithms_conf() which will
7075 always load it have also been added.
7078 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7079 Adjust NIDs and EVP layer.
7080 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7082 *) Config modules support in openssl utility.
7084 Most commands now load modules from the config file,
7085 though in a few (such as version) this isn't done
7086 because it couldn't be used for anything.
7088 In the case of ca and req the config file used is
7089 the same as the utility itself: that is the -config
7090 command line option can be used to specify an
7094 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7095 use "openssl_conf" if filename is NULL use default openssl config file.
7098 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7099 config section name. Add a new flag to tolerate a missing config file
7100 and move code to CONF_modules_load_file().
7103 *) Support for crypto accelerator cards from Accelerated Encryption
7104 Processing, www.aep.ie. (Use engine 'aep')
7105 The support was copied from 0.9.6c [engine] and adapted/corrected
7106 to work with the new engine framework.
7107 [AEP Inc. and Richard Levitte]
7109 *) Support for SureWare crypto accelerator cards from Baltimore
7110 Technologies. (Use engine 'sureware')
7111 The support was copied from 0.9.6c [engine] and adapted
7112 to work with the new engine framework.
7115 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7116 make the newer ENGINE framework commands for the CHIL engine work.
7117 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7119 *) Make it possible to produce shared libraries on ReliantUNIX.
7120 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7122 *) Add the configuration target debug-linux-ppro.
7123 Make 'openssl rsa' use the general key loading routines
7124 implemented in apps.c, and make those routines able to
7125 handle the key format FORMAT_NETSCAPE and the variant
7127 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7129 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7130 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7132 *) Add -keyform to rsautl, and document -engine.
7133 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7135 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7136 BIO_R_NO_SUCH_FILE error code rather than the generic
7137 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7140 *) Add new functions
7142 ERR_peek_last_error_line
7143 ERR_peek_last_error_line_data.
7144 These are similar to
7147 ERR_peek_error_line_data,
7148 but report on the latest error recorded rather than the first one
7149 still in the error queue.
7150 [Ben Laurie, Bodo Moeller]
7152 *) default_algorithms option in ENGINE config module. This allows things
7154 default_algorithms = ALL
7155 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7158 *) Preliminary ENGINE config module.
7161 *) New experimental application configuration code.
7164 *) Change the AES code to follow the same name structure as all other
7165 symmetric ciphers, and behave the same way. Move everything to
7166 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7167 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7169 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7170 [Ben Laurie and Theo de Raadt]
7172 *) Add option to output public keys in req command.
7173 [Massimiliano Pala madwolf@openca.org]
7175 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7176 (up to about 10% better than before for P-192 and P-224).
7179 *) New functions/macros
7181 SSL_CTX_set_msg_callback(ctx, cb)
7182 SSL_CTX_set_msg_callback_arg(ctx, arg)
7183 SSL_set_msg_callback(ssl, cb)
7184 SSL_set_msg_callback_arg(ssl, arg)
7186 to request calling a callback function
7188 void cb(int write_p, int version, int content_type,
7189 const void *buf, size_t len, SSL *ssl, void *arg)
7191 whenever a protocol message has been completely received
7192 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7193 protocol version according to which the SSL library interprets
7194 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7195 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7196 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7197 specification (change_cipher_spec(20), alert(21), handshake(22)).
7198 'buf' and 'len' point to the actual message, 'ssl' to the
7199 SSL object, and 'arg' is the application-defined value set by
7200 SSL[_CTX]_set_msg_callback_arg().
7202 'openssl s_client' and 'openssl s_server' have new '-msg' options
7203 to enable a callback that displays all protocol messages.
7206 *) Change the shared library support so shared libraries are built as
7207 soon as the corresponding static library is finished, and thereby get
7208 openssl and the test programs linked against the shared library.
7209 This still only happens when the keyword "shard" has been given to
7210 the configuration scripts.
7212 NOTE: shared library support is still an experimental thing, and
7213 backward binary compatibility is still not guaranteed.
7214 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7216 *) Add support for Subject Information Access extension.
7217 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7219 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7220 additional bytes when new memory had to be allocated, not just
7221 when reusing an existing buffer.
7224 *) New command line and configuration option 'utf8' for the req command.
7225 This allows field values to be specified as UTF8 strings.
7228 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7229 runs for the former and machine-readable output for the latter.
7232 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7233 of the e-mail address in the DN (i.e., it will go into a certificate
7234 extension only). The new configuration file option 'email_in_dn = no'
7235 has the same effect.
7236 [Massimiliano Pala madwolf@openca.org]
7238 *) Change all functions with names starting with des_ to be starting
7239 with DES_ instead. Add wrappers that are compatible with libdes,
7240 but are named _ossl_old_des_*. Finally, add macros that map the
7241 des_* symbols to the corresponding _ossl_old_des_* if libdes
7242 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7243 desired, the des_* symbols will be mapped to DES_*, with one
7246 Since we provide two compatibility mappings, the user needs to
7247 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7248 compatibility is desired. The default (i.e., when that macro
7249 isn't defined) is OpenSSL 0.9.6c compatibility.
7251 There are also macros that enable and disable the support of old
7252 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7253 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7254 are defined, the default will apply: to support the old des routines.
7256 In either case, one must include openssl/des.h to get the correct
7257 definitions. Do not try to just include openssl/des_old.h, that
7260 NOTE: This is a major break of an old API into a new one. Software
7261 authors are encouraged to switch to the DES_ style functions. Some
7262 time in the future, des_old.h and the libdes compatibility functions
7263 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7264 default), and then completely removed.
7267 *) Test for certificates which contain unsupported critical extensions.
7268 If such a certificate is found during a verify operation it is
7269 rejected by default: this behaviour can be overridden by either
7270 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7271 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7272 X509_supported_extension() has also been added which returns 1 if a
7273 particular extension is supported.
7276 *) Modify the behaviour of EVP cipher functions in similar way to digests
7277 to retain compatibility with existing code.
7280 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7281 compatibility with existing code. In particular the 'ctx' parameter does
7282 not have to be to be initialized before the call to EVP_DigestInit() and
7283 it is tidied up after a call to EVP_DigestFinal(). New function
7284 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7285 EVP_MD_CTX_copy() changed to not require the destination to be
7286 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7287 requires the destination to be valid.
7289 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7290 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7293 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7294 so that complete 'Handshake' protocol structures are kept in memory
7295 instead of overwriting 'msg_type' and 'length' with 'body' data.
7298 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7299 [Massimo Santin via Richard Levitte]
7301 *) Major restructuring to the underlying ENGINE code. This includes
7302 reduction of linker bloat, separation of pure "ENGINE" manipulation
7303 (initialisation, etc) from functionality dealing with implementations
7304 of specific crypto interfaces. This change also introduces integrated
7305 support for symmetric ciphers and digest implementations - so ENGINEs
7306 can now accelerate these by providing EVP_CIPHER and EVP_MD
7307 implementations of their own. This is detailed in crypto/engine/README
7308 as it couldn't be adequately described here. However, there are a few
7309 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7310 were changed in the original introduction of ENGINE code have now
7311 reverted back - the hooking from this code to ENGINE is now a good
7312 deal more passive and at run-time, operations deal directly with
7313 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7314 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7315 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7316 they were not being used by the framework as there is no concept of a
7317 BIGNUM_METHOD and they could not be generalised to the new
7318 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7319 ENGINE_cpy() has been removed as it cannot be consistently defined in
7323 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7326 *) Change mkdef.pl to sort symbols that get the same entry number,
7327 and make sure the automatically generated functions ERR_load_*
7328 become part of libeay.num as well.
7331 *) New function SSL_renegotiate_pending(). This returns true once
7332 renegotiation has been requested (either SSL_renegotiate() call
7333 or HelloRequest/ClientHello received from the peer) and becomes
7334 false once a handshake has been completed.
7335 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7336 sends a HelloRequest, but does not ensure that a handshake takes
7337 place. SSL_renegotiate_pending() is useful for checking if the
7338 client has followed the request.)
7341 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7342 By default, clients may request session resumption even during
7343 renegotiation (if session ID contexts permit); with this option,
7344 session resumption is possible only in the first handshake.
7346 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7347 more bits available for options that should not be part of
7348 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7351 *) Add some demos for certificate and certificate request creation.
7354 *) Make maximum certificate chain size accepted from the peer application
7355 settable (SSL*_get/set_max_cert_list()), as proposed by
7356 "Douglas E. Engert" <deengert@anl.gov>.
7359 *) Add support for shared libraries for Unixware-7
7360 (Boyd Lynn Gerber <gerberb@zenez.com>).
7363 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7364 be done prior to destruction. Use this to unload error strings from
7365 ENGINEs that load their own error strings. NB: This adds two new API
7366 functions to "get" and "set" this destroy handler in an ENGINE.
7369 *) Alter all existing ENGINE implementations (except "openssl" and
7370 "openbsd") to dynamically instantiate their own error strings. This
7371 makes them more flexible to be built both as statically-linked ENGINEs
7372 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7373 Also, add stub code to each that makes building them as self-contained
7374 shared-libraries easier (see README.ENGINE).
7377 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7378 implementations into applications that are completely implemented in
7379 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7380 commands that can be used to configure what shared-library to load and
7381 to control aspects of the way it is handled. Also, made an update to
7382 the README.ENGINE file that brings its information up-to-date and
7383 provides some information and instructions on the "dynamic" ENGINE
7384 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7387 *) Make it possible to unload ranges of ERR strings with a new
7388 "ERR_unload_strings" function.
7391 *) Add a copy() function to EVP_MD.
7394 *) Make EVP_MD routines take a context pointer instead of just the
7395 md_data void pointer.
7398 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7399 that the digest can only process a single chunk of data
7400 (typically because it is provided by a piece of
7401 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7402 is only going to provide a single chunk of data, and hence the
7403 framework needn't accumulate the data for oneshot drivers.
7406 *) As with "ERR", make it possible to replace the underlying "ex_data"
7407 functions. This change also alters the storage and management of global
7408 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7409 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7410 index counters. The API functions that use this state have been changed
7411 to take a "class_index" rather than pointers to the class's local STACK
7412 and counter, and there is now an API function to dynamically create new
7413 classes. This centralisation allows us to (a) plug a lot of the
7414 thread-safety problems that existed, and (b) makes it possible to clean
7415 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7416 such data would previously have always leaked in application code and
7417 workarounds were in place to make the memory debugging turn a blind eye
7418 to it. Application code that doesn't use this new function will still
7419 leak as before, but their memory debugging output will announce it now
7420 rather than letting it slide.
7422 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7423 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7424 has a return value to indicate success or failure.
7427 *) Make it possible to replace the underlying "ERR" functions such that the
7428 global state (2 LHASH tables and 2 locks) is only used by the "default"
7429 implementation. This change also adds two functions to "get" and "set"
7430 the implementation prior to it being automatically set the first time
7431 any other ERR function takes place. Ie. an application can call "get",
7432 pass the return value to a module it has just loaded, and that module
7433 can call its own "set" function using that value. This means the
7434 module's "ERR" operations will use (and modify) the error state in the
7435 application and not in its own statically linked copy of OpenSSL code.
7438 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7439 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7440 the operation, and provides a more encapsulated way for external code
7441 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7442 to use these functions rather than manually incrementing the counts.
7444 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7447 *) Add EVP test program.
7450 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7453 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7454 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7455 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7456 These allow a CRL to be built without having to access X509_CRL fields
7457 directly. Modify 'ca' application to use new functions.
7460 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7461 bug workarounds. Rollback attack detection is a security feature.
7462 The problem will only arise on OpenSSL servers when TLSv1 is not
7463 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7464 Software authors not wanting to support TLSv1 will have special reasons
7465 for their choice and can explicitly enable this option.
7466 [Bodo Moeller, Lutz Jaenicke]
7468 *) Rationalise EVP so it can be extended: don't include a union of
7469 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7470 (similar to those existing for EVP_CIPHER_CTX).
7475 EVP_MD_CTX_init(&md); /* new function call */
7476 EVP_DigestInit(&md, EVP_sha1());
7477 EVP_DigestUpdate(&md, in, len);
7478 EVP_DigestFinal(&md, out, NULL);
7479 EVP_MD_CTX_cleanup(&md); /* new function call */
7483 *) Make DES key schedule conform to the usual scheme, as well as
7484 correcting its structure. This means that calls to DES functions
7485 now have to pass a pointer to a des_key_schedule instead of a
7486 plain des_key_schedule (which was actually always a pointer
7489 des_key_schedule ks;
7491 des_set_key_checked(..., &ks);
7492 des_ncbc_encrypt(..., &ks, ...);
7494 (Note that a later change renames 'des_...' into 'DES_...'.)
7497 *) Initial reduction of linker bloat: the use of some functions, such as
7498 PEM causes large amounts of unused functions to be linked in due to
7499 poor organisation. For example pem_all.c contains every PEM function
7500 which has a knock on effect of linking in large amounts of (unused)
7501 ASN1 code. Grouping together similar functions and splitting unrelated
7502 functions prevents this.
7505 *) Cleanup of EVP macros.
7508 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7509 correct _ecb suffix.
7512 *) Add initial OCSP responder support to ocsp application. The
7513 revocation information is handled using the text based index
7514 use by the ca application. The responder can either handle
7515 requests generated internally, supplied in files (for example
7516 via a CGI script) or using an internal minimal server.
7519 *) Add configuration choices to get zlib compression for TLS.
7522 *) Changes to Kerberos SSL for RFC 2712 compliance:
7523 1. Implemented real KerberosWrapper, instead of just using
7524 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7525 2. Implemented optional authenticator field of KerberosWrapper.
7527 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7528 and authenticator structs; see crypto/krb5/.
7530 Generalized Kerberos calls to support multiple Kerberos libraries.
7531 [Vern Staats <staatsvr@asc.hpc.mil>,
7532 Jeffrey Altman <jaltman@columbia.edu>
7533 via Richard Levitte]
7535 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7536 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7537 values for each of the key sizes rather than having just
7538 parameters (and 'speed' generating keys each time).
7541 *) Speed up EVP routines.
7544 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7545 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7546 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7547 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7549 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7550 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7551 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7554 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7556 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7559 *) Added the OS2-EMX target.
7560 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7562 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7563 to support NCONF routines in extension code. New function CONF_set_nconf()
7564 to allow functions which take an NCONF to also handle the old LHASH
7565 structure: this means that the old CONF compatible routines can be
7566 retained (in particular wrt extensions) without having to duplicate the
7567 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7570 *) Enhance the general user interface with mechanisms for inner control
7571 and with possibilities to have yes/no kind of prompts.
7574 *) Change all calls to low level digest routines in the library and
7575 applications to use EVP. Add missing calls to HMAC_cleanup() and
7576 don't assume HMAC_CTX can be copied using memcpy().
7577 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7579 *) Add the possibility to control engines through control names but with
7580 arbitrary arguments instead of just a string.
7581 Change the key loaders to take a UI_METHOD instead of a callback
7582 function pointer. NOTE: this breaks binary compatibility with earlier
7583 versions of OpenSSL [engine].
7584 Adapt the nCipher code for these new conditions and add a card insertion
7588 *) Enhance the general user interface with mechanisms to better support
7589 dialog box interfaces, application-defined prompts, the possibility
7590 to use defaults (for example default passwords from somewhere else)
7591 and interrupts/cancellations.
7594 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7595 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7598 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7599 tidy up some unnecessarily weird code in 'sk_new()').
7600 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7602 *) Change the key loading routines for ENGINEs to use the same kind
7603 callback (pem_password_cb) as all other routines that need this
7607 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7608 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7609 than this minimum value is recommended.
7612 *) New random seeder for OpenVMS, using the system process statistics
7613 that are easily reachable.
7616 *) Windows apparently can't transparently handle global
7617 variables defined in DLLs. Initialisations such as:
7619 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7621 won't compile. This is used by the any applications that need to
7622 declare their own ASN1 modules. This was fixed by adding the option
7623 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7624 needed for static libraries under Win32.
7627 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7628 setting of purpose and trust fields. New X509_STORE trust and
7629 purpose functions and tidy up setting in other SSL functions.
7632 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7633 structure. These are inherited by X509_STORE_CTX when it is
7634 initialised. This allows various defaults to be set in the
7635 X509_STORE structure (such as flags for CRL checking and custom
7636 purpose or trust settings) for functions which only use X509_STORE_CTX
7637 internally such as S/MIME.
7639 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7640 trust settings if they are not set in X509_STORE. This allows X509_STORE
7641 purposes and trust (in S/MIME for example) to override any set by default.
7643 Add command line options for CRL checking to smime, s_client and s_server
7647 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7648 are set then the CRL is looked up in the X509_STORE structure and
7649 its validity and signature checked, then if the certificate is found
7650 in the CRL the verify fails with a revoked error.
7652 Various new CRL related callbacks added to X509_STORE_CTX structure.
7654 Command line options added to 'verify' application to support this.
7656 This needs some additional work, such as being able to handle multiple
7657 CRLs with different times, extension based lookup (rather than just
7658 by subject name) and ultimately more complete V2 CRL extension
7662 *) Add a general user interface API (crypto/ui/). This is designed
7663 to replace things like des_read_password and friends (backward
7664 compatibility functions using this new API are provided).
7665 The purpose is to remove prompting functions from the DES code
7666 section as well as provide for prompting through dialog boxes in
7667 a window system and the like.
7670 *) Add "ex_data" support to ENGINE so implementations can add state at a
7671 per-structure level rather than having to store it globally.
7674 *) Make it possible for ENGINE structures to be copied when retrieved by
7675 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7676 This causes the "original" ENGINE structure to act like a template,
7677 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7678 operational state can be localised to each ENGINE structure, despite the
7679 fact they all share the same "methods". New ENGINE structures returned in
7680 this case have no functional references and the return value is the single
7681 structural reference. This matches the single structural reference returned
7682 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7686 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7687 needs to match any other type at all we need to manually clear the
7691 *) Changes to the "openssl engine" utility to include;
7692 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7693 about an ENGINE's available control commands.
7694 - executing control commands from command line arguments using the
7695 '-pre' and '-post' switches. '-post' is only used if '-t' is
7696 specified and the ENGINE is successfully initialised. The syntax for
7697 the individual commands are colon-separated, for example;
7698 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7701 *) New dynamic control command support for ENGINEs. ENGINEs can now
7702 declare their own commands (numbers), names (strings), descriptions,
7703 and input types for run-time discovery by calling applications. A
7704 subset of these commands are implicitly classed as "executable"
7705 depending on their input type, and only these can be invoked through
7706 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7707 can be based on user input, config files, etc). The distinction is
7708 that "executable" commands cannot return anything other than a boolean
7709 result and can only support numeric or string input, whereas some
7710 discoverable commands may only be for direct use through
7711 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7712 pointers, or other custom uses. The "executable" commands are to
7713 support parameterisations of ENGINE behaviour that can be
7714 unambiguously defined by ENGINEs and used consistently across any
7715 OpenSSL-based application. Commands have been added to all the
7716 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7717 control over shared-library paths without source code alterations.
7720 *) Changed all ENGINE implementations to dynamically allocate their
7721 ENGINEs rather than declaring them statically. Apart from this being
7722 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7723 this also allows the implementations to compile without using the
7724 internal engine_int.h header.
7727 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7728 'const' value. Any code that should be able to modify a RAND_METHOD
7729 should already have non-const pointers to it (ie. they should only
7730 modify their own ones).
7733 *) Made a variety of little tweaks to the ENGINE code.
7734 - "atalla" and "ubsec" string definitions were moved from header files
7735 to C code. "nuron" string definitions were placed in variables
7736 rather than hard-coded - allowing parameterisation of these values
7737 later on via ctrl() commands.
7738 - Removed unused "#if 0"'d code.
7739 - Fixed engine list iteration code so it uses ENGINE_free() to release
7740 structural references.
7741 - Constified the RAND_METHOD element of ENGINE structures.
7742 - Constified various get/set functions as appropriate and added
7743 missing functions (including a catch-all ENGINE_cpy that duplicates
7744 all ENGINE values onto a new ENGINE except reference counts/state).
7745 - Removed NULL parameter checks in get/set functions. Setting a method
7746 or function to NULL is a way of cancelling out a previously set
7747 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7748 and doesn't justify the extra error symbols and code.
7749 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7750 flags from engine_int.h to engine.h.
7751 - Changed prototypes for ENGINE handler functions (init(), finish(),
7752 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7755 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7756 to the algorithm using long division. The binary algorithm can be
7757 used only if the modulus is odd. On 32-bit systems, it is faster
7758 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7759 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7760 up to 450 bits. In 64-bit environments, the binary algorithm
7761 appears to be advantageous for much longer moduli; here we use it
7762 for moduli up to 2048 bits.
7765 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7766 could not support the combine flag in choice fields.
7769 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7770 extensions from a certificate request to the certificate.
7773 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7774 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7775 file: this allows the display of the certificate about to be
7776 signed to be customised, to allow certain fields to be included
7777 or excluded and extension details. The old system didn't display
7778 multicharacter strings properly, omitted fields not in the policy
7779 and couldn't display additional details such as extensions.
7782 *) Function EC_POINTs_mul for multiple scalar multiplication
7783 of an arbitrary number of elliptic curve points
7784 \sum scalars[i]*points[i],
7785 optionally including the generator defined for the EC_GROUP:
7786 scalar*generator + \sum scalars[i]*points[i].
7788 EC_POINT_mul is a simple wrapper function for the typical case
7789 that the point list has just one item (besides the optional
7793 *) First EC_METHODs for curves over GF(p):
7795 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7796 operations and provides various method functions that can also
7797 operate with faster implementations of modular arithmetic.
7799 EC_GFp_mont_method() reuses most functions that are part of
7800 EC_GFp_simple_method, but uses Montgomery arithmetic.
7802 [Bodo Moeller; point addition and point doubling
7803 implementation directly derived from source code provided by
7804 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7806 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7807 crypto/ec/ec_lib.c):
7809 Curves are EC_GROUP objects (with an optional group generator)
7810 based on EC_METHODs that are built into the library.
7812 Points are EC_POINT objects based on EC_GROUP objects.
7814 Most of the framework would be able to handle curves over arbitrary
7815 finite fields, but as there are no obvious types for fields other
7816 than GF(p), some functions are limited to that for now.
7819 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7820 that the file contains a complete HTTP response.
7823 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7824 change the def and num file printf format specifier from "%-40sXXX"
7825 to "%-39s XXX". The latter will always guarantee a space after the
7826 field while the former will cause them to run together if the field
7827 is 40 of more characters long.
7830 *) Constify the cipher and digest 'method' functions and structures
7831 and modify related functions to take constant EVP_MD and EVP_CIPHER
7835 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7836 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7839 *) Modify EVP_Digest*() routines so they now return values. Although the
7840 internal software routines can never fail additional hardware versions
7844 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7846 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7847 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7850 ERR_R_NESTED_ASN1_ERROR
7852 ERR_R_MISSING_ASN1_EOS
7853 were 4 .. 9, conflicting with
7854 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7856 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7857 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7859 Add new error code 'ERR_R_INTERNAL_ERROR'.
7862 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7866 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7867 sets the subject name for a new request or supersedes the
7868 subject name in a given request. Formats that can be parsed are
7869 'CN=Some Name, OU=myOU, C=IT'
7871 'CN=Some Name/OU=myOU/C=IT'.
7873 Add options '-batch' and '-verbose' to 'openssl req'.
7874 [Massimiliano Pala <madwolf@hackmasters.net>]
7876 *) Introduce the possibility to access global variables through
7877 functions on platform were that's the best way to handle exporting
7878 global variables in shared libraries. To enable this functionality,
7879 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7880 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7881 is normally done by Configure or something similar).
7883 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7884 in the source file (foo.c) like this:
7886 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7887 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7889 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7890 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7892 OPENSSL_DECLARE_GLOBAL(int,foo);
7893 #define foo OPENSSL_GLOBAL_REF(foo)
7894 OPENSSL_DECLARE_GLOBAL(double,bar);
7895 #define bar OPENSSL_GLOBAL_REF(bar)
7897 The #defines are very important, and therefore so is including the
7898 header file everywhere where the defined globals are used.
7900 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7901 of ASN.1 items, but that structure is a bit different.
7903 The largest change is in util/mkdef.pl which has been enhanced with
7904 better and easier to understand logic to choose which symbols should
7905 go into the Windows .def files as well as a number of fixes and code
7906 cleanup (among others, algorithm keywords are now sorted
7907 lexicographically to avoid constant rewrites).
7910 *) In BN_div() keep a copy of the sign of 'num' before writing the
7911 result to 'rm' because if rm==num the value will be overwritten
7912 and produce the wrong result if 'num' is negative: this caused
7913 problems with BN_mod() and BN_nnmod().
7916 *) Function OCSP_request_verify(). This checks the signature on an
7917 OCSP request and verifies the signer certificate. The signer
7918 certificate is just checked for a generic purpose and OCSP request
7922 *) Add OCSP_check_validity() function to check the validity of OCSP
7923 responses. OCSP responses are prepared in real time and may only
7924 be a few seconds old. Simply checking that the current time lies
7925 between thisUpdate and nextUpdate max reject otherwise valid responses
7926 caused by either OCSP responder or client clock inaccuracy. Instead
7927 we allow thisUpdate and nextUpdate to fall within a certain period of
7928 the current time. The age of the response can also optionally be
7929 checked. Two new options -validity_period and -status_age added to
7933 *) If signature or public key algorithm is unrecognized print out its
7934 OID rather that just UNKNOWN.
7937 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7938 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7939 ID to be generated from the issuer certificate alone which can then be
7940 passed to OCSP_id_issuer_cmp().
7943 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7944 ASN1 modules to export functions returning ASN1_ITEM pointers
7945 instead of the ASN1_ITEM structures themselves. This adds several
7946 new macros which allow the underlying ASN1 function/structure to
7947 be accessed transparently. As a result code should not use ASN1_ITEM
7948 references directly (such as &X509_it) but instead use the relevant
7949 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7950 use of the new ASN1 code on platforms where exporting structures
7951 is problematical (for example in shared libraries) but exporting
7952 functions returning pointers to structures is not.
7955 *) Add support for overriding the generation of SSL/TLS session IDs.
7956 These callbacks can be registered either in an SSL_CTX or per SSL.
7957 The purpose of this is to allow applications to control, if they wish,
7958 the arbitrary values chosen for use as session IDs, particularly as it
7959 can be useful for session caching in multiple-server environments. A
7960 command-line switch for testing this (and any client code that wishes
7961 to use such a feature) has been added to "s_server".
7962 [Geoff Thorpe, Lutz Jaenicke]
7964 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7965 of the form '#if defined(...) || defined(...) || ...' and
7966 '#if !defined(...) && !defined(...) && ...'. This also avoids
7967 the growing number of special cases it was previously handling.
7970 *) Make all configuration macros available for application by making
7971 sure they are available in opensslconf.h, by giving them names starting
7972 with "OPENSSL_" to avoid conflicts with other packages and by making
7973 sure e_os2.h will cover all platform-specific cases together with
7975 Additionally, it is now possible to define configuration/platform-
7976 specific names (called "system identities"). In the C code, these
7977 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7978 macro with the name beginning with "OPENSSL_SYS_", which is determined
7979 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7983 *) New option -set_serial to 'req' and 'x509' this allows the serial
7984 number to use to be specified on the command line. Previously self
7985 signed certificates were hard coded with serial number 0 and the
7986 CA options of 'x509' had to use a serial number in a file which was
7990 *) New options to 'ca' utility to support V2 CRL entry extensions.
7991 Currently CRL reason, invalidity date and hold instruction are
7992 supported. Add new CRL extensions to V3 code and some new objects.
7995 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7996 disable standard block padding (aka PKCS#5 padding) in the EVP
7997 API, which was previously mandatory. This means that the data is
7998 not padded in any way and so the total length much be a multiple
7999 of the block size, otherwise an error occurs.
8002 *) Initial (incomplete) OCSP SSL support.
8005 *) New function OCSP_parse_url(). This splits up a URL into its host,
8006 port and path components: primarily to parse OCSP URLs. New -url
8007 option to ocsp utility.
8010 *) New nonce behavior. The return value of OCSP_check_nonce() now
8011 reflects the various checks performed. Applications can decide
8012 whether to tolerate certain situations such as an absent nonce
8013 in a response when one was present in a request: the ocsp application
8014 just prints out a warning. New function OCSP_add1_basic_nonce()
8015 this is to allow responders to include a nonce in a response even if
8016 the request is nonce-less.
8019 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8020 skipped when using openssl x509 multiple times on a single input file,
8021 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8024 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8025 set string type: to handle setting ASN1_TIME structures. Fix ca
8026 utility to correctly initialize revocation date of CRLs.
8029 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8030 the clients preferred ciphersuites and rather use its own preferences.
8031 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8032 Internet Explorer by ensuring unchanged hash method during stepup.
8033 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8036 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8037 to aes and add a new 'exist' option to print out symbols that don't
8041 *) Additional options to ocsp utility to allow flags to be set and
8042 additional certificates supplied.
8045 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8046 OCSP client a number of certificate to only verify the response
8050 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8051 handle the new API. Currently only ECB, CBC modes supported. Add new
8054 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8055 Encryption Standard (AES) Ciphersuites for Transport Layer
8056 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8057 not enabled by default and were not part of the "ALL" ciphersuite
8058 alias because they were not yet official; they could be
8059 explicitly requested by specifying the "AESdraft" ciphersuite
8060 group alias. In the final release of OpenSSL 0.9.7, the group
8061 alias is called "AES" and is part of "ALL".)
8062 [Ben Laurie, Steve Henson, Bodo Moeller]
8064 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8065 request to response.
8068 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8069 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8070 extract information from a certificate request. OCSP_response_create()
8071 creates a response and optionally adds a basic response structure.
8072 OCSP_basic_add1_status() adds a complete single response to a basic
8073 response and returns the OCSP_SINGLERESP structure just added (to allow
8074 extensions to be included for example). OCSP_basic_add1_cert() adds a
8075 certificate to a basic response and OCSP_basic_sign() signs a basic
8076 response with various flags. New helper functions ASN1_TIME_check()
8077 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8078 (converts ASN1_TIME to GeneralizedTime).
8081 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8082 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8083 structure from a certificate. X509_pubkey_digest() digests the public_key
8084 contents: this is used in various key identifiers.
8087 *) Make sk_sort() tolerate a NULL argument.
8088 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8090 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8091 passed by the function are trusted implicitly. If any of them signed the
8092 response then it is assumed to be valid and is not verified.
8095 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8096 to data. This was previously part of the PKCS7 ASN1 code. This
8097 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8098 [Steve Henson, reported by Kenneth R. Robinette
8099 <support@securenetterm.com>]
8101 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8102 routines: without these tracing memory leaks is very painful.
8103 Fix leaks in PKCS12 and PKCS7 routines.
8106 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8107 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8108 effectively meant GeneralizedTime would never be used. Now it
8109 is initialised to -1 but X509_time_adj() now has to check the value
8110 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8111 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8112 [Steve Henson, reported by Kenneth R. Robinette
8113 <support@securenetterm.com>]
8115 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8116 result in a zero length in the ASN1_INTEGER structure which was
8117 not consistent with the structure when d2i_ASN1_INTEGER() was used
8118 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8119 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8120 where it did not print out a minus for negative ASN1_INTEGER.
8123 *) Add summary printout to ocsp utility. The various functions which
8124 convert status values to strings have been renamed to:
8125 OCSP_response_status_str(), OCSP_cert_status_str() and
8126 OCSP_crl_reason_str() and are no longer static. New options
8127 to verify nonce values and to disable verification. OCSP response
8128 printout format cleaned up.
8131 *) Add additional OCSP certificate checks. These are those specified
8132 in RFC2560. This consists of two separate checks: the CA of the
8133 certificate being checked must either be the OCSP signer certificate
8134 or the issuer of the OCSP signer certificate. In the latter case the
8135 OCSP signer certificate must contain the OCSP signing extended key
8136 usage. This check is performed by attempting to match the OCSP
8137 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8138 in the OCSP_CERTID structures of the response.
8141 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8142 and related routines. This uses the standard OpenSSL certificate
8143 verify routines to perform initial checks (just CA validity) and
8144 to obtain the certificate chain. Then additional checks will be
8145 performed on the chain. Currently the root CA is checked to see
8146 if it is explicitly trusted for OCSP signing. This is used to set
8147 a root CA as a global signing root: that is any certificate that
8148 chains to that CA is an acceptable OCSP signing certificate.
8151 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8152 extensions from a separate configuration file.
8153 As when reading extensions from the main configuration file,
8154 the '-extensions ...' option may be used for specifying the
8156 [Massimiliano Pala <madwolf@comune.modena.it>]
8158 *) New OCSP utility. Allows OCSP requests to be generated or
8159 read. The request can be sent to a responder and the output
8160 parsed, outputed or printed in text form. Not complete yet:
8161 still needs to check the OCSP response validity.
8164 *) New subcommands for 'openssl ca':
8165 'openssl ca -status <serial>' prints the status of the cert with
8166 the given serial number (according to the index file).
8167 'openssl ca -updatedb' updates the expiry status of certificates
8169 [Massimiliano Pala <madwolf@comune.modena.it>]
8171 *) New '-newreq-nodes' command option to CA.pl. This is like
8172 '-newreq', but calls 'openssl req' with the '-nodes' option
8173 so that the resulting key is not encrypted.
8174 [Damien Miller <djm@mindrot.org>]
8176 *) New configuration for the GNU Hurd.
8177 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8179 *) Initial code to implement OCSP basic response verify. This
8180 is currently incomplete. Currently just finds the signer's
8181 certificate and verifies the signature on the response.
8184 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8185 value of OPENSSLDIR. This is available via the new '-d' option
8186 to 'openssl version', and is also included in 'openssl version -a'.
8189 *) Allowing defining memory allocation callbacks that will be given
8190 file name and line number information in additional arguments
8191 (a const char* and an int). The basic functionality remains, as
8192 well as the original possibility to just replace malloc(),
8193 realloc() and free() by functions that do not know about these
8194 additional arguments. To register and find out the current
8195 settings for extended allocation functions, the following
8196 functions are provided:
8198 CRYPTO_set_mem_ex_functions
8199 CRYPTO_set_locked_mem_ex_functions
8200 CRYPTO_get_mem_ex_functions
8201 CRYPTO_get_locked_mem_ex_functions
8203 These work the same way as CRYPTO_set_mem_functions and friends.
8204 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8205 extended allocation function is enabled.
8206 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8207 a conventional allocation function is enabled.
8208 [Richard Levitte, Bodo Moeller]
8210 *) Finish off removing the remaining LHASH function pointer casts.
8211 There should no longer be any prototype-casting required when using
8212 the LHASH abstraction, and any casts that remain are "bugs". See
8213 the callback types and macros at the head of lhash.h for details
8214 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8217 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8218 If /dev/[u]random devices are not available or do not return enough
8219 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8221 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8222 /etc/entropy will be queried once each in this sequence, querying stops
8223 when enough entropy was collected without querying more sockets.
8226 *) Change the Unix RAND_poll() variant to be able to poll several
8227 random devices, as specified by DEVRANDOM, until a sufficient amount
8228 of data has been collected. We spend at most 10 ms on each file
8229 (select timeout) and read in non-blocking mode. DEVRANDOM now
8230 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8231 (previously it was just the string "/dev/urandom"), so on typical
8232 platforms the 10 ms delay will never occur.
8233 Also separate out the Unix variant to its own file, rand_unix.c.
8234 For VMS, there's a currently-empty rand_vms.c.
8237 *) Move OCSP client related routines to ocsp_cl.c. These
8238 provide utility functions which an application needing
8239 to issue a request to an OCSP responder and analyse the
8240 response will typically need: as opposed to those which an
8241 OCSP responder itself would need which will be added later.
8243 OCSP_request_sign() signs an OCSP request with an API similar
8244 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8245 response. OCSP_response_get1_basic() extracts basic response
8246 from response. OCSP_resp_find_status(): finds and extracts status
8247 information from an OCSP_CERTID structure (which will be created
8248 when the request structure is built). These are built from lower
8249 level functions which work on OCSP_SINGLERESP structures but
8250 won't normally be used unless the application wishes to examine
8251 extensions in the OCSP response for example.
8253 Replace nonce routines with a pair of functions.
8254 OCSP_request_add1_nonce() adds a nonce value and optionally
8255 generates a random value. OCSP_check_nonce() checks the
8256 validity of the nonce in an OCSP response.
8259 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8260 This doesn't copy the supplied OCSP_CERTID and avoids the
8261 need to free up the newly created id. Change return type
8262 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8263 This can then be used to add extensions to the request.
8264 Deleted OCSP_request_new(), since most of its functionality
8265 is now in OCSP_REQUEST_new() (and the case insensitive name
8266 clash) apart from the ability to set the request name which
8267 will be added elsewhere.
8270 *) Update OCSP API. Remove obsolete extensions argument from
8271 various functions. Extensions are now handled using the new
8272 OCSP extension code. New simple OCSP HTTP function which
8273 can be used to send requests and parse the response.
8276 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8277 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8278 uses the special reorder version of SET OF to sort the attributes
8279 and reorder them to match the encoded order. This resolves a long
8280 standing problem: a verify on a PKCS7 structure just after signing
8281 it used to fail because the attribute order did not match the
8282 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8283 it uses the received order. This is necessary to tolerate some broken
8284 software that does not order SET OF. This is handled by encoding
8285 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8286 to produce the required SET OF.
8289 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8290 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8291 files to get correct declarations of the ASN.1 item variables.
8294 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8295 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8296 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8297 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8298 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8299 ASN1_ITEM and no wrapper functions.
8302 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8303 replace the old function pointer based I/O routines. Change most of
8304 the *_d2i_bio() and *_d2i_fp() functions to use these.
8307 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8308 lines, recognize more "algorithms" that can be deselected, and make
8309 it complain about algorithm deselection that isn't recognised.
8312 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8313 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8314 to use new functions. Add NO_ASN1_OLD which can be set to remove
8315 some old style ASN1 functions: this can be used to determine if old
8316 code will still work when these eventually go away.
8319 *) New extension functions for OCSP structures, these follow the
8320 same conventions as certificates and CRLs.
8323 *) New function X509V3_add1_i2d(). This automatically encodes and
8324 adds an extension. Its behaviour can be customised with various
8325 flags to append, replace or delete. Various wrappers added for
8326 certificates and CRLs.
8329 *) Fix to avoid calling the underlying ASN1 print routine when
8330 an extension cannot be parsed. Correct a typo in the
8331 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8334 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8335 entries for variables.
8338 *) Add functionality to apps/openssl.c for detecting locking
8339 problems: As the program is single-threaded, all we have
8340 to do is register a locking callback using an array for
8341 storing which locks are currently held by the program.
8344 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8345 SSL_get_ex_data_X509_STORE_idx(), which is used in
8346 ssl_verify_cert_chain() and thus can be called at any time
8347 during TLS/SSL handshakes so that thread-safety is essential.
8348 Unfortunately, the ex_data design is not at all suited
8349 for multi-threaded use, so it probably should be abolished.
8352 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8353 [Broadcom, tweaked and integrated by Geoff Thorpe]
8355 *) Move common extension printing code to new function
8356 X509V3_print_extensions(). Reorganise OCSP print routines and
8357 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8360 *) New function X509_signature_print() to remove duplication in some
8364 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8365 set (this was treated exactly the same as SET OF previously). This
8366 is used to reorder the STACK representing the structure to match the
8367 encoding. This will be used to get round a problem where a PKCS7
8368 structure which was signed could not be verified because the STACK
8369 order did not reflect the encoded order.
8372 *) Reimplement the OCSP ASN1 module using the new code.
8375 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8376 for its ASN1 operations. The old style function pointers still exist
8377 for now but they will eventually go away.
8380 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8381 completely replaces the old ASN1 functionality with a table driven
8382 encoder and decoder which interprets an ASN1_ITEM structure describing
8383 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8384 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8385 has also been converted to the new form.
8388 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8389 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8390 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8391 for negative moduli.
8394 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8395 of not touching the result's sign bit.
8398 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8402 *) Changed the LHASH code to use prototypes for callbacks, and created
8403 macros to declare and implement thin (optionally static) functions
8404 that provide type-safety and avoid function pointer casting for the
8405 type-specific callbacks.
8408 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8410 [Veers Staats <staatsvr@asc.hpc.mil>,
8411 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8413 *) Reformat the FAQ so the different questions and answers can be divided
8414 in sections depending on the subject.
8417 *) Have the zlib compression code load ZLIB.DLL dynamically under
8421 *) New function BN_mod_sqrt for computing square roots modulo a prime
8422 (using the probabilistic Tonelli-Shanks algorithm unless
8423 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8424 be handled deterministically).
8425 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8427 *) Make BN_mod_inverse faster by explicitly handling small quotients
8428 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8429 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8432 *) New function BN_kronecker.
8435 *) Fix BN_gcd so that it works on negative inputs; the result is
8436 positive unless both parameters are zero.
8437 Previously something reasonably close to an infinite loop was
8438 possible because numbers could be growing instead of shrinking
8439 in the implementation of Euclid's algorithm.
8442 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8443 sign of the number in question.
8445 Fix BN_is_word(a,w) to work correctly for w == 0.
8447 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8448 because its test if the absolute value of 'a' equals 'w'.
8449 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8450 it exists mostly for use in the implementations of BN_is_zero(),
8451 BN_is_one(), and BN_is_word().
8454 *) New function BN_swap.
8457 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8458 the exponentiation functions are more likely to produce reasonable
8459 results on negative inputs.
8462 *) Change BN_mod_mul so that the result is always non-negative.
8463 Previously, it could be negative if one of the factors was negative;
8464 I don't think anyone really wanted that behaviour.
8467 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8468 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8469 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8470 and add new functions:
8479 BN_mod_lshift1_quick
8483 These functions always generate non-negative results.
8485 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8486 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8488 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8489 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8490 be reduced modulo m.
8491 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8494 The following entry accidentally appeared in the CHANGES file
8495 distributed with OpenSSL 0.9.7. The modifications described in
8496 it do *not* apply to OpenSSL 0.9.7.
8498 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8499 was actually never needed) and in BN_mul(). The removal in BN_mul()
8500 required a small change in bn_mul_part_recursive() and the addition
8501 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8502 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8503 bn_sub_words() and bn_add_words() except they take arrays with
8508 *) In 'openssl passwd', verify passwords read from the terminal
8509 unless the '-salt' option is used (which usually means that
8510 verification would just waste user's time since the resulting
8511 hash is going to be compared with some given password hash)
8512 or the new '-noverify' option is used.
8514 This is an incompatible change, but it does not affect
8515 non-interactive use of 'openssl passwd' (passwords on the command
8516 line, '-stdin' option, '-in ...' option) and thus should not
8520 *) Remove all references to RSAref, since there's no more need for it.
8523 *) Make DSO load along a path given through an environment variable
8524 (SHLIB_PATH) with shl_load().
8527 *) Constify the ENGINE code as a result of BIGNUM constification.
8528 Also constify the RSA code and most things related to it. In a
8529 few places, most notable in the depth of the ASN.1 code, ugly
8530 casts back to non-const were required (to be solved at a later
8534 *) Make it so the openssl application has all engines loaded by default.
8537 *) Constify the BIGNUM routines a little more.
8540 *) Add the following functions:
8542 ENGINE_load_cswift()
8544 ENGINE_load_atalla()
8546 ENGINE_load_builtin_engines()
8548 That way, an application can itself choose if external engines that
8549 are built-in in OpenSSL shall ever be used or not. The benefit is
8550 that applications won't have to be linked with libdl or other dso
8551 libraries unless it's really needed.
8553 Changed 'openssl engine' to load all engines on demand.
8554 Changed the engine header files to avoid the duplication of some
8555 declarations (they differed!).
8558 *) 'openssl engine' can now list capabilities.
8561 *) Better error reporting in 'openssl engine'.
8564 *) Never call load_dh_param(NULL) in s_server.
8567 *) Add engine application. It can currently list engines by name and
8568 identity, and test if they are actually available.
8571 *) Improve RPM specification file by forcing symbolic linking and making
8572 sure the installed documentation is also owned by root.root.
8573 [Damien Miller <djm@mindrot.org>]
8575 *) Give the OpenSSL applications more possibilities to make use of
8576 keys (public as well as private) handled by engines.
8579 *) Add OCSP code that comes from CertCo.
8582 *) Add VMS support for the Rijndael code.
8585 *) Added untested support for Nuron crypto accelerator.
8588 *) Add support for external cryptographic devices. This code was
8589 previously distributed separately as the "engine" branch.
8590 [Geoff Thorpe, Richard Levitte]
8592 *) Rework the filename-translation in the DSO code. It is now possible to
8593 have far greater control over how a "name" is turned into a filename
8594 depending on the operating environment and any oddities about the
8595 different shared library filenames on each system.
8598 *) Support threads on FreeBSD-elf in Configure.
8601 *) Fix for SHA1 assembly problem with MASM: it produces
8602 warnings about corrupt line number information when assembling
8603 with debugging information. This is caused by the overlapping
8605 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8608 NCONF_get_number() has no error checking at all. As a replacement,
8609 NCONF_get_number_e() is defined (_e for "error checking") and is
8610 promoted strongly. The old NCONF_get_number is kept around for
8611 binary backward compatibility.
8612 Make it possible for methods to load from something other than a BIO,
8613 by providing a function pointer that is given a name instead of a BIO.
8614 For example, this could be used to load configuration data from an
8618 *) Fix for non blocking accept BIOs. Added new I/O special reason
8619 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8620 with non blocking I/O was not possible because no retry code was
8621 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8625 *) Added the beginnings of Rijndael support.
8628 *) Fix for bug in DirectoryString mask setting. Add support for
8629 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8630 to allow certificate printing to more controllable, additional
8631 'certopt' option to 'x509' to allow new printing options to be
8635 *) Clean old EAY MD5 hack from e_os.h.
8638 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8640 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8641 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8642 [Joe Orton, Steve Henson]
8644 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8646 *) Fix additional bug revealed by the NISCC test suite:
8648 Stop bug triggering large recursion when presented with
8649 certain ASN.1 tags (CVE-2003-0851)
8652 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8654 *) Fix various bugs revealed by running the NISCC test suite:
8656 Stop out of bounds reads in the ASN1 code when presented with
8657 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8659 If verify callback ignores invalid public key errors don't try to check
8660 certificate signature with the NULL public key.
8664 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8665 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8669 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8670 extra data after the compression methods not only for TLS 1.0
8671 but also for SSL 3.0 (as required by the specification).
8672 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8674 *) Change X509_certificate_type() to mark the key as exported/exportable
8675 when it's 512 *bits* long, not 512 bytes.
8678 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8680 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8681 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8682 a protocol version number mismatch like a decryption error
8683 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8686 *) Turn on RSA blinding by default in the default implementation
8687 to avoid a timing attack. Applications that don't want it can call
8688 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8689 They would be ill-advised to do so in most cases.
8690 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8692 *) Change RSA blinding code so that it works when the PRNG is not
8693 seeded (in this case, the secret RSA exponent is abused as
8694 an unpredictable seed -- if it is not unpredictable, there
8695 is no point in blinding anyway). Make RSA blinding thread-safe
8696 by remembering the creator's thread ID in rsa->blinding and
8697 having all other threads use local one-time blinding factors
8698 (this requires more computation than sharing rsa->blinding, but
8699 avoids excessive locking; and if an RSA object is not shared
8700 between threads, blinding will still be very fast).
8703 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8705 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8706 via timing by performing a MAC computation even if incorrect
8707 block cipher padding has been found. This is a countermeasure
8708 against active attacks where the attacker has to distinguish
8709 between bad padding and a MAC verification error. (CVE-2003-0078)
8711 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8712 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8713 Martin Vuagnoux (EPFL, Ilion)]
8715 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8717 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8718 memory from it's contents. This is done with a counter that will
8719 place alternating values in each byte. This can be used to solve
8720 two issues: 1) the removal of calls to memset() by highly optimizing
8721 compilers, and 2) cleansing with other values than 0, since those can
8722 be read through on certain media, for example a swap space on disk.
8725 *) Bugfix: client side session caching did not work with external caching,
8726 because the session->cipher setting was not restored when reloading
8727 from the external cache. This problem was masked, when
8728 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8729 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8732 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8733 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8734 [Zeev Lieber <zeev-l@yahoo.com>]
8736 *) Undo an undocumented change introduced in 0.9.6e which caused
8737 repeated calls to OpenSSL_add_all_ciphers() and
8738 OpenSSL_add_all_digests() to be ignored, even after calling
8742 *) Change the default configuration reader to deal with last line not
8743 being properly terminated.
8746 *) Change X509_NAME_cmp() so it applies the special rules on handling
8747 DN values that are of type PrintableString, as well as RDNs of type
8748 emailAddress where the value has the type ia5String.
8749 [stefank@valicert.com via Richard Levitte]
8751 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8752 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8753 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8754 the bitwise-OR of the two for use by the majority of applications
8755 wanting this behaviour, and update the docs. The documented
8756 behaviour and actual behaviour were inconsistent and had been
8757 changing anyway, so this is more a bug-fix than a behavioural
8759 [Geoff Thorpe, diagnosed by Nadav Har'El]
8761 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8762 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8765 *) Fix initialization code race conditions in
8766 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8767 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8768 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8769 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8770 ssl2_get_cipher_by_char(),
8771 ssl3_get_cipher_by_char().
8772 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8774 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8775 the cached sessions are flushed, as the remove_cb() might use ex_data
8776 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8777 (see [openssl.org #212]).
8778 [Geoff Thorpe, Lutz Jaenicke]
8780 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8781 length, instead of the encoding length to d2i_ASN1_OBJECT.
8784 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8786 *) [In 0.9.6g-engine release:]
8787 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8788 [Lynn Gazis <lgazis@rainbow.com>]
8790 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8792 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8793 and get fix the header length calculation.
8794 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8795 Alon Kantor <alonk@checkpoint.com> (and others),
8798 *) Use proper error handling instead of 'assertions' in buffer
8799 overflow checks added in 0.9.6e. This prevents DoS (the
8800 assertions could call abort()).
8801 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8803 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8805 *) Add various sanity checks to asn1_get_length() to reject
8806 the ASN1 length bytes if they exceed sizeof(long), will appear
8807 negative or the content length exceeds the length of the
8809 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8811 *) Fix cipher selection routines: ciphers without encryption had no flags
8812 for the cipher strength set and where therefore not handled correctly
8813 by the selection routines (PR #130).
8816 *) Fix EVP_dsa_sha macro.
8820 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8821 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8822 that was added in OpenSSL 0.9.6d.
8824 As the countermeasure turned out to be incompatible with some
8825 broken SSL implementations, the new option is part of SSL_OP_ALL.
8826 SSL_OP_ALL is usually employed when compatibility with weird SSL
8827 implementations is desired (e.g. '-bugs' option to 's_client' and
8828 's_server'), so the new option is automatically set in many
8832 *) Changes in security patch:
8834 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8835 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8836 Air Force Materiel Command, USAF, under agreement number
8839 *) Add various sanity checks to asn1_get_length() to reject
8840 the ASN1 length bytes if they exceed sizeof(long), will appear
8841 negative or the content length exceeds the length of the
8842 supplied buffer. (CVE-2002-0659)
8843 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8845 *) Assertions for various potential buffer overflows, not known to
8847 [Ben Laurie (CHATS)]
8849 *) Various temporary buffers to hold ASCII versions of integers were
8850 too small for 64 bit platforms. (CVE-2002-0655)
8851 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8853 *) Remote buffer overflow in SSL3 protocol - an attacker could
8854 supply an oversized session ID to a client. (CVE-2002-0656)
8855 [Ben Laurie (CHATS)]
8857 *) Remote buffer overflow in SSL2 protocol - an attacker could
8858 supply an oversized client master key. (CVE-2002-0656)
8859 [Ben Laurie (CHATS)]
8861 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8863 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8864 encoded as NULL) with id-dsa-with-sha1.
8865 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8867 *) Check various X509_...() return values in apps/req.c.
8868 [Nils Larsch <nla@trustcenter.de>]
8870 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8871 an end-of-file condition would erroneously be flagged, when the CRLF
8872 was just at the end of a processed block. The bug was discovered when
8873 processing data through a buffering memory BIO handing the data to a
8874 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8875 <ptsekov@syntrex.com> and Nedelcho Stanev.
8878 *) Implement a countermeasure against a vulnerability recently found
8879 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8880 before application data chunks to avoid the use of known IVs
8881 with data potentially chosen by the attacker.
8884 *) Fix length checks in ssl3_get_client_hello().
8887 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8888 to prevent ssl3_read_internal() from incorrectly assuming that
8889 ssl3_read_bytes() found application data while handshake
8890 processing was enabled when in fact s->s3->in_read_app_data was
8891 merely automatically cleared during the initial handshake.
8892 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8894 *) Fix object definitions for Private and Enterprise: they were not
8895 recognized in their shortname (=lowercase) representation. Extend
8896 obj_dat.pl to issue an error when using undefined keywords instead
8897 of silently ignoring the problem (Svenning Sorensen
8898 <sss@sss.dnsalias.net>).
8901 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8902 generators, i.e. generators other than 2 and 5. (Previously, the
8903 code did not properly initialise the 'add' and 'rem' values to
8904 BN_generate_prime().)
8906 In the new general case, we do not insist that 'generator' is
8907 actually a primitive root: This requirement is rather pointless;
8908 a generator of the order-q subgroup is just as good, if not
8912 *) Map new X509 verification errors to alerts. Discovered and submitted by
8913 Tom Wu <tom@arcot.com>.
8916 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8917 returning non-zero before the data has been completely received
8918 when using non-blocking I/O.
8919 [Bodo Moeller; problem pointed out by John Hughes]
8921 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8922 [Ben Laurie, Lutz Jaenicke]
8924 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8925 Yoram Zahavi <YoramZ@gilian.com>).
8928 *) Add information about CygWin 1.3 and on, and preserve proper
8929 configuration for the versions before that.
8930 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8932 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8933 check whether we deal with a copy of a session and do not delete from
8934 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8935 <izhar@checkpoint.com>.
8938 *) Do not store session data into the internal session cache, if it
8939 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8940 flag is set). Proposed by Aslam <aslam@funk.com>.
8943 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8947 *) [In 0.9.6d-engine release:]
8948 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8949 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8951 *) Add the configuration target linux-s390x.
8952 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8954 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8955 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8956 variable as an indication that a ClientHello message has been
8957 received. As the flag value will be lost between multiple
8958 invocations of ssl3_accept when using non-blocking I/O, the
8959 function may not be aware that a handshake has actually taken
8960 place, thus preventing a new session from being added to the
8963 To avoid this problem, we now set s->new_session to 2 instead of
8964 using a local variable.
8965 [Lutz Jaenicke, Bodo Moeller]
8967 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8968 if the SSL_R_LENGTH_MISMATCH error is detected.
8969 [Geoff Thorpe, Bodo Moeller]
8971 *) New 'shared_ldflag' column in Configure platform table.
8974 *) Fix EVP_CIPHER_mode macro.
8975 ["Dan S. Camper" <dan@bti.net>]
8977 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8978 type, we must throw them away by setting rr->length to 0.
8979 [D P Chang <dpc@qualys.com>]
8981 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8983 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8984 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8985 worked incorrectly for those cases where range = 10..._2 and
8986 3*range is two bits longer than range.)
8989 *) Only add signing time to PKCS7 structures if it is not already
8993 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8994 OBJ_ld_ce should be OBJ_id_ce.
8995 Also some ip-pda OIDs in crypto/objects/objects.txt were
8996 incorrect (cf. RFC 3039).
8997 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8999 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9000 returns early because it has nothing to do.
9001 [Andy Schneider <andy.schneider@bjss.co.uk>]
9003 *) [In 0.9.6c-engine release:]
9004 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9005 [Andy Schneider <andy.schneider@bjss.co.uk>]
9007 *) [In 0.9.6c-engine release:]
9008 Add support for Cryptographic Appliance's keyserver technology.
9009 (Use engine 'keyclient')
9010 [Cryptographic Appliances and Geoff Thorpe]
9012 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9013 is called via tools/c89.sh because arguments have to be
9014 rearranged (all '-L' options must appear before the first object
9016 [Richard Shapiro <rshapiro@abinitio.com>]
9018 *) [In 0.9.6c-engine release:]
9019 Add support for Broadcom crypto accelerator cards, backported
9021 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9023 *) [In 0.9.6c-engine release:]
9024 Add support for SureWare crypto accelerator cards from
9025 Baltimore Technologies. (Use engine 'sureware')
9026 [Baltimore Technologies and Mark Cox]
9028 *) [In 0.9.6c-engine release:]
9029 Add support for crypto accelerator cards from Accelerated
9030 Encryption Processing, www.aep.ie. (Use engine 'aep')
9031 [AEP Inc. and Mark Cox]
9033 *) Add a configuration entry for gcc on UnixWare.
9034 [Gary Benson <gbenson@redhat.com>]
9036 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9037 messages are stored in a single piece (fixed-length part and
9038 variable-length part combined) and fix various bugs found on the way.
9041 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9042 instead. BIO_gethostbyname() does not know what timeouts are
9043 appropriate, so entries would stay in cache even when they have
9045 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9047 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9048 faced with a pathologically small ClientHello fragment that does
9049 not contain client_version: Instead of aborting with an error,
9050 simply choose the highest available protocol version (i.e.,
9051 TLS 1.0 unless it is disabled). In practice, ClientHello
9052 messages are never sent like this, but this change gives us
9053 strictly correct behaviour at least for TLS.
9056 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9057 never resets s->method to s->ctx->method when called from within
9058 one of the SSL handshake functions.
9059 [Bodo Moeller; problem pointed out by Niko Baric]
9061 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9062 (sent using the client's version number) if client_version is
9063 smaller than the protocol version in use. Also change
9064 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9065 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9066 the client will at least see that alert.
9069 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9073 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9074 client receives HelloRequest while in a handshake.
9075 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9077 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9078 should end in 'break', not 'goto end' which circumvents various
9079 cleanups done in state SSL_ST_OK. But session related stuff
9080 must be disabled for SSL_ST_OK in the case that we just sent a
9083 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9084 before just sending a HelloRequest.
9085 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9087 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9088 reveal whether illegal block cipher padding was found or a MAC
9089 verification error occurred. (Neither SSLerr() codes nor alerts
9090 are directly visible to potential attackers, but the information
9091 may leak via logfiles.)
9093 Similar changes are not required for the SSL 2.0 implementation
9094 because the number of padding bytes is sent in clear for SSL 2.0,
9095 and the extra bytes are just ignored. However ssl/s2_pkt.c
9096 failed to verify that the purported number of padding bytes is in
9100 *) Add OpenUNIX-8 support including shared libraries
9101 (Boyd Lynn Gerber <gerberb@zenez.com>).
9104 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9105 'wristwatch attack' using huge encoding parameters (cf.
9106 James H. Manger's CRYPTO 2001 paper). Note that the
9107 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9108 encoding parameters and hence was not vulnerable.
9111 *) BN_sqr() bug fix.
9112 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9114 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9115 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9116 followed by modular reduction.
9117 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9119 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9120 equivalent based on BN_pseudo_rand() instead of BN_rand().
9123 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9124 This function was broken, as the check for a new client hello message
9125 to handle SGC did not allow these large messages.
9126 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9129 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9132 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9133 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9136 *) Rework the configuration and shared library support for Tru64 Unix.
9137 The configuration part makes use of modern compiler features and
9138 still retains old compiler behavior for those that run older versions
9139 of the OS. The shared library support part includes a variant that
9140 uses the RPATH feature, and is available through the special
9141 configuration target "alpha-cc-rpath", which will never be selected
9143 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9145 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9146 with the same message size as in ssl3_get_certificate_request().
9147 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9148 messages might inadvertently be reject as too long.
9149 [Petr Lampa <lampa@fee.vutbr.cz>]
9151 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9154 *) Modified SSL library such that the verify_callback that has been set
9155 specificly for an SSL object with SSL_set_verify() is actually being
9156 used. Before the change, a verify_callback set with this function was
9157 ignored and the verify_callback() set in the SSL_CTX at the time of
9158 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9159 to allow the necessary settings.
9162 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9163 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9164 done automatically (in contradiction to the requirements of the C
9165 standard). This made problems when used from OpenSSH.
9168 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9169 dh->length and always used
9171 BN_rand_range(priv_key, dh->p).
9173 BN_rand_range() is not necessary for Diffie-Hellman, and this
9174 specific range makes Diffie-Hellman unnecessarily inefficient if
9175 dh->length (recommended exponent length) is much smaller than the
9176 length of dh->p. We could use BN_rand_range() if the order of
9177 the subgroup was stored in the DH structure, but we only have
9182 BN_rand(priv_key, l, ...)
9184 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9190 RSA_eay_public_encrypt
9191 RSA_eay_private_decrypt
9192 RSA_eay_private_encrypt (signing)
9193 RSA_eay_public_decrypt (signature verification)
9195 (default implementations for RSA_public_encrypt,
9196 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9197 always reject numbers >= n.
9200 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9201 to synchronize access to 'locking_thread'. This is necessary on
9202 systems where access to 'locking_thread' (an 'unsigned long'
9203 variable) is not atomic.
9206 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9207 *before* setting the 'crypto_lock_rand' flag. The previous code had
9208 a race condition if 0 is a valid thread ID.
9209 [Travis Vitek <vitek@roguewave.com>]
9211 *) Add support for shared libraries under Irix.
9212 [Albert Chin-A-Young <china@thewrittenword.com>]
9214 *) Add configuration option to build on Linux on both big-endian and
9216 [Ralf Baechle <ralf@uni-koblenz.de>]
9218 *) Add the possibility to create shared libraries on HP-UX.
9221 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9223 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9224 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9225 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9226 PRNG state recovery was possible based on the output of
9227 one PRNG request appropriately sized to gain knowledge on
9228 'md' followed by enough consecutive 1-byte PRNG requests
9229 to traverse all of 'state'.
9231 1. When updating 'md_local' (the current thread's copy of 'md')
9232 during PRNG output generation, hash all of the previous
9233 'md_local' value, not just the half used for PRNG output.
9235 2. Make the number of bytes from 'state' included into the hash
9236 independent from the number of PRNG bytes requested.
9238 The first measure alone would be sufficient to avoid
9239 Markku-Juhani's attack. (Actually it had never occurred
9240 to me that the half of 'md_local' used for chaining was the
9241 half from which PRNG output bytes were taken -- I had always
9242 assumed that the secret half would be used.) The second
9243 measure makes sure that additional data from 'state' is never
9244 mixed into 'md_local' in small portions; this heuristically
9245 further strengthens the PRNG.
9248 *) Fix crypto/bn/asm/mips3.s.
9251 *) When only the key is given to "enc", the IV is undefined. Print out
9252 an error message in this case.
9255 *) Handle special case when X509_NAME is empty in X509 printing routines.
9258 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9259 positive and less than q.
9262 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9263 used: it isn't thread safe and the add_lock_callback should handle
9265 [Paul Rose <Paul.Rose@bridge.com>]
9267 *) Verify that incoming data obeys the block size in
9268 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9272 [Ulf Möller, Bodo Möller]
9274 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9275 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9276 when fixing the server behaviour for backwards-compatible 'client
9277 hello' messages. (Note that the attack is impractical against
9278 SSL 3.0 and TLS 1.0 anyway because length and version checking
9279 means that the probability of guessing a valid ciphertext is
9280 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9283 Before 0.9.5, the countermeasure (hide the error by generating a
9284 random 'decryption result') did not work properly because
9285 ERR_clear_error() was missing, meaning that SSL_get_error() would
9286 detect the supposedly ignored error.
9288 Both problems are now fixed.
9291 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9292 (previously it was 1024).
9295 *) Fix for compatibility mode trust settings: ignore trust settings
9296 unless some valid trust or reject settings are present.
9299 *) Fix for blowfish EVP: its a variable length cipher.
9302 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9303 parameters in DSA public key structures and return an error in the
9304 DSA routines if parameters are absent.
9307 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9308 in the current directory if neither $RANDFILE nor $HOME was set.
9309 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9310 caused some confusion to Windows users who haven't defined $HOME.
9311 Thus RAND_file_name() is changed again: e_os.h can define a
9312 DEFAULT_HOME, which will be used if $HOME is not set.
9313 For Windows, we use "C:"; on other platforms, we still require
9314 environment variables.
9316 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9317 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9318 having multiple threads call RAND_poll() concurrently.
9321 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9322 combination of a flag and a thread ID variable.
9323 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9324 flag), *other* threads can enter ssleay_add_bytes without obeying
9325 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9326 that they do not hold after the first thread unsets add_do_not_lock).
9329 *) Change bctest again: '-x' expressions are not available in all
9333 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9335 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9336 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9338 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9339 the default extension for executables, if any. Also, make the perl
9340 scripts that use symlink() to test if it really exists and use "cp"
9341 if it doesn't. All this made OpenSSL compilable and installable in
9345 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9346 If SEQUENCE is length is indefinite just set c->slen to the total
9347 amount of data available.
9348 [Steve Henson, reported by shige@FreeBSD.org]
9349 [This change does not apply to 0.9.7.]
9351 *) Change bctest to avoid here-documents inside command substitution
9352 (workaround for FreeBSD /bin/sh bug).
9353 For compatibility with Ultrix, avoid shell functions (introduced
9354 in the bctest version that searches along $PATH).
9357 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9358 with des_encrypt() defined on some operating systems, like Solaris
9362 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9363 On the Importance of Eliminating Errors in Cryptographic
9364 Computations, J. Cryptology 14 (2001) 2, 101-119,
9365 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9368 *) MIPS assembler BIGNUM division bug fix.
9371 *) Disabled incorrect Alpha assembler code.
9374 *) Fix PKCS#7 decode routines so they correctly update the length
9375 after reading an EOC for the EXPLICIT tag.
9377 [This change does not apply to 0.9.7.]
9379 *) Fix bug in PKCS#12 key generation routines. This was triggered
9380 if a 3DES key was generated with a 0 initial byte. Include
9381 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9382 (but broken) behaviour.
9385 *) Enhance bctest to search for a working bc along $PATH and print
9387 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9389 *) Fix memory leaks in err.c: free err_data string if necessary;
9390 don't write to the wrong index in ERR_set_error_data.
9393 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9397 *) Replace rdtsc with _emit statements for VC++ version 5.
9398 [Jeremy Cooper <jeremy@baymoo.org>]
9400 *) Make it possible to reuse SSLv2 sessions.
9403 *) In copy_email() check for >= 0 as a return value for
9404 X509_NAME_get_index_by_NID() since 0 is a valid index.
9405 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9407 *) Avoid coredump with unsupported or invalid public keys by checking if
9408 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9409 PKCS7_verify() fails with non detached data.
9412 *) Don't use getenv in library functions when run as setuid/setgid.
9413 New function OPENSSL_issetugid().
9416 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9417 due to incorrect handling of multi-threading:
9419 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9421 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9423 3. Count how many times MemCheck_off() has been called so that
9424 nested use can be treated correctly. This also avoids
9425 inband-signalling in the previous code (which relied on the
9426 assumption that thread ID 0 is impossible).
9429 *) Add "-rand" option also to s_client and s_server.
9432 *) Fix CPU detection on Irix 6.x.
9433 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9434 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9436 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9439 [This change does not apply to 0.9.7.]
9441 *) Use the cached encoding of an X509_NAME structure rather than
9442 copying it. This is apparently the reason for the libsafe "errors"
9443 but the code is actually correct.
9446 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9447 Bleichenbacher's DSA attack.
9448 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9449 to be set and top=0 forces the highest bit to be set; top=-1 is new
9450 and leaves the highest bit random.
9451 [Ulf Moeller, Bodo Moeller]
9453 *) In the NCONF_...-based implementations for CONF_... queries
9454 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9455 a temporary CONF structure with the data component set to NULL
9456 (which gives segmentation faults in lh_retrieve).
9457 Instead, use NULL for the CONF pointer in CONF_get_string and
9458 CONF_get_number (which may use environment variables) and directly
9459 return NULL from CONF_get_section.
9462 *) Fix potential buffer overrun for EBCDIC.
9465 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9466 keyUsage if basicConstraints absent for a CA.
9469 *) Make SMIME_write_PKCS7() write mail header values with a format that
9470 is more generally accepted (no spaces before the semicolon), since
9471 some programs can't parse those values properly otherwise. Also make
9472 sure BIO's that break lines after each write do not create invalid
9476 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9477 macros previously used would not encode an empty SEQUENCE OF
9478 and break the signature.
9480 [This change does not apply to 0.9.7.]
9482 *) Zero the premaster secret after deriving the master secret in
9486 *) Add some EVP_add_digest_alias registrations (as found in
9487 OpenSSL_add_all_digests()) to SSL_library_init()
9488 aka OpenSSL_add_ssl_algorithms(). This provides improved
9489 compatibility with peers using X.509 certificates
9490 with unconventional AlgorithmIdentifier OIDs.
9493 *) Fix for Irix with NO_ASM.
9494 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9496 *) ./config script fixes.
9497 [Ulf Moeller, Richard Levitte]
9499 *) Fix 'openssl passwd -1'.
9502 *) Change PKCS12_key_gen_asc() so it can cope with non null
9503 terminated strings whose length is passed in the passlen
9504 parameter, for example from PEM callbacks. This was done
9505 by adding an extra length parameter to asc2uni().
9506 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9508 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9509 call failed, free the DSA structure.
9512 *) Fix to uni2asc() to cope with zero length Unicode strings.
9513 These are present in some PKCS#12 files.
9516 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9517 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9518 when writing a 32767 byte record.
9519 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9521 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9522 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9524 (RSA objects have a reference count access to which is protected
9525 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9526 so they are meant to be shared between threads.)
9527 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9528 "Reddie, Steven" <Steven.Reddie@ca.com>]
9530 *) Fix a deadlock in CRYPTO_mem_leaks().
9533 *) Use better test patterns in bntest.
9536 *) rand_win.c fix for Borland C.
9539 *) BN_rshift bugfix for n == 0.
9542 *) Add a 'bctest' script that checks for some known 'bc' bugs
9543 so that 'make test' does not abort just because 'bc' is broken.
9546 *) Store verify_result within SSL_SESSION also for client side to
9547 avoid potential security hole. (Re-used sessions on the client side
9548 always resulted in verify_result==X509_V_OK, not using the original
9549 result of the server certificate verification.)
9552 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9553 SSL3_RT_APPLICATION_DATA, return 0.
9554 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9558 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9559 releases, have been re-implemented by renaming the previous
9560 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9561 and ssl3_read_internal, respectively, and adding 'peek' parameters
9562 to them. The new ssl[23]_{read,peek} functions are calls to
9563 ssl[23]_read_internal with the 'peek' flag set appropriately.
9564 A 'peek' parameter has also been added to ssl3_read_bytes, which
9565 does the actual work for ssl3_read_internal.
9568 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9569 the method-specific "init()" handler. Also clean up ex_data after
9570 calling the method-specific "finish()" handler. Previously, this was
9571 happening the other way round.
9574 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9575 The previous value, 12, was not always sufficient for BN_mod_exp().
9578 *) Make sure that shared libraries get the internal name engine with
9579 the full version number and not just 0. This should mark the
9580 shared libraries as not backward compatible. Of course, this should
9581 be changed again when we can guarantee backward binary compatibility.
9584 *) Fix typo in get_cert_by_subject() in by_dir.c
9585 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9587 *) Rework the system to generate shared libraries:
9589 - Make note of the expected extension for the shared libraries and
9590 if there is a need for symbolic links from for example libcrypto.so.0
9591 to libcrypto.so.0.9.7. There is extended info in Configure for
9594 - Make as few rebuilds of the shared libraries as possible.
9596 - Still avoid linking the OpenSSL programs with the shared libraries.
9598 - When installing, install the shared libraries separately from the
9602 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9604 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9605 and not in SSL_clear because the latter is also used by the
9606 accept/connect functions; previously, the settings made by
9607 SSL_set_read_ahead would be lost during the handshake.
9608 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9610 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9611 Previously, it would create entries for disabled algorithms no
9615 *) Added several new manual pages for SSL_* function.
9618 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9620 *) In ssl23_get_client_hello, generate an error message when faced
9621 with an initial SSL 3.0/TLS record that is too small to contain the
9622 first two bytes of the ClientHello message, i.e. client_version.
9623 (Note that this is a pathologic case that probably has never happened
9624 in real life.) The previous approach was to use the version number
9625 from the record header as a substitute; but our protocol choice
9626 should not depend on that one because it is not authenticated
9627 by the Finished messages.
9630 *) More robust randomness gathering functions for Windows.
9631 [Jeffrey Altman <jaltman@columbia.edu>]
9633 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9634 not set then we don't setup the error code for issuer check errors
9635 to avoid possibly overwriting other errors which the callback does
9636 handle. If an application does set the flag then we assume it knows
9637 what it is doing and can handle the new informational codes
9641 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9642 a general "ANY" type, as such it should be able to decode anything
9643 including tagged types. However it didn't check the class so it would
9644 wrongly interpret tagged types in the same way as their universal
9645 counterpart and unknown types were just rejected. Changed so that the
9646 tagged and unknown types are handled in the same way as a SEQUENCE:
9647 that is the encoding is stored intact. There is also a new type
9648 "V_ASN1_OTHER" which is used when the class is not universal, in this
9649 case we have no idea what the actual type is so we just lump them all
9653 *) On VMS, stdout may very well lead to a file that is written to
9654 in a record-oriented fashion. That means that every write() will
9655 write a separate record, which will be read separately by the
9656 programs trying to read from it. This can be very confusing.
9658 The solution is to put a BIO filter in the way that will buffer
9659 text until a linefeed is reached, and then write everything a
9660 line at a time, so every record written will be an actual line,
9661 not chunks of lines and not (usually doesn't happen, but I've
9662 seen it once) several lines in one record. BIO_f_linebuffer() is
9665 Currently, it's a VMS-only method, because that's where it has
9666 been tested well enough.
9669 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9670 it can return incorrect results.
9671 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9672 but it was in 0.9.6-beta[12].)
9675 *) Disable the check for content being present when verifying detached
9676 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9677 include zero length content when signing messages.
9680 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9681 BIO_ctrl (for BIO pairs).
9684 *) Add DSO method for VMS.
9687 *) Bug fix: Montgomery multiplication could produce results with the
9691 *) Add RPM specification openssl.spec and modify it to build three
9692 packages. The default package contains applications, application
9693 documentation and run-time libraries. The devel package contains
9694 include files, static libraries and function documentation. The
9695 doc package contains the contents of the doc directory. The original
9696 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9699 *) Add a large number of documentation files for many SSL routines.
9700 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9702 *) Add a configuration entry for Sony News 4.
9703 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9705 *) Don't set the two most significant bits to one when generating a
9706 random number < q in the DSA library.
9709 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9710 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9711 the underlying transport is blocking) if a handshake took place.
9712 (The default behaviour is needed by applications such as s_client
9713 and s_server that use select() to determine when to use SSL_read;
9714 but for applications that know in advance when to expect data, it
9715 just makes things more complicated.)
9718 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9722 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9723 work better on such systems.
9724 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9726 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9727 Update PKCS12_parse() so it copies the friendlyName and the
9728 keyid to the certificates aux info.
9731 *) Fix bug in PKCS7_verify() which caused an infinite loop
9732 if there was more than one signature.
9733 [Sven Uszpelkat <su@celocom.de>]
9735 *) Major change in util/mkdef.pl to include extra information
9736 about each symbol, as well as presenting variables as well
9737 as functions. This change means that there's n more need
9738 to rebuild the .num files when some algorithms are excluded.
9741 *) Allow the verify time to be set by an application,
9742 rather than always using the current time.
9745 *) Phase 2 verify code reorganisation. The certificate
9746 verify code now looks up an issuer certificate by a
9747 number of criteria: subject name, authority key id
9748 and key usage. It also verifies self signed certificates
9749 by the same criteria. The main comparison function is
9750 X509_check_issued() which performs these checks.
9752 Lot of changes were necessary in order to support this
9753 without completely rewriting the lookup code.
9755 Authority and subject key identifier are now cached.
9757 The LHASH 'certs' is X509_STORE has now been replaced
9758 by a STACK_OF(X509_OBJECT). This is mainly because an
9759 LHASH can't store or retrieve multiple objects with
9760 the same hash value.
9762 As a result various functions (which were all internal
9763 use only) have changed to handle the new X509_STORE
9764 structure. This will break anything that messed round
9765 with X509_STORE internally.
9767 The functions X509_STORE_add_cert() now checks for an
9768 exact match, rather than just subject name.
9770 The X509_STORE API doesn't directly support the retrieval
9771 of multiple certificates matching a given criteria, however
9772 this can be worked round by performing a lookup first
9773 (which will fill the cache with candidate certificates)
9774 and then examining the cache for matches. This is probably
9775 the best we can do without throwing out X509_LOOKUP
9776 entirely (maybe later...).
9778 The X509_VERIFY_CTX structure has been enhanced considerably.
9780 All certificate lookup operations now go via a get_issuer()
9781 callback. Although this currently uses an X509_STORE it
9782 can be replaced by custom lookups. This is a simple way
9783 to bypass the X509_STORE hackery necessary to make this
9784 work and makes it possible to use more efficient techniques
9785 in future. A very simple version which uses a simple
9786 STACK for its trusted certificate store is also provided
9787 using X509_STORE_CTX_trusted_stack().
9789 The verify_cb() and verify() callbacks now have equivalents
9790 in the X509_STORE_CTX structure.
9792 X509_STORE_CTX also has a 'flags' field which can be used
9793 to customise the verify behaviour.
9796 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9797 excludes S/MIME capabilities.
9800 *) When a certificate request is read in keep a copy of the
9801 original encoding of the signed data and use it when outputting
9802 again. Signatures then use the original encoding rather than
9803 a decoded, encoded version which may cause problems if the
9804 request is improperly encoded.
9807 *) For consistency with other BIO_puts implementations, call
9808 buffer_write(b, ...) directly in buffer_puts instead of calling
9811 In BIO_puts, increment b->num_write as in BIO_write.
9812 [Peter.Sylvester@EdelWeb.fr]
9814 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9815 BN_zero, we may not return a BIGNUM with an array consisting of
9819 *) Avoid calling abort() from within the library when problems are
9820 detected, except if preprocessor symbols have been defined
9821 (such as REF_CHECK, BN_DEBUG etc.).
9824 *) New openssl application 'rsautl'. This utility can be
9825 used for low level RSA operations. DER public key
9826 BIO/fp routines also added.
9829 *) New Configure entry and patches for compiling on QNX 4.
9830 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9832 *) A demo state-machine implementation was sponsored by
9833 Nuron (http://www.nuron.com/) and is now available in
9834 demos/state_machine.
9837 *) New options added to the 'dgst' utility for signature
9838 generation and verification.
9841 *) Unrecognized PKCS#7 content types are now handled via a
9842 catch all ASN1_TYPE structure. This allows unsupported
9843 types to be stored as a "blob" and an application can
9844 encode and decode it manually.
9847 *) Fix various signed/unsigned issues to make a_strex.c
9849 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9851 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9852 length if passed a buffer. ASN1_INTEGER_to_BN failed
9853 if passed a NULL BN and its argument was negative.
9854 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9856 *) Modification to PKCS#7 encoding routines to output definite
9857 length encoding. Since currently the whole structures are in
9858 memory there's not real point in using indefinite length
9859 constructed encoding. However if OpenSSL is compiled with
9860 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9863 *) Added BIO_vprintf() and BIO_vsnprintf().
9866 *) Added more prefixes to parse for in the strings written
9867 through a logging bio, to cover all the levels that are available
9868 through syslog. The prefixes are now:
9870 PANIC, EMERG, EMR => LOG_EMERG
9871 ALERT, ALR => LOG_ALERT
9872 CRIT, CRI => LOG_CRIT
9873 ERROR, ERR => LOG_ERR
9874 WARNING, WARN, WAR => LOG_WARNING
9875 NOTICE, NOTE, NOT => LOG_NOTICE
9876 INFO, INF => LOG_INFO
9877 DEBUG, DBG => LOG_DEBUG
9879 and as before, if none of those prefixes are present at the
9880 beginning of the string, LOG_ERR is chosen.
9882 On Win32, the LOG_* levels are mapped according to this:
9884 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9885 LOG_WARNING => EVENTLOG_WARNING_TYPE
9886 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9890 *) Made it possible to reconfigure with just the configuration
9891 argument "reconf" or "reconfigure". The command line arguments
9892 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9893 and are retrieved from there when reconfiguring.
9897 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9899 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9902 *) The obj_dat.pl script was messing up the sorting of object
9903 names. The reason was that it compared the quoted version
9904 of strings as a result "OCSP" > "OCSP Signing" because
9905 " > SPACE. Changed script to store unquoted versions of
9906 names and add quotes on output. It was also omitting some
9907 names from the lookup table if they were given a default
9908 value (that is if SN is missing it is given the same
9909 value as LN and vice versa), these are now added on the
9910 grounds that if an object has a name we should be able to
9911 look it up. Finally added warning output when duplicate
9912 short or long names are found.
9915 *) Changes needed for Tandem NSK.
9916 [Scott Uroff <scott@xypro.com>]
9918 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9919 RSA_padding_check_SSLv23(), special padding was never detected
9920 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9921 version rollback attacks was not effective.
9923 In s23_clnt.c, don't use special rollback-attack detection padding
9924 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9925 client; similarly, in s23_srvr.c, don't do the rollback check if
9926 SSL 2.0 is the only protocol enabled in the server.
9929 *) Make it possible to get hexdumps of unprintable data with 'openssl
9930 asn1parse'. By implication, the functions ASN1_parse_dump() and
9931 BIO_dump_indent() are added.
9934 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9935 these print out strings and name structures based on various
9936 flags including RFC2253 support and proper handling of
9937 multibyte characters. Added options to the 'x509' utility
9938 to allow the various flags to be set.
9941 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9942 Also change the functions X509_cmp_current_time() and
9943 X509_gmtime_adj() work with an ASN1_TIME structure,
9944 this will enable certificates using GeneralizedTime in validity
9945 dates to be checked.
9948 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9949 negative public key encodings) on by default,
9950 NO_NEG_PUBKEY_BUG can be set to disable it.
9953 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9954 content octets. An i2c_ASN1_OBJECT is unnecessary because
9955 the encoding can be trivially obtained from the structure.
9958 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9959 not read locks (CRYPTO_r_[un]lock).
9962 *) A first attempt at creating official support for shared
9963 libraries through configuration. I've kept it so the
9964 default is static libraries only, and the OpenSSL programs
9965 are always statically linked for now, but there are
9966 preparations for dynamic linking in place.
9967 This has been tested on Linux and Tru64.
9970 *) Randomness polling function for Win9x, as described in:
9971 Peter Gutmann, Software Generation of Practically Strong
9975 *) Fix so PRNG is seeded in req if using an already existing
9979 *) New options to smime application. -inform and -outform
9980 allow alternative formats for the S/MIME message including
9981 PEM and DER. The -content option allows the content to be
9982 specified separately. This should allow things like Netscape
9983 form signing output easier to verify.
9986 *) Fix the ASN1 encoding of tags using the 'long form'.
9989 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9990 STRING types. These convert content octets to and from the
9991 underlying type. The actual tag and length octets are
9992 already assumed to have been read in and checked. These
9993 are needed because all other string types have virtually
9994 identical handling apart from the tag. By having versions
9995 of the ASN1 functions that just operate on content octets
9996 IMPLICIT tagging can be handled properly. It also allows
9997 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9998 and ASN1_INTEGER are identical apart from the tag.
10001 *) Change the handling of OID objects as follows:
10003 - New object identifiers are inserted in objects.txt, following
10004 the syntax given in objects.README.
10005 - objects.pl is used to process obj_mac.num and create a new
10007 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10010 This is currently kind of a hack, and the perl code in objects.pl
10011 isn't very elegant, but it works as I intended. The simplest way
10012 to check that it worked correctly is to look in obj_dat.h and
10013 check the array nid_objs and make sure the objects haven't moved
10014 around (this is important!). Additions are OK, as well as
10015 consistent name changes.
10018 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10021 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10022 The given file adds to whatever has already been seeded into the
10023 random pool through the RANDFILE configuration file option or
10024 environment variable, or the default random state file.
10027 *) mkstack.pl now sorts each macro group into lexical order.
10028 Previously the output order depended on the order the files
10029 appeared in the directory, resulting in needless rewriting
10033 *) Patches to make OpenSSL compile under Win32 again. Mostly
10034 work arounds for the VC++ problem that it treats func() as
10035 func(void). Also stripped out the parts of mkdef.pl that
10036 added extra typesafe functions: these no longer exist.
10039 *) Reorganisation of the stack code. The macros are now all
10040 collected in safestack.h . Each macro is defined in terms of
10041 a "stack macro" of the form SKM_<name>(type, a, b). The
10042 DEBUG_SAFESTACK is now handled in terms of function casts,
10043 this has the advantage of retaining type safety without the
10044 use of additional functions. If DEBUG_SAFESTACK is not defined
10045 then the non typesafe macros are used instead. Also modified the
10046 mkstack.pl script to handle the new form. Needs testing to see
10047 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10048 the default if no major problems. Similar behaviour for ASN1_SET_OF
10049 and PKCS12_STACK_OF.
10052 *) When some versions of IIS use the 'NET' form of private key the
10053 key derivation algorithm is different. Normally MD5(password) is
10054 used as a 128 bit RC4 key. In the modified case
10055 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10056 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10057 as the old Netscape_RSA functions except they have an additional
10058 'sgckey' parameter which uses the modified algorithm. Also added
10059 an -sgckey command line option to the rsa utility. Thanks to
10060 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10061 algorithm to openssl-dev.
10064 *) The evp_local.h macros were using 'c.##kname' which resulted in
10065 invalid expansion on some systems (SCO 5.0.5 for example).
10066 Corrected to 'c.kname'.
10067 [Phillip Porch <root@theporch.com>]
10069 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10070 a STACK of email addresses from a certificate or request, these look
10071 in the subject name and the subject alternative name extensions and
10072 omit any duplicate addresses.
10075 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10076 This makes DSA verification about 2 % faster.
10079 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10080 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10081 plus overhead for 1024 bit moduli).
10082 This makes exponentiations about 0.5 % faster for 1024 bit
10083 exponents (as measured by "openssl speed rsa2048").
10086 *) Rename memory handling macros to avoid conflicts with other
10088 Malloc => OPENSSL_malloc
10089 Malloc_locked => OPENSSL_malloc_locked
10090 Realloc => OPENSSL_realloc
10091 Free => OPENSSL_free
10094 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10095 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10098 *) CygWin32 support.
10099 [John Jarvie <jjarvie@newsguy.com>]
10101 *) The type-safe stack code has been rejigged. It is now only compiled
10102 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10103 by default all type-specific stack functions are "#define"d back to
10104 standard stack functions. This results in more streamlined output
10105 but retains the type-safety checking possibilities of the original
10109 *) The STACK code has been cleaned up, and certain type declarations
10110 that didn't make a lot of sense have been brought in line. This has
10111 also involved a cleanup of sorts in safestack.h to more correctly
10112 map type-safe stack functions onto their plain stack counterparts.
10113 This work has also resulted in a variety of "const"ifications of
10114 lots of the code, especially "_cmp" operations which should normally
10115 be prototyped with "const" parameters anyway.
10118 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10119 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10120 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10121 where all of 'md' is used each time the PRNG is used, but 'state'
10122 is used only indexed by a cyclic counter. As entropy may not be
10123 well distributed from the beginning, 'md' is important as a
10124 chaining variable. However, the output function chains only half
10125 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10126 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10127 in all of 'state' being rewritten, with the new values depending
10128 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10131 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10132 the handshake is continued after ssl_verify_cert_chain();
10133 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10134 can lead to 'unexplainable' connection aborts later.
10135 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10137 *) Major EVP API cipher revision.
10138 Add hooks for extra EVP features. This allows various cipher
10139 parameters to be set in the EVP interface. Support added for variable
10140 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10141 setting of RC2 and RC5 parameters.
10143 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10146 Remove lots of duplicated code from the EVP library. For example *every*
10147 cipher init() function handles the 'iv' in the same way according to the
10148 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10149 for CFB and OFB modes they zero ctx->num.
10151 New functionality allows removal of S/MIME code RC2 hack.
10153 Most of the routines have the same form and so can be declared in terms
10156 By shifting this to the top level EVP_CipherInit() it can be removed from
10157 all individual ciphers. If the cipher wants to handle IVs or keys
10158 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10161 Change lots of functions like EVP_EncryptUpdate() to now return a
10162 value: although software versions of the algorithms cannot fail
10163 any installed hardware versions can.
10166 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10167 this option is set, tolerate broken clients that send the negotiated
10168 protocol version number instead of the requested protocol version
10172 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10173 i.e. non-zero for export ciphersuites, zero otherwise.
10174 Previous versions had this flag inverted, inconsistent with
10175 rsa_tmp_cb (..._TMP_RSA_CB).
10176 [Bodo Moeller; problem reported by Amit Chopra]
10178 *) Add missing DSA library text string. Work around for some IIS
10179 key files with invalid SEQUENCE encoding.
10182 *) Add a document (doc/standards.txt) that list all kinds of standards
10183 and so on that are implemented in OpenSSL.
10186 *) Enhance c_rehash script. Old version would mishandle certificates
10187 with the same subject name hash and wouldn't handle CRLs at all.
10188 Added -fingerprint option to crl utility, to support new c_rehash
10192 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10195 *) Fix for SSL server purpose checking. Server checking was
10196 rejecting certificates which had extended key usage present
10197 but no ssl client purpose.
10198 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10200 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10201 is a little unclear about how a blank password is handled.
10202 Since the password in encoded as a BMPString with terminating
10203 double NULL a zero length password would end up as just the
10204 double NULL. However no password at all is different and is
10205 handled differently in the PKCS#12 key generation code. NS
10206 treats a blank password as zero length. MSIE treats it as no
10207 password on export: but it will try both on import. We now do
10208 the same: PKCS12_parse() tries zero length and no password if
10209 the password is set to "" or NULL (NULL is now a valid password:
10210 it wasn't before) as does the pkcs12 application.
10213 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10214 perror when PEM_read_bio_X509_REQ fails, the error message must
10215 be obtained from the error queue.
10218 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10219 it in ERR_remove_state if appropriate, and change ERR_get_state
10220 accordingly to avoid race conditions (this is necessary because
10221 thread_hash is no longer constant once set).
10224 *) Bugfix for linux-elf makefile.one.
10227 *) RSA_get_default_method() will now cause a default
10228 RSA_METHOD to be chosen if one doesn't exist already.
10229 Previously this was only set during a call to RSA_new()
10230 or RSA_new_method(NULL) meaning it was possible for
10231 RSA_get_default_method() to return NULL.
10234 *) Added native name translation to the existing DSO code
10235 that will convert (if the flag to do so is set) filenames
10236 that are sufficiently small and have no path information
10237 into a canonical native form. Eg. "blah" converted to
10238 "libblah.so" or "blah.dll" etc.
10241 *) New function ERR_error_string_n(e, buf, len) which is like
10242 ERR_error_string(e, buf), but writes at most 'len' bytes
10243 including the 0 terminator. For ERR_error_string_n, 'buf'
10245 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10247 *) CONF library reworked to become more general. A new CONF
10248 configuration file reader "class" is implemented as well as a
10249 new functions (NCONF_*, for "New CONF") to handle it. The now
10250 old CONF_* functions are still there, but are reimplemented to
10251 work in terms of the new functions. Also, a set of functions
10252 to handle the internal storage of the configuration data is
10253 provided to make it easier to write new configuration file
10254 reader "classes" (I can definitely see something reading a
10255 configuration file in XML format, for example), called _CONF_*,
10256 or "the configuration storage API"...
10258 The new configuration file reading functions are:
10260 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10261 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10263 NCONF_default, NCONF_WIN32
10265 NCONF_dump_fp, NCONF_dump_bio
10267 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10268 NCONF_new creates a new CONF object. This works in the same way
10269 as other interfaces in OpenSSL, like the BIO interface.
10270 NCONF_dump_* dump the internal storage of the configuration file,
10271 which is useful for debugging. All other functions take the same
10272 arguments as the old CONF_* functions wth the exception of the
10273 first that must be a `CONF *' instead of a `LHASH *'.
10275 To make it easer to use the new classes with the old CONF_* functions,
10276 the function CONF_set_default_method is provided.
10279 *) Add '-tls1' option to 'openssl ciphers', which was already
10280 mentioned in the documentation but had not been implemented.
10281 (This option is not yet really useful because even the additional
10282 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10285 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10286 OpenSSL-based applications) load shared libraries and bind to
10287 them in a portable way.
10288 [Geoff Thorpe, with contributions from Richard Levitte]
10290 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10292 *) Make sure _lrotl and _lrotr are only used with MSVC.
10294 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10295 (the default implementation of RAND_status).
10297 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10298 to '-clrext' (= clear extensions), as intended and documented.
10299 [Bodo Moeller; inconsistency pointed out by Michael Attili
10300 <attili@amaxo.com>]
10302 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10303 was larger than the MD block size.
10304 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10306 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10307 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10308 using the passed key: if the passed key was a private key the result
10309 of X509_print(), for example, would be to print out all the private key
10313 *) des_quad_cksum() byte order bug fix.
10314 [Ulf Möller, using the problem description in krb4-0.9.7, where
10315 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10317 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10319 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10321 *) For easily testing in shell scripts whether some command
10322 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10323 returns with exit code 0 iff no command of the given name is available.
10324 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10325 the output goes to stdout and nothing is printed to stderr.
10326 Additional arguments are always ignored.
10328 Since for each cipher there is a command of the same name,
10329 the 'no-cipher' compilation switches can be tested this way.
10331 ('openssl no-XXX' is not able to detect pseudo-commands such
10332 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10335 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10338 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10339 is set; it will be thrown away anyway because each handshake creates
10341 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10342 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10343 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10344 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10347 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10348 'Q' and 'R' lose their special meanings (quit/renegotiate).
10349 This is part of what -quiet does; unlike -quiet, -ign_eof
10350 does not suppress any output.
10353 *) Add compatibility options to the purpose and trust code. The
10354 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10355 accepts a certificate or CA, this was the previous behaviour,
10356 with all the associated security issues.
10358 X509_TRUST_COMPAT is the old trust behaviour: only and
10359 automatically trust self signed roots in certificate store. A
10360 new trust setting X509_TRUST_DEFAULT is used to specify that
10361 a purpose has no associated trust setting and it should instead
10362 use the value in the default purpose.
10365 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10366 and fix a memory leak.
10369 *) In util/mkerr.pl (which implements 'make errors'), preserve
10370 reason strings from the previous version of the .c file, as
10371 the default to have only downcase letters (and digits) in
10372 automatically generated reasons codes is not always appropriate.
10375 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10376 using strerror. Previously, ERR_reason_error_string() returned
10377 library names as reason strings for SYSerr; but SYSerr is a special
10378 case where small numbers are errno values, not library numbers.
10381 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10382 converts DSA parameters into DH parameters. (When creating parameters,
10383 DSA_generate_parameters is used.)
10386 *) Include 'length' (recommended exponent length) in C code generated
10387 by 'openssl dhparam -C'.
10390 *) The second argument to set_label in perlasm was already being used
10391 so couldn't be used as a "file scope" flag. Moved to third argument
10395 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10396 instead of RAND_bytes for encryption IVs and salts.
10399 *) Include RAND_status() into RAND_METHOD instead of implementing
10400 it only for md_rand.c Otherwise replacing the PRNG by calling
10401 RAND_set_rand_method would be impossible.
10404 *) Don't let DSA_generate_key() enter an infinite loop if the random
10405 number generation fails.
10408 *) New 'rand' application for creating pseudo-random output.
10411 *) Added configuration support for Linux/IA64
10412 [Rolf Haberrecker <rolf@suse.de>]
10414 *) Assembler module support for Mingw32.
10417 *) Shared library support for HPUX (in shlib/).
10418 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10420 *) Shared library support for Solaris gcc.
10421 [Lutz Behnke <behnke@trustcenter.de>]
10423 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10425 *) PKCS7_encrypt() was adding text MIME headers twice because they
10426 were added manually and by SMIME_crlf_copy().
10429 *) In bntest.c don't call BN_rand with zero bits argument.
10430 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10432 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10433 case was implemented. This caused BN_div_recp() to fail occasionally.
10436 *) Add an optional second argument to the set_label() in the perl
10437 assembly language builder. If this argument exists and is set
10438 to 1 it signals that the assembler should use a symbol whose
10439 scope is the entire file, not just the current function. This
10440 is needed with MASM which uses the format label:: for this scope.
10441 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10443 *) Change the ASN1 types so they are typedefs by default. Before
10444 almost all types were #define'd to ASN1_STRING which was causing
10445 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10449 *) Change names of new functions to the new get1/get0 naming
10450 convention: After 'get1', the caller owns a reference count
10451 and has to call ..._free; 'get0' returns a pointer to some
10452 data structure without incrementing reference counters.
10453 (Some of the existing 'get' functions increment a reference
10454 counter, some don't.)
10455 Similarly, 'set1' and 'add1' functions increase reference
10456 counters or duplicate objects.
10459 *) Allow for the possibility of temp RSA key generation failure:
10460 the code used to assume it always worked and crashed on failure.
10463 *) Fix potential buffer overrun problem in BIO_printf().
10464 [Ulf Möller, using public domain code by Patrick Powell; problem
10465 pointed out by David Sacerdote <das33@cornell.edu>]
10467 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10468 RAND_egd() and RAND_status(). In the command line application,
10469 the EGD socket can be specified like a seed file using RANDFILE
10473 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10474 Some CAs (e.g. Verisign) distribute certificates in this form.
10477 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10478 list to exclude them. This means that no special compilation option
10479 is needed to use anonymous DH: it just needs to be included in the
10483 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10484 EVP_MD_type. The old functionality is available in a new macro called
10485 EVP_MD_md(). Change code that uses it and update docs.
10488 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10489 where the 'void *' argument is replaced by a function pointer argument.
10490 Previously 'void *' was abused to point to functions, which works on
10491 many platforms, but is not correct. As these functions are usually
10492 called by macros defined in OpenSSL header files, most source code
10493 should work without changes.
10496 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10497 sections with information on -D... compiler switches used for
10498 compiling the library so that applications can see them. To enable
10499 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10500 must be defined. E.g.,
10501 #define OPENSSL_ALGORITHM_DEFINES
10502 #include <openssl/opensslconf.h>
10503 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10504 [Richard Levitte, Ulf and Bodo Möller]
10506 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10510 *) Change the 'other' type in certificate aux info to a STACK_OF
10511 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10512 the required ASN1 format: arbitrary types determined by an OID.
10515 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10516 argument to 'req'. This is not because the function is newer or
10517 better than others it just uses the work 'NEW' in the certificate
10518 request header lines. Some software needs this.
10521 *) Reorganise password command line arguments: now passwords can be
10522 obtained from various sources. Delete the PEM_cb function and make
10523 it the default behaviour: i.e. if the callback is NULL and the
10524 usrdata argument is not NULL interpret it as a null terminated pass
10525 phrase. If usrdata and the callback are NULL then the pass phrase
10526 is prompted for as usual.
10529 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10530 the support is automatically enabled. The resulting binaries will
10531 autodetect the card and use it if present.
10532 [Ben Laurie and Compaq Inc.]
10534 *) Work around for Netscape hang bug. This sends certificate request
10535 and server done in one record. Since this is perfectly legal in the
10536 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10537 the bugs/SSLv3 entry for more info.
10540 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10543 *) Add -rand argument to smime and pkcs12 applications and read/write
10547 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10550 *) Add command line password options to the remaining applications.
10553 *) Bug fix for BN_div_recp() for numerators with an even number of
10557 *) More tests in bntest.c, and changed test_bn output.
10560 *) ./config recognizes MacOS X now.
10563 *) Bug fix for BN_div() when the first words of num and divisor are
10564 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10567 *) Add support for various broken PKCS#8 formats, and command line
10568 options to produce them.
10571 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10572 get temporary BIGNUMs from a BN_CTX.
10575 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10579 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10580 include a #define from the old name to the new. The original intent
10581 was that statically linked binaries could for example just call
10582 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10583 link with digests. This never worked because SSLeay_add_all_digests()
10584 and SSLeay_add_all_ciphers() were in the same source file so calling
10585 one would link with the other. They are now in separate source files.
10588 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10591 *) Use a less unusual form of the Miller-Rabin primality test (it used
10592 a binary algorithm for exponentiation integrated into the Miller-Rabin
10593 loop, our standard modexp algorithms are faster).
10596 *) Support for the EBCDIC character set completed.
10597 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10599 *) Source code cleanups: use const where appropriate, eliminate casts,
10600 use void * instead of char * in lhash.
10603 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10604 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10605 this the server could overwrite ephemeral keys that the client
10609 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10610 using 50 iterations of the Rabin-Miller test.
10612 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10613 iterations of the Rabin-Miller test as required by the appendix
10614 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10615 As BN_is_prime_fasttest includes trial division, DSA parameter
10616 generation becomes much faster.
10618 This implies a change for the callback functions in DSA_is_prime
10619 and DSA_generate_parameters: The callback function is called once
10620 for each positive witness in the Rabin-Miller test, not just
10621 occasionally in the inner loop; and the parameters to the
10622 callback function now provide an iteration count for the outer
10623 loop rather than for the current invocation of the inner loop.
10624 DSA_generate_parameters additionally can call the callback
10625 function with an 'iteration count' of -1, meaning that a
10626 candidate has passed the trial division test (when q is generated
10627 from an application-provided seed, trial division is skipped).
10630 *) New function BN_is_prime_fasttest that optionally does trial
10631 division before starting the Rabin-Miller test and has
10632 an additional BN_CTX * argument (whereas BN_is_prime always
10633 has to allocate at least one BN_CTX).
10634 'callback(1, -1, cb_arg)' is called when a number has passed the
10635 trial division stage.
10638 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10642 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10645 *) New function BN_pseudo_rand().
10648 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10649 bignum version of BN_from_montgomery() with the working code from
10650 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10654 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10655 made it impossible to use the same SSL_SESSION data structure in
10656 SSL2 clients in multiple threads.
10659 *) The return value of RAND_load_file() no longer counts bytes obtained
10660 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10661 to seed the PRNG (previously an explicit byte count was required).
10662 [Ulf Möller, Bodo Möller]
10664 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10665 used (char *) instead of (void *) and had casts all over the place.
10668 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10671 *) Retain source code compatibility for BN_prime_checks macro:
10672 BN_is_prime(..., BN_prime_checks, ...) now uses
10673 BN_prime_checks_for_size to determine the appropriate number of
10674 Rabin-Miller iterations.
10677 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10678 DH_CHECK_P_NOT_SAFE_PRIME.
10679 (Check if this is true? OpenPGP calls them "strong".)
10682 *) Merge the functionality of "dh" and "gendh" programs into a new program
10683 "dhparam". The old programs are retained for now but will handle DH keys
10684 (instead of parameters) in future.
10687 *) Make the ciphers, s_server and s_client programs check the return values
10688 when a new cipher list is set.
10691 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10692 ciphers. Before when the 56bit ciphers were enabled the sorting was
10695 The syntax for the cipher sorting has been extended to support sorting by
10696 cipher-strength (using the strength_bits hard coded in the tables).
10697 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10699 Fix a bug in the cipher-command parser: when supplying a cipher command
10700 string with an "undefined" symbol (neither command nor alphanumeric
10701 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10702 an error is flagged.
10704 Due to the strength-sorting extension, the code of the
10705 ssl_create_cipher_list() function was completely rearranged. I hope that
10706 the readability was also increased :-)
10707 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10709 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10710 for the first serial number and places 2 in the serial number file. This
10711 avoids problems when the root CA is created with serial number zero and
10712 the first user certificate has the same issuer name and serial number
10716 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10717 the new code. Add documentation for this stuff.
10720 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10721 X509_*() to X509at_*() on the grounds that they don't handle X509
10722 structures and behave in an analogous way to the X509v3 functions:
10723 they shouldn't be called directly but wrapper functions should be used
10726 So we also now have some wrapper functions that call the X509at functions
10727 when passed certificate requests. (TO DO: similar things can be done with
10728 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10729 things. Some of these need some d2i or i2d and print functionality
10730 because they handle more complex structures.)
10733 *) Add missing #ifndefs that caused missing symbols when building libssl
10734 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10735 NO_RSA in ssl/s2*.c.
10736 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10738 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10739 has a return value which indicates the quality of the random data
10740 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10741 error queue. New function RAND_pseudo_bytes() generates output that is
10742 guaranteed to be unique but not unpredictable. RAND_add is like
10743 RAND_seed, but takes an extra argument for an entropy estimate
10744 (RAND_seed always assumes full entropy).
10747 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10748 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10749 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10750 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10751 false-positive rate of at most 2^-80 for random input.
10754 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10757 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10758 in the 0.9.5 release), this returns the chain
10759 from an X509_CTX structure with a dup of the stack and all
10760 the X509 reference counts upped: so the stack will exist
10761 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10764 Also make SSL_SESSION_print() print out the verify return
10768 *) Add manpage for the pkcs12 command. Also change the default
10769 behaviour so MAC iteration counts are used unless the new
10770 -nomaciter option is used. This improves file security and
10771 only older versions of MSIE (4.0 for example) need it.
10774 *) Honor the no-xxx Configure options when creating .DEF files.
10777 *) Add PKCS#10 attributes to field table: challengePassword,
10778 unstructuredName and unstructuredAddress. These are taken from
10779 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10780 international characters are used.
10782 More changes to X509_ATTRIBUTE code: allow the setting of types
10783 based on strings. Remove the 'loc' parameter when adding
10784 attributes because these will be a SET OF encoding which is sorted
10788 *) Initial changes to the 'req' utility to allow request generation
10789 automation. This will allow an application to just generate a template
10790 file containing all the field values and have req construct the
10793 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10794 used all over the place including certificate requests and PKCS#7
10795 structures. They are currently handled manually where necessary with
10796 some primitive wrappers for PKCS#7. The new functions behave in a
10797 manner analogous to the X509 extension functions: they allow
10798 attributes to be looked up by NID and added.
10800 Later something similar to the X509V3 code would be desirable to
10801 automatically handle the encoding, decoding and printing of the
10802 more complex types. The string types like challengePassword can
10803 be handled by the string table functions.
10805 Also modified the multi byte string table handling. Now there is
10806 a 'global mask' which masks out certain types. The table itself
10807 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10808 is useful when for example there is only one permissible type
10809 (as in countryName) and using the mask might result in no valid
10813 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10814 SSL_get_peer_finished to allow applications to obtain the latest
10815 Finished messages sent to the peer or expected from the peer,
10816 respectively. (SSL_get_peer_finished is usually the Finished message
10817 actually received from the peer, otherwise the protocol will be aborted.)
10819 As the Finished message are message digests of the complete handshake
10820 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10821 be used for external authentication procedures when the authentication
10822 provided by SSL/TLS is not desired or is not enough.
10825 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10826 the host supports BWX extension and if Compaq C is present on the
10827 $PATH. Just exploiting of the BWX extension results in 20-30%
10828 performance kick for some algorithms, e.g. DES and RC4 to mention
10829 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10833 *) Add support for MS "fast SGC". This is arguably a violation of the
10834 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10835 weak crypto and after checking the certificate is SGC a second one
10836 with strong crypto. MS SGC stops the first handshake after receiving
10837 the server certificate message and sends a second client hello. Since
10838 a server will typically do all the time consuming operations before
10839 expecting any further messages from the client (server key exchange
10840 is the most expensive) there is little difference between the two.
10842 To get OpenSSL to support MS SGC we have to permit a second client
10843 hello message after we have sent server done. In addition we have to
10844 reset the MAC if we do get this second client hello.
10847 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10848 if a DER encoded private key is RSA or DSA traditional format. Changed
10849 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10850 format DER encoded private key. Newer code should use PKCS#8 format which
10851 has the key type encoded in the ASN1 structure. Added DER private key
10852 support to pkcs8 application.
10855 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10856 ciphersuites has been selected (as required by the SSL 3/TLS 1
10857 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10858 is set, we interpret this as a request to violate the specification
10859 (the worst that can happen is a handshake failure, and 'correct'
10860 behaviour would result in a handshake failure anyway).
10863 *) In SSL_CTX_add_session, take into account that there might be multiple
10864 SSL_SESSION structures with the same session ID (e.g. when two threads
10865 concurrently obtain them from an external cache).
10866 The internal cache can handle only one SSL_SESSION with a given ID,
10867 so if there's a conflict, we now throw out the old one to achieve
10871 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10872 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10873 some routines that use cipher OIDs: some ciphers do not have OIDs
10874 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10878 *) Simplify the trust setting structure and code. Now we just have
10879 two sequences of OIDs for trusted and rejected settings. These will
10880 typically have values the same as the extended key usage extension
10881 and any application specific purposes.
10883 The trust checking code now has a default behaviour: it will just
10884 check for an object with the same NID as the passed id. Functions can
10885 be provided to override either the default behaviour or the behaviour
10886 for a given id. SSL client, server and email already have functions
10887 in place for compatibility: they check the NID and also return "trusted"
10888 if the certificate is self signed.
10891 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10892 traditional format into an EVP_PKEY structure.
10895 *) Add a password callback function PEM_cb() which either prompts for
10896 a password if usr_data is NULL or otherwise assumes it is a null
10897 terminated password. Allow passwords to be passed on command line
10898 environment or config files in a few more utilities.
10901 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10902 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10903 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10904 Update documentation.
10907 *) Support for ASN1 "NULL" type. This could be handled before by using
10908 ASN1_TYPE but there wasn't any function that would try to read a NULL
10909 and produce an error if it couldn't. For compatibility we also have
10910 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10911 don't allocate anything because they don't need to.
10914 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10916 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10918 *) Rebuild of the memory allocation routines used by OpenSSL code and
10919 possibly others as well. The purpose is to make an interface that
10920 provide hooks so anyone can build a separate set of allocation and
10921 deallocation routines to be used by OpenSSL, for example memory
10922 pool implementations, or something else, which was previously hard
10923 since Malloc(), Realloc() and Free() were defined as macros having
10924 the values malloc, realloc and free, respectively (except for Win32
10925 compilations). The same is provided for memory debugging code.
10926 OpenSSL already comes with functionality to find memory leaks, but
10927 this gives people a chance to debug other memory problems.
10929 With these changes, a new set of functions and macros have appeared:
10931 CRYPTO_set_mem_debug_functions() [F]
10932 CRYPTO_get_mem_debug_functions() [F]
10933 CRYPTO_dbg_set_options() [F]
10934 CRYPTO_dbg_get_options() [F]
10935 CRYPTO_malloc_debug_init() [M]
10937 The memory debug functions are NULL by default, unless the library
10938 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10939 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10940 gives the standard debugging functions that come with OpenSSL) or
10941 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10942 provided by the library user) must be used. When the standard
10943 debugging functions are used, CRYPTO_dbg_set_options can be used to
10944 request additional information:
10945 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10946 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10948 Also, things like CRYPTO_set_mem_functions will always give the
10949 expected result (the new set of functions is used for allocation
10950 and deallocation) at all times, regardless of platform and compiler
10953 To finish it up, some functions that were never use in any other
10954 way than through macros have a new API and new semantic:
10956 CRYPTO_dbg_malloc()
10957 CRYPTO_dbg_realloc()
10960 All macros of value have retained their old syntax.
10961 [Richard Levitte and Bodo Moeller]
10963 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10964 ordering of SMIMECapabilities wasn't in "strength order" and there
10965 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10969 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10970 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10971 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10973 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10974 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10975 functionality to handle multipart/signed properly) and a utility
10976 called 'smime' to call all this stuff. This is based on code I
10977 originally wrote for Celo who have kindly allowed it to be
10978 included in OpenSSL.
10981 *) Add variants des_set_key_checked and des_set_key_unchecked of
10982 des_set_key (aka des_key_sched). Global variable des_check_key
10983 decides which of these is called by des_set_key; this way
10984 des_check_key behaves as it always did, but applications and
10985 the library itself, which was buggy for des_check_key == 1,
10986 have a cleaner way to pick the version they need.
10989 *) New function PKCS12_newpass() which changes the password of a
10993 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10994 dynamic mix. In both cases the ids can be used as an index into the
10995 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10996 functions so they accept a list of the field values and the
10997 application doesn't need to directly manipulate the X509_TRUST
11001 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11005 *) Modify the way the V3 extension code looks up extensions. This now
11006 works in a similar way to the object code: we have some "standard"
11007 extensions in a static table which is searched with OBJ_bsearch()
11008 and the application can add dynamic ones if needed. The file
11009 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11010 updated whenever a new extension is added to the core code and kept
11011 in ext_nid order. There is a simple program 'tabtest.c' which checks
11012 this. New extensions are not added too often so this file can readily
11013 be maintained manually.
11015 There are two big advantages in doing things this way. The extensions
11016 can be looked up immediately and no longer need to be "added" using
11017 X509V3_add_standard_extensions(): this function now does nothing.
11018 [Side note: I get *lots* of email saying the extension code doesn't
11019 work because people forget to call this function]
11020 Also no dynamic allocation is done unless new extensions are added:
11021 so if we don't add custom extensions there is no need to call
11022 X509V3_EXT_cleanup().
11025 *) Modify enc utility's salting as follows: make salting the default. Add a
11026 magic header, so unsalted files fail gracefully instead of just decrypting
11027 to garbage. This is because not salting is a big security hole, so people
11028 should be discouraged from doing it.
11031 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11032 digest to be passed on the command line but it only used this
11033 parameter when signing a certificate. Modified so all relevant
11034 operations are affected by the digest parameter including the
11035 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11036 DSA key was used because it didn't fix the digest.
11039 *) Initial certificate chain verify code. Currently tests the untrusted
11040 certificates for consistency with the verify purpose (which is set
11041 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11043 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11044 this is because it will reject chains with invalid extensions whereas
11045 every previous version of OpenSSL and SSLeay made no checks at all.
11047 Trust code: checks the root CA for the relevant trust settings. Trust
11048 settings have an initial value consistent with the verify purpose: e.g.
11049 if the verify purpose is for SSL client use it expects the CA to be
11050 trusted for SSL client use. However the default value can be changed to
11051 permit custom trust settings: one example of this would be to only trust
11052 certificates from a specific "secure" set of CAs.
11054 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11055 which should be used for version portability: especially since the
11056 verify structure is likely to change more often now.
11058 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11059 to set them. If not set then assume SSL clients will verify SSL servers
11062 Two new options to the verify program: -untrusted allows a set of
11063 untrusted certificates to be passed in and -purpose which sets the
11064 intended purpose of the certificate. If a purpose is set then the
11065 new chain verify code is used to check extension consistency.
11068 *) Support for the authority information access extension.
11071 *) Modify RSA and DSA PEM read routines to transparently handle
11072 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11073 public keys in a format compatible with certificate
11074 SubjectPublicKeyInfo structures. Unfortunately there were already
11075 functions called *_PublicKey_* which used various odd formats so
11076 these are retained for compatibility: however the DSA variants were
11077 never in a public release so they have been deleted. Changed dsa/rsa
11078 utilities to handle the new format: note no releases ever handled public
11079 keys so we should be OK.
11081 The primary motivation for this change is to avoid the same fiasco
11082 that dogs private keys: there are several incompatible private key
11083 formats some of which are standard and some OpenSSL specific and
11084 require various evil hacks to allow partial transparent handling and
11085 even then it doesn't work with DER formats. Given the option anything
11086 other than PKCS#8 should be dumped: but the other formats have to
11087 stay in the name of compatibility.
11089 With public keys and the benefit of hindsight one standard format
11090 is used which works with EVP_PKEY, RSA or DSA structures: though
11091 it clearly returns an error if you try to read the wrong kind of key.
11093 Added a -pubkey option to the 'x509' utility to output the public key.
11094 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11095 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11096 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11097 that do the same as the EVP_PKEY_assign_*() except they up the
11098 reference count of the added key (they don't "swallow" the
11102 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11103 CRLs would fail if the file contained no certificates or no CRLs:
11104 added a new function to read in both types and return the number
11105 read: this means that if none are read it will be an error. The
11106 DER versions of the certificate and CRL reader would always fail
11107 because it isn't possible to mix certificates and CRLs in DER format
11108 without choking one or the other routine. Changed this to just read
11109 a certificate: this is the best we can do. Also modified the code
11110 in apps/verify.c to take notice of return codes: it was previously
11111 attempting to read in certificates from NULL pointers and ignoring
11112 any errors: this is one reason why the cert and CRL reader seemed
11113 to work. It doesn't check return codes from the default certificate
11114 routines: these may well fail if the certificates aren't installed.
11117 *) Code to support otherName option in GeneralName.
11120 *) First update to verify code. Change the verify utility
11121 so it warns if it is passed a self signed certificate:
11122 for consistency with the normal behaviour. X509_verify
11123 has been modified to it will now verify a self signed
11124 certificate if *exactly* the same certificate appears
11125 in the store: it was previously impossible to trust a
11126 single self signed certificate. This means that:
11127 openssl verify ss.pem
11128 now gives a warning about a self signed certificate but
11129 openssl verify -CAfile ss.pem ss.pem
11133 *) For servers, store verify_result in SSL_SESSION data structure
11134 (and add it to external session representation).
11135 This is needed when client certificate verifications fails,
11136 but an application-provided verification callback (set by
11137 SSL_CTX_set_cert_verify_callback) allows accepting the session
11138 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11139 but returns 1): When the session is reused, we have to set
11140 ssl->verify_result to the appropriate error code to avoid
11142 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11144 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11145 case in PKCS7_dataInit() where the signed PKCS7 structure
11146 didn't contain any existing data because it was being created.
11147 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11149 *) Add a salt to the key derivation routines in enc.c. This
11150 forms the first 8 bytes of the encrypted file. Also add a
11151 -S option to allow a salt to be input on the command line.
11154 *) New function X509_cmp(). Oddly enough there wasn't a function
11155 to compare two certificates. We do this by working out the SHA1
11156 hash and comparing that. X509_cmp() will be needed by the trust
11160 *) SSL_get1_session() is like SSL_get_session(), but increments
11161 the reference count in the SSL_SESSION returned.
11162 [Geoff Thorpe <geoff@eu.c2.net>]
11164 *) Fix for 'req': it was adding a null to request attributes.
11165 Also change the X509_LOOKUP and X509_INFO code to handle
11166 certificate auxiliary information.
11169 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11173 *) Add the possibility to add extra information to the memory leak
11174 detecting output, to form tracebacks, showing from where each
11175 allocation was originated: CRYPTO_push_info("constant string") adds
11176 the string plus current file name and line number to a per-thread
11177 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11178 is like calling CYRPTO_pop_info() until the stack is empty.
11179 Also updated memory leak detection code to be multi-thread-safe.
11182 *) Add options -text and -noout to pkcs7 utility and delete the
11183 encryption options which never did anything. Update docs.
11186 *) Add options to some of the utilities to allow the pass phrase
11187 to be included on either the command line (not recommended on
11188 OSes like Unix) or read from the environment. Update the
11189 manpages and fix a few bugs.
11192 *) Add a few manpages for some of the openssl commands.
11195 *) Fix the -revoke option in ca. It was freeing up memory twice,
11196 leaking and not finding already revoked certificates.
11199 *) Extensive changes to support certificate auxiliary information.
11200 This involves the use of X509_CERT_AUX structure and X509_AUX
11201 functions. An X509_AUX function such as PEM_read_X509_AUX()
11202 can still read in a certificate file in the usual way but it
11203 will also read in any additional "auxiliary information". By
11204 doing things this way a fair degree of compatibility can be
11205 retained: existing certificates can have this information added
11206 using the new 'x509' options.
11208 Current auxiliary information includes an "alias" and some trust
11209 settings. The trust settings will ultimately be used in enhanced
11210 certificate chain verification routines: currently a certificate
11211 can only be trusted if it is self signed and then it is trusted
11215 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11216 The problem was that one of the replacement routines had not been working
11217 since SSLeay releases. For now the offending routine has been replaced
11218 with non-optimised assembler. Even so, this now gives around 95%
11219 performance improvement for 1024 bit RSA signs.
11222 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11223 handling. Most clients have the effective key size in bits equal to
11224 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11225 A few however don't do this and instead use the size of the decrypted key
11226 to determine the RC2 key length and the AlgorithmIdentifier to determine
11227 the effective key length. In this case the effective key length can still
11228 be 40 bits but the key length can be 168 bits for example. This is fixed
11229 by manually forcing an RC2 key into the EVP_PKEY structure because the
11230 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11231 the key length and effective key length are equal.
11234 *) Add a bunch of functions that should simplify the creation of
11235 X509_NAME structures. Now you should be able to do:
11236 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11237 and have it automatically work out the correct field type and fill in
11238 the structures. The more adventurous can try:
11239 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11240 and it will (hopefully) work out the correct multibyte encoding.
11243 *) Change the 'req' utility to use the new field handling and multibyte
11244 copy routines. Before the DN field creation was handled in an ad hoc
11245 way in req, ca, and x509 which was rather broken and didn't support
11246 BMPStrings or UTF8Strings. Since some software doesn't implement
11247 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11248 using the dirstring_type option. See the new comment in the default
11249 openssl.cnf for more info.
11252 *) Make crypto/rand/md_rand.c more robust:
11253 - Assure unique random numbers after fork().
11254 - Make sure that concurrent threads access the global counter and
11255 md serializably so that we never lose entropy in them
11256 or use exactly the same state in multiple threads.
11257 Access to the large state is not always serializable because
11258 the additional locking could be a performance killer, and
11259 md should be large enough anyway.
11262 *) New file apps/app_rand.c with commonly needed functionality
11263 for handling the random seed file.
11265 Use the random seed file in some applications that previously did not:
11267 dsaparam -genkey (which also ignored its '-rand' option),
11270 x509 (when signing).
11271 Except on systems with /dev/urandom, it is crucial to have a random
11272 seed file at least for key creation, DSA signing, and for DH exchanges;
11273 for RSA signatures we could do without one.
11275 gendh and gendsa (unlike genrsa) used to read only the first byte
11276 of each file listed in the '-rand' option. The function as previously
11277 found in genrsa is now in app_rand.c and is used by all programs
11278 that support '-rand'.
11281 *) In RAND_write_file, use mode 0600 for creating files;
11282 don't just chmod when it may be too late.
11285 *) Report an error from X509_STORE_load_locations
11286 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11289 *) New function ASN1_mbstring_copy() this copies a string in either
11290 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11291 into an ASN1_STRING type. A mask of permissible types is passed
11292 and it chooses the "minimal" type to use or an error if not type
11296 *) Add function equivalents to the various macros in asn1.h. The old
11297 macros are retained with an M_ prefix. Code inside the library can
11298 use the M_ macros. External code (including the openssl utility)
11299 should *NOT* in order to be "shared library friendly".
11302 *) Add various functions that can check a certificate's extensions
11303 to see if it usable for various purposes such as SSL client,
11304 server or S/MIME and CAs of these types. This is currently
11305 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11306 verification. Also added a -purpose flag to x509 utility to
11307 print out all the purposes.
11310 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11314 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11315 for, obtain and decode and extension and obtain its critical flag.
11316 This allows all the necessary extension code to be handled in a
11317 single function call.
11320 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11321 platforms. See crypto/rc4/rc4_enc.c for further details.
11324 *) New -noout option to asn1parse. This causes no output to be produced
11325 its main use is when combined with -strparse and -out to extract data
11326 from a file (which may not be in ASN.1 format).
11329 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11330 when producing the local key id.
11331 [Richard Levitte <levitte@stacken.kth.se>]
11333 *) New option -dhparam in s_server. This allows a DH parameter file to be
11334 stated explicitly. If it is not stated then it tries the first server
11335 certificate file. The previous behaviour hard coded the filename
11339 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11340 a public key to be input or output. For example:
11341 openssl rsa -in key.pem -pubout -out pubkey.pem
11342 Also added necessary DSA public key functions to handle this.
11345 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11346 in the message. This was handled by allowing
11347 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11348 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11350 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11351 to the end of the strings whereas this didn't. This would cause problems
11352 if strings read with d2i_ASN1_bytes() were later modified.
11353 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11355 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11356 data and it contains EOF it will end up returning an error. This is
11357 caused by input 46 bytes long. The cause is due to the way base64
11358 BIOs find the start of base64 encoded data. They do this by trying a
11359 trial decode on each line until they find one that works. When they
11360 do a flag is set and it starts again knowing it can pass all the
11361 data directly through the decoder. Unfortunately it doesn't reset
11362 the context it uses. This means that if EOF is reached an attempt
11363 is made to pass two EOFs through the context and this causes the
11364 resulting error. This can also cause other problems as well. As is
11365 usual with these problems it takes *ages* to find and the fix is
11366 trivial: move one line.
11367 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11369 *) Ugly workaround to get s_client and s_server working under Windows. The
11370 old code wouldn't work because it needed to select() on sockets and the
11371 tty (for keypresses and to see if data could be written). Win32 only
11372 supports select() on sockets so we select() with a 1s timeout on the
11373 sockets and then see if any characters are waiting to be read, if none
11374 are present then we retry, we also assume we can always write data to
11375 the tty. This isn't nice because the code then blocks until we've
11376 received a complete line of data and it is effectively polling the
11377 keyboard at 1s intervals: however it's quite a bit better than not
11378 working at all :-) A dedicated Windows application might handle this
11379 with an event loop for example.
11382 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11383 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11384 will be called when RSA_sign() and RSA_verify() are used. This is useful
11385 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11386 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11387 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11388 This necessitated the support of an extra signature type NID_md5_sha1
11389 for SSL signatures and modifications to the SSL library to use it instead
11390 of calling RSA_public_decrypt() and RSA_private_encrypt().
11393 *) Add new -verify -CAfile and -CApath options to the crl program, these
11394 will lookup a CRL issuers certificate and verify the signature in a
11395 similar way to the verify program. Tidy up the crl program so it
11396 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11397 less strict. It will now permit CRL extensions even if it is not
11398 a V2 CRL: this will allow it to tolerate some broken CRLs.
11401 *) Initialize all non-automatic variables each time one of the openssl
11402 sub-programs is started (this is necessary as they may be started
11403 multiple times from the "OpenSSL>" prompt).
11404 [Lennart Bang, Bodo Moeller]
11406 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11407 removing all other RSA functionality (this is what NO_RSA does). This
11408 is so (for example) those in the US can disable those operations covered
11409 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11413 *) Non-copying interface to BIO pairs.
11414 (still largely untested)
11417 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11418 ASCII string. This was handled independently in various places before.
11421 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11422 UTF8 strings a character at a time.
11425 *) Use client_version from client hello to select the protocol
11426 (s23_srvr.c) and for RSA client key exchange verification
11427 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11430 *) Add various utility functions to handle SPKACs, these were previously
11431 handled by poking round in the structure internals. Added new function
11432 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11433 print, verify and generate SPKACs. Based on an original idea from
11434 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11437 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11440 *) Allow the config file extension section to be overwritten on the
11441 command line. Based on an original idea from Massimiliano Pala
11442 <madwolf@comune.modena.it>. The new option is called -extensions
11443 and can be applied to ca, req and x509. Also -reqexts to override
11444 the request extensions in req and -crlexts to override the crl extensions
11448 *) Add new feature to the SPKAC handling in ca. Now you can include
11449 the same field multiple times by preceding it by "XXXX." for example:
11452 this is the same syntax as used in the req config file.
11455 *) Allow certificate extensions to be added to certificate requests. These
11456 are specified in a 'req_extensions' option of the req section of the
11457 config file. They can be printed out with the -text option to req but
11458 are otherwise ignored at present.
11461 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11462 data read consists of only the final block it would not decrypted because
11463 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11464 A misplaced 'break' also meant the decrypted final block might not be
11465 copied until the next read.
11468 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11469 a few extra parameters to the DH structure: these will be useful if
11470 for example we want the value of 'q' or implement X9.42 DH.
11473 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11474 provides hooks that allow the default DSA functions or functions on a
11475 "per key" basis to be replaced. This allows hardware acceleration and
11476 hardware key storage to be handled without major modification to the
11477 library. Also added low level modexp hooks and CRYPTO_EX structure and
11478 associated functions.
11481 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11482 as "read only": it can't be written to and the buffer it points to will
11483 not be freed. Reading from a read only BIO is much more efficient than
11484 a normal memory BIO. This was added because there are several times when
11485 an area of memory needs to be read from a BIO. The previous method was
11486 to create a memory BIO and write the data to it, this results in two
11487 copies of the data and an O(n^2) reading algorithm. There is a new
11488 function BIO_new_mem_buf() which creates a read only memory BIO from
11489 an area of memory. Also modified the PKCS#7 routines to use read only
11493 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11494 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11495 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11496 but a retry condition occurred while trying to read the rest.
11499 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11500 NID_pkcs7_encrypted by default: this was wrong since this should almost
11501 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11502 the encrypted data type: this is a more sensible place to put it and it
11503 allows the PKCS#12 code to be tidied up that duplicated this
11507 *) Changed obj_dat.pl script so it takes its input and output files on
11508 the command line. This should avoid shell escape redirection problems
11512 *) Initial support for certificate extension requests, these are included
11513 in things like Xenroll certificate requests. Included functions to allow
11514 extensions to be obtained and added.
11517 *) -crlf option to s_client and s_server for sending newlines as
11518 CRLF (as required by many protocols).
11521 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11523 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11524 [Ralf S. Engelschall]
11526 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11527 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11529 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11533 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11534 DH parameters/keys (q is lost during that conversion, but the resulting
11535 DH parameters contain its length).
11537 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11538 much faster than DH_generate_parameters (which creates parameters
11539 where p = 2*q + 1), and also the smaller q makes DH computations
11540 much more efficient (160-bit exponentiation instead of 1024-bit
11541 exponentiation); so this provides a convenient way to support DHE
11542 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11543 utter importance to use
11544 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11546 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11547 when such DH parameters are used, because otherwise small subgroup
11548 attacks may become possible!
11551 *) Avoid memory leak in i2d_DHparams.
11554 *) Allow the -k option to be used more than once in the enc program:
11555 this allows the same encrypted message to be read by multiple recipients.
11558 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11559 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11560 it will always use the numerical form of the OID, even if it has a short
11564 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11565 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11566 otherwise bn_mod_exp was called. In the case of hardware keys for example
11567 no private key components need be present and it might store extra data
11568 in the RSA structure, which cannot be accessed from bn_mod_exp.
11569 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11570 private key operations.
11573 *) Added support for SPARC Linux.
11576 *) pem_password_cb function type incompatibly changed from
11577 typedef int pem_password_cb(char *buf, int size, int rwflag);
11579 ....(char *buf, int size, int rwflag, void *userdata);
11580 so that applications can pass data to their callbacks:
11581 The PEM[_ASN1]_{read,write}... functions and macros now take an
11582 additional void * argument, which is just handed through whenever
11583 the password callback is called.
11584 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11586 New function SSL_CTX_set_default_passwd_cb_userdata.
11588 Compatibility note: As many C implementations push function arguments
11589 onto the stack in reverse order, the new library version is likely to
11590 interoperate with programs that have been compiled with the old
11591 pem_password_cb definition (PEM_whatever takes some data that
11592 happens to be on the stack as its last argument, and the callback
11593 just ignores this garbage); but there is no guarantee whatsoever that
11596 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11597 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11598 problems not only on Windows, but also on some Unix platforms.
11599 To avoid problematic command lines, these definitions are now in an
11600 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11601 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11604 *) MIPS III/IV assembler module is reimplemented.
11607 *) More DES library cleanups: remove references to srand/rand and
11608 delete an unused file.
11611 *) Add support for the free Netwide assembler (NASM) under Win32,
11612 since not many people have MASM (ml) and it can be hard to obtain.
11613 This is currently experimental but it seems to work OK and pass all
11614 the tests. Check out INSTALL.W32 for info.
11617 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11618 without temporary keys kept an extra copy of the server key,
11619 and connections with temporary keys did not free everything in case
11623 *) New function RSA_check_key and new openssl rsa option -check
11624 for verifying the consistency of RSA keys.
11625 [Ulf Moeller, Bodo Moeller]
11627 *) Various changes to make Win32 compile work:
11628 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11629 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11630 comparison" warnings.
11631 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11634 *) Add a debugging option to PKCS#5 v2 key generation function: when
11635 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11636 derived keys are printed to stderr.
11639 *) Copy the flags in ASN1_STRING_dup().
11640 [Roman E. Pavlov <pre@mo.msk.ru>]
11642 *) The x509 application mishandled signing requests containing DSA
11643 keys when the signing key was also DSA and the parameters didn't match.
11645 It was supposed to omit the parameters when they matched the signing key:
11646 the verifying software was then supposed to automatically use the CA's
11647 parameters if they were absent from the end user certificate.
11649 Omitting parameters is no longer recommended. The test was also
11650 the wrong way round! This was probably due to unusual behaviour in
11651 EVP_cmp_parameters() which returns 1 if the parameters match.
11652 This meant that parameters were omitted when they *didn't* match and
11653 the certificate was useless. Certificates signed with 'ca' didn't have
11655 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11657 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11658 The interface is as follows:
11659 Applications can use
11660 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11661 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11662 "off" is now the default.
11663 The library internally uses
11664 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11665 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11666 to disable memory-checking temporarily.
11668 Some inconsistent states that previously were possible (and were
11669 even the default) are now avoided.
11671 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11672 with each memory chunk allocated; this is occasionally more helpful
11673 than just having a counter.
11675 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11677 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11681 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11682 which largely parallels "options", but is for changing API behaviour,
11683 whereas "options" are about protocol behaviour.
11684 Initial "mode" flags are:
11686 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11687 a single record has been written.
11688 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11689 retries use the same buffer location.
11690 (But all of the contents must be
11694 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11697 *) Fix problems with no-hmac etc.
11698 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11700 *) New functions RSA_get_default_method(), RSA_set_method() and
11701 RSA_get_method(). These allows replacement of RSA_METHODs without having
11702 to mess around with the internals of an RSA structure.
11705 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11706 Also really enable memory leak checks in openssl.c and in some
11708 [Chad C. Mulligan, Bodo Moeller]
11710 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11711 up the length of negative integers. This has now been simplified to just
11712 store the length when it is first determined and use it later, rather
11713 than trying to keep track of where data is copied and updating it to
11715 [Steve Henson, reported by Brien Wheeler
11716 <bwheeler@authentica-security.com>]
11718 *) Add a new function PKCS7_signatureVerify. This allows the verification
11719 of a PKCS#7 signature but with the signing certificate passed to the
11720 function itself. This contrasts with PKCS7_dataVerify which assumes the
11721 certificate is present in the PKCS#7 structure. This isn't always the
11722 case: certificates can be omitted from a PKCS#7 structure and be
11723 distributed by "out of band" means (such as a certificate database).
11726 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11727 function prototypes in pem.h, also change util/mkdef.pl to add the
11728 necessary function names.
11731 *) mk1mf.pl (used by Windows builds) did not properly read the
11732 options set by Configure in the top level Makefile, and Configure
11733 was not even able to write more than one option correctly.
11734 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11737 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11738 file to be loaded from a BIO or FILE pointer. The BIO version will
11739 for example allow memory BIOs to contain config info.
11742 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11743 Whoever hopes to achieve shared-library compatibility across versions
11744 must use this, not the compile-time macro.
11745 (Exercise 0.9.4: Which is the minimum library version required by
11747 Note: All this applies only to multi-threaded programs, others don't
11751 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11752 through a BIO pair triggered the default case, i.e.
11753 SSLerr(...,SSL_R_UNKNOWN_STATE).
11756 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11757 can use the SSL library even if none of the specific BIOs is
11761 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11762 for the encoded length.
11763 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11765 *) Add initial documentation of the X509V3 functions.
11768 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11769 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11770 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11771 secure PKCS#8 private key format with a high iteration count.
11774 *) Fix determination of Perl interpreter: A perl or perl5
11775 _directory_ in $PATH was also accepted as the interpreter.
11776 [Ralf S. Engelschall]
11778 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11779 wrong with it but it was very old and did things like calling
11780 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11781 unusual formatting.
11784 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11785 to use the new extension code.
11788 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11789 with macros. This should make it easier to change their form, add extra
11790 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11794 *) Add to configuration table a new entry that can specify an alternative
11795 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11796 according to Mark Crispin <MRC@Panda.COM>.
11800 *) DES CBC did not update the IV. Weird.
11803 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11804 Changing the behaviour of the former might break existing programs --
11805 where IV updating is needed, des_ncbc_encrypt can be used.
11808 *) When bntest is run from "make test" it drives bc to check its
11809 calculations, as well as internally checking them. If an internal check
11810 fails, it needs to cause bc to give a non-zero result or make test carries
11811 on without noticing the failure. Fixed.
11814 *) DES library cleanups.
11817 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11818 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11819 ciphers. NOTE: although the key derivation function has been verified
11820 against some published test vectors it has not been extensively tested
11821 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11825 *) Instead of "mkdir -p", which is not fully portable, use new
11826 Perl script "util/mkdir-p.pl".
11829 *) Rewrite the way password based encryption (PBE) is handled. It used to
11830 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11831 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11832 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11833 the 'parameter' field of the AlgorithmIdentifier is passed to the
11834 underlying key generation function so it must do its own ASN1 parsing.
11835 This has also changed the EVP_PBE_CipherInit() function which now has a
11836 'parameter' argument instead of literal salt and iteration count values
11837 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11840 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11841 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11842 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11843 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11844 value was just used as a "magic string" and not used directly its
11845 value doesn't matter.
11848 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11852 *) "linux-sparc64" configuration (ultrapenguin).
11853 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11854 "linux-sparc" configuration.
11855 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11857 *) config now generates no-xxx options for missing ciphers.
11860 *) Support the EBCDIC character set (work in progress).
11861 File ebcdic.c not yet included because it has a different license.
11862 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11864 *) Support BS2000/OSD-POSIX.
11865 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11867 *) Make callbacks for key generation use void * instead of char *.
11870 *) Make S/MIME samples compile (not yet tested).
11873 *) Additional typesafe stacks.
11876 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11880 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11882 *) New configuration variant "sco5-gcc".
11884 *) Updated some demos.
11885 [Sean O Riordain, Wade Scholine]
11887 *) Add missing BIO_free at exit of pkcs12 application.
11890 *) Fix memory leak in conf.c.
11893 *) Updates for Win32 to assembler version of MD5.
11896 *) Set #! path to perl in apps/der_chop to where we found it
11897 instead of using a fixed path.
11900 *) SHA library changes for irix64-mips4-cc.
11903 *) Improvements for VMS support.
11907 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11909 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11910 This also avoids the problems with SC4.2 and unpatched SC5.
11911 [Andy Polyakov <appro@fy.chalmers.se>]
11913 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11914 These are required because of the typesafe stack would otherwise break
11915 existing code. If old code used a structure member which used to be STACK
11916 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11917 sk_num or sk_value it would produce an error because the num, data members
11918 are not present in STACK_OF. Now it just produces a warning. sk_set
11919 replaces the old method of assigning a value to sk_value
11920 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11921 that does this will no longer work (and should use sk_set instead) but
11922 this could be regarded as a "questionable" behaviour anyway.
11925 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11926 correctly handle encrypted S/MIME data.
11929 *) Change type of various DES function arguments from des_cblock
11930 (which means, in function argument declarations, pointer to char)
11931 to des_cblock * (meaning pointer to array with 8 char elements),
11932 which allows the compiler to do more typechecking; it was like
11933 that back in SSLeay, but with lots of ugly casts.
11935 Introduce new type const_des_cblock.
11938 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11939 problems: find RecipientInfo structure that matches recipient certificate
11940 and initialise the ASN1 structures properly based on passed cipher.
11943 *) Belatedly make the BN tests actually check the results.
11946 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11947 to and from BNs: it was completely broken. New compilation option
11948 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11949 key elements as negative integers.
11952 *) Reorganize and speed up MD5.
11953 [Andy Polyakov <appro@fy.chalmers.se>]
11956 [Richard Levitte <richard@levitte.org>]
11958 *) New option -out to asn1parse to allow the parsed structure to be
11959 output to a file. This is most useful when combined with the -strparse
11960 option to examine the output of things like OCTET STRINGS.
11963 *) Make SSL library a little more fool-proof by not requiring any longer
11964 that SSL_set_{accept,connect}_state be called before
11965 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11966 in many applications because usually everything *appeared* to work as
11967 intended anyway -- now it really works as intended).
11970 *) Move openssl.cnf out of lib/.
11973 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11974 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11975 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11976 [Ralf S. Engelschall]
11978 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11979 handle PKCS#7 enveloped data properly.
11980 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11982 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11983 copying pointers. The cert_st handling is changed by this in
11984 various ways (and thus what used to be known as ctx->default_cert
11985 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11986 any longer when s->cert does not give us what we need).
11987 ssl_cert_instantiate becomes obsolete by this change.
11988 As soon as we've got the new code right (possibly it already is?),
11989 we have solved a couple of bugs of the earlier code where s->cert
11990 was used as if it could not have been shared with other SSL structures.
11992 Note that using the SSL API in certain dirty ways now will result
11993 in different behaviour than observed with earlier library versions:
11994 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11995 does not influence s as it used to.
11997 In order to clean up things more thoroughly, inside SSL_SESSION
11998 we don't use CERT any longer, but a new structure SESS_CERT
11999 that holds per-session data (if available); currently, this is
12000 the peer's certificate chain and, for clients, the server's certificate
12001 and temporary key. CERT holds only those values that can have
12002 meaningful defaults in an SSL_CTX.
12005 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12006 from the internal representation. Various PKCS#7 fixes: remove some
12007 evil casts and set the enc_dig_alg field properly based on the signing
12011 *) Allow PKCS#12 password to be set from the command line or the
12012 environment. Let 'ca' get its config file name from the environment
12013 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12017 *) Allow certificate policies extension to use an IA5STRING for the
12018 organization field. This is contrary to the PKIX definition but
12019 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12023 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12024 without disallowing inline assembler and the like for non-pedantic builds.
12027 *) Support Borland C++ builder.
12028 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12030 *) Support Mingw32.
12033 *) SHA-1 cleanups and performance enhancements.
12034 [Andy Polyakov <appro@fy.chalmers.se>]
12036 *) Sparc v8plus assembler for the bignum library.
12037 [Andy Polyakov <appro@fy.chalmers.se>]
12039 *) Accept any -xxx and +xxx compiler options in Configure.
12042 *) Update HPUX configuration.
12045 *) Add missing sk_<type>_unshift() function to safestack.h
12046 [Ralf S. Engelschall]
12048 *) New function SSL_CTX_use_certificate_chain_file that sets the
12049 "extra_cert"s in addition to the certificate. (This makes sense
12050 only for "PEM" format files, as chains as a whole are not
12054 *) Support verify_depth from the SSL API.
12055 x509_vfy.c had what can be considered an off-by-one-error:
12056 Its depth (which was not part of the external interface)
12057 was actually counting the number of certificates in a chain;
12058 now it really counts the depth.
12061 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12062 instead of X509err, which often resulted in confusing error
12063 messages since the error codes are not globally unique
12064 (e.g. an alleged error in ssl3_accept when a certificate
12065 didn't match the private key).
12067 *) New function SSL_CTX_set_session_id_context that allows to set a default
12068 value (so that you don't need SSL_set_session_id_context for each
12069 connection using the SSL_CTX).
12072 *) OAEP decoding bug fix.
12075 *) Support INSTALL_PREFIX for package builders, as proposed by
12079 *) New Configure options "threads" and "no-threads". For systems
12080 where the proper compiler options are known (currently Solaris
12081 and Linux), "threads" is the default.
12084 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12087 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12088 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12089 such as /usr/local/bin.
12092 *) "make linux-shared" to build shared libraries.
12093 [Niels Poppe <niels@netbox.org>]
12095 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12098 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12099 extension adding in x509 utility.
12102 *) Remove NOPROTO sections and error code comments.
12105 *) Partial rewrite of the DEF file generator to now parse the ANSI
12109 *) New Configure options --prefix=DIR and --openssldir=DIR.
12112 *) Complete rewrite of the error code script(s). It is all now handled
12113 by one script at the top level which handles error code gathering,
12114 header rewriting and C source file generation. It should be much better
12115 than the old method: it now uses a modified version of Ulf's parser to
12116 read the ANSI prototypes in all header files (thus the old K&R definitions
12117 aren't needed for error creation any more) and do a better job of
12118 translating function codes into names. The old 'ASN1 error code imbedded
12119 in a comment' is no longer necessary and it doesn't use .err files which
12120 have now been deleted. Also the error code call doesn't have to appear all
12121 on one line (which resulted in some large lines...).
12124 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12127 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12128 0 (which usually indicates a closed connection), but continue reading.
12131 *) Fix some race conditions.
12134 *) Add support for CRL distribution points extension. Add Certificate
12135 Policies and CRL distribution points documentation.
12138 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12141 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12142 8 of keying material. Merlin has also confirmed interop with this fix
12143 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12144 [Merlin Hughes <merlin@baltimore.ie>]
12146 *) Fix lots of warnings.
12147 [Richard Levitte <levitte@stacken.kth.se>]
12149 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12150 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12151 [Richard Levitte <levitte@stacken.kth.se>]
12153 *) Fix problems with sizeof(long) == 8.
12154 [Andy Polyakov <appro@fy.chalmers.se>]
12156 *) Change functions to ANSI C.
12159 *) Fix typos in error codes.
12160 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12162 *) Remove defunct assembler files from Configure.
12165 *) SPARC v8 assembler BIGNUM implementation.
12166 [Andy Polyakov <appro@fy.chalmers.se>]
12168 *) Support for Certificate Policies extension: both print and set.
12169 Various additions to support the r2i method this uses.
12172 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12173 return a const string when you are expecting an allocated buffer.
12176 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12177 types DirectoryString and DisplayText.
12180 *) Add code to allow r2i extensions to access the configuration database,
12181 add an LHASH database driver and add several ctx helper functions.
12184 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12185 fail when they extended the size of a BIGNUM.
12188 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12189 support typesafe stack.
12192 *) Fix typo in SSL_[gs]et_options().
12193 [Nils Frostberg <nils@medcom.se>]
12195 *) Delete various functions and files that belonged to the (now obsolete)
12196 old X509V3 handling code.
12199 *) New Configure option "rsaref".
12202 *) Don't auto-generate pem.h.
12205 *) Introduce type-safe ASN.1 SETs.
12208 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12209 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12211 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12212 that links with OpenSSL (well at least cause lots of warnings), but fear
12213 not: the conversion is trivial, and it eliminates loads of evil casts. A
12214 few STACKed things have been converted already. Feel free to convert more.
12215 In the fullness of time, I'll do away with the STACK type altogether.
12218 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12219 specified in <certfile> by updating the entry in the index.txt file.
12220 This way one no longer has to edit the index.txt file manually for
12221 revoking a certificate. The -revoke option does the gory details now.
12222 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12224 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12225 `-text' option at all and this way the `-noout -text' combination was
12226 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12227 [Ralf S. Engelschall]
12229 *) Make sure a corresponding plain text error message exists for the
12230 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12231 verify callback function determined that a certificate was revoked.
12232 [Ralf S. Engelschall]
12234 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12235 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12236 all available ciphers including rc5, which was forgotten until now.
12237 In order to let the testing shell script know which algorithms
12238 are available, a new (up to now undocumented) command
12239 "openssl list-cipher-commands" is used.
12242 *) Bugfix: s_client occasionally would sleep in select() when
12243 it should have checked SSL_pending() first.
12246 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12247 the raw DSA values prior to ASN.1 encoding.
12250 *) Tweaks to Configure
12251 [Niels Poppe <niels@netbox.org>]
12253 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12257 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12260 *) New config option to avoid instructions that are illegal on the 80386.
12261 The default code is faster, but requires at least a 486.
12264 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12265 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12266 same as SSL2_VERSION anyway.
12269 *) New "-showcerts" option for s_client.
12272 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12273 application. Various cleanups and fixes.
12276 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12277 modify error routines to work internally. Add error codes and PBE init
12278 to library startup routines.
12281 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12282 packing functions to asn1 and evp. Changed function names and error
12283 codes along the way.
12286 *) PKCS12 integration: and so it begins... First of several patches to
12287 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12288 objects to objects.h
12291 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12292 and display support for Thawte strong extranet extension.
12295 *) Add LinuxPPC support.
12296 [Jeff Dubrule <igor@pobox.org>]
12298 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12299 bn_div_words in alpha.s.
12300 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12302 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12303 OAEP isn't supported when OpenSSL is built with RSAref.
12304 [Ulf Moeller <ulf@fitug.de>]
12306 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12307 so they no longer are missing under -DNOPROTO.
12308 [Soren S. Jorvang <soren@t.dk>]
12311 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12313 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12314 doesn't work when the session is reused. Coming soon!
12317 *) Fix a security hole, that allows sessions to be reused in the wrong
12318 context thus bypassing client cert protection! All software that uses
12319 client certs and session caches in multiple contexts NEEDS PATCHING to
12320 allow session reuse! A fuller solution is in the works.
12321 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12323 *) Some more source tree cleanups (removed obsolete files
12324 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12325 permission on "config" script to be executable) and a fix for the INSTALL
12327 [Ulf Moeller <ulf@fitug.de>]
12329 *) Remove some legacy and erroneous uses of malloc, free instead of
12331 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12333 *) Make rsa_oaep_test return non-zero on error.
12334 [Ulf Moeller <ulf@fitug.de>]
12336 *) Add support for native Solaris shared libraries. Configure
12337 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12338 if someone would make that last step automatic.
12339 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12341 *) ctx_size was not built with the right compiler during "make links". Fixed.
12344 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12345 except NULL ciphers". This means the default cipher list will no longer
12346 enable NULL ciphers. They need to be specifically enabled e.g. with
12347 the string "DEFAULT:eNULL".
12350 *) Fix to RSA private encryption routines: if p < q then it would
12351 occasionally produce an invalid result. This will only happen with
12352 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12355 *) Be less restrictive and allow also `perl util/perlpath.pl
12356 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12357 because this way one can also use an interpreter named `perl5' (which is
12358 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12359 installed as `perl').
12360 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12362 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12363 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12365 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12366 advapi32.lib to Win32 build and change the pem test comparison
12367 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12368 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12369 and crypto/des/ede_cbcm_enc.c.
12372 *) DES quad checksum was broken on big-endian architectures. Fixed.
12375 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12376 Win32 test batch file so it (might) work again. The Win32 test batch file
12377 is horrible: I feel ill....
12380 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12381 in e_os.h. Audit of header files to check ANSI and non ANSI
12382 sections: 10 functions were absent from non ANSI section and not exported
12383 from Windows DLLs. Fixed up libeay.num for new functions.
12386 *) Make `openssl version' output lines consistent.
12387 [Ralf S. Engelschall]
12389 *) Fix Win32 symbol export lists for BIO functions: Added
12390 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12391 to ms/libeay{16,32}.def.
12392 [Ralf S. Engelschall]
12394 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12395 fine under Unix and passes some trivial tests I've now added. But the
12396 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12397 added to make sure no one expects that this stuff really works in the
12398 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12399 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12401 [Ralf S. Engelschall]
12403 *) Fix the generation of two part addresses in perl.
12404 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12406 *) Add config entry for Linux on MIPS.
12407 [John Tobey <jtobey@channel1.com>]
12409 *) Make links whenever Configure is run, unless we are on Windoze.
12412 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12413 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12417 *) Add a useful kludge to allow package maintainers to specify compiler and
12418 other platforms details on the command line without having to patch the
12419 Configure script everytime: One now can use ``perl Configure
12420 <id>:<details>'', i.e. platform ids are allowed to have details appended
12421 to them (separated by colons). This is treated as there would be a static
12422 pre-configured entry in Configure's %table under key <id> with value
12423 <details> and ``perl Configure <id>'' is called. So, when you want to
12424 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12425 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12426 now, which overrides the FreeBSD-elf entry on-the-fly.
12427 [Ralf S. Engelschall]
12429 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12432 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12433 on the `perl Configure ...' command line. This way one can compile
12434 OpenSSL libraries with Position Independent Code (PIC) which is needed
12435 for linking it into DSOs.
12436 [Ralf S. Engelschall]
12438 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12442 *) Cleaned up the LICENSE document: The official contact for any license
12443 questions now is the OpenSSL core team under openssl-core@openssl.org.
12444 And add a paragraph about the dual-license situation to make sure people
12445 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12446 to the OpenSSL toolkit.
12447 [Ralf S. Engelschall]
12449 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12450 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12451 Additionally cleaned up the `make links' target: Remove unnecessary
12452 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12453 to speed processing and no longer clutter the display with confusing
12454 stuff. Instead only the actually done links are displayed.
12455 [Ralf S. Engelschall]
12457 *) Permit null encryption ciphersuites, used for authentication only. It used
12458 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12459 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12463 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12464 signed attributes when verifying signatures (this would break them),
12465 the detached data encoding was wrong and public keys obtained using
12466 X509_get_pubkey() weren't freed.
12469 *) Add text documentation for the BUFFER functions. Also added a work around
12470 to a Win95 console bug. This was triggered by the password read stuff: the
12471 last character typed gets carried over to the next fread(). If you were
12472 generating a new cert request using 'req' for example then the last
12473 character of the passphrase would be CR which would then enter the first
12477 *) Added the new `Includes OpenSSL Cryptography Software' button as
12478 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12479 button and can be used by applications based on OpenSSL to show the
12480 relationship to the OpenSSL project.
12481 [Ralf S. Engelschall]
12483 *) Remove confusing variables in function signatures in files
12484 ssl/ssl_lib.c and ssl/ssl.h.
12485 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12487 *) Don't install bss_file.c under PREFIX/include/
12488 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12490 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12491 functions that return function pointers and has support for NT specific
12492 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12493 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12494 unsigned to signed types: this was killing the Win32 compile.
12497 *) Add new certificate file to stack functions,
12498 SSL_add_dir_cert_subjects_to_stack() and
12499 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12500 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12501 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12502 This means that Apache-SSL and similar packages don't have to mess around
12503 to add as many CAs as they want to the preferred list.
12506 *) Experiment with doxygen documentation. Currently only partially applied to
12508 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12509 openssl.doxy as the configuration file.
12512 *) Get rid of remaining C++-style comments which strict C compilers hate.
12513 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12515 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12516 compiled in by default: it has problems with large keys.
12519 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12520 DH private keys and/or callback functions which directly correspond to
12521 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12522 is needed for applications which have to configure certificates on a
12523 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12525 For the RSA certificate situation is makes no difference, but
12526 for the DSA certificate situation this fixes the "no shared cipher"
12527 problem where the OpenSSL cipher selection procedure failed because the
12528 temporary keys were not overtaken from the context and the API provided
12529 no way to reconfigure them.
12530 The new functions now let applications reconfigure the stuff and they
12531 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12532 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12533 non-public-API function ssl_cert_instantiate() is used as a helper
12534 function and also to reduce code redundancy inside ssl_rsa.c.
12535 [Ralf S. Engelschall]
12537 *) Move s_server -dcert and -dkey options out of the undocumented feature
12538 area because they are useful for the DSA situation and should be
12539 recognized by the users.
12540 [Ralf S. Engelschall]
12542 *) Fix the cipher decision scheme for export ciphers: the export bits are
12543 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12544 SSL_EXP_MASK. So, the original variable has to be used instead of the
12545 already masked variable.
12546 [Richard Levitte <levitte@stacken.kth.se>]
12548 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12549 [Richard Levitte <levitte@stacken.kth.se>]
12551 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12552 from `int' to `unsigned int' because it's a length and initialized by
12553 EVP_DigestFinal() which expects an `unsigned int *'.
12554 [Richard Levitte <levitte@stacken.kth.se>]
12556 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12557 script. Instead use the usual Shell->Perl transition trick.
12558 [Ralf S. Engelschall]
12560 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12561 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12562 -noout -modulus' as it's already the case for `openssl rsa -noout
12563 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12564 currently the public key is printed (a decision which was already done by
12565 `openssl dsa -modulus' in the past) which serves a similar purpose.
12566 Additionally the NO_RSA no longer completely removes the whole -modulus
12567 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12569 [Ralf S. Engelschall]
12571 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12572 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12573 [Arne Ansper <arne@ats.cyber.ee>]
12575 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12576 to be added. Now both 'req' and 'ca' can use new objects defined in the
12580 *) Add cool BIO that does syslog (or event log on NT).
12581 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12583 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12584 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12585 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12586 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12589 *) Add preliminary config info for new extension code.
12592 *) Make RSA_NO_PADDING really use no padding.
12593 [Ulf Moeller <ulf@fitug.de>]
12595 *) Generate errors when private/public key check is done.
12598 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12599 for some CRL extensions and new objects added.
12602 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12603 key usage extension and fuller support for authority key id.
12606 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12607 padding method for RSA, which is recommended for new applications in PKCS
12608 #1 v2.0 (RFC 2437, October 1998).
12609 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12610 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12611 against Bleichbacher's attack on RSA.
12612 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12615 *) Updates to the new SSL compression code
12616 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12618 *) Fix so that the version number in the master secret, when passed
12619 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12620 (because the server will not accept higher), that the version number
12621 is 0x03,0x01, not 0x03,0x00
12622 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12624 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12625 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12626 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12629 *) Support for RAW extensions where an arbitrary extension can be
12630 created by including its DER encoding. See apps/openssl.cnf for
12634 *) Make sure latest Perl versions don't interpret some generated C array
12635 code as Perl array code in the crypto/err/err_genc.pl script.
12636 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12638 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12639 not many people have the assembler. Various Win32 compilation fixes and
12640 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12641 build instructions.
12644 *) Modify configure script 'Configure' to automatically create crypto/date.h
12645 file under Win32 and also build pem.h from pem.org. New script
12646 util/mkfiles.pl to create the MINFO file on environments that can't do a
12647 'make files': perl util/mkfiles.pl >MINFO should work.
12650 *) Major rework of DES function declarations, in the pursuit of correctness
12651 and purity. As a result, many evil casts evaporated, and some weirdness,
12652 too. You may find this causes warnings in your code. Zapping your evil
12653 casts will probably fix them. Mostly.
12656 *) Fix for a typo in asn1.h. Bug fix to object creation script
12657 obj_dat.pl. It considered a zero in an object definition to mean
12658 "end of object": none of the objects in objects.h have any zeros
12659 so it wasn't spotted.
12660 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12662 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12663 Masking (CBCM). In the absence of test vectors, the best I have been able
12664 to do is check that the decrypt undoes the encrypt, so far. Send me test
12665 vectors if you have them.
12668 *) Correct calculation of key length for export ciphers (too much space was
12669 allocated for null ciphers). This has not been tested!
12672 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12673 message is now correct (it understands "crypto" and "ssl" on its
12674 command line). There is also now an "update" option. This will update
12675 the util/ssleay.num and util/libeay.num files with any new functions.
12677 perl util/mkdef.pl crypto ssl update
12678 it will update them.
12681 *) Overhauled the Perl interface (perl/*):
12682 - ported BN stuff to OpenSSL's different BN library
12683 - made the perl/ source tree CVS-aware
12684 - renamed the package from SSLeay to OpenSSL (the files still contain
12685 their history because I've copied them in the repository)
12686 - removed obsolete files (the test scripts will be replaced
12687 by better Test::Harness variants in the future)
12688 [Ralf S. Engelschall]
12690 *) First cut for a very conservative source tree cleanup:
12691 1. merge various obsolete readme texts into doc/ssleay.txt
12692 where we collect the old documents and readme texts.
12693 2. remove the first part of files where I'm already sure that we no
12694 longer need them because of three reasons: either they are just temporary
12695 files which were left by Eric or they are preserved original files where
12696 I've verified that the diff is also available in the CVS via "cvs diff
12697 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12698 the crypto/md/ stuff).
12699 [Ralf S. Engelschall]
12701 *) More extension code. Incomplete support for subject and issuer alt
12702 name, issuer and authority key id. Change the i2v function parameters
12703 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12704 what that's for :-) Fix to ASN1 macro which messed up
12705 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12708 *) Preliminary support for ENUMERATED type. This is largely copied from the
12712 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12713 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12715 *) Make sure `make rehash' target really finds the `openssl' program.
12716 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12718 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12719 like to hear about it if this slows down other processors.
12722 *) Add CygWin32 platform information to Configure script.
12723 [Alan Batie <batie@aahz.jf.intel.com>]
12725 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12726 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12728 *) New program nseq to manipulate netscape certificate sequences
12731 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12735 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12736 but the BN code had some problems that would cause failures when
12737 doing certificate verification and some other functions.
12738 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12740 *) Add ASN1 and PEM code to support netscape certificate sequences.
12743 *) Add ASN1 and PEM code to support netscape certificate sequences.
12746 *) Add several PKIX and private extended key usage OIDs.
12749 *) Modify the 'ca' program to handle the new extension code. Modify
12750 openssl.cnf for new extension format, add comments.
12753 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12754 and add a sample to openssl.cnf so req -x509 now adds appropriate
12758 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12759 error code, add initial support to X509_print() and x509 application.
12762 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12763 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12764 stuff is currently isolated and isn't even compiled yet.
12767 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12768 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12769 Removed the versions check from X509 routines when loading extensions:
12770 this allows certain broken certificates that don't set the version
12771 properly to be processed.
12774 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12775 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12776 can still be regenerated with "make depend".
12779 *) Spelling mistake in C version of CAST-128.
12780 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12782 *) Changes to the error generation code. The perl script err-code.pl
12783 now reads in the old error codes and retains the old numbers, only
12784 adding new ones if necessary. It also only changes the .err files if new
12785 codes are added. The makefiles have been modified to only insert errors
12786 when needed (to avoid needlessly modifying header files). This is done
12787 by only inserting errors if the .err file is newer than the auto generated
12788 C file. To rebuild all the error codes from scratch (the old behaviour)
12789 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12790 or delete all the .err files.
12793 *) CAST-128 was incorrectly implemented for short keys. The C version has
12794 been fixed, but is untested. The assembler versions are also fixed, but
12795 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12796 to regenerate it if needed.
12797 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12798 Hagino <itojun@kame.net>]
12800 *) File was opened incorrectly in randfile.c.
12801 [Ulf Möller <ulf@fitug.de>]
12803 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12804 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12805 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12806 al: it's just almost always a UTCTime. Note this patch adds new error
12807 codes so do a "make errors" if there are problems.
12810 *) Correct Linux 1 recognition in config.
12811 [Ulf Möller <ulf@fitug.de>]
12813 *) Remove pointless MD5 hash when using DSA keys in ca.
12814 [Anonymous <nobody@replay.com>]
12816 *) Generate an error if given an empty string as a cert directory. Also
12817 generate an error if handed NULL (previously returned 0 to indicate an
12818 error, but didn't set one).
12819 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12821 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12824 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12825 parameters. This was causing a warning which killed off the Win32 compile.
12828 *) Remove C++ style comments from crypto/bn/bn_local.h.
12829 [Neil Costigan <neil.costigan@celocom.com>]
12831 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12832 based on a text string, looking up short and long names and finally
12833 "dot" format. The "dot" format stuff didn't work. Added new function
12834 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12835 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12836 OID is not part of the table.
12839 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12840 X509_LOOKUP_by_alias().
12843 *) Sort openssl functions by name.
12846 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12847 encryption from sample DSA keys (in case anyone is interested the password
12851 *) Make _all_ *_free functions accept a NULL pointer.
12852 [Frans Heymans <fheymans@isaserver.be>]
12854 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12856 [Anonymous <nobody@replay.com>]
12858 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12859 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12861 *) Don't blow it for numeric -newkey arguments to apps/req.
12862 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12864 *) Temp key "for export" tests were wrong in s3_srvr.c.
12865 [Anonymous <nobody@replay.com>]
12867 *) Add prototype for temp key callback functions
12868 SSL_CTX_set_tmp_{rsa,dh}_callback().
12871 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12872 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12875 *) X509_name_add_entry() freed the wrong thing after an error.
12876 [Arne Ansper <arne@ats.cyber.ee>]
12878 *) rsa_eay.c would attempt to free a NULL context.
12879 [Arne Ansper <arne@ats.cyber.ee>]
12881 *) BIO_s_socket() had a broken should_retry() on Windoze.
12882 [Arne Ansper <arne@ats.cyber.ee>]
12884 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12885 [Arne Ansper <arne@ats.cyber.ee>]
12887 *) Make sure the already existing X509_STORE->depth variable is initialized
12888 in X509_STORE_new(), but document the fact that this variable is still
12889 unused in the certificate verification process.
12890 [Ralf S. Engelschall]
12892 *) Fix the various library and apps files to free up pkeys obtained from
12893 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12896 *) Fix reference counting in X509_PUBKEY_get(). This makes
12897 demos/maurice/example2.c work, amongst others, probably.
12898 [Steve Henson and Ben Laurie]
12900 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12901 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12902 are no longer created. This way we have a single and consistent command
12903 line interface `openssl <command>', similar to `cvs <command>'.
12904 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12906 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12907 BIT STRING wrapper always have zero unused bits.
12910 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12913 *) Make the top-level INSTALL documentation easier to understand.
12916 *) Makefiles updated to exit if an error occurs in a sub-directory
12917 make (including if user presses ^C) [Paul Sutton]
12919 *) Make Montgomery context stuff explicit in RSA data structure.
12922 *) Fix build order of pem and err to allow for generated pem.h.
12925 *) Fix renumbering bug in X509_NAME_delete_entry().
12928 *) Enhanced the err-ins.pl script so it makes the error library number
12929 global and can add a library name. This is needed for external ASN1 and
12930 other error libraries.
12933 *) Fixed sk_insert which never worked properly.
12936 *) Fix ASN1 macros so they can handle indefinite length constructed
12937 EXPLICIT tags. Some non standard certificates use these: they can now
12941 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12942 into a single doc/ssleay.txt bundle. This way the information is still
12943 preserved but no longer messes up this directory. Now it's new room for
12944 the new set of documentation files.
12945 [Ralf S. Engelschall]
12947 *) SETs were incorrectly DER encoded. This was a major pain, because they
12948 shared code with SEQUENCEs, which aren't coded the same. This means that
12949 almost everything to do with SETs or SEQUENCEs has either changed name or
12950 number of arguments.
12951 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12953 *) Fix test data to work with the above.
12956 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12957 was already fixed by Eric for 0.9.1 it seems.
12958 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12960 *) Autodetect FreeBSD3.
12963 *) Fix various bugs in Configure. This affects the following platforms:
12967 unixware-2.0-pentium
12971 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12972 before they are needed.
12975 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12979 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12981 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12982 changed SSLeay to OpenSSL in version strings.
12983 [Ralf S. Engelschall]
12985 *) Some fixups to the top-level documents.
12988 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12989 because the symlink to include/ was missing.
12990 [Ralf S. Engelschall]
12992 *) Incorporated the popular no-RSA/DSA-only patches
12993 which allow to compile a RSA-free SSLeay.
12994 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12996 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12997 when "ssleay" is still not found.
12998 [Ralf S. Engelschall]
13000 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13001 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13003 *) Updated the README file.
13004 [Ralf S. Engelschall]
13006 *) Added various .cvsignore files in the CVS repository subdirs
13007 to make a "cvs update" really silent.
13008 [Ralf S. Engelschall]
13010 *) Recompiled the error-definition header files and added
13011 missing symbols to the Win32 linker tables.
13012 [Ralf S. Engelschall]
13014 *) Cleaned up the top-level documents;
13015 o new files: CHANGES and LICENSE
13016 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13017 o merged COPYRIGHT into LICENSE
13018 o removed obsolete TODO file
13019 o renamed MICROSOFT to INSTALL.W32
13020 [Ralf S. Engelschall]
13022 *) Removed dummy files from the 0.9.1b source tree:
13023 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13024 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13025 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13026 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13027 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13028 [Ralf S. Engelschall]
13030 *) Added various platform portability fixes.
13033 *) The Genesis of the OpenSSL rpject:
13034 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13035 Young and Tim J. Hudson created while they were working for C2Net until
13037 [The OpenSSL Project]
13040 Changes between 0.9.0b and 0.9.1b [not released]
13042 *) Updated a few CA certificates under certs/
13045 *) Changed some BIGNUM api stuff.
13048 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13049 DGUX x86, Linux Alpha, etc.
13052 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13053 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13057 *) Add -strparse option to asn1pars program which parses nested
13059 [Dr Stephen Henson <shenson@bigfoot.com>]
13061 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13064 *) DSA fix for "ca" program.
13067 *) Added "-genkey" option to "dsaparam" program.
13070 *) Added RIPE MD160 (rmd160) message digest.
13073 *) Added -a (all) option to "ssleay version" command.
13076 *) Added PLATFORM define which is the id given to Configure.
13079 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13082 *) Extended the ASN.1 parser routines.
13085 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13088 *) Added a BN_CTX to the BN library.
13091 *) Fixed the weak key values in DES library
13094 *) Changed API in EVP library for cipher aliases.
13097 *) Added support for RC2/64bit cipher.
13100 *) Converted the lhash library to the crypto/mem.c functions.
13103 *) Added more recognized ASN.1 object ids.
13106 *) Added more RSA padding checks for SSL/TLS.
13109 *) Added BIO proxy/filter functionality.
13112 *) Added extra_certs to SSL_CTX which can be used
13113 send extra CA certificates to the client in the CA cert chain sending
13114 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13117 *) Now Fortezza is denied in the authentication phase because
13118 this is key exchange mechanism is not supported by SSLeay at all.
13121 *) Additional PKCS1 checks.
13124 *) Support the string "TLSv1" for all TLS v1 ciphers.
13127 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13128 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13131 *) Fixed a few memory leaks.
13134 *) Fixed various code and comment typos.
13137 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13138 bytes sent in the client random.
13139 [Edward Bishop <ebishop@spyglass.com>]