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 1.1.0i and 1.1.1 [xx XXX xxxx]
12 *) Add SM2 base algorithm support.
15 *) s390x assembly pack: add (improved) hardware-support for the following
16 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
17 aes-cfb/cfb8, aes-ecb.
20 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
21 parameter is no longer accepted, as it leads to a corrupt table. NULL
22 pem_str is reserved for alias entries only.
25 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
26 step for prime curves. The new implementation is based on formulae from
27 differential addition-and-doubling in homogeneous projective coordinates
28 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
29 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
30 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
31 to work in projective coordinates.
32 [Billy Bob Brumley, Nicola Tuveri]
34 *) Change generating and checking of primes so that the error rate of not
35 being prime depends on the intended use based on the size of the input.
36 For larger primes this will result in more rounds of Miller-Rabin.
37 The maximal error rate for primes with more than 1080 bits is lowered
39 [Kurt Roeckx, Annie Yousar]
41 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
44 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
45 moving between systems, and to avoid confusion when a Windows build is
46 done with mingw vs with MSVC. For POSIX installs, there's still a
47 symlink or copy named 'tsget' to avoid that confusion as well.
50 *) Revert blinding in ECDSA sign and instead make problematic addition
51 length-invariant. Switch even to fixed-length Montgomery multiplication.
54 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
55 step for binary curves. The new implementation is based on formulae from
56 differential addition-and-doubling in mixed Lopez-Dahab projective
57 coordinates, modified to independently blind the operands.
58 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
60 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
61 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
62 EC_METHODs to implement their own specialized "ladder step", to take
63 advantage of more favorable coordinate systems or more efficient
64 differential addition-and-doubling algorithms.
65 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
67 *) Modified the random device based seed sources to keep the relevant
68 file descriptors open rather than reopening them on each access.
69 This allows such sources to operate in a chroot() jail without
70 the associated device nodes being available. This behaviour can be
71 controlled using RAND_keep_random_devices_open().
74 *) Numerous side-channel attack mitigations have been applied. This may have
75 performance impacts for some algorithms for the benefit of improved
76 security. Specific changes are noted in this change log by their respective
80 *) AIX shared library support overhaul. Switch to AIX "natural" way of
81 handling shared libraries, which means collecting shared objects of
82 different versions and bitnesses in one common archive. This allows to
83 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
84 doesn't affect the way 3rd party applications are linked, only how
85 multi-version installation is managed.
88 *) Make ec_group_do_inverse_ord() more robust and available to other
89 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
90 mitigations are applied to the fallback BN_mod_inverse().
91 When using this function rather than BN_mod_inverse() directly, new
92 EC cryptosystem implementations are then safer-by-default.
95 *) Add coordinate blinding for EC_POINT and implement projective
96 coordinate blinding for generic prime curves as a countermeasure to
97 chosen point SCA attacks.
98 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
100 *) Add blinding to ECDSA and DSA signatures to protect against side channel
101 attacks discovered by Keegan Ryan (NCC Group).
104 *) Enforce checking in the pkeyutl command line app to ensure that the input
105 length does not exceed the maximum supported digest length when performing
106 a sign, verify or verifyrecover operation.
109 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
110 I/O in combination with something like select() or poll() will hang. This
111 can be turned off again using SSL_CTX_clear_mode().
112 Many applications do not properly handle non-application data records, and
113 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
114 around the problems in those applications, but can also break some.
115 It's recommended to read the manpages about SSL_read(), SSL_write(),
116 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
117 SSL_CTX_set_read_ahead() again.
120 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
121 now allow empty (zero character) pass phrases.
124 *) Apply blinding to binary field modular inversion and remove patent
125 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
128 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
129 binary and prime elliptic curves.
132 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
133 constant time fixed point multiplication.
136 *) Revise elliptic curve scalar multiplication with timing attack
137 defenses: ec_wNAF_mul redirects to a constant time implementation
138 when computing fixed point and variable point multiplication (which
139 in OpenSSL are mostly used with secret scalars in keygen, sign,
140 ECDH derive operations).
141 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
144 *) Updated CONTRIBUTING
147 *) Updated DRBG / RAND to request nonce and additional low entropy
148 randomness from the system.
149 [Matthias St. Pierre]
151 *) Updated 'openssl rehash' to use OpenSSL consistent default.
154 *) Moved the load of the ssl_conf module to libcrypto, which helps
155 loading engines that libssl uses before libssl is initialised.
158 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
161 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
162 [Ingo Schwarze, Rich Salz]
164 *) Added output of accepting IP address and port for 'openssl s_server'
167 *) Added a new API for TLSv1.3 ciphersuites:
168 SSL_CTX_set_ciphersuites()
169 SSL_set_ciphersuites()
172 *) Memory allocation failures consistenly add an error to the error
176 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
177 in libcrypto when run as setuid/setgid.
180 *) Load any config file by default when libssl is used.
183 *) Added new public header file <openssl/rand_drbg.h> and documentation
184 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
185 [Matthias St. Pierre]
187 *) QNX support removed (cannot find contributors to get their approval
188 for the license change).
191 *) TLSv1.3 replay protection for early data has been implemented. See the
192 SSL_read_early_data() man page for further details.
195 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
196 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
197 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
198 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
199 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
200 configuration has been separated out. See the ciphers man page or the
201 SSL_CTX_set_ciphersuites() man page for more information.
204 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
205 in responder mode now supports the new "-multi" option, which
206 spawns the specified number of child processes to handle OCSP
207 requests. The "-timeout" option now also limits the OCSP
208 responder's patience to wait to receive the full client request
209 on a newly accepted connection. Child processes are respawned
210 as needed, and the CA index file is automatically reloaded
211 when changed. This makes it possible to run the "ocsp" responder
212 as a long-running service, making the OpenSSL CA somewhat more
213 feature-complete. In this mode, most diagnostic messages logged
214 after entering the event loop are logged via syslog(3) rather than
218 *) Added support for X448 and Ed448. Heavily based on original work by
222 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
223 objects loaded. This adds the functions OSSL_STORE_expect() and
224 OSSL_STORE_find() as well as needed tools to construct searches and
225 get the search data out of them.
228 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
229 version of OpenSSL should review their configuration settings to ensure
230 that they are still appropriate for TLSv1.3. For further information see:
231 https://wiki.openssl.org/index.php/TLS1.3
234 *) Grand redesign of the OpenSSL random generator
236 The default RAND method now utilizes an AES-CTR DRBG according to
237 NIST standard SP 800-90Ar1. The new random generator is essentially
238 a port of the default random generator from the OpenSSL FIPS 2.0
239 object module. It is a hybrid deterministic random bit generator
240 using an AES-CTR bit stream and which seeds and reseeds itself
241 automatically using trusted system entropy sources.
243 Some of its new features are:
244 o Support for multiple DRBG instances with seed chaining.
245 o The default RAND method makes use of a DRBG.
246 o There is a public and private DRBG instance.
247 o The DRBG instances are fork-safe.
248 o Keep all global DRBG instances on the secure heap if it is enabled.
249 o The public and private DRBG instance are per thread for lock free
251 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
253 *) Changed Configure so it only says what it does and doesn't dump
254 so much data. Instead, ./configdata.pm should be used as a script
255 to display all sorts of configuration data.
258 *) Added processing of "make variables" to Configure.
261 *) Added SHA512/224 and SHA512/256 algorithm support.
264 *) The last traces of Netware support, first removed in 1.1.0, have
268 *) Get rid of Makefile.shared, and in the process, make the processing
269 of certain files (rc.obj, or the .def/.map/.opt files produced from
270 the ordinal files) more visible and hopefully easier to trace and
271 debug (or make silent).
274 *) Make it possible to have environment variable assignments as
275 arguments to config / Configure.
278 *) Add multi-prime RSA (RFC 8017) support.
281 *) Add SM3 implemented according to GB/T 32905-2016
282 [ Jack Lloyd <jack.lloyd@ribose.com>,
283 Ronald Tse <ronald.tse@ribose.com>,
284 Erick Borsboom <erick.borsboom@ribose.com> ]
286 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
287 as documented in RFC6066.
288 Based on a patch from Tomasz Moń
289 [Filipe Raimundo da Silva]
291 *) Add SM4 implemented according to GB/T 32907-2016.
292 [ Jack Lloyd <jack.lloyd@ribose.com>,
293 Ronald Tse <ronald.tse@ribose.com>,
294 Erick Borsboom <erick.borsboom@ribose.com> ]
296 *) Reimplement -newreq-nodes and ERR_error_string_n; the
297 original author does not agree with the license change.
300 *) Add ARIA AEAD TLS support.
303 *) Some macro definitions to support VS6 have been removed. Visual
304 Studio 6 has not worked since 1.1.0
307 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
308 without clearing the errors.
311 *) Add "atfork" functions. If building on a system that without
312 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
313 requirements. The RAND facility now uses/requires this.
319 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
320 not possible to disable entirely. However, it's still possible to
321 disable the console reading UI method, UI_OpenSSL() (use UI_null()
324 To disable, configure with 'no-ui-console'. 'no-ui' is still
325 possible to use as an alias. Check at compile time with the
326 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
327 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
330 *) Add a STORE module, which implements a uniform and URI based reader of
331 stores that can contain keys, certificates, CRLs and numerous other
332 objects. The main API is loosely based on a few stdio functions,
333 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
334 OSSL_STORE_error and OSSL_STORE_close.
335 The implementation uses backends called "loaders" to implement arbitrary
336 URI schemes. There is one built in "loader" for the 'file' scheme.
339 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
340 then adjusted to work on FreeBSD 8.4 as well.
341 Enable by configuring with 'enable-devcryptoeng'. This is done by default
342 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
345 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
346 util/mkerr.pl, which is adapted to allow those prefixes, leading to
347 error code calls like this:
349 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
351 With this change, we claim the namespaces OSSL and OPENSSL in a manner
352 that can be encoded in C. For the foreseeable future, this will only
354 [Richard Levitte and Tim Hudson]
356 *) Removed BSD cryptodev engine.
359 *) Add a build target 'build_all_generated', to build all generated files
360 and only that. This can be used to prepare everything that requires
361 things like perl for a system that lacks perl and then move everything
362 to that system and do the rest of the build there.
365 *) In the UI interface, make it possible to duplicate the user data. This
366 can be used by engines that need to retain the data for a longer time
367 than just the call where this user data is passed.
370 *) Ignore the '-named_curve auto' value for compatibility of applications
372 [Tomas Mraz <tmraz@fedoraproject.org>]
374 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
375 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
376 alerts across multiple records (some of which could be empty). In practice
377 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
378 prohibts this altogether and other libraries (BoringSSL, NSS) do not
379 support this at all. Supporting it adds significant complexity to the
380 record layer, and its removal is unlikely to cause inter-operability
384 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
385 with Z. These are meant to replace LONG and ZLONG and to be size safe.
386 The use of LONG and ZLONG is discouraged and scheduled for deprecation
390 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
391 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
392 [Richard Levitte, Andy Polyakov]
394 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
398 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
399 platform rather than 'mingw'.
402 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
403 success if they are asked to add an object which already exists
404 in the store. This change cascades to other functions which load
405 certificates and CRLs.
408 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
409 facilitate stack unwinding even from assembly subroutines.
412 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
413 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
416 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
417 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
418 which is the minimum version we support.
421 *) Certificate time validation (X509_cmp_time) enforces stricter
422 compliance with RFC 5280. Fractional seconds and timezone offsets
423 are no longer allowed.
426 *) Add support for ARIA
429 *) s_client will now send the Server Name Indication (SNI) extension by
430 default unless the new "-noservername" option is used. The server name is
431 based on the host provided to the "-connect" option unless overridden by
435 *) Add support for SipHash
438 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
439 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
440 prevent issues where no progress is being made and the peer continually
441 sends unrecognised record types, using up resources processing them.
444 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
445 using the algorithm defined in
446 https://www.akkadia.org/drepper/SHA-crypt.txt
449 *) Heartbeat support has been removed; the ABI is changed for now.
450 [Richard Levitte, Rich Salz]
452 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
455 *) The RSA "null" method, which was partially supported to avoid patent
456 issues, has been replaced to always returns NULL.
460 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
462 *) Client DoS due to large DH parameter
464 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
465 malicious server can send a very large prime value to the client. This will
466 cause the client to spend an unreasonably long period of time generating a
467 key for this prime resulting in a hang until the client has finished. This
468 could be exploited in a Denial Of Service attack.
470 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
474 *) Cache timing vulnerability in RSA Key Generation
476 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
477 a cache timing side channel attack. An attacker with sufficient access to
478 mount cache timing attacks during the RSA key generation process could
479 recover the private key.
481 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
482 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
486 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
487 parameter is no longer accepted, as it leads to a corrupt table. NULL
488 pem_str is reserved for alias entries only.
491 *) Revert blinding in ECDSA sign and instead make problematic addition
492 length-invariant. Switch even to fixed-length Montgomery multiplication.
495 *) Change generating and checking of primes so that the error rate of not
496 being prime depends on the intended use based on the size of the input.
497 For larger primes this will result in more rounds of Miller-Rabin.
498 The maximal error rate for primes with more than 1080 bits is lowered
500 [Kurt Roeckx, Annie Yousar]
502 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
505 *) Add blinding to ECDSA and DSA signatures to protect against side channel
506 attacks discovered by Keegan Ryan (NCC Group).
509 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
510 now allow empty (zero character) pass phrases.
513 *) Certificate time validation (X509_cmp_time) enforces stricter
514 compliance with RFC 5280. Fractional seconds and timezone offsets
515 are no longer allowed.
518 *) Fixed a text canonicalisation bug in CMS
520 Where a CMS detached signature is used with text content the text goes
521 through a canonicalisation process first prior to signing or verifying a
522 signature. This process strips trailing space at the end of lines, converts
523 line terminators to CRLF and removes additional trailing line terminators
524 at the end of a file. A bug in the canonicalisation process meant that
525 some characters, such as form-feed, were incorrectly treated as whitespace
526 and removed. This is contrary to the specification (RFC5485). This fix
527 could mean that detached text data signed with an earlier version of
528 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
529 signed with a fixed OpenSSL may fail to verify with an earlier version of
530 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
531 and use the "-binary" flag (for the "cms" command line application) or set
532 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
535 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
537 *) Constructed ASN.1 types with a recursive definition could exceed the stack
539 Constructed ASN.1 types with a recursive definition (such as can be found
540 in PKCS7) could eventually exceed the stack given malicious input with
541 excessive recursion. This could result in a Denial Of Service attack. There
542 are no such structures used within SSL/TLS that come from untrusted sources
543 so this is considered safe.
545 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
550 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
552 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
553 effectively reduced to only comparing the least significant bit of each
554 byte. This allows an attacker to forge messages that would be considered as
555 authenticated in an amount of tries lower than that guaranteed by the
556 security claims of the scheme. The module can only be compiled by the
557 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
559 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
564 *) Add a build target 'build_all_generated', to build all generated files
565 and only that. This can be used to prepare everything that requires
566 things like perl for a system that lacks perl and then move everything
567 to that system and do the rest of the build there.
570 *) Backport SSL_OP_NO_RENGOTIATION
572 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
573 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
574 changes this is no longer possible in 1.1.0. Therefore the new
575 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
576 1.1.0 to provide equivalent functionality.
578 Note that if an application built against 1.1.0h headers (or above) is run
579 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
580 accepted but nothing will happen, i.e. renegotiation will not be prevented.
583 *) Removed the OS390-Unix config target. It relied on a script that doesn't
587 *) rsaz_1024_mul_avx2 overflow bug on x86_64
589 There is an overflow bug in the AVX2 Montgomery multiplication procedure
590 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
591 Analysis suggests that attacks against RSA and DSA as a result of this
592 defect would be very difficult to perform and are not believed likely.
593 Attacks against DH1024 are considered just feasible, because most of the
594 work necessary to deduce information about a private key may be performed
595 offline. The amount of resources required for such an attack would be
596 significant. However, for an attack on TLS to be meaningful, the server
597 would have to share the DH1024 private key among multiple clients, which is
598 no longer an option since CVE-2016-0701.
600 This only affects processors that support the AVX2 but not ADX extensions
601 like Intel Haswell (4th generation).
603 This issue was reported to OpenSSL by David Benjamin (Google). The issue
604 was originally found via the OSS-Fuzz project.
608 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
610 *) bn_sqrx8x_internal carry bug on x86_64
612 There is a carry propagating bug in the x86_64 Montgomery squaring
613 procedure. No EC algorithms are affected. Analysis suggests that attacks
614 against RSA and DSA as a result of this defect would be very difficult to
615 perform and are not believed likely. Attacks against DH are considered just
616 feasible (although very difficult) because most of the work necessary to
617 deduce information about a private key may be performed offline. The amount
618 of resources required for such an attack would be very significant and
619 likely only accessible to a limited number of attackers. An attacker would
620 additionally need online access to an unpatched system using the target
621 private key in a scenario with persistent DH parameters and a private
622 key that is shared between multiple clients.
624 This only affects processors that support the BMI1, BMI2 and ADX extensions
625 like Intel Broadwell (5th generation) and later or AMD Ryzen.
627 This issue was reported to OpenSSL by the OSS-Fuzz project.
631 *) Malformed X.509 IPAddressFamily could cause OOB read
633 If an X.509 certificate has a malformed IPAddressFamily extension,
634 OpenSSL could do a one-byte buffer overread. The most likely result
635 would be an erroneous display of the certificate in text format.
637 This issue was reported to OpenSSL by the OSS-Fuzz project.
641 Changes between 1.1.0e and 1.1.0f [25 May 2017]
643 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
644 platform rather than 'mingw'.
647 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
648 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
649 which is the minimum version we support.
652 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
654 *) Encrypt-Then-Mac renegotiation crash
656 During a renegotiation handshake if the Encrypt-Then-Mac extension is
657 negotiated where it was not in the original handshake (or vice-versa) then
658 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
659 and servers are affected.
661 This issue was reported to OpenSSL by Joe Orton (Red Hat).
665 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
667 *) Truncated packet could crash via OOB read
669 If one side of an SSL/TLS path is running on a 32-bit host and a specific
670 cipher is being used, then a truncated packet can cause that host to
671 perform an out-of-bounds read, usually resulting in a crash.
673 This issue was reported to OpenSSL by Robert Święcki of Google.
677 *) Bad (EC)DHE parameters cause a client crash
679 If a malicious server supplies bad parameters for a DHE or ECDHE key
680 exchange then this can result in the client attempting to dereference a
681 NULL pointer leading to a client crash. This could be exploited in a Denial
684 This issue was reported to OpenSSL by Guido Vranken.
688 *) BN_mod_exp may produce incorrect results on x86_64
690 There is a carry propagating bug in the x86_64 Montgomery squaring
691 procedure. No EC algorithms are affected. Analysis suggests that attacks
692 against RSA and DSA as a result of this defect would be very difficult to
693 perform and are not believed likely. Attacks against DH are considered just
694 feasible (although very difficult) because most of the work necessary to
695 deduce information about a private key may be performed offline. The amount
696 of resources required for such an attack would be very significant and
697 likely only accessible to a limited number of attackers. An attacker would
698 additionally need online access to an unpatched system using the target
699 private key in a scenario with persistent DH parameters and a private
700 key that is shared between multiple clients. For example this can occur by
701 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
702 similar to CVE-2015-3193 but must be treated as a separate problem.
704 This issue was reported to OpenSSL by the OSS-Fuzz project.
708 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
710 *) ChaCha20/Poly1305 heap-buffer-overflow
712 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
713 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
714 crash. This issue is not considered to be exploitable beyond a DoS.
716 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
720 *) CMS Null dereference
722 Applications parsing invalid CMS structures can crash with a NULL pointer
723 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
724 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
725 structure callback if an attempt is made to free certain invalid encodings.
726 Only CHOICE structures using a callback which do not handle NULL value are
729 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
733 *) Montgomery multiplication may produce incorrect results
735 There is a carry propagating bug in the Broadwell-specific Montgomery
736 multiplication procedure that handles input lengths divisible by, but
737 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
738 and DH private keys are impossible. This is because the subroutine in
739 question is not used in operations with the private key itself and an input
740 of the attacker's direct choice. Otherwise the bug can manifest itself as
741 transient authentication and key negotiation failures or reproducible
742 erroneous outcome of public-key operations with specially crafted input.
743 Among EC algorithms only Brainpool P-512 curves are affected and one
744 presumably can attack ECDH key negotiation. Impact was not analyzed in
745 detail, because pre-requisites for attack are considered unlikely. Namely
746 multiple clients have to choose the curve in question and the server has to
747 share the private key among them, neither of which is default behaviour.
748 Even then only clients that chose the curve will be affected.
750 This issue was publicly reported as transient failures and was not
751 initially recognized as a security issue. Thanks to Richard Morgan for
752 providing reproducible case.
756 *) Removed automatic addition of RPATH in shared libraries and executables,
757 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
760 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
762 *) Fix Use After Free for large message sizes
764 The patch applied to address CVE-2016-6307 resulted in an issue where if a
765 message larger than approx 16k is received then the underlying buffer to
766 store the incoming message is reallocated and moved. Unfortunately a
767 dangling pointer to the old location is left which results in an attempt to
768 write to the previously freed location. This is likely to result in a
769 crash, however it could potentially lead to execution of arbitrary code.
771 This issue only affects OpenSSL 1.1.0a.
773 This issue was reported to OpenSSL by Robert Święcki.
777 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
779 *) OCSP Status Request extension unbounded memory growth
781 A malicious client can send an excessively large OCSP Status Request
782 extension. If that client continually requests renegotiation, sending a
783 large OCSP Status Request extension each time, then there will be unbounded
784 memory growth on the server. This will eventually lead to a Denial Of
785 Service attack through memory exhaustion. Servers with a default
786 configuration are vulnerable even if they do not support OCSP. Builds using
787 the "no-ocsp" build time option are not affected.
789 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
793 *) SSL_peek() hang on empty record
795 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
796 sends an empty record. This could be exploited by a malicious peer in a
797 Denial Of Service attack.
799 This issue was reported to OpenSSL by Alex Gaynor.
803 *) Excessive allocation of memory in tls_get_message_header() and
804 dtls1_preprocess_fragment()
806 A (D)TLS message includes 3 bytes for its length in the header for the
807 message. This would allow for messages up to 16Mb in length. Messages of
808 this length are excessive and OpenSSL includes a check to ensure that a
809 peer is sending reasonably sized messages in order to avoid too much memory
810 being consumed to service a connection. A flaw in the logic of version
811 1.1.0 means that memory for the message is allocated too early, prior to
812 the excessive message length check. Due to way memory is allocated in
813 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
814 to service a connection. This could lead to a Denial of Service through
815 memory exhaustion. However, the excessive message length check still takes
816 place, and this would cause the connection to immediately fail. Assuming
817 that the application calls SSL_free() on the failed connection in a timely
818 manner then the 21Mb of allocated memory will then be immediately freed
819 again. Therefore the excessive memory allocation will be transitory in
820 nature. This then means that there is only a security impact if:
822 1) The application does not call SSL_free() in a timely manner in the event
823 that the connection fails
825 2) The application is working in a constrained environment where there is
826 very little free memory
828 3) The attacker initiates multiple connection attempts such that there are
829 multiple connections in a state where memory has been allocated for the
830 connection; SSL_free() has not yet been called; and there is insufficient
831 memory to service the multiple requests.
833 Except in the instance of (1) above any Denial Of Service is likely to be
834 transitory because as soon as the connection fails the memory is
835 subsequently freed again in the SSL_free() call. However there is an
836 increased risk during this period of application crashes due to the lack of
837 memory - which would then mean a more serious Denial of Service.
839 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
840 (CVE-2016-6307 and CVE-2016-6308)
843 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
844 had to be removed. Primary reason is that vendor assembler can't
845 assemble our modules with -KPIC flag. As result it, assembly
846 support, was not even available as option. But its lack means
847 lack of side-channel resistant code, which is incompatible with
848 security by todays standards. Fortunately gcc is readily available
849 prepackaged option, which we firmly point at...
852 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
854 *) Windows command-line tool supports UTF-8 opt-in option for arguments
855 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
856 (to any value) allows Windows user to access PKCS#12 file generated
857 with Windows CryptoAPI and protected with non-ASCII password, as well
858 as files generated under UTF-8 locale on Linux also protected with
862 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
863 have been disabled by default and removed from DEFAULT, just like RC4.
864 See the RC4 item below to re-enable both.
867 *) The method for finding the storage location for the Windows RAND seed file
868 has changed. First we check %RANDFILE%. If that is not set then we check
869 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
870 all else fails we fall back to C:\.
873 *) The EVP_EncryptUpdate() function has had its return type changed from void
874 to int. A return of 0 indicates and error while a return of 1 indicates
878 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
879 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
880 off the constant time implementation for RSA, DSA and DH have been made
881 no-ops and deprecated.
884 *) Windows RAND implementation was simplified to only get entropy by
885 calling CryptGenRandom(). Various other RAND-related tickets
887 [Joseph Wylie Yandle, Rich Salz]
889 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
890 and OPENSSL_LH_, respectively. The old names are available
891 with API compatibility. They new names are now completely documented.
894 *) Unify TYPE_up_ref(obj) methods signature.
895 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
896 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
897 int (instead of void) like all others TYPE_up_ref() methods.
898 So now these methods also check the return value of CRYPTO_atomic_add(),
899 and the validity of object reference counter.
900 [fdasilvayy@gmail.com]
902 *) With Windows Visual Studio builds, the .pdb files are installed
903 alongside the installed libraries and executables. For a static
904 library installation, ossl_static.pdb is the associate compiler
905 generated .pdb file to be used when linking programs.
908 *) Remove openssl.spec. Packaging files belong with the packagers.
911 *) Automatic Darwin/OSX configuration has had a refresh, it will now
912 recognise x86_64 architectures automatically. You can still decide
913 to build for a different bitness with the environment variable
914 KERNEL_BITS (can be 32 or 64), for example:
916 KERNEL_BITS=32 ./config
920 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
921 256 bit AES and HMAC with SHA256.
924 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
927 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
930 *) To enable users to have their own config files and build file templates,
931 Configure looks in the directory indicated by the environment variable
932 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
933 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
934 name and is used as is.
937 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
938 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
939 X509_CERT_FILE_CTX was removed.
942 *) "shared" builds are now the default. To create only static libraries use
943 the "no-shared" Configure option.
946 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
947 All of these option have not worked for some while and are fundamental
951 *) Make various cleanup routines no-ops and mark them as deprecated. Most
952 global cleanup functions are no longer required because they are handled
953 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
954 Explicitly de-initing can cause problems (e.g. where a library that uses
955 OpenSSL de-inits, but an application is still using it). The affected
956 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
957 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
958 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
962 *) --strict-warnings no longer enables runtime debugging options
963 such as REF_DEBUG. Instead, debug options are automatically
964 enabled with '--debug' builds.
965 [Andy Polyakov, Emilia Käsper]
967 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
968 have been moved out of the public header files. New functions for managing
969 these have been added.
972 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
973 objects have been moved out of the public header files. New
974 functions for managing these have been added.
977 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
978 have been moved out of the public header files. New functions for managing
979 these have been added.
982 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
983 moved out of the public header files. New functions for managing these
987 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
990 *) Removed the mk1mf build scripts.
993 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
994 it is always safe to #include a header now.
997 *) Removed the aged BC-32 config and all its supporting scripts
1000 *) Removed support for Ultrix, Netware, and OS/2.
1003 *) Add support for HKDF.
1004 [Alessandro Ghedini]
1006 *) Add support for blake2b and blake2s
1009 *) Added support for "pipelining". Ciphers that have the
1010 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1011 encryptions/decryptions simultaneously. There are currently no built-in
1012 ciphers with this property but the expectation is that engines will be able
1013 to offer it to significantly improve throughput. Support has been extended
1014 into libssl so that multiple records for a single connection can be
1015 processed in one go (for >=TLS 1.1).
1018 *) Added the AFALG engine. This is an async capable engine which is able to
1019 offload work to the Linux kernel. In this initial version it only supports
1020 AES128-CBC. The kernel must be version 4.1.0 or greater.
1023 *) OpenSSL now uses a new threading API. It is no longer necessary to
1024 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1025 are two supported threading models: pthreads and windows threads. It is
1026 also possible to configure OpenSSL at compile time for "no-threads". The
1027 old threading API should no longer be used. The functions have been
1028 replaced with "no-op" compatibility macros.
1029 [Alessandro Ghedini, Matt Caswell]
1031 *) Modify behavior of ALPN to invoke callback after SNI/servername
1032 callback, such that updates to the SSL_CTX affect ALPN.
1035 *) Add SSL_CIPHER queries for authentication and key-exchange.
1038 *) Changes to the DEFAULT cipherlist:
1039 - Prefer (EC)DHE handshakes over plain RSA.
1040 - Prefer AEAD ciphers over legacy ciphers.
1041 - Prefer ECDSA over RSA when both certificates are available.
1042 - Prefer TLSv1.2 ciphers/PRF.
1043 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1047 *) Change the ECC default curve list to be this, in order: x25519,
1048 secp256r1, secp521r1, secp384r1.
1051 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1052 disabled by default. They can be re-enabled using the
1053 enable-weak-ssl-ciphers option to Configure.
1056 *) If the server has ALPN configured, but supports no protocols that the
1057 client advertises, send a fatal "no_application_protocol" alert.
1058 This behaviour is SHALL in RFC 7301, though it isn't universally
1059 implemented by other servers.
1062 *) Add X25519 support.
1063 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1064 for public and private key encoding using the format documented in
1065 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1066 key generation and key derivation.
1068 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1072 *) Deprecate SRP_VBASE_get_by_user.
1073 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1074 In order to fix an unavoidable memory leak (CVE-2016-0798),
1075 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1076 seed, even if the seed is configured.
1078 Users should use SRP_VBASE_get1_by_user instead. Note that in
1079 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1080 also that even though configuring the SRP seed attempts to hide
1081 invalid usernames by continuing the handshake with fake
1082 credentials, this behaviour is not constant time and no strong
1083 guarantees are made that the handshake is indistinguishable from
1084 that of a valid user.
1087 *) Configuration change; it's now possible to build dynamic engines
1088 without having to build shared libraries and vice versa. This
1089 only applies to the engines in engines/, those in crypto/engine/
1090 will always be built into libcrypto (i.e. "static").
1092 Building dynamic engines is enabled by default; to disable, use
1093 the configuration option "disable-dynamic-engine".
1095 The only requirements for building dynamic engines are the
1096 presence of the DSO module and building with position independent
1097 code, so they will also automatically be disabled if configuring
1098 with "disable-dso" or "disable-pic".
1100 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1101 are also taken away from openssl/opensslconf.h, as they are
1105 *) Configuration change; if there is a known flag to compile
1106 position independent code, it will always be applied on the
1107 libcrypto and libssl object files, and never on the application
1108 object files. This means other libraries that use routines from
1109 libcrypto / libssl can be made into shared libraries regardless
1110 of how OpenSSL was configured.
1112 If this isn't desirable, the configuration options "disable-pic"
1113 or "no-pic" can be used to disable the use of PIC. This will
1114 also disable building shared libraries and dynamic engines.
1117 *) Removed JPAKE code. It was experimental and has no wide use.
1120 *) The INSTALL_PREFIX Makefile variable has been renamed to
1121 DESTDIR. That makes for less confusion on what this variable
1122 is for. Also, the configuration option --install_prefix is
1126 *) Heartbeat for TLS has been removed and is disabled by default
1127 for DTLS; configure with enable-heartbeats. Code that uses the
1128 old #define's might need to be updated.
1129 [Emilia Käsper, Rich Salz]
1131 *) Rename REF_CHECK to REF_DEBUG.
1134 *) New "unified" build system
1136 The "unified" build system is aimed to be a common system for all
1137 platforms we support. With it comes new support for VMS.
1139 This system builds supports building in a different directory tree
1140 than the source tree. It produces one Makefile (for unix family
1141 or lookalikes), or one descrip.mms (for VMS).
1143 The source of information to make the Makefile / descrip.mms is
1144 small files called 'build.info', holding the necessary
1145 information for each directory with source to compile, and a
1146 template in Configurations, like unix-Makefile.tmpl or
1149 With this change, the library names were also renamed on Windows
1150 and on VMS. They now have names that are closer to the standard
1151 on Unix, and include the major version number, and in certain
1152 cases, the architecture they are built for. See "Notes on shared
1153 libraries" in INSTALL.
1155 We rely heavily on the perl module Text::Template.
1158 *) Added support for auto-initialisation and de-initialisation of the library.
1159 OpenSSL no longer requires explicit init or deinit routines to be called,
1160 except in certain circumstances. See the OPENSSL_init_crypto() and
1161 OPENSSL_init_ssl() man pages for further information.
1164 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1165 "peer" argument is now expected to be a BIO_ADDR object.
1167 *) Rewrite of BIO networking library. The BIO library lacked consistent
1168 support of IPv6, and adding it required some more extensive
1169 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1170 which hold all types of addresses and chains of address information.
1171 It also introduces a new API, with functions like BIO_socket,
1172 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1173 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1174 have been adapted accordingly.
1177 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1181 *) CRIME protection: disable compression by default, even if OpenSSL is
1182 compiled with zlib enabled. Applications can still enable compression
1183 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1184 using the SSL_CONF library to configure compression.
1187 *) The signature of the session callback configured with
1188 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1189 was explicitly marked as 'const unsigned char*' instead of
1193 *) Always DPURIFY. Remove the use of uninitialized memory in the
1194 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1197 *) Removed many obsolete configuration items, including
1198 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1199 MD2_CHAR, MD2_INT, MD2_LONG
1201 IDEA_SHORT, IDEA_LONG
1202 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1203 [Rich Salz, with advice from Andy Polyakov]
1205 *) Many BN internals have been moved to an internal header file.
1206 [Rich Salz with help from Andy Polyakov]
1208 *) Configuration and writing out the results from it has changed.
1209 Files such as Makefile include/openssl/opensslconf.h and are now
1210 produced through general templates, such as Makefile.in and
1211 crypto/opensslconf.h.in and some help from the perl module
1214 Also, the center of configuration information is no longer
1215 Makefile. Instead, Configure produces a perl module in
1216 configdata.pm which holds most of the config data (in the hash
1217 table %config), the target data that comes from the target
1218 configuration in one of the Configurations/*.conf files (in
1222 *) To clarify their intended purposes, the Configure options
1223 --prefix and --openssldir change their semantics, and become more
1224 straightforward and less interdependent.
1226 --prefix shall be used exclusively to give the location INSTALLTOP
1227 where programs, scripts, libraries, include files and manuals are
1228 going to be installed. The default is now /usr/local.
1230 --openssldir shall be used exclusively to give the default
1231 location OPENSSLDIR where certificates, private keys, CRLs are
1232 managed. This is also where the default openssl.cnf gets
1234 If the directory given with this option is a relative path, the
1235 values of both the --prefix value and the --openssldir value will
1236 be combined to become OPENSSLDIR.
1237 The default for --openssldir is INSTALLTOP/ssl.
1239 Anyone who uses --openssldir to specify where OpenSSL is to be
1240 installed MUST change to use --prefix instead.
1243 *) The GOST engine was out of date and therefore it has been removed. An up
1244 to date GOST engine is now being maintained in an external repository.
1245 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1246 support for GOST ciphersuites (these are only activated if a GOST engine
1250 *) EGD is no longer supported by default; use enable-egd when
1252 [Ben Kaduk and Rich Salz]
1254 *) The distribution now has Makefile.in files, which are used to
1255 create Makefile's when Configure is run. *Configure must be run
1256 before trying to build now.*
1259 *) The return value for SSL_CIPHER_description() for error conditions
1263 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1265 Obtaining and performing DNSSEC validation of TLSA records is
1266 the application's responsibility. The application provides
1267 the TLSA records of its choice to OpenSSL, and these are then
1268 used to authenticate the peer.
1270 The TLSA records need not even come from DNS. They can, for
1271 example, be used to implement local end-entity certificate or
1272 trust-anchor "pinning", where the "pin" data takes the form
1273 of TLSA records, which can augment or replace verification
1274 based on the usual WebPKI public certification authorities.
1277 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1278 continues to support deprecated interfaces in default builds.
1279 However, applications are strongly advised to compile their
1280 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1281 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1282 or the 1.1.0 releases.
1284 In environments in which all applications have been ported to
1285 not use any deprecated interfaces OpenSSL's Configure script
1286 should be used with the --api=1.1.0 option to entirely remove
1287 support for the deprecated features from the library and
1288 unconditionally disable them in the installed headers.
1289 Essentially the same effect can be achieved with the "no-deprecated"
1290 argument to Configure, except that this will always restrict
1291 the build to just the latest API, rather than a fixed API
1294 As applications are ported to future revisions of the API,
1295 they should update their compile-time OPENSSL_API_COMPAT define
1296 accordingly, but in most cases should be able to continue to
1297 compile with later releases.
1299 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1300 0x10000000L and 0x00908000L, respectively. However those
1301 versions did not support the OPENSSL_API_COMPAT feature, and
1302 so applications are not typically tested for explicit support
1303 of just the undeprecated features of either release.
1306 *) Add support for setting the minimum and maximum supported protocol.
1307 It can bet set via the SSL_set_min_proto_version() and
1308 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1309 MaxProtocol. It's recommended to use the new APIs to disable
1310 protocols instead of disabling individual protocols using
1311 SSL_set_options() or SSL_CONF's Protocol. This change also
1312 removes support for disabling TLS 1.2 in the OpenSSL TLS
1313 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1316 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1319 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1320 and integrates ECDSA and ECDH functionality into EC. Implementations can
1321 now redirect key generation and no longer need to convert to or from
1324 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1325 include the ec.h header file instead.
1328 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1329 ciphers who are no longer supported and drops support the ephemeral RSA key
1330 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1333 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1334 opaque. For HMAC_CTX, the following constructors and destructors
1337 HMAC_CTX *HMAC_CTX_new(void);
1338 void HMAC_CTX_free(HMAC_CTX *ctx);
1340 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1341 destroy such methods has been added. See EVP_MD_meth_new(3) and
1342 EVP_CIPHER_meth_new(3) for documentation.
1345 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1346 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1347 EVP_MD_CTX_reset() should be called instead to reinitialise
1348 an already created structure.
1349 2) For consistency with the majority of our object creators and
1350 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1351 EVP_MD_CTX_(new|free). The old names are retained as macros
1352 for deprecated builds.
1355 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1356 cryptographic operations to be performed asynchronously as long as an
1357 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1358 further details. Libssl has also had this capability integrated with the
1359 introduction of the new mode SSL_MODE_ASYNC and associated error
1360 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1361 pages. This work was developed in partnership with Intel Corp.
1364 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1365 always enabled now. If you want to disable the support you should
1366 exclude it using the list of supported ciphers. This also means that the
1367 "-no_ecdhe" option has been removed from s_server.
1370 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1371 SSL_{CTX_}set1_curves() which can set a list.
1374 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1375 curve you want to support using SSL_{CTX_}set1_curves().
1378 *) State machine rewrite. The state machine code has been significantly
1379 refactored in order to remove much duplication of code and solve issues
1380 with the old code (see ssl/statem/README for further details). This change
1381 does have some associated API changes. Notably the SSL_state() function
1382 has been removed and replaced by SSL_get_state which now returns an
1383 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1384 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1388 *) All instances of the string "ssleay" in the public API were replaced
1389 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1390 Some error codes related to internal RSA_eay API's were renamed.
1393 *) The demo files in crypto/threads were moved to demo/threads.
1396 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1398 [Matt Caswell, Rich Salz]
1400 *) New ASN.1 embed macro.
1402 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1403 structure is not allocated: it is part of the parent. That is instead of
1411 This reduces memory fragmentation and make it impossible to accidentally
1412 set a mandatory field to NULL.
1414 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1415 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1416 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1420 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1423 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1424 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1425 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1426 DES and RC4 ciphersuites.
1429 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1430 This changes the decoding behaviour for some invalid messages,
1431 though the change is mostly in the more lenient direction, and
1432 legacy behaviour is preserved as much as possible.
1435 *) Fix no-stdio build.
1436 [ David Woodhouse <David.Woodhouse@intel.com> and also
1437 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1439 *) New testing framework
1440 The testing framework has been largely rewritten and is now using
1441 perl and the perl modules Test::Harness and an extended variant of
1442 Test::More called OpenSSL::Test to do its work. All test scripts in
1443 test/ have been rewritten into test recipes, and all direct calls to
1444 executables in test/Makefile have become individual recipes using the
1445 simplified testing OpenSSL::Test::Simple.
1447 For documentation on our testing modules, do:
1449 perldoc test/testlib/OpenSSL/Test/Simple.pm
1450 perldoc test/testlib/OpenSSL/Test.pm
1454 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1455 are used; the latter aborts on memory leaks (usually checked on exit).
1456 Some undocumented "set malloc, etc., hooks" functions were removed
1457 and others were changed. All are now documented.
1460 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1462 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1464 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1465 from RFC4279, RFC4785, RFC5487, RFC5489.
1467 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1468 original RSA_PSK patch.
1471 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1472 era flag was never set throughout the codebase (only read). Also removed
1473 SSL3_FLAGS_POP_BUFFER which was only used if
1474 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1477 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1478 to be "oneline" instead of "compat".
1481 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1482 not aware of clients that still exhibit this bug, and the workaround
1483 hasn't been working properly for a while.
1486 *) The return type of BIO_number_read() and BIO_number_written() as well as
1487 the corresponding num_read and num_write members in the BIO structure has
1488 changed from unsigned long to uint64_t. On platforms where an unsigned
1489 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1493 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1494 OpenSSL without support for them. It also means that maintaining
1495 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1496 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1499 *) Removed support for the two export grade static DH ciphersuites
1500 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1501 were newly added (along with a number of other static DH ciphersuites) to
1502 1.0.2. However the two export ones have *never* worked since they were
1503 introduced. It seems strange in any case to be adding new export
1504 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1507 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1508 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1509 and turned into macros which simply call the new preferred function names
1510 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1511 should use the new names instead. Also as part of this change the ssl23.h
1512 header file has been removed.
1515 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1516 code and the associated standard is no longer considered fit-for-purpose.
1519 *) RT2547 was closed. When generating a private key, try to make the
1520 output file readable only by the owner. This behavior change might
1521 be noticeable when interacting with other software.
1523 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1527 *) Added HTTP GET support to the ocsp command.
1530 *) Changed default digest for the dgst and enc commands from MD5 to
1534 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1537 *) Added support for TLS extended master secret from
1538 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1539 initial patch which was a great help during development.
1542 *) All libssl internal structures have been removed from the public header
1543 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1544 now redundant). Users should not attempt to access internal structures
1545 directly. Instead they should use the provided API functions.
1548 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1549 Access to deprecated functions can be re-enabled by running config with
1550 "enable-deprecated". In addition applications wishing to use deprecated
1551 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1552 will, by default, disable some transitive includes that previously existed
1553 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1556 *) Added support for OCB mode. OpenSSL has been granted a patent license
1557 compatible with the OpenSSL license for use of OCB. Details are available
1558 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1559 for OCB can be removed by calling config with no-ocb.
1562 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1563 compatible client hello.
1566 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1567 done while fixing the error code for the key-too-small case.
1568 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1570 *) CA.sh has been removed; use CA.pl instead.
1573 *) Removed old DES API.
1576 *) Remove various unsupported platforms:
1582 Sinix/ReliantUNIX RM400
1587 16-bit platforms such as WIN16
1590 *) Clean up OPENSSL_NO_xxx #define's
1591 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1592 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1593 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1594 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1595 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1596 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1597 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1598 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1599 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1600 Remove MS_STATIC; it's a relic from platforms <32 bits.
1603 *) Cleaned up dead code
1604 Remove all but one '#ifdef undef' which is to be looked at.
1607 *) Clean up calling of xxx_free routines.
1608 Just like free(), fix most of the xxx_free routines to accept
1609 NULL. Remove the non-null checks from callers. Save much code.
1612 *) Add secure heap for storage of private keys (when possible).
1613 Add BIO_s_secmem(), CBIGNUM, etc.
1614 Contributed by Akamai Technologies under our Corporate CLA.
1617 *) Experimental support for a new, fast, unbiased prime candidate generator,
1618 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1619 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1621 *) New output format NSS in the sess_id command line tool. This allows
1622 exporting the session id and the master key in NSS keylog format.
1623 [Martin Kaiser <martin@kaiser.cx>]
1625 *) Harmonize version and its documentation. -f flag is used to display
1627 [mancha <mancha1@zoho.com>]
1629 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1630 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1631 [mancha <mancha1@zoho.com>]
1633 *) Fix some double frees. These are not thought to be exploitable.
1634 [mancha <mancha1@zoho.com>]
1636 *) A missing bounds check in the handling of the TLS heartbeat extension
1637 can be used to reveal up to 64k of memory to a connected client or
1640 Thanks for Neel Mehta of Google Security for discovering this bug and to
1641 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1642 preparing the fix (CVE-2014-0160)
1643 [Adam Langley, Bodo Moeller]
1645 *) Fix for the attack described in the paper "Recovering OpenSSL
1646 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1647 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1648 http://eprint.iacr.org/2014/140
1650 Thanks to Yuval Yarom and Naomi Benger for discovering this
1651 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1652 [Yuval Yarom and Naomi Benger]
1654 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1655 this fixes a limitation in previous versions of OpenSSL.
1658 *) Experimental encrypt-then-mac support.
1660 Experimental support for encrypt then mac from
1661 draft-gutmann-tls-encrypt-then-mac-02.txt
1663 To enable it set the appropriate extension number (0x42 for the test
1664 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1666 For non-compliant peers (i.e. just about everything) this should have no
1669 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1673 *) Add EVP support for key wrapping algorithms, to avoid problems with
1674 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1675 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1676 algorithms and include tests cases.
1679 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1683 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1684 MGF1 digest and OAEP label.
1687 *) Make openssl verify return errors.
1688 [Chris Palmer <palmer@google.com> and Ben Laurie]
1690 *) New function ASN1_TIME_diff to calculate the difference between two
1691 ASN1_TIME structures or one structure and the current time.
1694 *) Update fips_test_suite to support multiple command line options. New
1695 test to induce all self test errors in sequence and check expected
1699 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1700 sign or verify all in one operation.
1703 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1704 test programs and fips_test_suite. Includes functionality to parse
1705 the minimal script output of fipsalgest.pl directly.
1708 *) Add authorisation parameter to FIPS_module_mode_set().
1711 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1714 *) Use separate DRBG fields for internal and external flags. New function
1715 FIPS_drbg_health_check() to perform on demand health checking. Add
1716 generation tests to fips_test_suite with reduced health check interval to
1717 demonstrate periodic health checking. Add "nodh" option to
1718 fips_test_suite to skip very slow DH test.
1721 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1725 *) More extensive health check for DRBG checking many more failure modes.
1726 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1727 combination: call this in fips_test_suite.
1730 *) Add support for canonical generation of DSA parameter 'g'. See
1733 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1734 POST to handle HMAC cases.
1737 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1738 to return numerical and string versions of the FIPS module number.
1741 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1742 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1743 outside the validated module in the FIPS capable OpenSSL.
1746 *) Minor change to DRBG entropy callback semantics. In some cases
1747 there is no multiple of the block length between min_len and
1748 max_len. Allow the callback to return more than max_len bytes
1749 of entropy but discard any extra: it is the callback's responsibility
1750 to ensure that the extra data discarded does not impact the
1751 requested amount of entropy.
1754 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1755 information in FIPS186-3, SP800-57 and SP800-131A.
1758 *) CCM support via EVP. Interface is very similar to GCM case except we
1759 must supply all data in one chunk (i.e. no update, final) and the
1760 message length must be supplied if AAD is used. Add algorithm test
1764 *) Initial version of POST overhaul. Add POST callback to allow the status
1765 of POST to be monitored and/or failures induced. Modify fips_test_suite
1766 to use callback. Always run all selftests even if one fails.
1769 *) XTS support including algorithm test driver in the fips_gcmtest program.
1770 Note: this does increase the maximum key length from 32 to 64 bytes but
1771 there should be no binary compatibility issues as existing applications
1772 will never use XTS mode.
1775 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1776 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1777 performs algorithm blocking for unapproved PRNG types. Also do not
1778 set PRNG type in FIPS_mode_set(): leave this to the application.
1779 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1780 the standard OpenSSL PRNG: set additional data to a date time vector.
1783 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1784 This shouldn't present any incompatibility problems because applications
1785 shouldn't be using these directly and any that are will need to rethink
1786 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1789 *) Extensive self tests and health checking required by SP800-90 DRBG.
1790 Remove strength parameter from FIPS_drbg_instantiate and always
1791 instantiate at maximum supported strength.
1794 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1797 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1800 *) New function DH_compute_key_padded() to compute a DH key and pad with
1801 leading zeroes if needed: this complies with SP800-56A et al.
1804 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1805 anything, incomplete, subject to change and largely untested at present.
1808 *) Modify fipscanisteronly build option to only build the necessary object
1809 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1812 *) Add experimental option FIPSSYMS to give all symbols in
1813 fipscanister.o and FIPS or fips prefix. This will avoid
1814 conflicts with future versions of OpenSSL. Add perl script
1815 util/fipsas.pl to preprocess assembly language source files
1816 and rename any affected symbols.
1819 *) Add selftest checks and algorithm block of non-fips algorithms in
1820 FIPS mode. Remove DES2 from selftests.
1823 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1824 return internal method without any ENGINE dependencies. Add new
1825 tiny fips sign and verify functions.
1828 *) New build option no-ec2m to disable characteristic 2 code.
1831 *) New build option "fipscanisteronly". This only builds fipscanister.o
1832 and (currently) associated fips utilities. Uses the file Makefile.fips
1833 instead of Makefile.org as the prototype.
1836 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1837 Update fips_gcmtest to use IV generator.
1840 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1841 setting output buffer to NULL. The *Final function must be
1842 called although it will not retrieve any additional data. The tag
1843 can be set or retrieved with a ctrl. The IV length is by default 12
1844 bytes (96 bits) but can be set to an alternative value. If the IV
1845 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1849 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1850 underlying do_cipher function handles all cipher semantics itself
1851 including padding and finalisation. This is useful if (for example)
1852 an ENGINE cipher handles block padding itself. The behaviour of
1853 do_cipher is subtly changed if this flag is set: the return value
1854 is the number of characters written to the output buffer (zero is
1855 no longer an error code) or a negative error code. Also if the
1856 input buffer is NULL and length 0 finalisation should be performed.
1859 *) If a candidate issuer certificate is already part of the constructed
1860 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1863 *) Improve forward-security support: add functions
1865 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1866 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1868 for use by SSL/TLS servers; the callback function will be called whenever a
1869 new session is created, and gets to decide whether the session may be
1870 cached to make it resumable (return 0) or not (return 1). (As by the
1871 SSL/TLS protocol specifications, the session_id sent by the server will be
1872 empty to indicate that the session is not resumable; also, the server will
1873 not generate RFC 4507 (RFC 5077) session tickets.)
1875 A simple reasonable callback implementation is to return is_forward_secure.
1876 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1877 by the SSL/TLS server library, indicating whether it can provide forward
1879 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1881 *) New -verify_name option in command line utilities to set verification
1885 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1886 Add CMAC pkey methods.
1889 *) Experimental renegotiation in s_server -www mode. If the client
1890 browses /reneg connection is renegotiated. If /renegcert it is
1891 renegotiated requesting a certificate.
1894 *) Add an "external" session cache for debugging purposes to s_server. This
1895 should help trace issues which normally are only apparent in deployed
1896 multi-process servers.
1899 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1900 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1901 BIO_set_cipher() and some obscure PEM functions were changed so they
1902 can now return an error. The RAND changes required a change to the
1903 RAND_METHOD structure.
1906 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1907 a gcc attribute to warn if the result of a function is ignored. This
1908 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1909 whose return value is often ignored.
1912 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1913 These allow SCTs (signed certificate timestamps) to be requested and
1914 validated when establishing a connection.
1915 [Rob Percival <robpercival@google.com>]
1917 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1919 *) Prevent padding oracle in AES-NI CBC MAC check
1921 A MITM attacker can use a padding oracle attack to decrypt traffic
1922 when the connection uses an AES CBC cipher and the server support
1925 This issue was introduced as part of the fix for Lucky 13 padding
1926 attack (CVE-2013-0169). The padding check was rewritten to be in
1927 constant time by making sure that always the same bytes are read and
1928 compared against either the MAC or padding bytes. But it no longer
1929 checked that there was enough data to have both the MAC and padding
1932 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1936 *) Fix EVP_EncodeUpdate overflow
1938 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1939 Base64 encoding of binary data. If an attacker is able to supply very large
1940 amounts of input data then a length check can overflow resulting in a heap
1943 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1944 the PEM_write_bio* family of functions. These are mainly used within the
1945 OpenSSL command line applications, so any application which processes data
1946 from an untrusted source and outputs it as a PEM file should be considered
1947 vulnerable to this issue. User applications that call these APIs directly
1948 with large amounts of untrusted data may also be vulnerable.
1950 This issue was reported by Guido Vranken.
1954 *) Fix EVP_EncryptUpdate overflow
1956 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1957 is able to supply very large amounts of input data after a previous call to
1958 EVP_EncryptUpdate() with a partial block then a length check can overflow
1959 resulting in a heap corruption. Following an analysis of all OpenSSL
1960 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1961 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1962 the first called function after an EVP_EncryptInit(), and therefore that
1963 specific call must be safe. The second form is where the length passed to
1964 EVP_EncryptUpdate() can be seen from the code to be some small value and
1965 therefore there is no possibility of an overflow. Since all instances are
1966 one of these two forms, it is believed that there can be no overflows in
1967 internal code due to this problem. It should be noted that
1968 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1969 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1970 of these calls have also been analysed too and it is believed there are no
1971 instances in internal usage where an overflow could occur.
1973 This issue was reported by Guido Vranken.
1977 *) Prevent ASN.1 BIO excessive memory allocation
1979 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1980 a short invalid encoding can cause allocation of large amounts of memory
1981 potentially consuming excessive resources or exhausting memory.
1983 Any application parsing untrusted data through d2i BIO functions is
1984 affected. The memory based functions such as d2i_X509() are *not* affected.
1985 Since the memory based functions are used by the TLS library, TLS
1986 applications are not affected.
1988 This issue was reported by Brian Carpenter.
1994 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1995 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1996 in arbitrary stack data being returned in the buffer.
1998 This issue was reported by Guido Vranken.
2002 *) Modify behavior of ALPN to invoke callback after SNI/servername
2003 callback, such that updates to the SSL_CTX affect ALPN.
2006 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2010 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2011 methods are enabled and ssl2 is disabled the methods return NULL.
2014 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2016 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2017 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2018 provide any "EXPORT" or "LOW" strength ciphers.
2021 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2022 is by default disabled at build-time. Builds that are not configured with
2023 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2024 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2025 will need to explicitly call either of:
2027 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2029 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2031 as appropriate. Even if either of those is used, or the application
2032 explicitly uses the version-specific SSLv2_method() or its client and
2033 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2034 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2035 ciphers, and SSLv2 56-bit DES are no longer available.
2039 *) Fix a double-free in DSA code
2041 A double free bug was discovered when OpenSSL parses malformed DSA private
2042 keys and could lead to a DoS attack or memory corruption for applications
2043 that receive DSA private keys from untrusted sources. This scenario is
2046 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2051 *) Disable SRP fake user seed to address a server memory leak.
2053 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2055 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2056 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2057 was changed to ignore the "fake user" SRP seed, even if the seed
2060 Users should use SRP_VBASE_get1_by_user instead. Note that in
2061 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2062 also that even though configuring the SRP seed attempts to hide
2063 invalid usernames by continuing the handshake with fake
2064 credentials, this behaviour is not constant time and no strong
2065 guarantees are made that the handshake is indistinguishable from
2066 that of a valid user.
2070 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2072 In the BN_hex2bn function the number of hex digits is calculated using an
2073 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2074 large values of |i| this can result in |bn_expand| not allocating any
2075 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2076 field as NULL leading to a subsequent NULL ptr deref. For very large values
2077 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2078 In this case memory is allocated to the internal BIGNUM data field, but it
2079 is insufficiently sized leading to heap corruption. A similar issue exists
2080 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2081 is ever called by user applications with very large untrusted hex/dec data.
2082 This is anticipated to be a rare occurrence.
2084 All OpenSSL internal usage of these functions use data that is not expected
2085 to be untrusted, e.g. config file data or application command line
2086 arguments. If user developed applications generate config file data based
2087 on untrusted data then it is possible that this could also lead to security
2088 consequences. This is also anticipated to be rare.
2090 This issue was reported to OpenSSL by Guido Vranken.
2094 *) Fix memory issues in BIO_*printf functions
2096 The internal |fmtstr| function used in processing a "%s" format string in
2097 the BIO_*printf functions could overflow while calculating the length of a
2098 string and cause an OOB read when printing very long strings.
2100 Additionally the internal |doapr_outch| function can attempt to write to an
2101 OOB memory location (at an offset from the NULL pointer) in the event of a
2102 memory allocation failure. In 1.0.2 and below this could be caused where
2103 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2104 could be in processing a very long "%s" format string. Memory leaks can
2107 The first issue may mask the second issue dependent on compiler behaviour.
2108 These problems could enable attacks where large amounts of untrusted data
2109 is passed to the BIO_*printf functions. If applications use these functions
2110 in this way then they could be vulnerable. OpenSSL itself uses these
2111 functions when printing out human-readable dumps of ASN.1 data. Therefore
2112 applications that print this data could be vulnerable if the data is from
2113 untrusted sources. OpenSSL command line applications could also be
2114 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2115 as command line arguments.
2117 Libssl is not considered directly vulnerable. Additionally certificates etc
2118 received via remote connections via libssl are also unlikely to be able to
2119 trigger these issues because of message size limits enforced within libssl.
2121 This issue was reported to OpenSSL Guido Vranken.
2125 *) Side channel attack on modular exponentiation
2127 A side-channel attack was found which makes use of cache-bank conflicts on
2128 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2129 of RSA keys. The ability to exploit this issue is limited as it relies on
2130 an attacker who has control of code in a thread running on the same
2131 hyper-threaded core as the victim thread which is performing decryptions.
2133 This issue was reported to OpenSSL by Yuval Yarom, The University of
2134 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2135 Nadia Heninger, University of Pennsylvania with more information at
2136 http://cachebleed.info.
2140 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2141 if no keysize is specified with default_bits. This fixes an
2142 omission in an earlier change that changed all RSA/DSA key generation
2143 apps to use 2048 bits by default.
2146 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2147 *) DH small subgroups
2149 Historically OpenSSL only ever generated DH parameters based on "safe"
2150 primes. More recently (in version 1.0.2) support was provided for
2151 generating X9.42 style parameter files such as those required for RFC 5114
2152 support. The primes used in such files may not be "safe". Where an
2153 application is using DH configured with parameters based on primes that are
2154 not "safe" then an attacker could use this fact to find a peer's private
2155 DH exponent. This attack requires that the attacker complete multiple
2156 handshakes in which the peer uses the same private DH exponent. For example
2157 this could be used to discover a TLS server's private DH exponent if it's
2158 reusing the private DH exponent or it's using a static DH ciphersuite.
2160 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2161 TLS. It is not on by default. If the option is not set then the server
2162 reuses the same private DH exponent for the life of the server process and
2163 would be vulnerable to this attack. It is believed that many popular
2164 applications do set this option and would therefore not be at risk.
2166 The fix for this issue adds an additional check where a "q" parameter is
2167 available (as is the case in X9.42 based parameters). This detects the
2168 only known attack, and is the only possible defense for static DH
2169 ciphersuites. This could have some performance impact.
2171 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2172 default and cannot be disabled. This could have some performance impact.
2174 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2178 *) SSLv2 doesn't block disabled ciphers
2180 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2181 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2182 been disabled, provided that the SSLv2 protocol was not also disabled via
2185 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2186 and Sebastian Schinzel.
2190 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2192 *) BN_mod_exp may produce incorrect results on x86_64
2194 There is a carry propagating bug in the x86_64 Montgomery squaring
2195 procedure. No EC algorithms are affected. Analysis suggests that attacks
2196 against RSA and DSA as a result of this defect would be very difficult to
2197 perform and are not believed likely. Attacks against DH are considered just
2198 feasible (although very difficult) because most of the work necessary to
2199 deduce information about a private key may be performed offline. The amount
2200 of resources required for such an attack would be very significant and
2201 likely only accessible to a limited number of attackers. An attacker would
2202 additionally need online access to an unpatched system using the target
2203 private key in a scenario with persistent DH parameters and a private
2204 key that is shared between multiple clients. For example this can occur by
2205 default in OpenSSL DHE based SSL/TLS ciphersuites.
2207 This issue was reported to OpenSSL by Hanno Böck.
2211 *) Certificate verify crash with missing PSS parameter
2213 The signature verification routines will crash with a NULL pointer
2214 dereference if presented with an ASN.1 signature using the RSA PSS
2215 algorithm and absent mask generation function parameter. Since these
2216 routines are used to verify certificate signature algorithms this can be
2217 used to crash any certificate verification operation and exploited in a
2218 DoS attack. Any application which performs certificate verification is
2219 vulnerable including OpenSSL clients and servers which enable client
2222 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2226 *) X509_ATTRIBUTE memory leak
2228 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2229 memory. This structure is used by the PKCS#7 and CMS routines so any
2230 application which reads PKCS#7 or CMS data from untrusted sources is
2231 affected. SSL/TLS is not affected.
2233 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2238 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2239 This changes the decoding behaviour for some invalid messages,
2240 though the change is mostly in the more lenient direction, and
2241 legacy behaviour is preserved as much as possible.
2244 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2246 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2248 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2250 *) Alternate chains certificate forgery
2252 During certificate verification, OpenSSL will attempt to find an
2253 alternative certificate chain if the first attempt to build such a chain
2254 fails. An error in the implementation of this logic can mean that an
2255 attacker could cause certain checks on untrusted certificates to be
2256 bypassed, such as the CA flag, enabling them to use a valid leaf
2257 certificate to act as a CA and "issue" an invalid certificate.
2259 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2263 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2265 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2266 incompatibility in the handling of HMAC. The previous ABI has now been
2270 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2272 *) Malformed ECParameters causes infinite loop
2274 When processing an ECParameters structure OpenSSL enters an infinite loop
2275 if the curve specified is over a specially malformed binary polynomial
2278 This can be used to perform denial of service against any
2279 system which processes public keys, certificate requests or
2280 certificates. This includes TLS clients and TLS servers with
2281 client authentication enabled.
2283 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2287 *) Exploitable out-of-bounds read in X509_cmp_time
2289 X509_cmp_time does not properly check the length of the ASN1_TIME
2290 string and can read a few bytes out of bounds. In addition,
2291 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2294 An attacker can use this to craft malformed certificates and CRLs of
2295 various sizes and potentially cause a segmentation fault, resulting in
2296 a DoS on applications that verify certificates or CRLs. TLS clients
2297 that verify CRLs are affected. TLS clients and servers with client
2298 authentication enabled may be affected if they use custom verification
2301 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2302 independently by Hanno Böck.
2306 *) PKCS7 crash with missing EnvelopedContent
2308 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2309 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2310 with missing content and trigger a NULL pointer dereference on parsing.
2312 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2313 structures from untrusted sources are affected. OpenSSL clients and
2314 servers are not affected.
2316 This issue was reported to OpenSSL by Michal Zalewski (Google).
2320 *) CMS verify infinite loop with unknown hash function
2322 When verifying a signedData message the CMS code can enter an infinite loop
2323 if presented with an unknown hash function OID. This can be used to perform
2324 denial of service against any system which verifies signedData messages using
2326 This issue was reported to OpenSSL by Johannes Bauer.
2330 *) Race condition handling NewSessionTicket
2332 If a NewSessionTicket is received by a multi-threaded client when attempting to
2333 reuse a previous ticket then a race condition can occur potentially leading to
2334 a double free of the ticket data.
2338 *) Only support 256-bit or stronger elliptic curves with the
2339 'ecdh_auto' setting (server) or by default (client). Of supported
2340 curves, prefer P-256 (both).
2343 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2345 *) ClientHello sigalgs DoS fix
2347 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2348 invalid signature algorithms extension a NULL pointer dereference will
2349 occur. This can be exploited in a DoS attack against the server.
2351 This issue was was reported to OpenSSL by David Ramos of Stanford
2354 [Stephen Henson and Matt Caswell]
2356 *) Multiblock corrupted pointer fix
2358 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2359 feature only applies on 64 bit x86 architecture platforms that support AES
2360 NI instructions. A defect in the implementation of "multiblock" can cause
2361 OpenSSL's internal write buffer to become incorrectly set to NULL when
2362 using non-blocking IO. Typically, when the user application is using a
2363 socket BIO for writing, this will only result in a failed connection.
2364 However if some other BIO is used then it is likely that a segmentation
2365 fault will be triggered, thus enabling a potential DoS attack.
2367 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2371 *) Segmentation fault in DTLSv1_listen fix
2373 The DTLSv1_listen function is intended to be stateless and processes the
2374 initial ClientHello from many peers. It is common for user code to loop
2375 over the call to DTLSv1_listen until a valid ClientHello is received with
2376 an associated cookie. A defect in the implementation of DTLSv1_listen means
2377 that state is preserved in the SSL object from one invocation to the next
2378 that can lead to a segmentation fault. Errors processing the initial
2379 ClientHello can trigger this scenario. An example of such an error could be
2380 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2383 This issue was reported to OpenSSL by Per Allansson.
2387 *) Segmentation fault in ASN1_TYPE_cmp fix
2389 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2390 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2391 certificate signature algorithm consistency this can be used to crash any
2392 certificate verification operation and exploited in a DoS attack. Any
2393 application which performs certificate verification is vulnerable including
2394 OpenSSL clients and servers which enable client authentication.
2398 *) Segmentation fault for invalid PSS parameters fix
2400 The signature verification routines will crash with a NULL pointer
2401 dereference if presented with an ASN.1 signature using the RSA PSS
2402 algorithm and invalid parameters. Since these routines are used to verify
2403 certificate signature algorithms this can be used to crash any
2404 certificate verification operation and exploited in a DoS attack. Any
2405 application which performs certificate verification is vulnerable including
2406 OpenSSL clients and servers which enable client authentication.
2408 This issue was was reported to OpenSSL by Brian Carpenter.
2412 *) ASN.1 structure reuse memory corruption fix
2414 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2415 memory corruption via an invalid write. Such reuse is and has been
2416 strongly discouraged and is believed to be rare.
2418 Applications that parse structures containing CHOICE or ANY DEFINED BY
2419 components may be affected. Certificate parsing (d2i_X509 and related
2420 functions) are however not affected. OpenSSL clients and servers are
2425 *) PKCS7 NULL pointer dereferences fix
2427 The PKCS#7 parsing code does not handle missing outer ContentInfo
2428 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2429 missing content and trigger a NULL pointer dereference on parsing.
2431 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2432 otherwise parse PKCS#7 structures from untrusted sources are
2433 affected. OpenSSL clients and servers are not affected.
2435 This issue was reported to OpenSSL by Michal Zalewski (Google).
2439 *) DoS via reachable assert in SSLv2 servers fix
2441 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2442 servers that both support SSLv2 and enable export cipher suites by sending
2443 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2445 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2446 (OpenSSL development team).
2450 *) Empty CKE with client auth and DHE fix
2452 If client auth is used then a server can seg fault in the event of a DHE
2453 ciphersuite being selected and a zero length ClientKeyExchange message
2454 being sent by the client. This could be exploited in a DoS attack.
2458 *) Handshake with unseeded PRNG fix
2460 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2461 with an unseeded PRNG. The conditions are:
2462 - The client is on a platform where the PRNG has not been seeded
2463 automatically, and the user has not seeded manually
2464 - A protocol specific client method version has been used (i.e. not
2465 SSL_client_methodv23)
2466 - A ciphersuite is used that does not require additional random data from
2467 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2469 If the handshake succeeds then the client random that has been used will
2470 have been generated from a PRNG with insufficient entropy and therefore the
2471 output may be predictable.
2473 For example using the following command with an unseeded openssl will
2474 succeed on an unpatched platform:
2476 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2480 *) Use After Free following d2i_ECPrivatekey error fix
2482 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2483 could cause a use after free condition. This, in turn, could cause a double
2484 free in several private key parsing functions (such as d2i_PrivateKey
2485 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2486 for applications that receive EC private keys from untrusted
2487 sources. This scenario is considered rare.
2489 This issue was discovered by the BoringSSL project and fixed in their
2494 *) X509_to_X509_REQ NULL pointer deref fix
2496 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2497 the certificate key is invalid. This function is rarely used in practice.
2499 This issue was discovered by Brian Carpenter.
2503 *) Removed the export ciphers from the DEFAULT ciphers
2506 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2508 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2509 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2510 So far those who have to target multiple platforms would compromise
2511 and argue that binary targeting say ARMv5 would still execute on
2512 ARMv8. "Universal" build resolves this compromise by providing
2513 near-optimal performance even on newer platforms.
2516 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2517 (other platforms pending).
2518 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2520 *) Add support for the SignedCertificateTimestampList certificate and
2521 OCSP response extensions from RFC6962.
2524 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2525 for corner cases. (Certain input points at infinity could lead to
2526 bogus results, with non-infinity inputs mapped to infinity too.)
2529 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2530 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2531 common cases are optimized and there still is room for further
2532 improvements. Vector Permutation AES for Altivec is also added.
2535 *) Add support for little-endian ppc64 Linux target.
2536 [Marcelo Cerri (IBM)]
2538 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2539 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2540 are optimized and there still is room for further improvements.
2541 Both 32- and 64-bit modes are supported.
2542 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2544 *) Improved ARMv7 NEON support.
2547 *) Support for SPARC Architecture 2011 crypto extensions, first
2548 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2549 SHA256/512, MD5, GHASH and modular exponentiation.
2550 [Andy Polyakov, David Miller]
2552 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2554 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2556 *) Support for new and upcoming Intel processors, including AVX2,
2557 BMI and SHA ISA extensions. This includes additional "stitched"
2558 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2561 This work was sponsored by Intel Corp.
2564 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2565 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2566 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2569 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2570 this fixes a limitation in previous versions of OpenSSL.
2573 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2574 MGF1 digest and OAEP label.
2577 *) Add EVP support for key wrapping algorithms, to avoid problems with
2578 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2579 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2580 algorithms and include tests cases.
2583 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2585 [Douglas E. Engert, Steve Henson]
2587 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2588 difference in days and seconds between two tm or ASN1_TIME structures.
2591 *) Add -rev test option to s_server to just reverse order of characters
2592 received by client and send back to server. Also prints an abbreviated
2593 summary of the connection parameters.
2596 *) New option -brief for s_client and s_server to print out a brief summary
2597 of connection parameters.
2600 *) Add callbacks for arbitrary TLS extensions.
2601 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2603 *) New option -crl_download in several openssl utilities to download CRLs
2604 from CRLDP extension in certificates.
2607 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2610 *) New function X509_CRL_diff to generate a delta CRL from the difference
2611 of two full CRLs. Add support to "crl" utility.
2614 *) New functions to set lookup_crls function and to retrieve
2615 X509_STORE from X509_STORE_CTX.
2618 *) Print out deprecated issuer and subject unique ID fields in
2622 *) Extend OCSP I/O functions so they can be used for simple general purpose
2623 HTTP as well as OCSP. New wrapper function which can be used to download
2624 CRLs using the OCSP API.
2627 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2630 *) SSL_CONF* functions. These provide a common framework for application
2631 configuration using configuration files or command lines.
2634 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2635 message callback and prints the results. Needs compile time option
2636 "enable-ssl-trace". New options to s_client and s_server to enable
2640 *) New ctrl and macro to retrieve supported points extensions.
2641 Print out extension in s_server and s_client.
2644 *) New functions to retrieve certificate signature and signature
2648 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2652 *) New Suite B modes for TLS code. These use and enforce the requirements
2653 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2654 only use Suite B curves. The Suite B modes can be set by using the
2655 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2658 *) New chain verification flags for Suite B levels of security. Check
2659 algorithms are acceptable when flags are set in X509_verify_cert.
2662 *) Make tls1_check_chain return a set of flags indicating checks passed
2663 by a certificate chain. Add additional tests to handle client
2664 certificates: checks for matching certificate type and issuer name
2668 *) If an attempt is made to use a signature algorithm not in the peer
2669 preference list abort the handshake. If client has no suitable
2670 signature algorithms in response to a certificate request do not
2671 use the certificate.
2674 *) If server EC tmp key is not in client preference list abort handshake.
2677 *) Add support for certificate stores in CERT structure. This makes it
2678 possible to have different stores per SSL structure or one store in
2679 the parent SSL_CTX. Include distinct stores for certificate chain
2680 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2681 to build and store a certificate chain in CERT structure: returning
2682 an error if the chain cannot be built: this will allow applications
2683 to test if a chain is correctly configured.
2685 Note: if the CERT based stores are not set then the parent SSL_CTX
2686 store is used to retain compatibility with existing behaviour.
2690 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2691 mask based on the current session, check mask when sending client
2692 hello and checking the requested ciphersuite.
2695 *) New ctrls to retrieve and set certificate types in a certificate
2696 request message. Print out received values in s_client. If certificate
2697 types is not set with custom values set sensible values based on
2698 supported signature algorithms.
2701 *) Support for distinct client and server supported signature algorithms.
2704 *) Add certificate callback. If set this is called whenever a certificate
2705 is required by client or server. An application can decide which
2706 certificate chain to present based on arbitrary criteria: for example
2707 supported signature algorithms. Add very simple example to s_server.
2708 This fixes many of the problems and restrictions of the existing client
2709 certificate callback: for example you can now clear an existing
2710 certificate and specify the whole chain.
2713 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2714 the certificate can be used for (if anything). Set valid_flags field
2715 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2716 to have similar checks in it.
2718 Add new "cert_flags" field to CERT structure and include a "strict mode".
2719 This enforces some TLS certificate requirements (such as only permitting
2720 certificate signature algorithms contained in the supported algorithms
2721 extension) which some implementations ignore: this option should be used
2722 with caution as it could cause interoperability issues.
2725 *) Update and tidy signature algorithm extension processing. Work out
2726 shared signature algorithms based on preferences and peer algorithms
2727 and print them out in s_client and s_server. Abort handshake if no
2728 shared signature algorithms.
2731 *) Add new functions to allow customised supported signature algorithms
2732 for SSL and SSL_CTX structures. Add options to s_client and s_server
2736 *) New function SSL_certs_clear() to delete all references to certificates
2737 from an SSL structure. Before this once a certificate had been added
2738 it couldn't be removed.
2741 *) Integrate hostname, email address and IP address checking with certificate
2742 verification. New verify options supporting checking in openssl utility.
2745 *) Fixes and wildcard matching support to hostname and email checking
2746 functions. Add manual page.
2747 [Florian Weimer (Red Hat Product Security Team)]
2749 *) New functions to check a hostname email or IP address against a
2750 certificate. Add options x509 utility to print results of checks against
2754 *) Fix OCSP checking.
2755 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2757 *) Initial experimental support for explicitly trusted non-root CAs.
2758 OpenSSL still tries to build a complete chain to a root but if an
2759 intermediate CA has a trust setting included that is used. The first
2760 setting is used: whether to trust (e.g., -addtrust option to the x509
2764 *) Add -trusted_first option which attempts to find certificates in the
2765 trusted store even if an untrusted chain is also supplied.
2768 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2769 platform support for Linux and Android.
2772 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2775 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2776 When in FIPS mode the approved implementations are used as normal,
2777 when not in FIPS mode the internal unapproved versions are used instead.
2778 This means that the FIPS capable OpenSSL isn't forced to use the
2779 (often lower performance) FIPS implementations outside FIPS mode.
2782 *) Transparently support X9.42 DH parameters when calling
2783 PEM_read_bio_DHparameters. This means existing applications can handle
2784 the new parameter format automatically.
2787 *) Initial experimental support for X9.42 DH parameter format: mainly
2788 to support use of 'q' parameter for RFC5114 parameters.
2791 *) Add DH parameters from RFC5114 including test data to dhtest.
2794 *) Support for automatic EC temporary key parameter selection. If enabled
2795 the most preferred EC parameters are automatically used instead of
2796 hardcoded fixed parameters. Now a server just has to call:
2797 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2798 support ECDH and use the most appropriate parameters.
2801 *) Enhance and tidy EC curve and point format TLS extension code. Use
2802 static structures instead of allocation if default values are used.
2803 New ctrls to set curves we wish to support and to retrieve shared curves.
2804 Print out shared curves in s_server. New options to s_server and s_client
2805 to set list of supported curves.
2808 *) New ctrls to retrieve supported signature algorithms and
2809 supported curve values as an array of NIDs. Extend openssl utility
2810 to print out received values.
2813 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2814 between NIDs and the more common NIST names such as "P-256". Enhance
2815 ecparam utility and ECC method to recognise the NIST names for curves.
2818 *) Enhance SSL/TLS certificate chain handling to support different
2819 chains for each certificate instead of one chain in the parent SSL_CTX.
2822 *) Support for fixed DH ciphersuite client authentication: where both
2823 server and client use DH certificates with common parameters.
2826 *) Support for fixed DH ciphersuites: those requiring DH server
2830 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2832 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2833 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2834 X509_CINF_get_signature were reverted post internal team review.
2836 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2838 *) Build fixes for the Windows and OpenVMS platforms
2839 [Matt Caswell and Richard Levitte]
2841 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2843 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2844 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2845 dereference. This could lead to a Denial Of Service attack. Thanks to
2846 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2850 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2851 dtls1_buffer_record function under certain conditions. In particular this
2852 could occur if an attacker sent repeated DTLS records with the same
2853 sequence number but for the next epoch. The memory leak could be exploited
2854 by an attacker in a Denial of Service attack through memory exhaustion.
2855 Thanks to Chris Mueller for reporting this issue.
2859 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2860 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2861 method would be set to NULL which could later result in a NULL pointer
2862 dereference. Thanks to Frank Schmirler for reporting this issue.
2866 *) Abort handshake if server key exchange message is omitted for ephemeral
2869 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2870 reporting this issue.
2874 *) Remove non-export ephemeral RSA code on client and server. This code
2875 violated the TLS standard by allowing the use of temporary RSA keys in
2876 non-export ciphersuites and could be used by a server to effectively
2877 downgrade the RSA key length used to a value smaller than the server
2878 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2879 INRIA or reporting this issue.
2883 *) Fixed issue where DH client certificates are accepted without verification.
2884 An OpenSSL server will accept a DH certificate for client authentication
2885 without the certificate verify message. This effectively allows a client to
2886 authenticate without the use of a private key. This only affects servers
2887 which trust a client certificate authority which issues certificates
2888 containing DH keys: these are extremely rare and hardly ever encountered.
2889 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2894 *) Ensure that the session ID context of an SSL is updated when its
2895 SSL_CTX is updated via SSL_set_SSL_CTX.
2897 The session ID context is typically set from the parent SSL_CTX,
2898 and can vary with the CTX.
2901 *) Fix various certificate fingerprint issues.
2903 By using non-DER or invalid encodings outside the signed portion of a
2904 certificate the fingerprint can be changed without breaking the signature.
2905 Although no details of the signed portion of the certificate can be changed
2906 this can cause problems with some applications: e.g. those using the
2907 certificate fingerprint for blacklists.
2909 1. Reject signatures with non zero unused bits.
2911 If the BIT STRING containing the signature has non zero unused bits reject
2912 the signature. All current signature algorithms require zero unused bits.
2914 2. Check certificate algorithm consistency.
2916 Check the AlgorithmIdentifier inside TBS matches the one in the
2917 certificate signature. NB: this will result in signature failure
2918 errors for some broken certificates.
2920 Thanks to Konrad Kraszewski from Google for reporting this issue.
2922 3. Check DSA/ECDSA signatures use DER.
2924 Re-encode DSA/ECDSA signatures and compare with the original received
2925 signature. Return an error if there is a mismatch.
2927 This will reject various cases including garbage after signature
2928 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2929 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2930 (negative or with leading zeroes).
2932 Further analysis was conducted and fixes were developed by Stephen Henson
2933 of the OpenSSL core team.
2938 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2939 results on some platforms, including x86_64. This bug occurs at random
2940 with a very low probability, and is not known to be exploitable in any
2941 way, though its exact impact is difficult to determine. Thanks to Pieter
2942 Wuille (Blockstream) who reported this issue and also suggested an initial
2943 fix. Further analysis was conducted by the OpenSSL development team and
2944 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2945 the OpenSSL core team.
2949 *) Do not resume sessions on the server if the negotiated protocol
2950 version does not match the session's version. Resuming with a different
2951 version, while not strictly forbidden by the RFC, is of questionable
2952 sanity and breaks all known clients.
2953 [David Benjamin, Emilia Käsper]
2955 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2956 early CCS messages during renegotiation. (Note that because
2957 renegotiation is encrypted, this early CCS was not exploitable.)
2960 *) Tighten client-side session ticket handling during renegotiation:
2961 ensure that the client only accepts a session ticket if the server sends
2962 the extension anew in the ServerHello. Previously, a TLS client would
2963 reuse the old extension state and thus accept a session ticket if one was
2964 announced in the initial ServerHello.
2966 Similarly, ensure that the client requires a session ticket if one
2967 was advertised in the ServerHello. Previously, a TLS client would
2968 ignore a missing NewSessionTicket message.
2971 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2973 *) SRTP Memory Leak.
2975 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2976 sends a carefully crafted handshake message, to cause OpenSSL to fail
2977 to free up to 64k of memory causing a memory leak. This could be
2978 exploited in a Denial Of Service attack. This issue affects OpenSSL
2979 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2980 whether SRTP is used or configured. Implementations of OpenSSL that
2981 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2983 The fix was developed by the OpenSSL team.
2987 *) Session Ticket Memory Leak.
2989 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2990 integrity of that ticket is first verified. In the event of a session
2991 ticket integrity check failing, OpenSSL will fail to free memory
2992 causing a memory leak. By sending a large number of invalid session
2993 tickets an attacker could exploit this issue in a Denial Of Service
2998 *) Build option no-ssl3 is incomplete.
3000 When OpenSSL is configured with "no-ssl3" as a build option, servers
3001 could accept and complete a SSL 3.0 handshake, and clients could be
3002 configured to send them.
3004 [Akamai and the OpenSSL team]
3006 *) Add support for TLS_FALLBACK_SCSV.
3007 Client applications doing fallback retries should call
3008 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3010 [Adam Langley, Bodo Moeller]
3012 *) Add additional DigestInfo checks.
3014 Re-encode DigestInto in DER and check against the original when
3015 verifying RSA signature: this will reject any improperly encoded
3016 DigestInfo structures.
3018 Note: this is a precautionary measure and no attacks are currently known.
3022 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3024 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3025 SRP code can be overrun an internal buffer. Add sanity check that
3026 g, A, B < N to SRP code.
3028 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3029 Group for discovering this issue.
3033 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3034 TLS 1.0 instead of higher protocol versions when the ClientHello message
3035 is badly fragmented. This allows a man-in-the-middle attacker to force a
3036 downgrade to TLS 1.0 even if both the server and the client support a
3037 higher protocol version, by modifying the client's TLS records.
3039 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3040 researching this issue.
3044 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3045 to a denial of service attack. A malicious server can crash the client
3046 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3047 ciphersuite and sending carefully crafted handshake messages.
3049 Thanks to Felix Gröbert (Google) for discovering and researching this
3054 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3055 to leak memory. This can be exploited through a Denial of Service attack.
3056 Thanks to Adam Langley for discovering and researching this issue.
3060 *) An attacker can force openssl to consume large amounts of memory whilst
3061 processing DTLS handshake messages. This can be exploited through a
3062 Denial of Service attack.
3063 Thanks to Adam Langley for discovering and researching this issue.
3067 *) An attacker can force an error condition which causes openssl to crash
3068 whilst processing DTLS packets due to memory being freed twice. This
3069 can be exploited through a Denial of Service attack.
3070 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3075 *) If a multithreaded client connects to a malicious server using a resumed
3076 session and the server sends an ec point format extension it could write
3077 up to 255 bytes to freed memory.
3079 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3084 *) A malicious server can crash an OpenSSL client with a null pointer
3085 dereference (read) by specifying an SRP ciphersuite even though it was not
3086 properly negotiated with the client. This can be exploited through a
3087 Denial of Service attack.
3089 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3090 discovering and researching this issue.
3094 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3095 X509_name_oneline, X509_name_print_ex et al. to leak some information
3096 from the stack. Applications may be affected if they echo pretty printing
3097 output to the attacker.
3099 Thanks to Ivan Fratric (Google) for discovering this issue.
3101 [Emilia Käsper, and Steve Henson]
3103 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3104 for corner cases. (Certain input points at infinity could lead to
3105 bogus results, with non-infinity inputs mapped to infinity too.)
3108 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3110 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3111 handshake can force the use of weak keying material in OpenSSL
3112 SSL/TLS clients and servers.
3114 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3115 researching this issue. (CVE-2014-0224)
3116 [KIKUCHI Masashi, Steve Henson]
3118 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3119 OpenSSL DTLS client the code can be made to recurse eventually crashing
3122 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3124 [Imre Rad, Steve Henson]
3126 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3127 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3128 client or server. This is potentially exploitable to run arbitrary
3129 code on a vulnerable client or server.
3131 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3132 [Jüri Aedla, Steve Henson]
3134 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3135 are subject to a denial of service attack.
3137 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3138 this issue. (CVE-2014-3470)
3139 [Felix Gröbert, Ivan Fratric, Steve Henson]
3141 *) Harmonize version and its documentation. -f flag is used to display
3143 [mancha <mancha1@zoho.com>]
3145 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3146 in i2d_ECPrivateKey.
3147 [mancha <mancha1@zoho.com>]
3149 *) Fix some double frees. These are not thought to be exploitable.
3150 [mancha <mancha1@zoho.com>]
3152 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3154 *) A missing bounds check in the handling of the TLS heartbeat extension
3155 can be used to reveal up to 64k of memory to a connected client or
3158 Thanks for Neel Mehta of Google Security for discovering this bug and to
3159 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3160 preparing the fix (CVE-2014-0160)
3161 [Adam Langley, Bodo Moeller]
3163 *) Fix for the attack described in the paper "Recovering OpenSSL
3164 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3165 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3166 http://eprint.iacr.org/2014/140
3168 Thanks to Yuval Yarom and Naomi Benger for discovering this
3169 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3170 [Yuval Yarom and Naomi Benger]
3172 *) TLS pad extension: draft-agl-tls-padding-03
3174 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3175 TLS client Hello record length value would otherwise be > 255 and
3176 less that 512 pad with a dummy extension containing zeroes so it
3177 is at least 512 bytes long.
3179 [Adam Langley, Steve Henson]
3181 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3183 *) Fix for TLS record tampering bug. A carefully crafted invalid
3184 handshake could crash OpenSSL with a NULL pointer exception.
3185 Thanks to Anton Johansson for reporting this issues.
3188 *) Keep original DTLS digest and encryption contexts in retransmission
3189 structures so we can use the previous session parameters if they need
3190 to be resent. (CVE-2013-6450)
3193 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3194 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3195 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3196 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3197 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3198 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3199 [Rob Stradling, Adam Langley]
3201 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3203 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3204 supporting platforms or when small records were transferred.
3205 [Andy Polyakov, Steve Henson]
3207 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3209 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3211 This addresses the flaw in CBC record processing discovered by
3212 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3213 at: http://www.isg.rhul.ac.uk/tls/
3215 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3216 Security Group at Royal Holloway, University of London
3217 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3218 Emilia Käsper for the initial patch.
3220 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3222 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3223 ciphersuites which can be exploited in a denial of service attack.
3224 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3225 and detecting this bug and to Wolfgang Ettlinger
3226 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3230 *) Return an error when checking OCSP signatures when key is NULL.
3231 This fixes a DoS attack. (CVE-2013-0166)
3234 *) Make openssl verify return errors.
3235 [Chris Palmer <palmer@google.com> and Ben Laurie]
3237 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3238 the right response is stapled. Also change SSL_get_certificate()
3239 so it returns the certificate actually sent.
3240 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3241 [Rob Stradling <rob.stradling@comodo.com>]
3243 *) Fix possible deadlock when decoding public keys.
3246 *) Don't use TLS 1.0 record version number in initial client hello
3250 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3252 *) Sanity check record length before skipping explicit IV in TLS
3253 1.2, 1.1 and DTLS to fix DoS attack.
3255 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3256 fuzzing as a service testing platform.
3260 *) Initialise tkeylen properly when encrypting CMS messages.
3261 Thanks to Solar Designer of Openwall for reporting this issue.
3264 *) In FIPS mode don't try to use composite ciphers as they are not
3268 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3270 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3271 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3272 mean any application compiled against OpenSSL 1.0.0 headers setting
3273 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3274 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3275 0x10000000L Any application which was previously compiled against
3276 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3277 will need to be recompiled as a result. Letting be results in
3278 inability to disable specifically TLS 1.1 and in client context,
3279 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3282 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3283 disable just protocol X, but all protocols above X *if* there are
3284 protocols *below* X still enabled. In more practical terms it means
3285 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3286 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3287 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3291 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3293 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3294 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3295 in CRYPTO_realloc_clean.
3297 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3298 issue and to Adam Langley <agl@chromium.org> for fixing it.
3300 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3302 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3305 *) Workarounds for some broken servers that "hang" if a client hello
3306 record length exceeds 255 bytes.
3308 1. Do not use record version number > TLS 1.0 in initial client
3309 hello: some (but not all) hanging servers will now work.
3310 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3311 the number of ciphers sent in the client hello. This should be
3312 set to an even number, such as 50, for example by passing:
3313 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3314 Most broken servers should now work.
3315 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3316 TLS 1.2 client support entirely.
3319 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3322 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3324 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3325 STRING form instead of a DigestInfo.
3328 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3329 and the RSA_sign/RSA_verify functions. This was made more apparent when
3330 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3331 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3332 the correct format in RSA_verify so both forms transparently work.
3335 *) Some servers which support TLS 1.0 can choke if we initially indicate
3336 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3337 encrypted premaster secret. As a workaround use the maximum permitted
3338 client version in client hello, this should keep such servers happy
3339 and still work with previous versions of OpenSSL.
3342 *) Add support for TLS/DTLS heartbeats.
3343 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3345 *) Add support for SCTP.
3346 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3348 *) Improved PRNG seeding for VOS.
3349 [Paul Green <Paul.Green@stratus.com>]
3351 *) Extensive assembler packs updates, most notably:
3353 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3354 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3355 - x86_64: bit-sliced AES implementation;
3356 - ARM: NEON support, contemporary platforms optimizations;
3357 - s390x: z196 support;
3358 - *: GHASH and GF(2^m) multiplication implementations;
3362 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3363 (removal of unnecessary code)
3364 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3366 *) Add TLS key material exporter from RFC 5705.
3369 *) Add DTLS-SRTP negotiation from RFC 5764.
3372 *) Add Next Protocol Negotiation,
3373 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3374 disabled with a no-npn flag to config or Configure. Code donated
3376 [Adam Langley <agl@google.com> and Ben Laurie]
3378 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3379 NIST-P256, NIST-P521, with constant-time single point multiplication on
3380 typical inputs. Compiler support for the nonstandard type __uint128_t is
3381 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3382 Code made available under Apache License version 2.0.
3384 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3385 line to include this in your build of OpenSSL, and run "make depend" (or
3386 "make update"). This enables the following EC_METHODs:
3388 EC_GFp_nistp224_method()
3389 EC_GFp_nistp256_method()
3390 EC_GFp_nistp521_method()
3392 EC_GROUP_new_by_curve_name() will automatically use these (while
3393 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3395 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3397 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3398 all platforms. Move ssize_t definition from e_os.h to the public
3399 header file e_os2.h as it now appears in public header file cms.h
3402 *) New -sigopt option to the ca, req and x509 utilities. Additional
3403 signature parameters can be passed using this option and in
3407 *) Add RSA PSS signing function. This will generate and set the
3408 appropriate AlgorithmIdentifiers for PSS based on those in the
3409 corresponding EVP_MD_CTX structure. No application support yet.
3412 *) Support for companion algorithm specific ASN1 signing routines.
3413 New function ASN1_item_sign_ctx() signs a pre-initialised
3414 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3415 the appropriate parameters.
3418 *) Add new algorithm specific ASN1 verification initialisation function
3419 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3420 handling will be the same no matter what EVP_PKEY_METHOD is used.
3421 Add a PSS handler to support verification of PSS signatures: checked
3422 against a number of sample certificates.
3425 *) Add signature printing for PSS. Add PSS OIDs.
3426 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3428 *) Add algorithm specific signature printing. An individual ASN1 method
3429 can now print out signatures instead of the standard hex dump.
3431 More complex signatures (e.g. PSS) can print out more meaningful
3432 information. Include DSA version that prints out the signature
3436 *) Password based recipient info support for CMS library: implementing
3440 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3441 neatly separates the code into cipher and PBE sections and is required
3442 for some algorithms that split PBES2 into separate pieces (such as
3443 password based CMS).
3446 *) Session-handling fixes:
3447 - Fix handling of connections that are resuming with a session ID,
3448 but also support Session Tickets.
3449 - Fix a bug that suppressed issuing of a new ticket if the client
3450 presented a ticket with an expired session.
3451 - Try to set the ticket lifetime hint to something reasonable.
3452 - Make tickets shorter by excluding irrelevant information.
3453 - On the client side, don't ignore renewed tickets.
3454 [Adam Langley, Bodo Moeller (Google)]
3456 *) Fix PSK session representation.
3459 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3461 This work was sponsored by Intel.
3464 *) Add GCM support to TLS library. Some custom code is needed to split
3465 the IV between the fixed (from PRF) and explicit (from TLS record)
3466 portions. This adds all GCM ciphersuites supported by RFC5288 and
3467 RFC5289. Generalise some AES* cipherstrings to include GCM and
3468 add a special AESGCM string for GCM only.
3471 *) Expand range of ctrls for AES GCM. Permit setting invocation
3472 field on decrypt and retrieval of invocation field only on encrypt.
3475 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3476 As required by RFC5289 these ciphersuites cannot be used if for
3477 versions of TLS earlier than 1.2.
3480 *) For FIPS capable OpenSSL interpret a NULL default public key method
3481 as unset and return the appropriate default but do *not* set the default.
3482 This means we can return the appropriate method in applications that
3483 switch between FIPS and non-FIPS modes.
3486 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3487 ENGINE is used then we cannot handle that in the FIPS module so we
3488 keep original code iff non-FIPS operations are allowed.
3491 *) Add -attime option to openssl utilities.
3492 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3494 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3497 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3498 FIPS EC methods unconditionally for now.
3501 *) New build option no-ec2m to disable characteristic 2 code.
3504 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3505 all cases can be covered as some introduce binary incompatibilities.
3508 *) Redirect RSA operations to FIPS module including keygen,
3509 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3512 *) Add similar low level API blocking to ciphers.
3515 *) Low level digest APIs are not approved in FIPS mode: any attempt
3516 to use these will cause a fatal error. Applications that *really* want
3517 to use them can use the private_* version instead.
3520 *) Redirect cipher operations to FIPS module for FIPS builds.
3523 *) Redirect digest operations to FIPS module for FIPS builds.
3526 *) Update build system to add "fips" flag which will link in fipscanister.o
3527 for static and shared library builds embedding a signature if needed.
3530 *) Output TLS supported curves in preference order instead of numerical
3531 order. This is currently hardcoded for the highest order curves first.
3532 This should be configurable so applications can judge speed vs strength.
3535 *) Add TLS v1.2 server support for client authentication.
3538 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3542 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3543 FIPS modules versions.
3546 *) Add TLS v1.2 client side support for client authentication. Keep cache
3547 of handshake records longer as we don't know the hash algorithm to use
3548 until after the certificate request message is received.
3551 *) Initial TLS v1.2 client support. Add a default signature algorithms
3552 extension including all the algorithms we support. Parse new signature
3553 format in client key exchange. Relax some ECC signing restrictions for
3554 TLS v1.2 as indicated in RFC5246.
3557 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3558 to new signature format when needed using client digest preference.
3559 All server ciphersuites should now work correctly in TLS v1.2. No client
3560 support yet and no support for client certificates.
3563 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3564 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3565 ciphersuites. At present only RSA key exchange ciphersuites work with
3566 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3567 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3568 and version checking.
3571 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3572 with this defined it will not be affected by any changes to ssl internal
3573 structures. Add several utility functions to allow openssl application
3574 to work with OPENSSL_NO_SSL_INTERN defined.
3577 *) A long standing patch to add support for SRP from EdelWeb (Peter
3578 Sylvester and Christophe Renou) was integrated.
3579 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3580 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3583 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3586 *) Permit abbreviated handshakes when renegotiating using the function
3587 SSL_renegotiate_abbreviated().
3588 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3590 *) Add call to ENGINE_register_all_complete() to
3591 ENGINE_load_builtin_engines(), so some implementations get used
3592 automatically instead of needing explicit application support.
3595 *) Add support for TLS key exporter as described in RFC5705.
3596 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3598 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3599 a few changes are required:
3601 Add SSL_OP_NO_TLSv1_1 flag.
3602 Add TLSv1_1 methods.
3603 Update version checking logic to handle version 1.1.
3604 Add explicit IV handling (ported from DTLS code).
3605 Add command line options to s_client/s_server.
3608 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3610 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3611 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3612 content decryption and always return the same error. Note: this attack
3613 needs on average 2^20 messages so it only affects automated senders. The
3614 old behaviour can be re-enabled in the CMS code by setting the
3615 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3616 an MMA defence is not necessary.
3617 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3618 this issue. (CVE-2012-0884)
3621 *) Fix CVE-2011-4619: make sure we really are receiving a
3622 client hello before rejecting multiple SGC restarts. Thanks to
3623 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3626 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3628 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3629 Thanks to Antonio Martin, Enterprise Secure Access Research and
3630 Development, Cisco Systems, Inc. for discovering this bug and
3631 preparing a fix. (CVE-2012-0050)
3634 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3636 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3637 of the Vaudenay padding oracle attack on CBC mode encryption
3638 which enables an efficient plaintext recovery attack against
3639 the OpenSSL implementation of DTLS. Their attack exploits timing
3640 differences arising during decryption processing. A research
3641 paper describing this attack can be found at:
3642 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3643 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3644 Security Group at Royal Holloway, University of London
3645 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3646 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3647 for preparing the fix. (CVE-2011-4108)
3648 [Robin Seggelmann, Michael Tuexen]
3650 *) Clear bytes used for block padding of SSL 3.0 records.
3652 [Adam Langley (Google)]
3654 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3655 Kadianakis <desnacked@gmail.com> for discovering this issue and
3656 Adam Langley for preparing the fix. (CVE-2011-4619)
3657 [Adam Langley (Google)]
3659 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3660 [Andrey Kulikov <amdeich@gmail.com>]
3662 *) Prevent malformed RFC3779 data triggering an assertion failure.
3663 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3664 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3665 [Rob Austein <sra@hactrn.net>]
3667 *) Improved PRNG seeding for VOS.
3668 [Paul Green <Paul.Green@stratus.com>]
3670 *) Fix ssl_ciph.c set-up race.
3671 [Adam Langley (Google)]
3673 *) Fix spurious failures in ecdsatest.c.
3674 [Emilia Käsper (Google)]
3676 *) Fix the BIO_f_buffer() implementation (which was mixing different
3677 interpretations of the '..._len' fields).
3678 [Adam Langley (Google)]
3680 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3681 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3682 threads won't reuse the same blinding coefficients.
3684 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3685 lock to call BN_BLINDING_invert_ex, and avoids one use of
3686 BN_BLINDING_update for each BN_BLINDING structure (previously,
3687 the last update always remained unused).
3688 [Emilia Käsper (Google)]
3690 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3691 [Bob Buckholz (Google)]
3693 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3695 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3696 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3697 [Kaspar Brand <ossl@velox.ch>]
3699 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3700 for multi-threaded use of ECDH. (CVE-2011-3210)
3701 [Adam Langley (Google)]
3703 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3706 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3707 signature public key algorithm by using OID xref utilities instead.
3708 Before this you could only use some ECC ciphersuites with SHA1 only.
3711 *) Add protection against ECDSA timing attacks as mentioned in the paper
3712 by Billy Bob Brumley and Nicola Tuveri, see:
3714 http://eprint.iacr.org/2011/232.pdf
3716 [Billy Bob Brumley and Nicola Tuveri]
3718 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3720 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3721 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3723 *) Fix bug in string printing code: if *any* escaping is enabled we must
3724 escape the escape character (backslash) or the resulting string is
3728 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3730 *) Disable code workaround for ancient and obsolete Netscape browsers
3731 and servers: an attacker can use it in a ciphersuite downgrade attack.
3732 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3735 *) Fixed J-PAKE implementation error, originally discovered by
3736 Sebastien Martini, further info and confirmation from Stefan
3737 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3740 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3742 *) Fix extension code to avoid race conditions which can result in a buffer
3743 overrun vulnerability: resumed sessions must not be modified as they can
3744 be shared by multiple threads. CVE-2010-3864
3747 *) Fix WIN32 build system to correctly link an ENGINE directory into
3751 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3753 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3755 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3757 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3759 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3760 context. The operation can be customised via the ctrl mechanism in
3761 case ENGINEs want to include additional functionality.
3764 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3767 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3768 output hashes compatible with older versions of OpenSSL.
3769 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3771 *) Fix compression algorithm handling: if resuming a session use the
3772 compression algorithm of the resumed session instead of determining
3773 it from client hello again. Don't allow server to change algorithm.
3776 *) Add load_crls() function to apps tidying load_certs() too. Add option
3777 to verify utility to allow additional CRLs to be included.
3780 *) Update OCSP request code to permit adding custom headers to the request:
3781 some responders need this.
3784 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3786 [Julia Lawall <julia@diku.dk>]
3788 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3789 needlessly dereferenced structures, used obsolete functions and
3790 didn't handle all updated verify codes correctly.
3793 *) Disable MD2 in the default configuration.
3796 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3797 indicate the initial BIO being pushed or popped. This makes it possible
3798 to determine whether the BIO is the one explicitly called or as a result
3799 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3800 it handles reference counts correctly and doesn't zero out the I/O bio
3801 when it is not being explicitly popped. WARNING: applications which
3802 included workarounds for the old buggy behaviour will need to be modified
3803 or they could free up already freed BIOs.
3806 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3807 renaming to all platforms (within the 0.9.8 branch, this was
3808 done conditionally on Netware platforms to avoid a name clash).
3809 [Guenter <lists@gknw.net>]
3811 *) Add ECDHE and PSK support to DTLS.
3812 [Michael Tuexen <tuexen@fh-muenster.de>]
3814 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3818 *) Add "missing" function EVP_MD_flags() (without this the only way to
3819 retrieve a digest flags is by accessing the structure directly. Update
3820 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3821 or cipher is registered as in the "from" argument. Print out all
3822 registered digests in the dgst usage message instead of manually
3823 attempting to work them out.
3826 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3827 this allows the use of compression and extensions. Change default cipher
3828 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3829 by default unless an application cipher string requests it.
3832 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3833 key ids to find matching certificates and keys but some PKCS#12 files
3834 don't follow the (somewhat unwritten) rules and this strategy fails.
3835 Now just gather all certificates together and the first private key
3836 then look for the first certificate that matches the key.
3839 *) Support use of registered digest and cipher names for dgst and cipher
3840 commands instead of having to add each one as a special case. So now
3847 openssl dgst -sha256 foo
3849 and this works for ENGINE based algorithms too.
3853 *) Update Gost ENGINE to support parameter files.
3854 [Victor B. Wagner <vitus@cryptocom.ru>]
3856 *) Support GeneralizedTime in ca utility.
3857 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3859 *) Enhance the hash format used for certificate directory links. The new
3860 form uses the canonical encoding (meaning equivalent names will work
3861 even if they aren't identical) and uses SHA1 instead of MD5. This form
3862 is incompatible with the older format and as a result c_rehash should
3863 be used to rebuild symbolic links.
3866 *) Make PKCS#8 the default write format for private keys, replacing the
3867 traditional format. This form is standardised, more secure and doesn't
3868 include an implicit MD5 dependency.
3871 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3872 committed to OpenSSL should pass this lot as a minimum.
3875 *) Add session ticket override functionality for use by EAP-FAST.
3876 [Jouni Malinen <j@w1.fi>]
3878 *) Modify HMAC functions to return a value. Since these can be implemented
3879 in an ENGINE errors can occur.
3882 *) Type-checked OBJ_bsearch_ex.
3885 *) Type-checked OBJ_bsearch. Also some constification necessitated
3886 by type-checking. Still to come: TXT_DB, bsearch(?),
3887 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3891 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3892 seconds to a tm structure directly, instead of going through OS
3893 specific date routines. This avoids any issues with OS routines such
3894 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3895 and X509_time_adj_ex() to cover the extended range. The existing
3896 X509_time_adj() is still usable and will no longer have any date issues.
3899 *) Delta CRL support. New use deltas option which will attempt to locate
3900 and search any appropriate delta CRLs available.
3902 This work was sponsored by Google.
3905 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3906 code and add additional score elements. Validate alternate CRL paths
3907 as part of the CRL checking and indicate a new error "CRL path validation
3908 error" in this case. Applications wanting additional details can use
3909 the verify callback and check the new "parent" field. If this is not
3910 NULL CRL path validation is taking place. Existing applications won't
3911 see this because it requires extended CRL support which is off by
3914 This work was sponsored by Google.
3917 *) Support for freshest CRL extension.
3919 This work was sponsored by Google.
3922 *) Initial indirect CRL support. Currently only supported in the CRLs
3923 passed directly and not via lookup. Process certificate issuer
3924 CRL entry extension and lookup CRL entries by bother issuer name
3925 and serial number. Check and process CRL issuer entry in IDP extension.
3927 This work was sponsored by Google.
3930 *) Add support for distinct certificate and CRL paths. The CRL issuer
3931 certificate is validated separately in this case. Only enabled if
3932 an extended CRL support flag is set: this flag will enable additional
3933 CRL functionality in future.
3935 This work was sponsored by Google.
3938 *) Add support for policy mappings extension.
3940 This work was sponsored by Google.
3943 *) Fixes to pathlength constraint, self issued certificate handling,
3944 policy processing to align with RFC3280 and PKITS tests.
3946 This work was sponsored by Google.
3949 *) Support for name constraints certificate extension. DN, email, DNS
3950 and URI types are currently supported.
3952 This work was sponsored by Google.
3955 *) To cater for systems that provide a pointer-based thread ID rather
3956 than numeric, deprecate the current numeric thread ID mechanism and
3957 replace it with a structure and associated callback type. This
3958 mechanism allows a numeric "hash" to be extracted from a thread ID in
3959 either case, and on platforms where pointers are larger than 'long',
3960 mixing is done to help ensure the numeric 'hash' is usable even if it
3961 can't be guaranteed unique. The default mechanism is to use "&errno"
3962 as a pointer-based thread ID to distinguish between threads.
3964 Applications that want to provide their own thread IDs should now use
3965 CRYPTO_THREADID_set_callback() to register a callback that will call
3966 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3968 Note that ERR_remove_state() is now deprecated, because it is tied
3969 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3970 to free the current thread's error state should be replaced by
3971 ERR_remove_thread_state(NULL).
3973 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3974 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3975 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3976 application was previously providing a numeric thread callback that
3977 was inappropriate for distinguishing threads, then uniqueness might
3978 have been obtained with &errno that happened immediately in the
3979 intermediate development versions of OpenSSL; this is no longer the
3980 case, the numeric thread callback will now override the automatic use
3982 [Geoff Thorpe, with help from Bodo Moeller]
3984 *) Initial support for different CRL issuing certificates. This covers a
3985 simple case where the self issued certificates in the chain exist and
3986 the real CRL issuer is higher in the existing chain.
3988 This work was sponsored by Google.
3991 *) Removed effectively defunct crypto/store from the build.
3994 *) Revamp of STACK to provide stronger type-checking. Still to come:
3995 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3996 ASN1_STRING, CONF_VALUE.
3999 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4000 RAM on SSL connections. This option can save about 34k per idle SSL.
4003 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4004 STACK, TXT_DB, bsearch, qsort.
4007 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4008 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4009 support for data, signedData, compressedData, digestedData and
4010 encryptedData, envelopedData types included. Scripts to check against
4011 RFC4134 examples draft and interop and consistency checks of many
4012 content types and variants.
4015 *) Add options to enc utility to support use of zlib compression BIO.
4018 *) Extend mk1mf to support importing of options and assembly language
4019 files from Configure script, currently only included in VC-WIN32.
4020 The assembly language rules can now optionally generate the source
4021 files from the associated perl scripts.
4024 *) Implement remaining functionality needed to support GOST ciphersuites.
4025 Interop testing has been performed using CryptoPro implementations.
4026 [Victor B. Wagner <vitus@cryptocom.ru>]
4028 *) s390x assembler pack.
4031 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4035 *) Implement Opaque PRF Input TLS extension as specified in
4036 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4037 official specification yet and no extension type assignment by
4038 IANA exists, this extension (for now) will have to be explicitly
4039 enabled when building OpenSSL by providing the extension number
4040 to use. For example, specify an option
4042 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4044 to the "config" or "Configure" script to enable the extension,
4045 assuming extension number 0x9527 (which is a completely arbitrary
4046 and unofficial assignment based on the MD5 hash of the Internet
4047 Draft). Note that by doing so, you potentially lose
4048 interoperability with other TLS implementations since these might
4049 be using the same extension number for other purposes.
4051 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4052 opaque PRF input value to use in the handshake. This will create
4053 an internal copy of the length-'len' string at 'src', and will
4054 return non-zero for success.
4056 To get more control and flexibility, provide a callback function
4059 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4060 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4064 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4067 Callback function 'cb' will be called in handshakes, and is
4068 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4069 Argument 'arg' is for application purposes (the value as given to
4070 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4071 be provided to the callback function). The callback function
4072 has to return non-zero to report success: usually 1 to use opaque
4073 PRF input just if possible, or 2 to enforce use of the opaque PRF
4074 input. In the latter case, the library will abort the handshake
4075 if opaque PRF input is not successfully negotiated.
4077 Arguments 'peerinput' and 'len' given to the callback function
4078 will always be NULL and 0 in the case of a client. A server will
4079 see the client's opaque PRF input through these variables if
4080 available (NULL and 0 otherwise). Note that if the server
4081 provides an opaque PRF input, the length must be the same as the
4082 length of the client's opaque PRF input.
4084 Note that the callback function will only be called when creating
4085 a new session (session resumption can resume whatever was
4086 previously negotiated), and will not be called in SSL 2.0
4087 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4088 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4089 for applications that need to enforce opaque PRF input.
4093 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4096 [Victor B. Wagner <vitus@cryptocom.ru>]
4098 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4099 RFC4507bis. The encrypted ticket format is an encrypted encoded
4100 SSL_SESSION structure, that way new session features are automatically
4103 If a client application caches session in an SSL_SESSION structure
4104 support is transparent because tickets are now stored in the encoded
4107 The SSL_CTX structure automatically generates keys for ticket
4108 protection in servers so again support should be possible
4109 with no application modification.
4111 If a client or server wishes to disable RFC4507 support then the option
4112 SSL_OP_NO_TICKET can be set.
4114 Add a TLS extension debugging callback to allow the contents of any client
4115 or server extensions to be examined.
4117 This work was sponsored by Google.
4120 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4121 OpenSSL should now compile cleanly on gcc 4.2
4122 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4124 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4125 support including streaming MAC support: this is required for GOST
4126 ciphersuite support.
4127 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4129 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4130 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4131 to output in BER and PEM format.
4134 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4135 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4136 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4137 ENGINE support for HMAC keys which are unextractable. New -mac and
4138 -macopt options to dgst utility.
4141 *) New option -sigopt to dgst utility. Update dgst to use
4142 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4143 alternative signing parameters such as X9.31 or PSS in the dgst
4147 *) Change ssl_cipher_apply_rule(), the internal function that does
4148 the work each time a ciphersuite string requests enabling
4149 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4150 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4151 the order of disabled ciphersuites such that those ciphersuites
4152 that most recently went from enabled to disabled not only stay
4153 in order with respect to each other, but also have higher priority
4154 than other disabled ciphersuites the next time ciphersuites are
4157 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4158 the same ciphersuites as with "HIGH" alone, but in a specific
4159 order where the PSK ciphersuites come first (since they are the
4160 most recently disabled ciphersuites when "HIGH" is parsed).
4162 Also, change ssl_create_cipher_list() (using this new
4163 functionality) such that between otherwise identical
4164 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4168 *) Change ssl_create_cipher_list() so that it automatically
4169 arranges the ciphersuites in reasonable order before starting
4170 to process the rule string. Thus, the definition for "DEFAULT"
4171 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4172 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4173 This makes it much easier to arrive at a reasonable default order
4174 in applications for which anonymous ciphers are OK (meaning
4175 that you can't actually use DEFAULT).
4176 [Bodo Moeller; suggested by Victor Duchovni]
4178 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4179 processing) into multiple integers instead of setting
4180 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4181 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4182 (These masks as well as the individual bit definitions are hidden
4183 away into the non-exported interface ssl/ssl_locl.h, so this
4184 change to the definition of the SSL_CIPHER structure shouldn't
4185 affect applications.) This give us more bits for each of these
4186 categories, so there is no longer a need to coagulate AES128 and
4187 AES256 into a single algorithm bit, and to coagulate Camellia128
4188 and Camellia256 into a single algorithm bit, which has led to all
4191 Thus, among other things, the kludge introduced in 0.9.7m and
4192 0.9.8e for masking out AES256 independently of AES128 or masking
4193 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4195 With the change, we also introduce new ciphersuite aliases that
4196 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4200 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4201 Use the leftmost N bytes of the signature input if the input is
4202 larger than the prime q (with N being the size in bytes of q).
4205 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4206 it yet and it is largely untested.
4209 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4212 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4213 some compilers (gcc 4.2 and later) reject their use. Safestack is
4214 reimplemented. Update ASN1 to avoid use of legacy functions.
4217 *) Win32/64 targets are linked with Winsock2.
4220 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4221 to external functions. This can be used to increase CRL handling
4222 efficiency especially when CRLs are very large by (for example) storing
4223 the CRL revoked certificates in a database.
4226 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4227 new CRLs added to a directory can be used. New command line option
4228 -verify_return_error to s_client and s_server. This causes real errors
4229 to be returned by the verify callback instead of carrying on no matter
4230 what. This reflects the way a "real world" verify callback would behave.
4233 *) GOST engine, supporting several GOST algorithms and public key formats.
4234 Kindly donated by Cryptocom.
4237 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4238 partitioned by DP are handled but no indirect CRL or reason partitioning
4239 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4240 selected via a scoring technique which handles IDP and AKID in CRLs.
4243 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4244 will ultimately be used for all verify operations: this will remove the
4245 X509_STORE dependency on certificate verification and allow alternative
4246 lookup methods. X509_STORE based implementations of these two callbacks.
4249 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4250 Modify get_crl() to find a valid (unexpired) CRL if possible.
4253 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4254 this would be called X509_CRL_cmp() but that name is already used by
4255 a function that just compares CRL issuer names. Cache several CRL
4256 extensions in X509_CRL structure and cache CRLDP in X509.
4259 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4260 this maps equivalent X509_NAME structures into a consistent structure.
4261 Name comparison can then be performed rapidly using memcmp().
4264 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4268 *) Allow digests to supply their own micalg string for S/MIME type using
4269 the ctrl EVP_MD_CTRL_MICALG.
4272 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4273 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4274 ctrl. It can then customise the structure before and/or after signing
4278 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4279 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4280 to free up any added signature OIDs.
4283 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4284 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4285 digest and cipher tables. New options added to openssl utility:
4286 list-message-digest-algorithms and list-cipher-algorithms.
4289 *) Change the array representation of binary polynomials: the list
4290 of degrees of non-zero coefficients is now terminated with -1.
4291 Previously it was terminated with 0, which was also part of the
4292 value; thus, the array representation was not applicable to
4293 polynomials where t^0 has coefficient zero. This change makes
4294 the array representation useful in a more general context.
4297 *) Various modifications and fixes to SSL/TLS cipher string
4298 handling. For ECC, the code now distinguishes between fixed ECDH
4299 with RSA certificates on the one hand and with ECDSA certificates
4300 on the other hand, since these are separate ciphersuites. The
4301 unused code for Fortezza ciphersuites has been removed.
4303 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4304 (not "ECDHE"). For consistency with the code for DH
4305 certificates, use of ECDH certificates is now considered ECDH
4306 authentication, not RSA or ECDSA authentication (the latter is
4307 merely the CA's signing algorithm and not actively used in the
4310 The temporary ciphersuite alias "ECCdraft" is no longer
4311 available, and ECC ciphersuites are no longer excluded from "ALL"
4312 and "DEFAULT". The following aliases now exist for RFC 4492
4313 ciphersuites, most of these by analogy with the DH case:
4315 kECDHr - ECDH cert, signed with RSA
4316 kECDHe - ECDH cert, signed with ECDSA
4317 kECDH - ECDH cert (signed with either RSA or ECDSA)
4318 kEECDH - ephemeral ECDH
4319 ECDH - ECDH cert or ephemeral ECDH
4325 AECDH - anonymous ECDH
4326 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4330 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4331 Use correct micalg parameters depending on digest(s) in signed message.
4334 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4335 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4338 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4339 an engine to register a method. Add ENGINE lookups for methods and
4340 functional reference processing.
4343 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4344 EVP_{Sign,Verify}* which allow an application to customise the signature
4348 *) New -resign option to smime utility. This adds one or more signers
4349 to an existing PKCS#7 signedData structure. Also -md option to use an
4350 alternative message digest algorithm for signing.
4353 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4354 create PKCS7 structures containing multiple signers. Update smime
4355 application to support multiple signers.
4358 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4362 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4363 Reorganize PBE internals to lookup from a static table using NIDs,
4364 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4365 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4366 PRF which will be automatically used with PBES2.
4369 *) Replace the algorithm specific calls to generate keys in "req" with the
4373 *) Update PKCS#7 enveloped data routines to use new API. This is now
4374 supported by any public key method supporting the encrypt operation. A
4375 ctrl is added to allow the public key algorithm to examine or modify
4376 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4380 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4381 a default digest type to use. In most cases this will be SHA1 but some
4382 algorithms (such as GOST) need to specify an alternative digest. The
4383 return value indicates how strong the preference is 1 means optional and
4384 2 is mandatory (that is it is the only supported type). Modify
4385 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4386 use the default md. Update openssl utilities to use the default digest
4387 type for signing if it is not explicitly indicated.
4390 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4391 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4392 signing method from the key type. This effectively removes the link
4393 between digests and public key types.
4396 *) Add an OID cross reference table and utility functions. Its purpose is to
4397 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4398 rsaEncryption. This will allow some of the algorithm specific hackery
4399 needed to use the correct OID to be removed.
4402 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4403 structures for PKCS7_sign(). They are now set up by the relevant public
4407 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4410 *) Add support for key derivation (agreement) in the API, DH method and
4414 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4415 public and private key formats. As a side effect these add additional
4416 command line functionality not previously available: DSA signatures can be
4417 generated and verified using pkeyutl and DH key support and generation in
4422 [Oliver Tappe <zooey@hirschkaefer.de>]
4424 *) New make target "install_html_docs" installs HTML renditions of the
4426 [Oliver Tappe <zooey@hirschkaefer.de>]
4428 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4429 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4430 support key and parameter generation and add initial key generation
4431 functionality for RSA.
4434 *) Add functions for main EVP_PKEY_method operations. The undocumented
4435 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4436 EVP_PKEY_{encrypt,decrypt}_old.
4439 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4440 key API, doesn't do much yet.
4443 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4444 public key algorithms. New option to openssl utility:
4445 "list-public-key-algorithms" to print out info.
4448 *) Implement the Supported Elliptic Curves Extension for
4449 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4452 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4453 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4456 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4457 utilities such as rsa, dsa, dsaparam etc except they process any key
4461 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4462 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4463 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4467 *) Initial support for pluggable public key ASN1.
4468 De-spaghettify the public key ASN1 handling. Move public and private
4469 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4470 algorithm specific handling to a single module within the relevant
4471 algorithm directory. Add functions to allow (near) opaque processing
4472 of public and private key structures.
4475 *) Implement the Supported Point Formats Extension for
4476 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4479 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4480 for the psk identity [hint] and the psk callback functions to the
4481 SSL_SESSION, SSL and SSL_CTX structure.
4484 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4488 SSL_CTX_use_psk_identity_hint
4489 SSL_get_psk_identity_hint
4490 SSL_get_psk_identity
4491 SSL_use_psk_identity_hint
4493 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4495 *) Add RFC 3161 compliant time stamp request creation, response generation
4496 and response verification functionality.
4497 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4499 *) Add initial support for TLS extensions, specifically for the server_name
4500 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4501 have new members for a host name. The SSL data structure has an
4502 additional member SSL_CTX *initial_ctx so that new sessions can be
4503 stored in that context to allow for session resumption, even after the
4504 SSL has been switched to a new SSL_CTX in reaction to a client's
4505 server_name extension.
4507 New functions (subject to change):
4509 SSL_get_servername()
4510 SSL_get_servername_type()
4513 New CTRL codes and macros (subject to change):
4515 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4516 - SSL_CTX_set_tlsext_servername_callback()
4517 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4518 - SSL_CTX_set_tlsext_servername_arg()
4519 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4521 openssl s_client has a new '-servername ...' option.
4523 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4524 '-key2 ...', '-servername_fatal' (subject to change). This allows
4525 testing the HostName extension for a specific single host name ('-cert'
4526 and '-key' remain fallbacks for handshakes without HostName
4527 negotiation). If the unrecognized_name alert has to be sent, this by
4528 default is a warning; it becomes fatal with the '-servername_fatal'
4531 [Peter Sylvester, Remy Allais, Christophe Renou]
4533 *) Whirlpool hash implementation is added.
4536 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4537 bn(64,32). Because of instruction set limitations it doesn't have
4538 any negative impact on performance. This was done mostly in order
4539 to make it possible to share assembler modules, such as bn_mul_mont
4540 implementations, between 32- and 64-bit builds without hassle.
4543 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4544 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4548 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4549 dedicated Montgomery multiplication procedure, is introduced.
4550 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4551 "64-bit" performance on certain 32-bit targets.
4554 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4555 in SSL structures. New SSL ctrl to set maximum send fragment size.
4556 Save memory by setting the I/O buffer sizes dynamically instead of
4557 using the maximum available value.
4560 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4561 in addition to the text details.
4564 *) Very, very preliminary EXPERIMENTAL support for printing of general
4565 ASN1 structures. This currently produces rather ugly output and doesn't
4566 handle several customised structures at all.
4569 *) Integrated support for PVK file format and some related formats such
4570 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4571 these in the 'rsa' and 'dsa' utilities.
4574 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4577 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4578 place for the (very old) "NETSCAPE" format certificates which are now
4579 handled using new ASN1 code equivalents.
4582 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4583 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4584 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4587 *) Modify CRL distribution points extension code to print out previously
4588 unsupported fields. Enhance extension setting code to allow setting of
4592 *) Add print and set support for Issuing Distribution Point CRL extension.
4595 *) Change 'Configure' script to enable Camellia by default.
4598 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4600 *) When rejecting SSL/TLS records due to an incorrect version number, never
4601 update s->server with a new major version number. As of
4602 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4603 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4604 the previous behavior could result in a read attempt at NULL when
4605 receiving specific incorrect SSL/TLS records once record payload
4606 protection is active. (CVE-2010-0740)
4607 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4609 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4610 could be crashed if the relevant tables were not present (e.g. chrooted).
4611 [Tomas Hoger <thoger@redhat.com>]
4613 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4615 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4616 [Martin Olsson, Neel Mehta]
4618 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4619 accommodate for stack sorting, always a write lock!).
4622 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4623 excessive delays in the RAND_poll(): over a minute. As a workaround
4624 include a time check in the inner Heap32Next loop too.
4627 *) The code that handled flushing of data in SSL/TLS originally used the
4628 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4629 the problem outlined in PR#1949. The fix suggested there however can
4630 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4631 of Apache). So instead simplify the code to flush unconditionally.
4632 This should be fine since flushing with no data to flush is a no op.
4635 *) Handle TLS versions 2.0 and later properly and correctly use the
4636 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4637 off ancient servers have a habit of sticking around for a while...
4640 *) Modify compression code so it frees up structures without using the
4641 ex_data callbacks. This works around a problem where some applications
4642 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4643 restarting) then use compression (e.g. SSL with compression) later.
4644 This results in significant per-connection memory leaks and
4645 has caused some security issues including CVE-2008-1678 and
4649 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4650 change when encrypting or decrypting.
4653 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4654 connect and renegotiate with servers which do not support RI.
4655 Until RI is more widely deployed this option is enabled by default.
4658 *) Add "missing" ssl ctrls to clear options and mode.
4661 *) If client attempts to renegotiate and doesn't support RI respond with
4662 a no_renegotiation alert as required by RFC5746. Some renegotiating
4663 TLS clients will continue a connection gracefully when they receive
4664 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4665 waiting for a server hello which it will never receive. Now we treat a
4666 received no_renegotiation alert as a fatal error. This is because
4667 applications requesting a renegotiation might well expect it to succeed
4668 and would have no code in place to handle the server denying it so the
4669 only safe thing to do is to terminate the connection.
4672 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4673 peer supports secure renegotiation and 0 otherwise. Print out peer
4674 renegotiation support in s_client/s_server.
4677 *) Replace the highly broken and deprecated SPKAC certification method with
4678 the updated NID creation version. This should correctly handle UTF8.
4681 *) Implement RFC5746. Re-enable renegotiation but require the extension
4682 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4683 turns out to be a bad idea. It has been replaced by
4684 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4685 SSL_CTX_set_options(). This is really not recommended unless you
4686 know what you are doing.
4687 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4689 *) Fixes to stateless session resumption handling. Use initial_ctx when
4690 issuing and attempting to decrypt tickets in case it has changed during
4691 servername handling. Use a non-zero length session ID when attempting
4692 stateless session resumption: this makes it possible to determine if
4693 a resumption has occurred immediately after receiving server hello
4694 (several places in OpenSSL subtly assume this) instead of later in
4698 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4699 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4700 fixes for a few places where the return code is not checked
4702 [Julia Lawall <julia@diku.dk>]
4704 *) Add --strict-warnings option to Configure script to include devteam
4705 warnings in other configurations.
4708 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4709 makes it possible to install openssl libraries in locations which
4710 have names other than "lib", for example "/usr/lib64" which some
4712 [Steve Henson, based on patch from Jeremy Utley]
4714 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4715 X690 8.9.12 and can produce some misleading textual output of OIDs.
4716 [Steve Henson, reported by Dan Kaminsky]
4718 *) Delete MD2 from algorithm tables. This follows the recommendation in
4719 several standards that it is not used in new applications due to
4720 several cryptographic weaknesses. For binary compatibility reasons
4721 the MD2 API is still compiled in by default.
4724 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4728 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4729 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4731 [Guenter <lists@gknw.net>]
4733 *) Fix the server certificate chain building code to use X509_verify_cert(),
4734 it used to have an ad-hoc builder which was unable to cope with anything
4735 other than a simple chain.
4736 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4738 *) Don't check self signed certificate signatures in X509_verify_cert()
4739 by default (a flag can override this): it just wastes time without
4740 adding any security. As a useful side effect self signed root CAs
4741 with non-FIPS digests are now usable in FIPS mode.
4744 *) In dtls1_process_out_of_seq_message() the check if the current message
4745 is already buffered was missing. For every new message was memory
4746 allocated, allowing an attacker to perform an denial of service attack
4747 with sending out of seq handshake messages until there is no memory
4748 left. Additionally every future message was buffered, even if the
4749 sequence number made no sense and would be part of another handshake.
4750 So only messages with sequence numbers less than 10 in advance will be
4751 buffered. (CVE-2009-1378)
4752 [Robin Seggelmann, discovered by Daniel Mentz]
4754 *) Records are buffered if they arrive with a future epoch to be
4755 processed after finishing the corresponding handshake. There is
4756 currently no limitation to this buffer allowing an attacker to perform
4757 a DOS attack with sending records with future epochs until there is no
4758 memory left. This patch adds the pqueue_size() function to determine
4759 the size of a buffer and limits the record buffer to 100 entries.
4761 [Robin Seggelmann, discovered by Daniel Mentz]
4763 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4764 parent structure is freed. (CVE-2009-1379)
4767 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4768 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4771 [Ilya O. <vrghost@gmail.com>]
4773 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4775 *) Disable renegotiation completely - this fixes a severe security
4776 problem (CVE-2009-3555) at the cost of breaking all
4777 renegotiation. Renegotiation can be re-enabled by setting
4778 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4779 run-time. This is really not recommended unless you know what
4783 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4785 *) Don't set val to NULL when freeing up structures, it is freed up by
4786 underlying code. If sizeof(void *) > sizeof(long) this can result in
4787 zeroing past the valid field. (CVE-2009-0789)
4788 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4790 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4791 checked correctly. This would allow some invalid signed attributes to
4792 appear to verify correctly. (CVE-2009-0591)
4793 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4795 *) Reject UniversalString and BMPString types with invalid lengths. This
4796 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4797 a legal length. (CVE-2009-0590)
4800 *) Set S/MIME signing as the default purpose rather than setting it
4801 unconditionally. This allows applications to override it at the store
4805 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4806 to handle some structures.
4809 *) Improve efficiency of mem_gets: don't search whole buffer each time
4811 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4813 *) New -hex option for openssl rand.
4816 *) Print out UTF8String and NumericString when parsing ASN1.
4819 *) Support NumericString type for name components.
4822 *) Allow CC in the environment to override the automatically chosen
4823 compiler. Note that nothing is done to ensure flags work with the
4827 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4829 *) Properly check EVP_VerifyFinal() and similar return values
4831 [Ben Laurie, Bodo Moeller, Google Security Team]
4833 *) Enable TLS extensions by default.
4836 *) Allow the CHIL engine to be loaded, whether the application is
4837 multithreaded or not. (This does not release the developer from the
4838 obligation to set up the dynamic locking callbacks.)
4839 [Sander Temme <sander@temme.net>]
4841 *) Use correct exit code if there is an error in dgst command.
4842 [Steve Henson; problem pointed out by Roland Dirlewanger]
4844 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4845 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4848 *) Add experimental JPAKE support, including demo authentication in
4849 s_client and s_server.
4852 *) Set the comparison function in v3_addr_canonize().
4853 [Rob Austein <sra@hactrn.net>]
4855 *) Add support for XMPP STARTTLS in s_client.
4856 [Philip Paeps <philip@freebsd.org>]
4858 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4859 to ensure that even with this option, only ciphersuites in the
4860 server's preference list will be accepted. (Note that the option
4861 applies only when resuming a session, so the earlier behavior was
4862 just about the algorithm choice for symmetric cryptography.)
4865 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4867 *) Fix NULL pointer dereference if a DTLS server received
4868 ChangeCipherSpec as first record (CVE-2009-1386).
4871 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4872 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4875 *) The fix in 0.9.8c that supposedly got rid of unsafe
4876 double-checked locking was incomplete for RSA blinding,
4877 addressing just one layer of what turns out to have been
4878 doubly unsafe triple-checked locking.
4880 So now fix this for real by retiring the MONT_HELPER macro
4881 in crypto/rsa/rsa_eay.c.
4883 [Bodo Moeller; problem pointed out by Marius Schilder]
4885 *) Various precautionary measures:
4887 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4889 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4890 (NB: This would require knowledge of the secret session ticket key
4891 to exploit, in which case you'd be SOL either way.)
4893 - Change bn_nist.c so that it will properly handle input BIGNUMs
4894 outside the expected range.
4896 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4899 [Neel Mehta, Bodo Moeller]
4901 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4902 the load fails. Useful for distros.
4903 [Ben Laurie and the FreeBSD team]
4905 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4908 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4911 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4913 This work was sponsored by Logica.
4916 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4917 keystores. Support for SSL/TLS client authentication too.
4918 Not compiled unless enable-capieng specified to Configure.
4920 This work was sponsored by Logica.
4923 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4924 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4925 attribute creation routines such as certificate requests and PKCS#12
4929 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4931 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4932 handshake which could lead to a client crash as found using the
4933 Codenomicon TLS test suite (CVE-2008-1672)
4934 [Steve Henson, Mark Cox]
4936 *) Fix double free in TLS server name extensions which could lead to
4937 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4940 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4942 Clear the error queue to ensure that error entries left from
4943 older function calls do not interfere with the correct operation.
4944 [Lutz Jaenicke, Erik de Castro Lopo]
4946 *) Remove root CA certificates of commercial CAs:
4948 The OpenSSL project does not recommend any specific CA and does not
4949 have any policy with respect to including or excluding any CA.
4950 Therefore it does not make any sense to ship an arbitrary selection
4951 of root CA certificates with the OpenSSL software.
4954 *) RSA OAEP patches to fix two separate invalid memory reads.
4955 The first one involves inputs when 'lzero' is greater than
4956 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4957 before the beginning of from). The second one involves inputs where
4958 the 'db' section contains nothing but zeroes (there is a one-byte
4959 invalid read after the end of 'db').
4960 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4962 *) Partial backport from 0.9.9-dev:
4964 Introduce bn_mul_mont (dedicated Montgomery multiplication
4965 procedure) as a candidate for BIGNUM assembler implementation.
4966 While 0.9.9-dev uses assembler for various architectures, only
4967 x86_64 is available by default here in the 0.9.8 branch, and
4968 32-bit x86 is available through a compile-time setting.
4970 To try the 32-bit x86 assembler implementation, use Configure
4971 option "enable-montasm" (which exists only for this backport).
4973 As "enable-montasm" for 32-bit x86 disclaims code stability
4974 anyway, in this constellation we activate additional code
4975 backported from 0.9.9-dev for further performance improvements,
4976 namely BN_from_montgomery_word. (To enable this otherwise,
4977 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4979 [Andy Polyakov (backport partially by Bodo Moeller)]
4981 *) Add TLS session ticket callback. This allows an application to set
4982 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4983 values. This is useful for key rollover for example where several key
4984 sets may exist with different names.
4987 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4988 This was broken until now in 0.9.8 releases, such that the only way
4989 a registered ENGINE could be used (assuming it initialises
4990 successfully on the host) was to explicitly set it as the default
4991 for the relevant algorithms. This is in contradiction with 0.9.7
4992 behaviour and the documentation. With this fix, when an ENGINE is
4993 registered into a given algorithm's table of implementations, the
4994 'uptodate' flag is reset so that auto-discovery will be used next
4995 time a new context for that algorithm attempts to select an
4997 [Ian Lister (tweaked by Geoff Thorpe)]
4999 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5000 implementation in the following ways:
5002 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5005 Lack of BER streaming support means one pass streaming processing is
5006 only supported if data is detached: setting the streaming flag is
5007 ignored for embedded content.
5009 CMS support is disabled by default and must be explicitly enabled
5010 with the enable-cms configuration option.
5013 *) Update the GMP engine glue to do direct copies between BIGNUM and
5014 mpz_t when openssl and GMP use the same limb size. Otherwise the
5015 existing "conversion via a text string export" trick is still used.
5016 [Paul Sheer <paulsheer@gmail.com>]
5018 *) Zlib compression BIO. This is a filter BIO which compressed and
5019 uncompresses any data passed through it.
5022 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5023 RFC3394 compatible AES key wrapping.
5026 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5027 sets string data without copying. X509_ALGOR_set0() and
5028 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5029 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5030 from an X509_ATTRIBUTE structure optionally checking it occurs only
5031 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5035 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5036 to get the expected BN_FLG_CONSTTIME behavior.
5037 [Bodo Moeller (Google)]
5041 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5042 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5043 - added some more tests to do_tests.pl
5044 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5045 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5046 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5047 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5048 - various changes to netware.pl to enable gcc-cross builds on Win32
5050 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5051 - various changes to fix missing prototype warnings
5052 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5053 - added AES, WHIRLPOOL and CPUID assembler code to build files
5054 - added missing AES assembler make rules to mk1mf.pl
5055 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5056 [Guenter Knauf <eflash@gmx.net>]
5058 *) Implement certificate status request TLS extension defined in RFC3546.
5059 A client can set the appropriate parameters and receive the encoded
5060 OCSP response via a callback. A server can query the supplied parameters
5061 and set the encoded OCSP response in the callback. Add simplified examples
5062 to s_client and s_server.
5065 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5067 *) Fix various bugs:
5068 + Binary incompatibility of ssl_ctx_st structure
5069 + DTLS interoperation with non-compliant servers
5070 + Don't call get_session_cb() without proposed session
5071 + Fix ia64 assembler code
5072 [Andy Polyakov, Steve Henson]
5074 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5076 *) DTLS Handshake overhaul. There were longstanding issues with
5077 OpenSSL DTLS implementation, which were making it impossible for
5078 RFC 4347 compliant client to communicate with OpenSSL server.
5079 Unfortunately just fixing these incompatibilities would "cut off"
5080 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5081 server keeps tolerating non RFC compliant syntax. The opposite is
5082 not true, 0.9.8f client can not communicate with earlier server.
5083 This update even addresses CVE-2007-4995.
5086 *) Changes to avoid need for function casts in OpenSSL: some compilers
5087 (gcc 4.2 and later) reject their use.
5088 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5091 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5092 RFC4507bis. The encrypted ticket format is an encrypted encoded
5093 SSL_SESSION structure, that way new session features are automatically
5096 If a client application caches session in an SSL_SESSION structure
5097 support is transparent because tickets are now stored in the encoded
5100 The SSL_CTX structure automatically generates keys for ticket
5101 protection in servers so again support should be possible
5102 with no application modification.
5104 If a client or server wishes to disable RFC4507 support then the option
5105 SSL_OP_NO_TICKET can be set.
5107 Add a TLS extension debugging callback to allow the contents of any client
5108 or server extensions to be examined.
5110 This work was sponsored by Google.
5113 *) Add initial support for TLS extensions, specifically for the server_name
5114 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5115 have new members for a host name. The SSL data structure has an
5116 additional member SSL_CTX *initial_ctx so that new sessions can be
5117 stored in that context to allow for session resumption, even after the
5118 SSL has been switched to a new SSL_CTX in reaction to a client's
5119 server_name extension.
5121 New functions (subject to change):
5123 SSL_get_servername()
5124 SSL_get_servername_type()
5127 New CTRL codes and macros (subject to change):
5129 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5130 - SSL_CTX_set_tlsext_servername_callback()
5131 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5132 - SSL_CTX_set_tlsext_servername_arg()
5133 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5135 openssl s_client has a new '-servername ...' option.
5137 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5138 '-key2 ...', '-servername_fatal' (subject to change). This allows
5139 testing the HostName extension for a specific single host name ('-cert'
5140 and '-key' remain fallbacks for handshakes without HostName
5141 negotiation). If the unrecognized_name alert has to be sent, this by
5142 default is a warning; it becomes fatal with the '-servername_fatal'
5145 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5147 *) Add AES and SSE2 assembly language support to VC++ build.
5150 *) Mitigate attack on final subtraction in Montgomery reduction.
5153 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5154 (which previously caused an internal error).
5157 *) Squeeze another 10% out of IGE mode when in != out.
5160 *) AES IGE mode speedup.
5161 [Dean Gaudet (Google)]
5163 *) Add the Korean symmetric 128-bit cipher SEED (see
5164 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5165 add SEED ciphersuites from RFC 4162:
5167 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5168 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5169 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5170 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5172 To minimize changes between patchlevels in the OpenSSL 0.9.8
5173 series, SEED remains excluded from compilation unless OpenSSL
5174 is configured with 'enable-seed'.
5175 [KISA, Bodo Moeller]
5177 *) Mitigate branch prediction attacks, which can be practical if a
5178 single processor is shared, allowing a spy process to extract
5179 information. For detailed background information, see
5180 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5181 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5182 and Necessary Software Countermeasures"). The core of the change
5183 are new versions BN_div_no_branch() and
5184 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5185 respectively, which are slower, but avoid the security-relevant
5186 conditional branches. These are automatically called by BN_div()
5187 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5188 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5189 remove a conditional branch.
5191 BN_FLG_CONSTTIME is the new name for the previous
5192 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5193 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5194 in the exponent causes BN_mod_exp_mont() to use the alternative
5195 implementation in BN_mod_exp_mont_consttime().) The old name
5196 remains as a deprecated alias.
5198 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5199 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5200 constant-time implementations for more than just exponentiation.
5201 Here too the old name is kept as a deprecated alias.
5203 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5204 the BN_BLINDING structure gets an independent copy of the
5205 modulus. This means that the previous "BIGNUM *m" argument to
5206 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5207 essentially becomes "const BIGNUM *m", although we can't actually
5208 change this in the header file before 0.9.9. It allows
5209 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5210 enable BN_FLG_CONSTTIME.
5212 [Matthew D Wood (Intel Corp)]
5214 *) In the SSL/TLS server implementation, be strict about session ID
5215 context matching (which matters if an application uses a single
5216 external cache for different purposes). Previously,
5217 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5218 set. This did ensure strict client verification, but meant that,
5219 with applications using a single external cache for quite
5220 different requirements, clients could circumvent ciphersuite
5221 restrictions for a given session ID context by starting a session
5222 in a different context.
5225 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5226 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5227 authentication-only ciphersuites.
5230 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5231 not complete and could lead to a possible single byte overflow
5232 (CVE-2007-5135) [Ben Laurie]
5234 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5236 *) Since AES128 and AES256 (and similarly Camellia128 and
5237 Camellia256) share a single mask bit in the logic of
5238 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5239 kludge to work properly if AES128 is available and AES256 isn't
5240 (or if Camellia128 is available and Camellia256 isn't).
5243 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5244 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5245 When a point or a seed is encoded in a BIT STRING, we need to
5246 prevent the removal of trailing zero bits to get the proper DER
5247 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5248 of a NamedBitList, for which trailing 0 bits need to be removed.)
5251 *) Have SSL/TLS server implementation tolerate "mismatched" record
5252 protocol version while receiving ClientHello even if the
5253 ClientHello is fragmented. (The server can't insist on the
5254 particular protocol version it has chosen before the ServerHello
5255 message has informed the client about his choice.)
5258 *) Add RFC 3779 support.
5259 [Rob Austein for ARIN, Ben Laurie]
5261 *) Load error codes if they are not already present instead of using a
5262 static variable. This allows them to be cleanly unloaded and reloaded.
5263 Improve header file function name parsing.
5266 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5267 or CAPABILITY handshake as required by RFCs.
5270 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5272 *) Introduce limits to prevent malicious keys being able to
5273 cause a denial of service. (CVE-2006-2940)
5274 [Steve Henson, Bodo Moeller]
5276 *) Fix ASN.1 parsing of certain invalid structures that can result
5277 in a denial of service. (CVE-2006-2937) [Steve Henson]
5279 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5280 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5282 *) Fix SSL client code which could crash if connecting to a
5283 malicious SSLv2 server. (CVE-2006-4343)
5284 [Tavis Ormandy and Will Drewry, Google Security Team]
5286 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5287 match only those. Before that, "AES256-SHA" would be interpreted
5288 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5289 the same strength classification in 0.9.7h) as we currently only
5290 have a single AES bit in the ciphersuite description bitmap.
5291 That change, however, also applied to ciphersuite strings such as
5292 "RC4-MD5" that intentionally matched multiple ciphersuites --
5293 namely, SSL 2.0 ciphersuites in addition to the more common ones
5294 from SSL 3.0/TLS 1.0.
5296 So we change the selection algorithm again: Naming an explicit
5297 ciphersuite selects this one ciphersuite, and any other similar
5298 ciphersuite (same bitmap) from *other* protocol versions.
5299 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5300 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5302 Since SSL 2.0 does not have any ciphersuites for which the
5303 128/256 bit distinction would be relevant, this works for now.
5304 The proper fix will be to use different bits for AES128 and
5305 AES256, which would have avoided the problems from the beginning;
5306 however, bits are scarce, so we can only do this in a new release
5307 (not just a patchlevel) when we can change the SSL_CIPHER
5308 definition to split the single 'unsigned long mask' bitmap into
5309 multiple values to extend the available space.
5313 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5315 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5316 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5318 *) Add AES IGE and biIGE modes.
5321 *) Change the Unix randomness entropy gathering to use poll() when
5322 possible instead of select(), since the latter has some
5323 undesirable limitations.
5324 [Darryl Miles via Richard Levitte and Bodo Moeller]
5326 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5327 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5328 cannot be implicitly activated as part of, e.g., the "AES" alias.
5329 However, please upgrade to OpenSSL 0.9.9[-dev] for
5330 non-experimental use of the ECC ciphersuites to get TLS extension
5331 support, which is required for curve and point format negotiation
5332 to avoid potential handshake problems.
5335 *) Disable rogue ciphersuites:
5337 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5338 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5339 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5341 The latter two were purportedly from
5342 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5345 Also deactivate the remaining ciphersuites from
5346 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5347 unofficial, and the ID has long expired.
5350 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5351 dual-core machines) and other potential thread-safety issues.
5354 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5355 versions), which is now available for royalty-free use
5356 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5357 Also, add Camellia TLS ciphersuites from RFC 4132.
5359 To minimize changes between patchlevels in the OpenSSL 0.9.8
5360 series, Camellia remains excluded from compilation unless OpenSSL
5361 is configured with 'enable-camellia'.
5364 *) Disable the padding bug check when compression is in use. The padding
5365 bug check assumes the first packet is of even length, this is not
5366 necessarily true if compression is enabled and can result in false
5367 positives causing handshake failure. The actual bug test is ancient
5368 code so it is hoped that implementations will either have fixed it by
5369 now or any which still have the bug do not support compression.
5372 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5374 *) When applying a cipher rule check to see if string match is an explicit
5375 cipher suite and only match that one cipher suite if it is.
5378 *) Link in manifests for VC++ if needed.
5379 [Austin Ziegler <halostatue@gmail.com>]
5381 *) Update support for ECC-based TLS ciphersuites according to
5382 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5383 TLS extensions, which are supported starting with the 0.9.9
5384 branch, not in the OpenSSL 0.9.8 branch).
5387 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5388 opaque EVP_CIPHER_CTX handling.
5391 *) Fixes and enhancements to zlib compression code. We now only use
5392 "zlib1.dll" and use the default __cdecl calling convention on Win32
5393 to conform with the standards mentioned here:
5394 http://www.zlib.net/DLL_FAQ.txt
5395 Static zlib linking now works on Windows and the new --with-zlib-include
5396 --with-zlib-lib options to Configure can be used to supply the location
5397 of the headers and library. Gracefully handle case where zlib library
5401 *) Several fixes and enhancements to the OID generation code. The old code
5402 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5403 handle numbers larger than ULONG_MAX, truncated printing and had a
5404 non standard OBJ_obj2txt() behaviour.
5407 *) Add support for building of engines under engine/ as shared libraries
5408 under VC++ build system.
5411 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5412 Hopefully, we will not see any false combination of paths any more.
5415 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5417 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5418 (part of SSL_OP_ALL). This option used to disable the
5419 countermeasure against man-in-the-middle protocol-version
5420 rollback in the SSL 2.0 server implementation, which is a bad
5421 idea. (CVE-2005-2969)
5423 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5424 for Information Security, National Institute of Advanced Industrial
5425 Science and Technology [AIST], Japan)]
5427 *) Add two function to clear and return the verify parameter flags.
5430 *) Keep cipherlists sorted in the source instead of sorting them at
5431 runtime, thus removing the need for a lock.
5434 *) Avoid some small subgroup attacks in Diffie-Hellman.
5435 [Nick Mathewson and Ben Laurie]
5437 *) Add functions for well-known primes.
5440 *) Extended Windows CE support.
5441 [Satoshi Nakamura and Andy Polyakov]
5443 *) Initialize SSL_METHOD structures at compile time instead of during
5444 runtime, thus removing the need for a lock.
5447 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5448 attempting to decrypt each encrypted key in turn. Add support to
5452 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5454 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5457 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5460 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5461 key into the same file any more.
5464 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5467 *) Add -utf8 command line and config file option to 'ca'.
5468 [Stefan <stf@udoma.org]
5470 *) Removed the macro des_crypt(), as it seems to conflict with some
5471 libraries. Use DES_crypt().
5474 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5475 involves renaming the source and generated shared-libs for
5476 both. The engines will accept the corrected or legacy ids
5477 ('ncipher' and '4758_cca' respectively) when binding. NB,
5478 this only applies when building 'shared'.
5479 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5481 *) Add attribute functions to EVP_PKEY structure. Modify
5482 PKCS12_create() to recognize a CSP name attribute and
5483 use it. Make -CSP option work again in pkcs12 utility.
5486 *) Add new functionality to the bn blinding code:
5487 - automatic re-creation of the BN_BLINDING parameters after
5488 a fixed number of uses (currently 32)
5489 - add new function for parameter creation
5490 - introduce flags to control the update behaviour of the
5491 BN_BLINDING parameters
5492 - hide BN_BLINDING structure
5493 Add a second BN_BLINDING slot to the RSA structure to improve
5494 performance when a single RSA object is shared among several
5498 *) Add support for DTLS.
5499 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5501 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5502 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5505 *) Remove buggy and incomplete DH cert support from
5506 ssl/ssl_rsa.c and ssl/s3_both.c
5509 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5510 the apps/openssl applications.
5513 *) Compile clean with "-Wall -Wmissing-prototypes
5514 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5515 DEBUG_SAFESTACK must also be set.
5518 *) Change ./Configure so that certain algorithms can be disabled by default.
5519 The new counterpiece to "no-xxx" is "enable-xxx".
5521 The patented RC5 and MDC2 algorithms will now be disabled unless
5522 "enable-rc5" and "enable-mdc2", respectively, are specified.
5524 (IDEA remains enabled despite being patented. This is because IDEA
5525 is frequently required for interoperability, and there is no license
5526 fee for non-commercial use. As before, "no-idea" can be used to
5527 avoid this algorithm.)
5531 *) Add processing of proxy certificates (see RFC 3820). This work was
5532 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5533 EGEE (Enabling Grids for E-science in Europe).
5536 *) RC4 performance overhaul on modern architectures/implementations, such
5537 as Intel P4, IA-64 and AMD64.
5540 *) New utility extract-section.pl. This can be used specify an alternative
5541 section number in a pod file instead of having to treat each file as
5542 a separate case in Makefile. This can be done by adding two lines to the
5545 =for comment openssl_section:XXX
5547 The blank line is mandatory.
5551 *) New arguments -certform, -keyform and -pass for s_client and s_server
5552 to allow alternative format key and certificate files and passphrase
5556 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5557 update associated structures and add various utility functions.
5559 Add new policy related verify parameters, include policy checking in
5560 standard verify code. Enhance 'smime' application with extra parameters
5561 to support policy checking and print out.
5564 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5565 Nehemiah processors. These extensions support AES encryption in hardware
5566 as well as RNG (though RNG support is currently disabled).
5567 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5569 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5572 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5573 [Andy Polyakov and a number of other people]
5575 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5576 implementation contributed by IBM.
5577 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5579 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5580 exponent rather than 'unsigned long'. There is a corresponding change to
5581 the new 'rsa_keygen' element of the RSA_METHOD structure.
5582 [Jelte Jansen, Geoff Thorpe]
5584 *) Functionality for creating the initial serial number file is now
5585 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5587 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5588 number file to 1, which is bound to cause problems. To avoid
5589 the problems while respecting compatibility between different 0.9.7
5590 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5591 CA.pl for serial number initialization. With the new release 0.9.8,
5592 we can fix the problem directly in the 'ca' utility.)
5595 *) Reduced header interdependencies by declaring more opaque objects in
5596 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5597 give fewer recursive includes, which could break lazy source code - so
5598 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5599 developers should define this symbol when building and using openssl to
5600 ensure they track the recommended behaviour, interfaces, [etc], but
5601 backwards-compatible behaviour prevails when this isn't defined.
5604 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5607 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5608 This will generate a random key of the appropriate length based on the
5609 cipher context. The EVP_CIPHER can provide its own random key generation
5610 routine to support keys of a specific form. This is used in the des and
5611 3des routines to generate a key of the correct parity. Update S/MIME
5612 code to use new functions and hence generate correct parity DES keys.
5613 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5614 valid (weak or incorrect parity).
5617 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5618 as looking them up. This is useful when the verified structure may contain
5619 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5620 present unless the new PKCS7_NO_CRL flag is asserted.
5623 *) Extend ASN1 oid configuration module. It now additionally accepts the
5626 shortName = some long name, 1.2.3.4
5629 *) Reimplemented the BN_CTX implementation. There is now no more static
5630 limitation on the number of variables it can handle nor the depth of the
5631 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5632 information can now expand as required, and rather than having a single
5633 static array of bignums, BN_CTX now uses a linked-list of such arrays
5634 allowing it to expand on demand whilst maintaining the usefulness of
5635 BN_CTX's "bundling".
5638 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5639 to allow all RSA operations to function using a single BN_CTX.
5642 *) Preliminary support for certificate policy evaluation and checking. This
5643 is initially intended to pass the tests outlined in "Conformance Testing
5644 of Relying Party Client Certificate Path Processing Logic" v1.07.
5647 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5648 remained unused and not that useful. A variety of other little bignum
5649 tweaks and fixes have also been made continuing on from the audit (see
5653 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5654 associated ASN1, EVP and SSL functions and old ASN1 macros.
5657 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5658 and this should never fail. So the return value from the use of
5659 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5660 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5663 *) BN_CTX_get() should return zero-valued bignums, providing the same
5664 initialised value as BN_new().
5665 [Geoff Thorpe, suggested by Ulf Möller]
5667 *) Support for inhibitAnyPolicy certificate extension.
5670 *) An audit of the BIGNUM code is underway, for which debugging code is
5671 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5672 is considered valid when processing BIGNUMs, and causes execution to
5673 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5674 further steps are taken to deliberately pollute unused data in BIGNUM
5675 structures to try and expose faulty code further on. For now, openssl will
5676 (in its default mode of operation) continue to tolerate the inconsistent
5677 forms that it has tolerated in the past, but authors and packagers should
5678 consider trying openssl and their own applications when compiled with
5679 these debugging symbols defined. It will help highlight potential bugs in
5680 their own code, and will improve the test coverage for OpenSSL itself. At
5681 some point, these tighter rules will become openssl's default to improve
5682 maintainability, though the assert()s and other overheads will remain only
5683 in debugging configurations. See bn.h for more details.
5684 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5686 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5687 that can only be obtained through BN_CTX_new() (which implicitly
5688 initialises it). The presence of this function only made it possible
5689 to overwrite an existing structure (and cause memory leaks).
5692 *) Because of the callback-based approach for implementing LHASH as a
5693 template type, lh_insert() adds opaque objects to hash-tables and
5694 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5695 to clean up those corresponding objects before destroying the hash table
5696 (and losing the object pointers). So some over-zealous constifications in
5697 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5698 objects as "const" and the lh_doall[_arg] callback wrappers are not
5699 prototyped to have "const" restrictions on the object pointers they are
5700 given (and so aren't required to cast them away any more).
5703 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5704 (speed) prefers to use its own implementation. The two implementations
5705 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5706 its object type properly exposed (MS_TM) instead of casting to/from "char
5707 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5708 aren't necessarily the greatest nomenclatures - but this is what was used
5709 internally to the implementation so I've used that for now.
5712 *) Ensure that deprecated functions do not get compiled when
5713 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5714 the self-tests were still using deprecated key-generation functions so
5715 these have been updated also.
5718 *) Reorganise PKCS#7 code to separate the digest location functionality
5719 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5720 New function PKCS7_set_digest() to set the digest type for PKCS#7
5721 digestedData type. Add additional code to correctly generate the
5722 digestedData type and add support for this type in PKCS7 initialization
5726 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5727 structure of type "other".
5730 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5731 sure the loop does correctly stop and breaking ("division by zero")
5732 modulus operations are not performed. The (pre-generated) prime
5733 table crypto/bn/bn_prime.h was already correct, but it could not be
5734 re-generated on some platforms because of the "division by zero"
5735 situation in the script.
5736 [Ralf S. Engelschall]
5738 *) Update support for ECC-based TLS ciphersuites according to
5739 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5740 SHA-1 now is only used for "small" curves (where the
5741 representation of a field element takes up to 24 bytes); for
5742 larger curves, the field element resulting from ECDH is directly
5743 used as premaster secret.
5744 [Douglas Stebila (Sun Microsystems Laboratories)]
5746 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5747 curve secp160r1 to the tests.
5748 [Douglas Stebila (Sun Microsystems Laboratories)]
5750 *) Add the possibility to load symbols globally with DSO.
5751 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5753 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5754 control of the error stack.
5757 *) Add support for STORE in ENGINE.
5760 *) Add the STORE type. The intention is to provide a common interface
5761 to certificate and key stores, be they simple file-based stores, or
5762 HSM-type store, or LDAP stores, or...
5763 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5766 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5767 pass a list of arguments to any function as well as provide a way
5768 for a function to pass data back to the caller.
5771 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5772 works like BUF_strdup() but can be used to duplicate a portion of
5773 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5777 *) Add the function sk_find_ex() which works like sk_find(), but will
5778 return an index to an element even if an exact match couldn't be
5779 found. The index is guaranteed to point at the element where the
5780 searched-for key would be inserted to preserve sorting order.
5783 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5784 takes an extra flags argument for optional functionality. Currently,
5785 the following flags are defined:
5787 OBJ_BSEARCH_VALUE_ON_NOMATCH
5788 This one gets OBJ_bsearch_ex() to return a pointer to the first
5789 element where the comparing function returns a negative or zero
5792 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5793 This one gets OBJ_bsearch_ex() to return a pointer to the first
5794 element where the comparing function returns zero. This is useful
5795 if there are more than one element where the comparing function
5799 *) Make it possible to create self-signed certificates with 'openssl ca'
5800 in such a way that the self-signed certificate becomes part of the
5801 CA database and uses the same mechanisms for serial number generation
5802 as all other certificate signing. The new flag '-selfsign' enables
5803 this functionality. Adapt CA.sh and CA.pl.in.
5806 *) Add functionality to check the public key of a certificate request
5807 against a given private. This is useful to check that a certificate
5808 request can be signed by that key (self-signing).
5811 *) Make it possible to have multiple active certificates with the same
5812 subject in the CA index file. This is done only if the keyword
5813 'unique_subject' is set to 'no' in the main CA section (default
5814 if 'CA_default') of the configuration file. The value is saved
5815 with the database itself in a separate index attribute file,
5816 named like the index file with '.attr' appended to the name.
5819 *) Generate multi-valued AVAs using '+' notation in config files for
5823 *) Support for nameConstraints certificate extension.
5826 *) Support for policyConstraints certificate extension.
5829 *) Support for policyMappings certificate extension.
5832 *) Make sure the default DSA_METHOD implementation only uses its
5833 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5834 and change its own handlers to be NULL so as to remove unnecessary
5835 indirection. This lets alternative implementations fallback to the
5836 default implementation more easily.
5839 *) Support for directoryName in GeneralName related extensions
5843 *) Make it possible to link applications using Makefile.shared.
5844 Make that possible even when linking against static libraries!
5847 *) Support for single pass processing for S/MIME signing. This now
5848 means that S/MIME signing can be done from a pipe, in addition
5849 cleartext signing (multipart/signed type) is effectively streaming
5850 and the signed data does not need to be all held in memory.
5852 This is done with a new flag PKCS7_STREAM. When this flag is set
5853 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5854 is done after the data is output (and digests calculated) in
5855 SMIME_write_PKCS7().
5858 *) Add full support for -rpath/-R, both in shared libraries and
5859 applications, at least on the platforms where it's known how
5863 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5864 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5865 will now compute a table of multiples of the generator that
5866 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5867 faster (notably in the case of a single point multiplication,
5868 scalar * generator).
5869 [Nils Larsch, Bodo Moeller]
5871 *) IPv6 support for certificate extensions. The various extensions
5872 which use the IP:a.b.c.d can now take IPv6 addresses using the
5873 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5877 *) Added an ENGINE that implements RSA by performing private key
5878 exponentiations with the GMP library. The conversions to and from
5879 GMP's mpz_t format aren't optimised nor are any montgomery forms
5880 cached, and on x86 it appears OpenSSL's own performance has caught up.
5881 However there are likely to be other architectures where GMP could
5882 provide a boost. This ENGINE is not built in by default, but it can be
5883 specified at Configure time and should be accompanied by the necessary
5884 linker additions, eg;
5885 ./config -DOPENSSL_USE_GMP -lgmp
5888 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5889 testing availability of engines with "-t" - the old behaviour is
5890 produced by increasing the feature's verbosity with "-tt".
5893 *) ECDSA routines: under certain error conditions uninitialized BN objects
5894 could be freed. Solution: make sure initialization is performed early
5895 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5899 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5900 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5901 software implementations. For DSA and DH, parameter generation can
5902 also be overridden by providing the appropriate method callbacks.
5905 *) Change the "progress" mechanism used in key-generation and
5906 primality testing to functions that take a new BN_GENCB pointer in
5907 place of callback/argument pairs. The new API functions have "_ex"
5908 postfixes and the older functions are reimplemented as wrappers for
5909 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5910 declarations of the old functions to help (graceful) attempts to
5911 migrate to the new functions. Also, the new key-generation API
5912 functions operate on a caller-supplied key-structure and return
5913 success/failure rather than returning a key or NULL - this is to
5914 help make "keygen" another member function of RSA_METHOD etc.
5916 Example for using the new callback interface:
5918 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5922 BN_GENCB_set(&my_cb, my_callback, my_arg);
5924 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5925 /* For the meaning of a, b in calls to my_callback(), see the
5926 * documentation of the function that calls the callback.
5927 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5928 * my_callback should return 1 if it wants BN_is_prime_ex()
5929 * to continue, or 0 to stop.
5934 *) Change the ZLIB compression method to be stateful, and make it
5935 available to TLS with the number defined in
5936 draft-ietf-tls-compression-04.txt.
5939 *) Add the ASN.1 structures and functions for CertificatePair, which
5940 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5942 CertificatePair ::= SEQUENCE {
5943 forward [0] Certificate OPTIONAL,
5944 reverse [1] Certificate OPTIONAL,
5945 -- at least one of the pair shall be present -- }
5947 Also implement the PEM functions to read and write certificate
5948 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5950 This needed to be defined, mostly for the sake of the LDAP
5951 attribute crossCertificatePair, but may prove useful elsewhere as
5955 *) Make it possible to inhibit symlinking of shared libraries in
5956 Makefile.shared, for Cygwin's sake.
5959 *) Extend the BIGNUM API by creating a function
5960 void BN_set_negative(BIGNUM *a, int neg);
5961 and a macro that behave like
5962 int BN_is_negative(const BIGNUM *a);
5964 to avoid the need to access 'a->neg' directly in applications.
5967 *) Implement fast modular reduction for pseudo-Mersenne primes
5968 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5969 EC_GROUP_new_curve_GFp() will now automatically use this
5971 [Nils Larsch <nla@trustcenter.de>]
5973 *) Add new lock type (CRYPTO_LOCK_BN).
5976 *) Change the ENGINE framework to automatically load engines
5977 dynamically from specific directories unless they could be
5978 found to already be built in or loaded. Move all the
5979 current engines except for the cryptodev one to a new
5981 The engines in engines/ are built as shared libraries if
5982 the "shared" options was given to ./Configure or ./config.
5983 Otherwise, they are inserted in libcrypto.a.
5984 /usr/local/ssl/engines is the default directory for dynamic
5985 engines, but that can be overridden at configure time through
5986 the usual use of --prefix and/or --openssldir, and at run
5987 time with the environment variable OPENSSL_ENGINES.
5988 [Geoff Thorpe and Richard Levitte]
5990 *) Add Makefile.shared, a helper makefile to build shared
5991 libraries. Adapt Makefile.org.
5994 *) Add version info to Win32 DLLs.
5995 [Peter 'Luna' Runestig" <peter@runestig.com>]
5997 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5998 can be added using this API to created arbitrary PKCS#12
5999 files while avoiding the low level API.
6001 New options to PKCS12_create(), key or cert can be NULL and
6002 will then be omitted from the output file. The encryption
6003 algorithm NIDs can be set to -1 for no encryption, the mac
6004 iteration count can be set to 0 to omit the mac.
6006 Enhance pkcs12 utility by making the -nokeys and -nocerts
6007 options work when creating a PKCS#12 file. New option -nomac
6008 to omit the mac, NONE can be set for an encryption algorithm.
6009 New code is modified to use the enhanced PKCS12_create()
6010 instead of the low level API.
6013 *) Extend ASN1 encoder to support indefinite length constructed
6014 encoding. This can output sequences tags and octet strings in
6015 this form. Modify pk7_asn1.c to support indefinite length
6016 encoding. This is experimental and needs additional code to
6017 be useful, such as an ASN1 bio and some enhanced streaming
6020 Extend template encode functionality so that tagging is passed
6021 down to the template encoder.
6024 *) Let 'openssl req' fail if an argument to '-newkey' is not
6025 recognized instead of using RSA as a default.
6028 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6029 As these are not official, they are not included in "ALL";
6030 the "ECCdraft" ciphersuite group alias can be used to select them.
6031 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6033 *) Add ECDH engine support.
6034 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6036 *) Add ECDH in new directory crypto/ecdh/.
6037 [Douglas Stebila (Sun Microsystems Laboratories)]
6039 *) Let BN_rand_range() abort with an error after 100 iterations
6040 without success (which indicates a broken PRNG).
6043 *) Change BN_mod_sqrt() so that it verifies that the input value
6044 is really the square of the return value. (Previously,
6045 BN_mod_sqrt would show GIGO behaviour.)
6048 *) Add named elliptic curves over binary fields from X9.62, SECG,
6049 and WAP/WTLS; add OIDs that were still missing.
6051 [Sheueling Chang Shantz and Douglas Stebila
6052 (Sun Microsystems Laboratories)]
6054 *) Extend the EC library for elliptic curves over binary fields
6055 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6058 EC_GF2m_simple_method
6062 EC_GROUP_new_curve_GF2m
6063 EC_GROUP_set_curve_GF2m
6064 EC_GROUP_get_curve_GF2m
6065 EC_POINT_set_affine_coordinates_GF2m
6066 EC_POINT_get_affine_coordinates_GF2m
6067 EC_POINT_set_compressed_coordinates_GF2m
6069 Point compression for binary fields is disabled by default for
6070 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6073 As binary polynomials are represented as BIGNUMs, various members
6074 of the EC_GROUP and EC_POINT data structures can be shared
6075 between the implementations for prime fields and binary fields;
6076 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6077 are essentially identical to their ..._GFp counterparts.
6078 (For simplicity, the '..._GFp' prefix has been dropped from
6079 various internal method names.)
6081 An internal 'field_div' method (similar to 'field_mul' and
6082 'field_sqr') has been added; this is used only for binary fields.
6084 [Sheueling Chang Shantz and Douglas Stebila
6085 (Sun Microsystems Laboratories)]
6087 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6088 through methods ('mul', 'precompute_mult').
6090 The generic implementations (now internally called 'ec_wNAF_mul'
6091 and 'ec_wNAF_precomputed_mult') remain the default if these
6092 methods are undefined.
6094 [Sheueling Chang Shantz and Douglas Stebila
6095 (Sun Microsystems Laboratories)]
6097 *) New function EC_GROUP_get_degree, which is defined through
6098 EC_METHOD. For curves over prime fields, this returns the bit
6099 length of the modulus.
6101 [Sheueling Chang Shantz and Douglas Stebila
6102 (Sun Microsystems Laboratories)]
6104 *) New functions EC_GROUP_dup, EC_POINT_dup.
6105 (These simply call ..._new and ..._copy).
6107 [Sheueling Chang Shantz and Douglas Stebila
6108 (Sun Microsystems Laboratories)]
6110 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6111 Polynomials are represented as BIGNUMs (where the sign bit is not
6112 used) in the following functions [macros]:
6115 BN_GF2m_sub [= BN_GF2m_add]
6116 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6117 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6118 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6120 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6121 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6122 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6123 BN_GF2m_cmp [= BN_ucmp]
6125 (Note that only the 'mod' functions are actually for fields GF(2^m).
6126 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6128 For some functions, an the irreducible polynomial defining a
6129 field can be given as an 'unsigned int[]' with strictly
6130 decreasing elements giving the indices of those bits that are set;
6131 i.e., p[] represents the polynomial
6132 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6134 p[0] > p[1] > ... > p[k] = 0.
6135 This applies to the following functions:
6140 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6141 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6143 BN_GF2m_mod_sqrt_arr
6144 BN_GF2m_mod_solve_quad_arr
6148 Conversion can be performed by the following functions:
6153 bntest.c has additional tests for binary polynomial arithmetic.
6155 Two implementations for BN_GF2m_mod_div() are available.
6156 The default algorithm simply uses BN_GF2m_mod_inv() and
6157 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6158 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6159 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6161 [Sheueling Chang Shantz and Douglas Stebila
6162 (Sun Microsystems Laboratories)]
6164 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6165 functionality is disabled at compile-time.
6166 [Douglas Stebila <douglas.stebila@sun.com>]
6168 *) Change default behaviour of 'openssl asn1parse' so that more
6169 information is visible when viewing, e.g., a certificate:
6171 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6172 mode the content of non-printable OCTET STRINGs is output in a
6173 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6174 avoid the appearance of a printable string.
6175 [Nils Larsch <nla@trustcenter.de>]
6177 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6179 EC_GROUP_set_asn1_flag()
6180 EC_GROUP_get_asn1_flag()
6181 EC_GROUP_set_point_conversion_form()
6182 EC_GROUP_get_point_conversion_form()
6183 These control ASN1 encoding details:
6184 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6185 has been set to OPENSSL_EC_NAMED_CURVE.
6186 - Points are encoded in uncompressed form by default; options for
6187 asn1_for are as for point2oct, namely
6188 POINT_CONVERSION_COMPRESSED
6189 POINT_CONVERSION_UNCOMPRESSED
6190 POINT_CONVERSION_HYBRID
6192 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6195 EC_GROUP_get0_seed()
6196 EC_GROUP_get_seed_len()
6197 This is used only for ASN1 purposes (so far).
6198 [Nils Larsch <nla@trustcenter.de>]
6200 *) Add 'field_type' member to EC_METHOD, which holds the NID
6201 of the appropriate field type OID. The new function
6202 EC_METHOD_get_field_type() returns this value.
6203 [Nils Larsch <nla@trustcenter.de>]
6208 EC_POINT_point2hex()
6209 EC_POINT_hex2point()
6210 providing useful interfaces to EC_POINT_point2oct() and
6211 EC_POINT_oct2point().
6212 [Nils Larsch <nla@trustcenter.de>]
6214 *) Change internals of the EC library so that the functions
6215 EC_GROUP_set_generator()
6216 EC_GROUP_get_generator()
6217 EC_GROUP_get_order()
6218 EC_GROUP_get_cofactor()
6219 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6220 to methods, which would lead to unnecessary code duplication when
6221 adding different types of curves.
6222 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6224 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6225 arithmetic, and such that modified wNAFs are generated
6226 (which avoid length expansion in many cases).
6229 *) Add a function EC_GROUP_check_discriminant() (defined via
6230 EC_METHOD) that verifies that the curve discriminant is non-zero.
6232 Add a function EC_GROUP_check() that makes some sanity tests
6233 on a EC_GROUP, its generator and order. This includes
6234 EC_GROUP_check_discriminant().
6235 [Nils Larsch <nla@trustcenter.de>]
6237 *) Add ECDSA in new directory crypto/ecdsa/.
6239 Add applications 'openssl ecparam' and 'openssl ecdsa'
6240 (these are based on 'openssl dsaparam' and 'openssl dsa').
6242 ECDSA support is also included in various other files across the
6243 library. Most notably,
6244 - 'openssl req' now has a '-newkey ecdsa:file' option;
6245 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6246 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6247 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6248 them suitable for ECDSA where domain parameters must be
6249 extracted before the specific public key;
6250 - ECDSA engine support has been added.
6251 [Nils Larsch <nla@trustcenter.de>]
6253 *) Include some named elliptic curves, and add OIDs from X9.62,
6254 SECG, and WAP/WTLS. Each curve can be obtained from the new
6256 EC_GROUP_new_by_curve_name(),
6257 and the list of available named curves can be obtained with
6258 EC_get_builtin_curves().
6259 Also add a 'curve_name' member to EC_GROUP objects, which can be
6261 EC_GROUP_set_curve_name()
6262 EC_GROUP_get_curve_name()
6263 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6265 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6266 was actually never needed) and in BN_mul(). The removal in BN_mul()
6267 required a small change in bn_mul_part_recursive() and the addition
6268 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6269 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6270 bn_sub_words() and bn_add_words() except they take arrays with
6274 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6276 *) Cleanse PEM buffers before freeing them since they may contain
6278 [Benjamin Bennett <ben@psc.edu>]
6280 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6281 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6282 authentication-only ciphersuites.
6285 *) Since AES128 and AES256 share a single mask bit in the logic of
6286 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6287 kludge to work properly if AES128 is available and AES256 isn't.
6290 *) Expand security boundary to match 1.1.1 module.
6293 *) Remove redundant features: hash file source, editing of test vectors
6294 modify fipsld to use external fips_premain.c signature.
6297 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6298 run algorithm test programs.
6301 *) Make algorithm test programs more tolerant of whitespace.
6304 *) Have SSL/TLS server implementation tolerate "mismatched" record
6305 protocol version while receiving ClientHello even if the
6306 ClientHello is fragmented. (The server can't insist on the
6307 particular protocol version it has chosen before the ServerHello
6308 message has informed the client about his choice.)
6311 *) Load error codes if they are not already present instead of using a
6312 static variable. This allows them to be cleanly unloaded and reloaded.
6315 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6317 *) Introduce limits to prevent malicious keys being able to
6318 cause a denial of service. (CVE-2006-2940)
6319 [Steve Henson, Bodo Moeller]
6321 *) Fix ASN.1 parsing of certain invalid structures that can result
6322 in a denial of service. (CVE-2006-2937) [Steve Henson]
6324 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6325 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6327 *) Fix SSL client code which could crash if connecting to a
6328 malicious SSLv2 server. (CVE-2006-4343)
6329 [Tavis Ormandy and Will Drewry, Google Security Team]
6331 *) Change ciphersuite string processing so that an explicit
6332 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6333 will no longer include "AES128-SHA"), and any other similar
6334 ciphersuite (same bitmap) from *other* protocol versions (so that
6335 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6336 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6337 changes from 0.9.8b and 0.9.8d.
6340 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6342 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6343 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6345 *) Change the Unix randomness entropy gathering to use poll() when
6346 possible instead of select(), since the latter has some
6347 undesirable limitations.
6348 [Darryl Miles via Richard Levitte and Bodo Moeller]
6350 *) Disable rogue ciphersuites:
6352 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6353 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6354 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6356 The latter two were purportedly from
6357 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6360 Also deactivate the remaining ciphersuites from
6361 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6362 unofficial, and the ID has long expired.
6365 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6366 dual-core machines) and other potential thread-safety issues.
6369 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6371 *) Adapt fipsld and the build system to link against the validated FIPS
6372 module in FIPS mode.
6375 *) Fixes for VC++ 2005 build under Windows.
6378 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6379 from a Windows bash shell such as MSYS. It is autodetected from the
6380 "config" script when run from a VC++ environment. Modify standard VC++
6381 build to use fipscanister.o from the GNU make build.
6384 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6386 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6387 The value now differs depending on if you build for FIPS or not.
6388 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6389 safely run with a non-FIPSed libcrypto, as it may crash because of
6390 the difference induced by this change.
6393 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6395 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6396 (part of SSL_OP_ALL). This option used to disable the
6397 countermeasure against man-in-the-middle protocol-version
6398 rollback in the SSL 2.0 server implementation, which is a bad
6399 idea. (CVE-2005-2969)
6401 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6402 for Information Security, National Institute of Advanced Industrial
6403 Science and Technology [AIST], Japan)]
6405 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6406 mainly for FIPS compliance and not fully integrated at this stage.
6409 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6410 the exponentiation using a fixed-length exponent. (Otherwise,
6411 the information leaked through timing could expose the secret key
6412 after many signatures; cf. Bleichenbacher's attack on DSA with
6416 *) Make a new fixed-window mod_exp implementation the default for
6417 RSA, DSA, and DH private-key operations so that the sequence of
6418 squares and multiplies and the memory access pattern are
6419 independent of the particular secret key. This will mitigate
6420 cache-timing and potential related attacks.
6422 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6423 and this is automatically used by BN_mod_exp_mont() if the new flag
6424 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6425 will use this BN flag for private exponents unless the flag
6426 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6427 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6429 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6431 *) Change the client implementation for SSLv23_method() and
6432 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6433 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6434 (Previously, the SSL 2.0 backwards compatible Client Hello
6435 message format would be used even with SSL_OP_NO_SSLv2.)
6438 *) Add support for smime-type MIME parameter in S/MIME messages which some
6442 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6443 a threadsafe manner. Modify rsa code to use new function and add calls
6444 to dsa and dh code (which had race conditions before).
6447 *) Include the fixed error library code in the C error file definitions
6448 instead of fixing them up at runtime. This keeps the error code
6449 structures constant.
6452 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6454 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6457 *) Fixes for newer kerberos headers. NB: the casts are needed because
6458 the 'length' field is signed on one version and unsigned on another
6459 with no (?) obvious way to tell the difference, without these VC++
6460 complains. Also the "definition" of FAR (blank) is no longer included
6461 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6462 some needed definitions.
6465 *) Undo Cygwin change.
6468 *) Added support for proxy certificates according to RFC 3820.
6469 Because they may be a security thread to unaware applications,
6470 they must be explicitly allowed in run-time. See
6471 docs/HOWTO/proxy_certificates.txt for further information.
6474 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6476 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6477 server and client random values. Previously
6478 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6479 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6481 This change has negligible security impact because:
6483 1. Server and client random values still have 24 bytes of pseudo random
6486 2. Server and client random values are sent in the clear in the initial
6489 3. The master secret is derived using the premaster secret (48 bytes in
6490 size for static RSA ciphersuites) as well as client server and random
6493 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6496 [Stephen Henson, reported by UK NISCC]
6498 *) Use Windows randomness collection on Cygwin.
6501 *) Fix hang in EGD/PRNGD query when communication socket is closed
6502 prematurely by EGD/PRNGD.
6503 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6505 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6508 *) Back-port of selected performance improvements from development
6509 branch, as well as improved support for PowerPC platforms.
6512 *) Add lots of checks for memory allocation failure, error codes to indicate
6513 failure and freeing up memory if a failure occurs.
6514 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6516 *) Add new -passin argument to dgst.
6519 *) Perform some character comparisons of different types in X509_NAME_cmp:
6520 this is needed for some certificates that re-encode DNs into UTF8Strings
6521 (in violation of RFC3280) and can't or won't issue name rollover
6525 *) Make an explicit check during certificate validation to see that
6526 the CA setting in each certificate on the chain is correct. As a
6527 side effect always do the following basic checks on extensions,
6528 not just when there's an associated purpose to the check:
6530 - if there is an unhandled critical extension (unless the user
6531 has chosen to ignore this fault)
6532 - if the path length has been exceeded (if one is set at all)
6533 - that certain extensions fit the associated purpose (if one has
6537 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6539 *) Avoid a race condition when CRLs are checked in a multi threaded
6540 environment. This would happen due to the reordering of the revoked
6541 entries during signature checking and serial number lookup. Now the
6542 encoding is cached and the serial number sort performed under a lock.
6543 Add new STACK function sk_is_sorted().
6546 *) Add Delta CRL to the extension code.
6549 *) Various fixes to s3_pkt.c so alerts are sent properly.
6550 [David Holmes <d.holmes@f5.com>]
6552 *) Reduce the chances of duplicate issuer name and serial numbers (in
6553 violation of RFC3280) using the OpenSSL certificate creation utilities.
6554 This is done by creating a random 64 bit value for the initial serial
6555 number when a serial number file is created or when a self signed
6556 certificate is created using 'openssl req -x509'. The initial serial
6557 number file is created using 'openssl x509 -next_serial' in CA.pl
6558 rather than being initialized to 1.
6561 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6563 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6564 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6565 [Joe Orton, Steve Henson]
6567 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6569 [Joe Orton, Steve Henson]
6571 *) Make it possible to have multiple active certificates with the same
6572 subject in the CA index file. This is done only if the keyword
6573 'unique_subject' is set to 'no' in the main CA section (default
6574 if 'CA_default') of the configuration file. The value is saved
6575 with the database itself in a separate index attribute file,
6576 named like the index file with '.attr' appended to the name.
6579 *) X509 verify fixes. Disable broken certificate workarounds when
6580 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6581 keyUsage extension present. Don't accept CRLs with unhandled critical
6582 extensions: since verify currently doesn't process CRL extensions this
6583 rejects a CRL with *any* critical extensions. Add new verify error codes
6587 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6588 A clarification of RFC2560 will require the use of OCTET STRINGs and
6589 some implementations cannot handle the current raw format. Since OpenSSL
6590 copies and compares OCSP nonces as opaque blobs without any attempt at
6591 parsing them this should not create any compatibility issues.
6594 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6595 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6596 this HMAC (and other) operations are several times slower than OpenSSL
6600 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6601 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6603 *) Use the correct content when signing type "other".
6606 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6608 *) Fix various bugs revealed by running the NISCC test suite:
6610 Stop out of bounds reads in the ASN1 code when presented with
6611 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6613 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6615 If verify callback ignores invalid public key errors don't try to check
6616 certificate signature with the NULL public key.
6620 *) New -ignore_err option in ocsp application to stop the server
6621 exiting on the first error in a request.
6624 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6625 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6629 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6630 extra data after the compression methods not only for TLS 1.0
6631 but also for SSL 3.0 (as required by the specification).
6632 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6634 *) Change X509_certificate_type() to mark the key as exported/exportable
6635 when it's 512 *bits* long, not 512 bytes.
6638 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6639 blocks during encryption.
6642 *) Various fixes to base64 BIO and non blocking I/O. On write
6643 flushes were not handled properly if the BIO retried. On read
6644 data was not being buffered properly and had various logic bugs.
6645 This also affects blocking I/O when the data being decoded is a
6649 *) Various S/MIME bugfixes and compatibility changes:
6650 output correct application/pkcs7 MIME type if
6651 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6652 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6653 of files as .eml work). Correctly handle very long lines in MIME
6657 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6659 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6660 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6661 a protocol version number mismatch like a decryption error
6662 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6665 *) Turn on RSA blinding by default in the default implementation
6666 to avoid a timing attack. Applications that don't want it can call
6667 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6668 They would be ill-advised to do so in most cases.
6669 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6671 *) Change RSA blinding code so that it works when the PRNG is not
6672 seeded (in this case, the secret RSA exponent is abused as
6673 an unpredictable seed -- if it is not unpredictable, there
6674 is no point in blinding anyway). Make RSA blinding thread-safe
6675 by remembering the creator's thread ID in rsa->blinding and
6676 having all other threads use local one-time blinding factors
6677 (this requires more computation than sharing rsa->blinding, but
6678 avoids excessive locking; and if an RSA object is not shared
6679 between threads, blinding will still be very fast).
6682 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6683 ENGINE as defaults for all supported algorithms irrespective of
6684 the 'flags' parameter. 'flags' is now honoured, so applications
6685 should make sure they are passing it correctly.
6688 *) Target "mingw" now allows native Windows code to be generated in
6689 the Cygwin environment as well as with the MinGW compiler.
6692 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6694 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6695 via timing by performing a MAC computation even if incorrect
6696 block cipher padding has been found. This is a countermeasure
6697 against active attacks where the attacker has to distinguish
6698 between bad padding and a MAC verification error. (CVE-2003-0078)
6700 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6701 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6702 Martin Vuagnoux (EPFL, Ilion)]
6704 *) Make the no-err option work as intended. The intention with no-err
6705 is not to have the whole error stack handling routines removed from
6706 libcrypto, it's only intended to remove all the function name and
6707 reason texts, thereby removing some of the footprint that may not
6708 be interesting if those errors aren't displayed anyway.
6710 NOTE: it's still possible for any application or module to have it's
6711 own set of error texts inserted. The routines are there, just not
6712 used by default when no-err is given.
6715 *) Add support for FreeBSD on IA64.
6716 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6718 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6719 Kerberos function mit_des_cbc_cksum(). Before this change,
6720 the value returned by DES_cbc_cksum() was like the one from
6721 mit_des_cbc_cksum(), except the bytes were swapped.
6722 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6724 *) Allow an application to disable the automatic SSL chain building.
6725 Before this a rather primitive chain build was always performed in
6726 ssl3_output_cert_chain(): an application had no way to send the
6727 correct chain if the automatic operation produced an incorrect result.
6729 Now the chain builder is disabled if either:
6731 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6733 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6735 The reasoning behind this is that an application would not want the
6736 auto chain building to take place if extra chain certificates are
6737 present and it might also want a means of sending no additional
6738 certificates (for example the chain has two certificates and the
6742 *) Add the possibility to build without the ENGINE framework.
6743 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6745 *) Under Win32 gmtime() can return NULL: check return value in
6746 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6749 *) DSA routines: under certain error conditions uninitialized BN objects
6750 could be freed. Solution: make sure initialization is performed early
6751 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6752 Nils Larsch <nla@trustcenter.de> via PR#459)
6755 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6756 checked on reconnect on the client side, therefore session resumption
6757 could still fail with a "ssl session id is different" error. This
6758 behaviour is masked when SSL_OP_ALL is used due to
6759 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6760 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6761 followup to PR #377.
6764 *) IA-32 assembler support enhancements: unified ELF targets, support
6765 for SCO/Caldera platforms, fix for Cygwin shared build.
6768 *) Add support for FreeBSD on sparc64. As a consequence, support for
6769 FreeBSD on non-x86 processors is separate from x86 processors on
6770 the config script, much like the NetBSD support.
6771 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6773 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6775 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6778 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6779 code (06) was taken as the first octet of the session ID and the last
6780 octet was ignored consequently. As a result SSLv2 client side session
6781 caching could not have worked due to the session ID mismatch between
6783 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6787 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6788 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6792 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6793 seems that in spite of existing for more than a year, many application
6794 author have done nothing to provide the necessary callbacks, which
6795 means that this particular engine will not work properly anywhere.
6796 This is a very unfortunate situation which forces us, in the name
6797 of usability, to give the hw_ncipher.c a static lock, which is part
6799 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6800 appear in 0.9.8 or later. We EXPECT application authors to have
6801 dealt properly with this when 0.9.8 is released (unless we actually
6802 make such changes in the libcrypto locking code that changes will
6803 have to be made anyway).
6806 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6807 octets have been read, EOF or an error occurs. Without this change
6808 some truncated ASN1 structures will not produce an error.
6811 *) Disable Heimdal support, since it hasn't been fully implemented.
6812 Still give the possibility to force the use of Heimdal, but with
6813 warnings and a request that patches get sent to openssl-dev.
6816 *) Add the VC-CE target, introduce the WINCE sysname, and add
6817 INSTALL.WCE and appropriate conditionals to make it build.
6818 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6820 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6821 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6822 edit numbers of the version.
6823 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6825 *) Introduce safe string copy and catenation functions
6826 (BUF_strlcpy() and BUF_strlcat()).
6827 [Ben Laurie (CHATS) and Richard Levitte]
6829 *) Avoid using fixed-size buffers for one-line DNs.
6830 [Ben Laurie (CHATS)]
6832 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6833 resizing buffers containing secrets, and use where appropriate.
6834 [Ben Laurie (CHATS)]
6836 *) Avoid using fixed size buffers for configuration file location.
6837 [Ben Laurie (CHATS)]
6839 *) Avoid filename truncation for various CA files.
6840 [Ben Laurie (CHATS)]
6842 *) Use sizeof in preference to magic numbers.
6843 [Ben Laurie (CHATS)]
6845 *) Avoid filename truncation in cert requests.
6846 [Ben Laurie (CHATS)]
6848 *) Add assertions to check for (supposedly impossible) buffer
6850 [Ben Laurie (CHATS)]
6852 *) Don't cache truncated DNS entries in the local cache (this could
6853 potentially lead to a spoofing attack).
6854 [Ben Laurie (CHATS)]
6856 *) Fix various buffers to be large enough for hex/decimal
6857 representations in a platform independent manner.
6858 [Ben Laurie (CHATS)]
6860 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6861 resizing buffers containing secrets, and use where appropriate.
6862 [Ben Laurie (CHATS)]
6864 *) Add BIO_indent() to avoid much slightly worrying code to do
6866 [Ben Laurie (CHATS)]
6868 *) Convert sprintf()/BIO_puts() to BIO_printf().
6869 [Ben Laurie (CHATS)]
6871 *) buffer_gets() could terminate with the buffer only half
6873 [Ben Laurie (CHATS)]
6875 *) Add assertions to prevent user-supplied crypto functions from
6876 overflowing internal buffers by having large block sizes, etc.
6877 [Ben Laurie (CHATS)]
6879 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6881 [Ben Laurie (CHATS)]
6883 *) Eliminate unused copy of key in RC4.
6884 [Ben Laurie (CHATS)]
6886 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6887 [Ben Laurie (CHATS)]
6889 *) Fix off-by-one error in EGD path.
6890 [Ben Laurie (CHATS)]
6892 *) If RANDFILE path is too long, ignore instead of truncating.
6893 [Ben Laurie (CHATS)]
6895 *) Eliminate unused and incorrectly sized X.509 structure
6897 [Ben Laurie (CHATS)]
6899 *) Eliminate unused and dangerous function knumber().
6900 [Ben Laurie (CHATS)]
6902 *) Eliminate unused and dangerous structure, KSSL_ERR.
6903 [Ben Laurie (CHATS)]
6905 *) Protect against overlong session ID context length in an encoded
6906 session object. Since these are local, this does not appear to be
6908 [Ben Laurie (CHATS)]
6910 *) Change from security patch (see 0.9.6e below) that did not affect
6911 the 0.9.6 release series:
6913 Remote buffer overflow in SSL3 protocol - an attacker could
6914 supply an oversized master key in Kerberos-enabled versions.
6916 [Ben Laurie (CHATS)]
6918 *) Change the SSL kerb5 codes to match RFC 2712.
6921 *) Make -nameopt work fully for req and add -reqopt switch.
6922 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6924 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6925 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6927 *) Make sure tests can be performed even if the corresponding algorithms
6928 have been removed entirely. This was also the last step to make
6929 OpenSSL compilable with DJGPP under all reasonable conditions.
6930 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6932 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6933 to allow version independent disabling of normally unselected ciphers,
6934 which may be activated as a side-effect of selecting a single cipher.
6936 (E.g., cipher list string "RSA" enables ciphersuites that are left
6937 out of "ALL" because they do not provide symmetric encryption.
6938 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6939 [Lutz Jaenicke, Bodo Moeller]
6941 *) Add appropriate support for separate platform-dependent build
6942 directories. The recommended way to make a platform-dependent
6943 build directory is the following (tested on Linux), maybe with
6946 # Place yourself outside of the OpenSSL source tree. In
6947 # this example, the environment variable OPENSSL_SOURCE
6948 # is assumed to contain the absolute OpenSSL source directory.
6949 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6950 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6951 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6952 mkdir -p `dirname $F`
6953 ln -s $OPENSSL_SOURCE/$F $F
6956 To be absolutely sure not to disturb the source tree, a "make clean"
6957 is a good thing. If it isn't successful, don't worry about it,
6958 it probably means the source directory is very clean.
6961 *) Make sure any ENGINE control commands make local copies of string
6962 pointers passed to them whenever necessary. Otherwise it is possible
6963 the caller may have overwritten (or deallocated) the original string
6964 data when a later ENGINE operation tries to use the stored values.
6965 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6967 *) Improve diagnostics in file reading and command-line digests.
6968 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6970 *) Add AES modes CFB and OFB to the object database. Correct an
6971 error in AES-CFB decryption.
6974 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6975 allows existing EVP_CIPHER_CTX structures to be reused after
6976 calling EVP_*Final(). This behaviour is used by encryption
6977 BIOs and some applications. This has the side effect that
6978 applications must explicitly clean up cipher contexts with
6979 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6982 *) Check the values of dna and dnb in bn_mul_recursive before calling
6983 bn_mul_comba (a non zero value means the a or b arrays do not contain
6984 n2 elements) and fallback to bn_mul_normal if either is not zero.
6987 *) Fix escaping of non-ASCII characters when using the -subj option
6988 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6991 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6992 form for "surname", serialNumber has no short form.
6993 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6994 therefore remove "mail" short name for "internet 7".
6995 The OID for unique identifiers in X509 certificates is
6996 x500UniqueIdentifier, not uniqueIdentifier.
6997 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7000 *) Add an "init" command to the ENGINE config module and auto initialize
7001 ENGINEs. Without any "init" command the ENGINE will be initialized
7002 after all ctrl commands have been executed on it. If init=1 the
7003 ENGINE is initialized at that point (ctrls before that point are run
7004 on the uninitialized ENGINE and after on the initialized one). If
7005 init=0 then the ENGINE will not be initialized at all.
7008 *) Fix the 'app_verify_callback' interface so that the user-defined
7009 argument is actually passed to the callback: In the
7010 SSL_CTX_set_cert_verify_callback() prototype, the callback
7011 declaration has been changed from
7014 int (*cb)(X509_STORE_CTX *,void *);
7015 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7016 i=s->ctx->app_verify_callback(&ctx)
7017 has been changed into
7018 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7020 To update applications using SSL_CTX_set_cert_verify_callback(),
7021 a dummy argument can be added to their callback functions.
7022 [D. K. Smetters <smetters@parc.xerox.com>]
7024 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7025 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7027 *) Add and OPENSSL_LOAD_CONF define which will cause
7028 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7029 This allows older applications to transparently support certain
7030 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7031 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7032 load the config file and OPENSSL_add_all_algorithms_conf() which will
7033 always load it have also been added.
7036 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7037 Adjust NIDs and EVP layer.
7038 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7040 *) Config modules support in openssl utility.
7042 Most commands now load modules from the config file,
7043 though in a few (such as version) this isn't done
7044 because it couldn't be used for anything.
7046 In the case of ca and req the config file used is
7047 the same as the utility itself: that is the -config
7048 command line option can be used to specify an
7052 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7053 use "openssl_conf" if filename is NULL use default openssl config file.
7056 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7057 config section name. Add a new flag to tolerate a missing config file
7058 and move code to CONF_modules_load_file().
7061 *) Support for crypto accelerator cards from Accelerated Encryption
7062 Processing, www.aep.ie. (Use engine 'aep')
7063 The support was copied from 0.9.6c [engine] and adapted/corrected
7064 to work with the new engine framework.
7065 [AEP Inc. and Richard Levitte]
7067 *) Support for SureWare crypto accelerator cards from Baltimore
7068 Technologies. (Use engine 'sureware')
7069 The support was copied from 0.9.6c [engine] and adapted
7070 to work with the new engine framework.
7073 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7074 make the newer ENGINE framework commands for the CHIL engine work.
7075 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7077 *) Make it possible to produce shared libraries on ReliantUNIX.
7078 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7080 *) Add the configuration target debug-linux-ppro.
7081 Make 'openssl rsa' use the general key loading routines
7082 implemented in apps.c, and make those routines able to
7083 handle the key format FORMAT_NETSCAPE and the variant
7085 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7087 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7088 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7090 *) Add -keyform to rsautl, and document -engine.
7091 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7093 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7094 BIO_R_NO_SUCH_FILE error code rather than the generic
7095 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7098 *) Add new functions
7100 ERR_peek_last_error_line
7101 ERR_peek_last_error_line_data.
7102 These are similar to
7105 ERR_peek_error_line_data,
7106 but report on the latest error recorded rather than the first one
7107 still in the error queue.
7108 [Ben Laurie, Bodo Moeller]
7110 *) default_algorithms option in ENGINE config module. This allows things
7112 default_algorithms = ALL
7113 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7116 *) Preliminary ENGINE config module.
7119 *) New experimental application configuration code.
7122 *) Change the AES code to follow the same name structure as all other
7123 symmetric ciphers, and behave the same way. Move everything to
7124 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7125 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7127 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7128 [Ben Laurie and Theo de Raadt]
7130 *) Add option to output public keys in req command.
7131 [Massimiliano Pala madwolf@openca.org]
7133 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7134 (up to about 10% better than before for P-192 and P-224).
7137 *) New functions/macros
7139 SSL_CTX_set_msg_callback(ctx, cb)
7140 SSL_CTX_set_msg_callback_arg(ctx, arg)
7141 SSL_set_msg_callback(ssl, cb)
7142 SSL_set_msg_callback_arg(ssl, arg)
7144 to request calling a callback function
7146 void cb(int write_p, int version, int content_type,
7147 const void *buf, size_t len, SSL *ssl, void *arg)
7149 whenever a protocol message has been completely received
7150 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7151 protocol version according to which the SSL library interprets
7152 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7153 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7154 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7155 specification (change_cipher_spec(20), alert(21), handshake(22)).
7156 'buf' and 'len' point to the actual message, 'ssl' to the
7157 SSL object, and 'arg' is the application-defined value set by
7158 SSL[_CTX]_set_msg_callback_arg().
7160 'openssl s_client' and 'openssl s_server' have new '-msg' options
7161 to enable a callback that displays all protocol messages.
7164 *) Change the shared library support so shared libraries are built as
7165 soon as the corresponding static library is finished, and thereby get
7166 openssl and the test programs linked against the shared library.
7167 This still only happens when the keyword "shard" has been given to
7168 the configuration scripts.
7170 NOTE: shared library support is still an experimental thing, and
7171 backward binary compatibility is still not guaranteed.
7172 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7174 *) Add support for Subject Information Access extension.
7175 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7177 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7178 additional bytes when new memory had to be allocated, not just
7179 when reusing an existing buffer.
7182 *) New command line and configuration option 'utf8' for the req command.
7183 This allows field values to be specified as UTF8 strings.
7186 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7187 runs for the former and machine-readable output for the latter.
7190 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7191 of the e-mail address in the DN (i.e., it will go into a certificate
7192 extension only). The new configuration file option 'email_in_dn = no'
7193 has the same effect.
7194 [Massimiliano Pala madwolf@openca.org]
7196 *) Change all functions with names starting with des_ to be starting
7197 with DES_ instead. Add wrappers that are compatible with libdes,
7198 but are named _ossl_old_des_*. Finally, add macros that map the
7199 des_* symbols to the corresponding _ossl_old_des_* if libdes
7200 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7201 desired, the des_* symbols will be mapped to DES_*, with one
7204 Since we provide two compatibility mappings, the user needs to
7205 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7206 compatibility is desired. The default (i.e., when that macro
7207 isn't defined) is OpenSSL 0.9.6c compatibility.
7209 There are also macros that enable and disable the support of old
7210 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7211 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7212 are defined, the default will apply: to support the old des routines.
7214 In either case, one must include openssl/des.h to get the correct
7215 definitions. Do not try to just include openssl/des_old.h, that
7218 NOTE: This is a major break of an old API into a new one. Software
7219 authors are encouraged to switch to the DES_ style functions. Some
7220 time in the future, des_old.h and the libdes compatibility functions
7221 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7222 default), and then completely removed.
7225 *) Test for certificates which contain unsupported critical extensions.
7226 If such a certificate is found during a verify operation it is
7227 rejected by default: this behaviour can be overridden by either
7228 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7229 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7230 X509_supported_extension() has also been added which returns 1 if a
7231 particular extension is supported.
7234 *) Modify the behaviour of EVP cipher functions in similar way to digests
7235 to retain compatibility with existing code.
7238 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7239 compatibility with existing code. In particular the 'ctx' parameter does
7240 not have to be to be initialized before the call to EVP_DigestInit() and
7241 it is tidied up after a call to EVP_DigestFinal(). New function
7242 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7243 EVP_MD_CTX_copy() changed to not require the destination to be
7244 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7245 requires the destination to be valid.
7247 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7248 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7251 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7252 so that complete 'Handshake' protocol structures are kept in memory
7253 instead of overwriting 'msg_type' and 'length' with 'body' data.
7256 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7257 [Massimo Santin via Richard Levitte]
7259 *) Major restructuring to the underlying ENGINE code. This includes
7260 reduction of linker bloat, separation of pure "ENGINE" manipulation
7261 (initialisation, etc) from functionality dealing with implementations
7262 of specific crypto interfaces. This change also introduces integrated
7263 support for symmetric ciphers and digest implementations - so ENGINEs
7264 can now accelerate these by providing EVP_CIPHER and EVP_MD
7265 implementations of their own. This is detailed in crypto/engine/README
7266 as it couldn't be adequately described here. However, there are a few
7267 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7268 were changed in the original introduction of ENGINE code have now
7269 reverted back - the hooking from this code to ENGINE is now a good
7270 deal more passive and at run-time, operations deal directly with
7271 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7272 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7273 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7274 they were not being used by the framework as there is no concept of a
7275 BIGNUM_METHOD and they could not be generalised to the new
7276 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7277 ENGINE_cpy() has been removed as it cannot be consistently defined in
7281 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7284 *) Change mkdef.pl to sort symbols that get the same entry number,
7285 and make sure the automatically generated functions ERR_load_*
7286 become part of libeay.num as well.
7289 *) New function SSL_renegotiate_pending(). This returns true once
7290 renegotiation has been requested (either SSL_renegotiate() call
7291 or HelloRequest/ClientHello received from the peer) and becomes
7292 false once a handshake has been completed.
7293 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7294 sends a HelloRequest, but does not ensure that a handshake takes
7295 place. SSL_renegotiate_pending() is useful for checking if the
7296 client has followed the request.)
7299 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7300 By default, clients may request session resumption even during
7301 renegotiation (if session ID contexts permit); with this option,
7302 session resumption is possible only in the first handshake.
7304 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7305 more bits available for options that should not be part of
7306 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7309 *) Add some demos for certificate and certificate request creation.
7312 *) Make maximum certificate chain size accepted from the peer application
7313 settable (SSL*_get/set_max_cert_list()), as proposed by
7314 "Douglas E. Engert" <deengert@anl.gov>.
7317 *) Add support for shared libraries for Unixware-7
7318 (Boyd Lynn Gerber <gerberb@zenez.com>).
7321 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7322 be done prior to destruction. Use this to unload error strings from
7323 ENGINEs that load their own error strings. NB: This adds two new API
7324 functions to "get" and "set" this destroy handler in an ENGINE.
7327 *) Alter all existing ENGINE implementations (except "openssl" and
7328 "openbsd") to dynamically instantiate their own error strings. This
7329 makes them more flexible to be built both as statically-linked ENGINEs
7330 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7331 Also, add stub code to each that makes building them as self-contained
7332 shared-libraries easier (see README.ENGINE).
7335 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7336 implementations into applications that are completely implemented in
7337 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7338 commands that can be used to configure what shared-library to load and
7339 to control aspects of the way it is handled. Also, made an update to
7340 the README.ENGINE file that brings its information up-to-date and
7341 provides some information and instructions on the "dynamic" ENGINE
7342 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7345 *) Make it possible to unload ranges of ERR strings with a new
7346 "ERR_unload_strings" function.
7349 *) Add a copy() function to EVP_MD.
7352 *) Make EVP_MD routines take a context pointer instead of just the
7353 md_data void pointer.
7356 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7357 that the digest can only process a single chunk of data
7358 (typically because it is provided by a piece of
7359 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7360 is only going to provide a single chunk of data, and hence the
7361 framework needn't accumulate the data for oneshot drivers.
7364 *) As with "ERR", make it possible to replace the underlying "ex_data"
7365 functions. This change also alters the storage and management of global
7366 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7367 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7368 index counters. The API functions that use this state have been changed
7369 to take a "class_index" rather than pointers to the class's local STACK
7370 and counter, and there is now an API function to dynamically create new
7371 classes. This centralisation allows us to (a) plug a lot of the
7372 thread-safety problems that existed, and (b) makes it possible to clean
7373 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7374 such data would previously have always leaked in application code and
7375 workarounds were in place to make the memory debugging turn a blind eye
7376 to it. Application code that doesn't use this new function will still
7377 leak as before, but their memory debugging output will announce it now
7378 rather than letting it slide.
7380 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7381 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7382 has a return value to indicate success or failure.
7385 *) Make it possible to replace the underlying "ERR" functions such that the
7386 global state (2 LHASH tables and 2 locks) is only used by the "default"
7387 implementation. This change also adds two functions to "get" and "set"
7388 the implementation prior to it being automatically set the first time
7389 any other ERR function takes place. Ie. an application can call "get",
7390 pass the return value to a module it has just loaded, and that module
7391 can call its own "set" function using that value. This means the
7392 module's "ERR" operations will use (and modify) the error state in the
7393 application and not in its own statically linked copy of OpenSSL code.
7396 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7397 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7398 the operation, and provides a more encapsulated way for external code
7399 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7400 to use these functions rather than manually incrementing the counts.
7402 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7405 *) Add EVP test program.
7408 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7411 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7412 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7413 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7414 These allow a CRL to be built without having to access X509_CRL fields
7415 directly. Modify 'ca' application to use new functions.
7418 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7419 bug workarounds. Rollback attack detection is a security feature.
7420 The problem will only arise on OpenSSL servers when TLSv1 is not
7421 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7422 Software authors not wanting to support TLSv1 will have special reasons
7423 for their choice and can explicitly enable this option.
7424 [Bodo Moeller, Lutz Jaenicke]
7426 *) Rationalise EVP so it can be extended: don't include a union of
7427 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7428 (similar to those existing for EVP_CIPHER_CTX).
7433 EVP_MD_CTX_init(&md); /* new function call */
7434 EVP_DigestInit(&md, EVP_sha1());
7435 EVP_DigestUpdate(&md, in, len);
7436 EVP_DigestFinal(&md, out, NULL);
7437 EVP_MD_CTX_cleanup(&md); /* new function call */
7441 *) Make DES key schedule conform to the usual scheme, as well as
7442 correcting its structure. This means that calls to DES functions
7443 now have to pass a pointer to a des_key_schedule instead of a
7444 plain des_key_schedule (which was actually always a pointer
7447 des_key_schedule ks;
7449 des_set_key_checked(..., &ks);
7450 des_ncbc_encrypt(..., &ks, ...);
7452 (Note that a later change renames 'des_...' into 'DES_...'.)
7455 *) Initial reduction of linker bloat: the use of some functions, such as
7456 PEM causes large amounts of unused functions to be linked in due to
7457 poor organisation. For example pem_all.c contains every PEM function
7458 which has a knock on effect of linking in large amounts of (unused)
7459 ASN1 code. Grouping together similar functions and splitting unrelated
7460 functions prevents this.
7463 *) Cleanup of EVP macros.
7466 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7467 correct _ecb suffix.
7470 *) Add initial OCSP responder support to ocsp application. The
7471 revocation information is handled using the text based index
7472 use by the ca application. The responder can either handle
7473 requests generated internally, supplied in files (for example
7474 via a CGI script) or using an internal minimal server.
7477 *) Add configuration choices to get zlib compression for TLS.
7480 *) Changes to Kerberos SSL for RFC 2712 compliance:
7481 1. Implemented real KerberosWrapper, instead of just using
7482 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7483 2. Implemented optional authenticator field of KerberosWrapper.
7485 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7486 and authenticator structs; see crypto/krb5/.
7488 Generalized Kerberos calls to support multiple Kerberos libraries.
7489 [Vern Staats <staatsvr@asc.hpc.mil>,
7490 Jeffrey Altman <jaltman@columbia.edu>
7491 via Richard Levitte]
7493 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7494 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7495 values for each of the key sizes rather than having just
7496 parameters (and 'speed' generating keys each time).
7499 *) Speed up EVP routines.
7502 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7503 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7504 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7505 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7507 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7508 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7509 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7512 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7514 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7517 *) Added the OS2-EMX target.
7518 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7520 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7521 to support NCONF routines in extension code. New function CONF_set_nconf()
7522 to allow functions which take an NCONF to also handle the old LHASH
7523 structure: this means that the old CONF compatible routines can be
7524 retained (in particular wrt extensions) without having to duplicate the
7525 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7528 *) Enhance the general user interface with mechanisms for inner control
7529 and with possibilities to have yes/no kind of prompts.
7532 *) Change all calls to low level digest routines in the library and
7533 applications to use EVP. Add missing calls to HMAC_cleanup() and
7534 don't assume HMAC_CTX can be copied using memcpy().
7535 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7537 *) Add the possibility to control engines through control names but with
7538 arbitrary arguments instead of just a string.
7539 Change the key loaders to take a UI_METHOD instead of a callback
7540 function pointer. NOTE: this breaks binary compatibility with earlier
7541 versions of OpenSSL [engine].
7542 Adapt the nCipher code for these new conditions and add a card insertion
7546 *) Enhance the general user interface with mechanisms to better support
7547 dialog box interfaces, application-defined prompts, the possibility
7548 to use defaults (for example default passwords from somewhere else)
7549 and interrupts/cancellations.
7552 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7553 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7556 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7557 tidy up some unnecessarily weird code in 'sk_new()').
7558 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7560 *) Change the key loading routines for ENGINEs to use the same kind
7561 callback (pem_password_cb) as all other routines that need this
7565 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7566 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7567 than this minimum value is recommended.
7570 *) New random seeder for OpenVMS, using the system process statistics
7571 that are easily reachable.
7574 *) Windows apparently can't transparently handle global
7575 variables defined in DLLs. Initialisations such as:
7577 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7579 won't compile. This is used by the any applications that need to
7580 declare their own ASN1 modules. This was fixed by adding the option
7581 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7582 needed for static libraries under Win32.
7585 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7586 setting of purpose and trust fields. New X509_STORE trust and
7587 purpose functions and tidy up setting in other SSL functions.
7590 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7591 structure. These are inherited by X509_STORE_CTX when it is
7592 initialised. This allows various defaults to be set in the
7593 X509_STORE structure (such as flags for CRL checking and custom
7594 purpose or trust settings) for functions which only use X509_STORE_CTX
7595 internally such as S/MIME.
7597 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7598 trust settings if they are not set in X509_STORE. This allows X509_STORE
7599 purposes and trust (in S/MIME for example) to override any set by default.
7601 Add command line options for CRL checking to smime, s_client and s_server
7605 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7606 are set then the CRL is looked up in the X509_STORE structure and
7607 its validity and signature checked, then if the certificate is found
7608 in the CRL the verify fails with a revoked error.
7610 Various new CRL related callbacks added to X509_STORE_CTX structure.
7612 Command line options added to 'verify' application to support this.
7614 This needs some additional work, such as being able to handle multiple
7615 CRLs with different times, extension based lookup (rather than just
7616 by subject name) and ultimately more complete V2 CRL extension
7620 *) Add a general user interface API (crypto/ui/). This is designed
7621 to replace things like des_read_password and friends (backward
7622 compatibility functions using this new API are provided).
7623 The purpose is to remove prompting functions from the DES code
7624 section as well as provide for prompting through dialog boxes in
7625 a window system and the like.
7628 *) Add "ex_data" support to ENGINE so implementations can add state at a
7629 per-structure level rather than having to store it globally.
7632 *) Make it possible for ENGINE structures to be copied when retrieved by
7633 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7634 This causes the "original" ENGINE structure to act like a template,
7635 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7636 operational state can be localised to each ENGINE structure, despite the
7637 fact they all share the same "methods". New ENGINE structures returned in
7638 this case have no functional references and the return value is the single
7639 structural reference. This matches the single structural reference returned
7640 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7644 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7645 needs to match any other type at all we need to manually clear the
7649 *) Changes to the "openssl engine" utility to include;
7650 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7651 about an ENGINE's available control commands.
7652 - executing control commands from command line arguments using the
7653 '-pre' and '-post' switches. '-post' is only used if '-t' is
7654 specified and the ENGINE is successfully initialised. The syntax for
7655 the individual commands are colon-separated, for example;
7656 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7659 *) New dynamic control command support for ENGINEs. ENGINEs can now
7660 declare their own commands (numbers), names (strings), descriptions,
7661 and input types for run-time discovery by calling applications. A
7662 subset of these commands are implicitly classed as "executable"
7663 depending on their input type, and only these can be invoked through
7664 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7665 can be based on user input, config files, etc). The distinction is
7666 that "executable" commands cannot return anything other than a boolean
7667 result and can only support numeric or string input, whereas some
7668 discoverable commands may only be for direct use through
7669 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7670 pointers, or other custom uses. The "executable" commands are to
7671 support parameterisations of ENGINE behaviour that can be
7672 unambiguously defined by ENGINEs and used consistently across any
7673 OpenSSL-based application. Commands have been added to all the
7674 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7675 control over shared-library paths without source code alterations.
7678 *) Changed all ENGINE implementations to dynamically allocate their
7679 ENGINEs rather than declaring them statically. Apart from this being
7680 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7681 this also allows the implementations to compile without using the
7682 internal engine_int.h header.
7685 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7686 'const' value. Any code that should be able to modify a RAND_METHOD
7687 should already have non-const pointers to it (ie. they should only
7688 modify their own ones).
7691 *) Made a variety of little tweaks to the ENGINE code.
7692 - "atalla" and "ubsec" string definitions were moved from header files
7693 to C code. "nuron" string definitions were placed in variables
7694 rather than hard-coded - allowing parameterisation of these values
7695 later on via ctrl() commands.
7696 - Removed unused "#if 0"'d code.
7697 - Fixed engine list iteration code so it uses ENGINE_free() to release
7698 structural references.
7699 - Constified the RAND_METHOD element of ENGINE structures.
7700 - Constified various get/set functions as appropriate and added
7701 missing functions (including a catch-all ENGINE_cpy that duplicates
7702 all ENGINE values onto a new ENGINE except reference counts/state).
7703 - Removed NULL parameter checks in get/set functions. Setting a method
7704 or function to NULL is a way of cancelling out a previously set
7705 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7706 and doesn't justify the extra error symbols and code.
7707 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7708 flags from engine_int.h to engine.h.
7709 - Changed prototypes for ENGINE handler functions (init(), finish(),
7710 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7713 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7714 to the algorithm using long division. The binary algorithm can be
7715 used only if the modulus is odd. On 32-bit systems, it is faster
7716 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7717 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7718 up to 450 bits. In 64-bit environments, the binary algorithm
7719 appears to be advantageous for much longer moduli; here we use it
7720 for moduli up to 2048 bits.
7723 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7724 could not support the combine flag in choice fields.
7727 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7728 extensions from a certificate request to the certificate.
7731 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7732 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7733 file: this allows the display of the certificate about to be
7734 signed to be customised, to allow certain fields to be included
7735 or excluded and extension details. The old system didn't display
7736 multicharacter strings properly, omitted fields not in the policy
7737 and couldn't display additional details such as extensions.
7740 *) Function EC_POINTs_mul for multiple scalar multiplication
7741 of an arbitrary number of elliptic curve points
7742 \sum scalars[i]*points[i],
7743 optionally including the generator defined for the EC_GROUP:
7744 scalar*generator + \sum scalars[i]*points[i].
7746 EC_POINT_mul is a simple wrapper function for the typical case
7747 that the point list has just one item (besides the optional
7751 *) First EC_METHODs for curves over GF(p):
7753 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7754 operations and provides various method functions that can also
7755 operate with faster implementations of modular arithmetic.
7757 EC_GFp_mont_method() reuses most functions that are part of
7758 EC_GFp_simple_method, but uses Montgomery arithmetic.
7760 [Bodo Moeller; point addition and point doubling
7761 implementation directly derived from source code provided by
7762 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7764 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7765 crypto/ec/ec_lib.c):
7767 Curves are EC_GROUP objects (with an optional group generator)
7768 based on EC_METHODs that are built into the library.
7770 Points are EC_POINT objects based on EC_GROUP objects.
7772 Most of the framework would be able to handle curves over arbitrary
7773 finite fields, but as there are no obvious types for fields other
7774 than GF(p), some functions are limited to that for now.
7777 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7778 that the file contains a complete HTTP response.
7781 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7782 change the def and num file printf format specifier from "%-40sXXX"
7783 to "%-39s XXX". The latter will always guarantee a space after the
7784 field while the former will cause them to run together if the field
7785 is 40 of more characters long.
7788 *) Constify the cipher and digest 'method' functions and structures
7789 and modify related functions to take constant EVP_MD and EVP_CIPHER
7793 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7794 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7797 *) Modify EVP_Digest*() routines so they now return values. Although the
7798 internal software routines can never fail additional hardware versions
7802 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7804 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7805 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7808 ERR_R_NESTED_ASN1_ERROR
7810 ERR_R_MISSING_ASN1_EOS
7811 were 4 .. 9, conflicting with
7812 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7814 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7815 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7817 Add new error code 'ERR_R_INTERNAL_ERROR'.
7820 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7824 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7825 sets the subject name for a new request or supersedes the
7826 subject name in a given request. Formats that can be parsed are
7827 'CN=Some Name, OU=myOU, C=IT'
7829 'CN=Some Name/OU=myOU/C=IT'.
7831 Add options '-batch' and '-verbose' to 'openssl req'.
7832 [Massimiliano Pala <madwolf@hackmasters.net>]
7834 *) Introduce the possibility to access global variables through
7835 functions on platform were that's the best way to handle exporting
7836 global variables in shared libraries. To enable this functionality,
7837 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7838 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7839 is normally done by Configure or something similar).
7841 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7842 in the source file (foo.c) like this:
7844 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7845 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7847 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7848 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7850 OPENSSL_DECLARE_GLOBAL(int,foo);
7851 #define foo OPENSSL_GLOBAL_REF(foo)
7852 OPENSSL_DECLARE_GLOBAL(double,bar);
7853 #define bar OPENSSL_GLOBAL_REF(bar)
7855 The #defines are very important, and therefore so is including the
7856 header file everywhere where the defined globals are used.
7858 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7859 of ASN.1 items, but that structure is a bit different.
7861 The largest change is in util/mkdef.pl which has been enhanced with
7862 better and easier to understand logic to choose which symbols should
7863 go into the Windows .def files as well as a number of fixes and code
7864 cleanup (among others, algorithm keywords are now sorted
7865 lexicographically to avoid constant rewrites).
7868 *) In BN_div() keep a copy of the sign of 'num' before writing the
7869 result to 'rm' because if rm==num the value will be overwritten
7870 and produce the wrong result if 'num' is negative: this caused
7871 problems with BN_mod() and BN_nnmod().
7874 *) Function OCSP_request_verify(). This checks the signature on an
7875 OCSP request and verifies the signer certificate. The signer
7876 certificate is just checked for a generic purpose and OCSP request
7880 *) Add OCSP_check_validity() function to check the validity of OCSP
7881 responses. OCSP responses are prepared in real time and may only
7882 be a few seconds old. Simply checking that the current time lies
7883 between thisUpdate and nextUpdate max reject otherwise valid responses
7884 caused by either OCSP responder or client clock inaccuracy. Instead
7885 we allow thisUpdate and nextUpdate to fall within a certain period of
7886 the current time. The age of the response can also optionally be
7887 checked. Two new options -validity_period and -status_age added to
7891 *) If signature or public key algorithm is unrecognized print out its
7892 OID rather that just UNKNOWN.
7895 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7896 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7897 ID to be generated from the issuer certificate alone which can then be
7898 passed to OCSP_id_issuer_cmp().
7901 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7902 ASN1 modules to export functions returning ASN1_ITEM pointers
7903 instead of the ASN1_ITEM structures themselves. This adds several
7904 new macros which allow the underlying ASN1 function/structure to
7905 be accessed transparently. As a result code should not use ASN1_ITEM
7906 references directly (such as &X509_it) but instead use the relevant
7907 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7908 use of the new ASN1 code on platforms where exporting structures
7909 is problematical (for example in shared libraries) but exporting
7910 functions returning pointers to structures is not.
7913 *) Add support for overriding the generation of SSL/TLS session IDs.
7914 These callbacks can be registered either in an SSL_CTX or per SSL.
7915 The purpose of this is to allow applications to control, if they wish,
7916 the arbitrary values chosen for use as session IDs, particularly as it
7917 can be useful for session caching in multiple-server environments. A
7918 command-line switch for testing this (and any client code that wishes
7919 to use such a feature) has been added to "s_server".
7920 [Geoff Thorpe, Lutz Jaenicke]
7922 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7923 of the form '#if defined(...) || defined(...) || ...' and
7924 '#if !defined(...) && !defined(...) && ...'. This also avoids
7925 the growing number of special cases it was previously handling.
7928 *) Make all configuration macros available for application by making
7929 sure they are available in opensslconf.h, by giving them names starting
7930 with "OPENSSL_" to avoid conflicts with other packages and by making
7931 sure e_os2.h will cover all platform-specific cases together with
7933 Additionally, it is now possible to define configuration/platform-
7934 specific names (called "system identities"). In the C code, these
7935 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7936 macro with the name beginning with "OPENSSL_SYS_", which is determined
7937 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7941 *) New option -set_serial to 'req' and 'x509' this allows the serial
7942 number to use to be specified on the command line. Previously self
7943 signed certificates were hard coded with serial number 0 and the
7944 CA options of 'x509' had to use a serial number in a file which was
7948 *) New options to 'ca' utility to support V2 CRL entry extensions.
7949 Currently CRL reason, invalidity date and hold instruction are
7950 supported. Add new CRL extensions to V3 code and some new objects.
7953 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7954 disable standard block padding (aka PKCS#5 padding) in the EVP
7955 API, which was previously mandatory. This means that the data is
7956 not padded in any way and so the total length much be a multiple
7957 of the block size, otherwise an error occurs.
7960 *) Initial (incomplete) OCSP SSL support.
7963 *) New function OCSP_parse_url(). This splits up a URL into its host,
7964 port and path components: primarily to parse OCSP URLs. New -url
7965 option to ocsp utility.
7968 *) New nonce behavior. The return value of OCSP_check_nonce() now
7969 reflects the various checks performed. Applications can decide
7970 whether to tolerate certain situations such as an absent nonce
7971 in a response when one was present in a request: the ocsp application
7972 just prints out a warning. New function OCSP_add1_basic_nonce()
7973 this is to allow responders to include a nonce in a response even if
7974 the request is nonce-less.
7977 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7978 skipped when using openssl x509 multiple times on a single input file,
7979 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7982 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7983 set string type: to handle setting ASN1_TIME structures. Fix ca
7984 utility to correctly initialize revocation date of CRLs.
7987 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7988 the clients preferred ciphersuites and rather use its own preferences.
7989 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7990 Internet Explorer by ensuring unchanged hash method during stepup.
7991 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7994 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7995 to aes and add a new 'exist' option to print out symbols that don't
7999 *) Additional options to ocsp utility to allow flags to be set and
8000 additional certificates supplied.
8003 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8004 OCSP client a number of certificate to only verify the response
8008 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8009 handle the new API. Currently only ECB, CBC modes supported. Add new
8012 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8013 Encryption Standard (AES) Ciphersuites for Transport Layer
8014 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8015 not enabled by default and were not part of the "ALL" ciphersuite
8016 alias because they were not yet official; they could be
8017 explicitly requested by specifying the "AESdraft" ciphersuite
8018 group alias. In the final release of OpenSSL 0.9.7, the group
8019 alias is called "AES" and is part of "ALL".)
8020 [Ben Laurie, Steve Henson, Bodo Moeller]
8022 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8023 request to response.
8026 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8027 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8028 extract information from a certificate request. OCSP_response_create()
8029 creates a response and optionally adds a basic response structure.
8030 OCSP_basic_add1_status() adds a complete single response to a basic
8031 response and returns the OCSP_SINGLERESP structure just added (to allow
8032 extensions to be included for example). OCSP_basic_add1_cert() adds a
8033 certificate to a basic response and OCSP_basic_sign() signs a basic
8034 response with various flags. New helper functions ASN1_TIME_check()
8035 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8036 (converts ASN1_TIME to GeneralizedTime).
8039 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8040 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8041 structure from a certificate. X509_pubkey_digest() digests the public_key
8042 contents: this is used in various key identifiers.
8045 *) Make sk_sort() tolerate a NULL argument.
8046 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8048 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8049 passed by the function are trusted implicitly. If any of them signed the
8050 response then it is assumed to be valid and is not verified.
8053 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8054 to data. This was previously part of the PKCS7 ASN1 code. This
8055 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8056 [Steve Henson, reported by Kenneth R. Robinette
8057 <support@securenetterm.com>]
8059 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8060 routines: without these tracing memory leaks is very painful.
8061 Fix leaks in PKCS12 and PKCS7 routines.
8064 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8065 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8066 effectively meant GeneralizedTime would never be used. Now it
8067 is initialised to -1 but X509_time_adj() now has to check the value
8068 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8069 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8070 [Steve Henson, reported by Kenneth R. Robinette
8071 <support@securenetterm.com>]
8073 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8074 result in a zero length in the ASN1_INTEGER structure which was
8075 not consistent with the structure when d2i_ASN1_INTEGER() was used
8076 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8077 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8078 where it did not print out a minus for negative ASN1_INTEGER.
8081 *) Add summary printout to ocsp utility. The various functions which
8082 convert status values to strings have been renamed to:
8083 OCSP_response_status_str(), OCSP_cert_status_str() and
8084 OCSP_crl_reason_str() and are no longer static. New options
8085 to verify nonce values and to disable verification. OCSP response
8086 printout format cleaned up.
8089 *) Add additional OCSP certificate checks. These are those specified
8090 in RFC2560. This consists of two separate checks: the CA of the
8091 certificate being checked must either be the OCSP signer certificate
8092 or the issuer of the OCSP signer certificate. In the latter case the
8093 OCSP signer certificate must contain the OCSP signing extended key
8094 usage. This check is performed by attempting to match the OCSP
8095 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8096 in the OCSP_CERTID structures of the response.
8099 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8100 and related routines. This uses the standard OpenSSL certificate
8101 verify routines to perform initial checks (just CA validity) and
8102 to obtain the certificate chain. Then additional checks will be
8103 performed on the chain. Currently the root CA is checked to see
8104 if it is explicitly trusted for OCSP signing. This is used to set
8105 a root CA as a global signing root: that is any certificate that
8106 chains to that CA is an acceptable OCSP signing certificate.
8109 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8110 extensions from a separate configuration file.
8111 As when reading extensions from the main configuration file,
8112 the '-extensions ...' option may be used for specifying the
8114 [Massimiliano Pala <madwolf@comune.modena.it>]
8116 *) New OCSP utility. Allows OCSP requests to be generated or
8117 read. The request can be sent to a responder and the output
8118 parsed, outputed or printed in text form. Not complete yet:
8119 still needs to check the OCSP response validity.
8122 *) New subcommands for 'openssl ca':
8123 'openssl ca -status <serial>' prints the status of the cert with
8124 the given serial number (according to the index file).
8125 'openssl ca -updatedb' updates the expiry status of certificates
8127 [Massimiliano Pala <madwolf@comune.modena.it>]
8129 *) New '-newreq-nodes' command option to CA.pl. This is like
8130 '-newreq', but calls 'openssl req' with the '-nodes' option
8131 so that the resulting key is not encrypted.
8132 [Damien Miller <djm@mindrot.org>]
8134 *) New configuration for the GNU Hurd.
8135 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8137 *) Initial code to implement OCSP basic response verify. This
8138 is currently incomplete. Currently just finds the signer's
8139 certificate and verifies the signature on the response.
8142 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8143 value of OPENSSLDIR. This is available via the new '-d' option
8144 to 'openssl version', and is also included in 'openssl version -a'.
8147 *) Allowing defining memory allocation callbacks that will be given
8148 file name and line number information in additional arguments
8149 (a const char* and an int). The basic functionality remains, as
8150 well as the original possibility to just replace malloc(),
8151 realloc() and free() by functions that do not know about these
8152 additional arguments. To register and find out the current
8153 settings for extended allocation functions, the following
8154 functions are provided:
8156 CRYPTO_set_mem_ex_functions
8157 CRYPTO_set_locked_mem_ex_functions
8158 CRYPTO_get_mem_ex_functions
8159 CRYPTO_get_locked_mem_ex_functions
8161 These work the same way as CRYPTO_set_mem_functions and friends.
8162 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8163 extended allocation function is enabled.
8164 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8165 a conventional allocation function is enabled.
8166 [Richard Levitte, Bodo Moeller]
8168 *) Finish off removing the remaining LHASH function pointer casts.
8169 There should no longer be any prototype-casting required when using
8170 the LHASH abstraction, and any casts that remain are "bugs". See
8171 the callback types and macros at the head of lhash.h for details
8172 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8175 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8176 If /dev/[u]random devices are not available or do not return enough
8177 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8179 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8180 /etc/entropy will be queried once each in this sequence, querying stops
8181 when enough entropy was collected without querying more sockets.
8184 *) Change the Unix RAND_poll() variant to be able to poll several
8185 random devices, as specified by DEVRANDOM, until a sufficient amount
8186 of data has been collected. We spend at most 10 ms on each file
8187 (select timeout) and read in non-blocking mode. DEVRANDOM now
8188 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8189 (previously it was just the string "/dev/urandom"), so on typical
8190 platforms the 10 ms delay will never occur.
8191 Also separate out the Unix variant to its own file, rand_unix.c.
8192 For VMS, there's a currently-empty rand_vms.c.
8195 *) Move OCSP client related routines to ocsp_cl.c. These
8196 provide utility functions which an application needing
8197 to issue a request to an OCSP responder and analyse the
8198 response will typically need: as opposed to those which an
8199 OCSP responder itself would need which will be added later.
8201 OCSP_request_sign() signs an OCSP request with an API similar
8202 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8203 response. OCSP_response_get1_basic() extracts basic response
8204 from response. OCSP_resp_find_status(): finds and extracts status
8205 information from an OCSP_CERTID structure (which will be created
8206 when the request structure is built). These are built from lower
8207 level functions which work on OCSP_SINGLERESP structures but
8208 won't normally be used unless the application wishes to examine
8209 extensions in the OCSP response for example.
8211 Replace nonce routines with a pair of functions.
8212 OCSP_request_add1_nonce() adds a nonce value and optionally
8213 generates a random value. OCSP_check_nonce() checks the
8214 validity of the nonce in an OCSP response.
8217 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8218 This doesn't copy the supplied OCSP_CERTID and avoids the
8219 need to free up the newly created id. Change return type
8220 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8221 This can then be used to add extensions to the request.
8222 Deleted OCSP_request_new(), since most of its functionality
8223 is now in OCSP_REQUEST_new() (and the case insensitive name
8224 clash) apart from the ability to set the request name which
8225 will be added elsewhere.
8228 *) Update OCSP API. Remove obsolete extensions argument from
8229 various functions. Extensions are now handled using the new
8230 OCSP extension code. New simple OCSP HTTP function which
8231 can be used to send requests and parse the response.
8234 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8235 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8236 uses the special reorder version of SET OF to sort the attributes
8237 and reorder them to match the encoded order. This resolves a long
8238 standing problem: a verify on a PKCS7 structure just after signing
8239 it used to fail because the attribute order did not match the
8240 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8241 it uses the received order. This is necessary to tolerate some broken
8242 software that does not order SET OF. This is handled by encoding
8243 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8244 to produce the required SET OF.
8247 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8248 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8249 files to get correct declarations of the ASN.1 item variables.
8252 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8253 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8254 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8255 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8256 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8257 ASN1_ITEM and no wrapper functions.
8260 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8261 replace the old function pointer based I/O routines. Change most of
8262 the *_d2i_bio() and *_d2i_fp() functions to use these.
8265 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8266 lines, recognize more "algorithms" that can be deselected, and make
8267 it complain about algorithm deselection that isn't recognised.
8270 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8271 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8272 to use new functions. Add NO_ASN1_OLD which can be set to remove
8273 some old style ASN1 functions: this can be used to determine if old
8274 code will still work when these eventually go away.
8277 *) New extension functions for OCSP structures, these follow the
8278 same conventions as certificates and CRLs.
8281 *) New function X509V3_add1_i2d(). This automatically encodes and
8282 adds an extension. Its behaviour can be customised with various
8283 flags to append, replace or delete. Various wrappers added for
8284 certificates and CRLs.
8287 *) Fix to avoid calling the underlying ASN1 print routine when
8288 an extension cannot be parsed. Correct a typo in the
8289 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8292 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8293 entries for variables.
8296 *) Add functionality to apps/openssl.c for detecting locking
8297 problems: As the program is single-threaded, all we have
8298 to do is register a locking callback using an array for
8299 storing which locks are currently held by the program.
8302 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8303 SSL_get_ex_data_X509_STORE_idx(), which is used in
8304 ssl_verify_cert_chain() and thus can be called at any time
8305 during TLS/SSL handshakes so that thread-safety is essential.
8306 Unfortunately, the ex_data design is not at all suited
8307 for multi-threaded use, so it probably should be abolished.
8310 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8311 [Broadcom, tweaked and integrated by Geoff Thorpe]
8313 *) Move common extension printing code to new function
8314 X509V3_print_extensions(). Reorganise OCSP print routines and
8315 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8318 *) New function X509_signature_print() to remove duplication in some
8322 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8323 set (this was treated exactly the same as SET OF previously). This
8324 is used to reorder the STACK representing the structure to match the
8325 encoding. This will be used to get round a problem where a PKCS7
8326 structure which was signed could not be verified because the STACK
8327 order did not reflect the encoded order.
8330 *) Reimplement the OCSP ASN1 module using the new code.
8333 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8334 for its ASN1 operations. The old style function pointers still exist
8335 for now but they will eventually go away.
8338 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8339 completely replaces the old ASN1 functionality with a table driven
8340 encoder and decoder which interprets an ASN1_ITEM structure describing
8341 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8342 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8343 has also been converted to the new form.
8346 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8347 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8348 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8349 for negative moduli.
8352 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8353 of not touching the result's sign bit.
8356 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8360 *) Changed the LHASH code to use prototypes for callbacks, and created
8361 macros to declare and implement thin (optionally static) functions
8362 that provide type-safety and avoid function pointer casting for the
8363 type-specific callbacks.
8366 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8368 [Veers Staats <staatsvr@asc.hpc.mil>,
8369 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8371 *) Reformat the FAQ so the different questions and answers can be divided
8372 in sections depending on the subject.
8375 *) Have the zlib compression code load ZLIB.DLL dynamically under
8379 *) New function BN_mod_sqrt for computing square roots modulo a prime
8380 (using the probabilistic Tonelli-Shanks algorithm unless
8381 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8382 be handled deterministically).
8383 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8385 *) Make BN_mod_inverse faster by explicitly handling small quotients
8386 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8387 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8390 *) New function BN_kronecker.
8393 *) Fix BN_gcd so that it works on negative inputs; the result is
8394 positive unless both parameters are zero.
8395 Previously something reasonably close to an infinite loop was
8396 possible because numbers could be growing instead of shrinking
8397 in the implementation of Euclid's algorithm.
8400 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8401 sign of the number in question.
8403 Fix BN_is_word(a,w) to work correctly for w == 0.
8405 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8406 because its test if the absolute value of 'a' equals 'w'.
8407 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8408 it exists mostly for use in the implementations of BN_is_zero(),
8409 BN_is_one(), and BN_is_word().
8412 *) New function BN_swap.
8415 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8416 the exponentiation functions are more likely to produce reasonable
8417 results on negative inputs.
8420 *) Change BN_mod_mul so that the result is always non-negative.
8421 Previously, it could be negative if one of the factors was negative;
8422 I don't think anyone really wanted that behaviour.
8425 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8426 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8427 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8428 and add new functions:
8437 BN_mod_lshift1_quick
8441 These functions always generate non-negative results.
8443 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8444 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8446 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8447 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8448 be reduced modulo m.
8449 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8452 The following entry accidentally appeared in the CHANGES file
8453 distributed with OpenSSL 0.9.7. The modifications described in
8454 it do *not* apply to OpenSSL 0.9.7.
8456 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8457 was actually never needed) and in BN_mul(). The removal in BN_mul()
8458 required a small change in bn_mul_part_recursive() and the addition
8459 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8460 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8461 bn_sub_words() and bn_add_words() except they take arrays with
8466 *) In 'openssl passwd', verify passwords read from the terminal
8467 unless the '-salt' option is used (which usually means that
8468 verification would just waste user's time since the resulting
8469 hash is going to be compared with some given password hash)
8470 or the new '-noverify' option is used.
8472 This is an incompatible change, but it does not affect
8473 non-interactive use of 'openssl passwd' (passwords on the command
8474 line, '-stdin' option, '-in ...' option) and thus should not
8478 *) Remove all references to RSAref, since there's no more need for it.
8481 *) Make DSO load along a path given through an environment variable
8482 (SHLIB_PATH) with shl_load().
8485 *) Constify the ENGINE code as a result of BIGNUM constification.
8486 Also constify the RSA code and most things related to it. In a
8487 few places, most notable in the depth of the ASN.1 code, ugly
8488 casts back to non-const were required (to be solved at a later
8492 *) Make it so the openssl application has all engines loaded by default.
8495 *) Constify the BIGNUM routines a little more.
8498 *) Add the following functions:
8500 ENGINE_load_cswift()
8502 ENGINE_load_atalla()
8504 ENGINE_load_builtin_engines()
8506 That way, an application can itself choose if external engines that
8507 are built-in in OpenSSL shall ever be used or not. The benefit is
8508 that applications won't have to be linked with libdl or other dso
8509 libraries unless it's really needed.
8511 Changed 'openssl engine' to load all engines on demand.
8512 Changed the engine header files to avoid the duplication of some
8513 declarations (they differed!).
8516 *) 'openssl engine' can now list capabilities.
8519 *) Better error reporting in 'openssl engine'.
8522 *) Never call load_dh_param(NULL) in s_server.
8525 *) Add engine application. It can currently list engines by name and
8526 identity, and test if they are actually available.
8529 *) Improve RPM specification file by forcing symbolic linking and making
8530 sure the installed documentation is also owned by root.root.
8531 [Damien Miller <djm@mindrot.org>]
8533 *) Give the OpenSSL applications more possibilities to make use of
8534 keys (public as well as private) handled by engines.
8537 *) Add OCSP code that comes from CertCo.
8540 *) Add VMS support for the Rijndael code.
8543 *) Added untested support for Nuron crypto accelerator.
8546 *) Add support for external cryptographic devices. This code was
8547 previously distributed separately as the "engine" branch.
8548 [Geoff Thorpe, Richard Levitte]
8550 *) Rework the filename-translation in the DSO code. It is now possible to
8551 have far greater control over how a "name" is turned into a filename
8552 depending on the operating environment and any oddities about the
8553 different shared library filenames on each system.
8556 *) Support threads on FreeBSD-elf in Configure.
8559 *) Fix for SHA1 assembly problem with MASM: it produces
8560 warnings about corrupt line number information when assembling
8561 with debugging information. This is caused by the overlapping
8563 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8566 NCONF_get_number() has no error checking at all. As a replacement,
8567 NCONF_get_number_e() is defined (_e for "error checking") and is
8568 promoted strongly. The old NCONF_get_number is kept around for
8569 binary backward compatibility.
8570 Make it possible for methods to load from something other than a BIO,
8571 by providing a function pointer that is given a name instead of a BIO.
8572 For example, this could be used to load configuration data from an
8576 *) Fix for non blocking accept BIOs. Added new I/O special reason
8577 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8578 with non blocking I/O was not possible because no retry code was
8579 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8583 *) Added the beginnings of Rijndael support.
8586 *) Fix for bug in DirectoryString mask setting. Add support for
8587 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8588 to allow certificate printing to more controllable, additional
8589 'certopt' option to 'x509' to allow new printing options to be
8593 *) Clean old EAY MD5 hack from e_os.h.
8596 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8598 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8599 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8600 [Joe Orton, Steve Henson]
8602 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8604 *) Fix additional bug revealed by the NISCC test suite:
8606 Stop bug triggering large recursion when presented with
8607 certain ASN.1 tags (CVE-2003-0851)
8610 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8612 *) Fix various bugs revealed by running the NISCC test suite:
8614 Stop out of bounds reads in the ASN1 code when presented with
8615 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8617 If verify callback ignores invalid public key errors don't try to check
8618 certificate signature with the NULL public key.
8622 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8623 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8627 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8628 extra data after the compression methods not only for TLS 1.0
8629 but also for SSL 3.0 (as required by the specification).
8630 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8632 *) Change X509_certificate_type() to mark the key as exported/exportable
8633 when it's 512 *bits* long, not 512 bytes.
8636 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8638 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8639 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8640 a protocol version number mismatch like a decryption error
8641 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8644 *) Turn on RSA blinding by default in the default implementation
8645 to avoid a timing attack. Applications that don't want it can call
8646 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8647 They would be ill-advised to do so in most cases.
8648 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8650 *) Change RSA blinding code so that it works when the PRNG is not
8651 seeded (in this case, the secret RSA exponent is abused as
8652 an unpredictable seed -- if it is not unpredictable, there
8653 is no point in blinding anyway). Make RSA blinding thread-safe
8654 by remembering the creator's thread ID in rsa->blinding and
8655 having all other threads use local one-time blinding factors
8656 (this requires more computation than sharing rsa->blinding, but
8657 avoids excessive locking; and if an RSA object is not shared
8658 between threads, blinding will still be very fast).
8661 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8663 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8664 via timing by performing a MAC computation even if incorrect
8665 block cipher padding has been found. This is a countermeasure
8666 against active attacks where the attacker has to distinguish
8667 between bad padding and a MAC verification error. (CVE-2003-0078)
8669 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8670 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8671 Martin Vuagnoux (EPFL, Ilion)]
8673 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8675 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8676 memory from it's contents. This is done with a counter that will
8677 place alternating values in each byte. This can be used to solve
8678 two issues: 1) the removal of calls to memset() by highly optimizing
8679 compilers, and 2) cleansing with other values than 0, since those can
8680 be read through on certain media, for example a swap space on disk.
8683 *) Bugfix: client side session caching did not work with external caching,
8684 because the session->cipher setting was not restored when reloading
8685 from the external cache. This problem was masked, when
8686 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8687 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8690 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8691 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8692 [Zeev Lieber <zeev-l@yahoo.com>]
8694 *) Undo an undocumented change introduced in 0.9.6e which caused
8695 repeated calls to OpenSSL_add_all_ciphers() and
8696 OpenSSL_add_all_digests() to be ignored, even after calling
8700 *) Change the default configuration reader to deal with last line not
8701 being properly terminated.
8704 *) Change X509_NAME_cmp() so it applies the special rules on handling
8705 DN values that are of type PrintableString, as well as RDNs of type
8706 emailAddress where the value has the type ia5String.
8707 [stefank@valicert.com via Richard Levitte]
8709 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8710 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8711 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8712 the bitwise-OR of the two for use by the majority of applications
8713 wanting this behaviour, and update the docs. The documented
8714 behaviour and actual behaviour were inconsistent and had been
8715 changing anyway, so this is more a bug-fix than a behavioural
8717 [Geoff Thorpe, diagnosed by Nadav Har'El]
8719 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8720 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8723 *) Fix initialization code race conditions in
8724 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8725 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8726 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8727 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8728 ssl2_get_cipher_by_char(),
8729 ssl3_get_cipher_by_char().
8730 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8732 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8733 the cached sessions are flushed, as the remove_cb() might use ex_data
8734 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8735 (see [openssl.org #212]).
8736 [Geoff Thorpe, Lutz Jaenicke]
8738 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8739 length, instead of the encoding length to d2i_ASN1_OBJECT.
8742 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8744 *) [In 0.9.6g-engine release:]
8745 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8746 [Lynn Gazis <lgazis@rainbow.com>]
8748 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8750 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8751 and get fix the header length calculation.
8752 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8753 Alon Kantor <alonk@checkpoint.com> (and others),
8756 *) Use proper error handling instead of 'assertions' in buffer
8757 overflow checks added in 0.9.6e. This prevents DoS (the
8758 assertions could call abort()).
8759 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8761 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8763 *) Add various sanity checks to asn1_get_length() to reject
8764 the ASN1 length bytes if they exceed sizeof(long), will appear
8765 negative or the content length exceeds the length of the
8767 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8769 *) Fix cipher selection routines: ciphers without encryption had no flags
8770 for the cipher strength set and where therefore not handled correctly
8771 by the selection routines (PR #130).
8774 *) Fix EVP_dsa_sha macro.
8778 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8779 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8780 that was added in OpenSSL 0.9.6d.
8782 As the countermeasure turned out to be incompatible with some
8783 broken SSL implementations, the new option is part of SSL_OP_ALL.
8784 SSL_OP_ALL is usually employed when compatibility with weird SSL
8785 implementations is desired (e.g. '-bugs' option to 's_client' and
8786 's_server'), so the new option is automatically set in many
8790 *) Changes in security patch:
8792 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8793 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8794 Air Force Materiel Command, USAF, under agreement number
8797 *) Add various sanity checks to asn1_get_length() to reject
8798 the ASN1 length bytes if they exceed sizeof(long), will appear
8799 negative or the content length exceeds the length of the
8800 supplied buffer. (CVE-2002-0659)
8801 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8803 *) Assertions for various potential buffer overflows, not known to
8805 [Ben Laurie (CHATS)]
8807 *) Various temporary buffers to hold ASCII versions of integers were
8808 too small for 64 bit platforms. (CVE-2002-0655)
8809 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8811 *) Remote buffer overflow in SSL3 protocol - an attacker could
8812 supply an oversized session ID to a client. (CVE-2002-0656)
8813 [Ben Laurie (CHATS)]
8815 *) Remote buffer overflow in SSL2 protocol - an attacker could
8816 supply an oversized client master key. (CVE-2002-0656)
8817 [Ben Laurie (CHATS)]
8819 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8821 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8822 encoded as NULL) with id-dsa-with-sha1.
8823 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8825 *) Check various X509_...() return values in apps/req.c.
8826 [Nils Larsch <nla@trustcenter.de>]
8828 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8829 an end-of-file condition would erroneously be flagged, when the CRLF
8830 was just at the end of a processed block. The bug was discovered when
8831 processing data through a buffering memory BIO handing the data to a
8832 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8833 <ptsekov@syntrex.com> and Nedelcho Stanev.
8836 *) Implement a countermeasure against a vulnerability recently found
8837 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8838 before application data chunks to avoid the use of known IVs
8839 with data potentially chosen by the attacker.
8842 *) Fix length checks in ssl3_get_client_hello().
8845 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8846 to prevent ssl3_read_internal() from incorrectly assuming that
8847 ssl3_read_bytes() found application data while handshake
8848 processing was enabled when in fact s->s3->in_read_app_data was
8849 merely automatically cleared during the initial handshake.
8850 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8852 *) Fix object definitions for Private and Enterprise: they were not
8853 recognized in their shortname (=lowercase) representation. Extend
8854 obj_dat.pl to issue an error when using undefined keywords instead
8855 of silently ignoring the problem (Svenning Sorensen
8856 <sss@sss.dnsalias.net>).
8859 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8860 generators, i.e. generators other than 2 and 5. (Previously, the
8861 code did not properly initialise the 'add' and 'rem' values to
8862 BN_generate_prime().)
8864 In the new general case, we do not insist that 'generator' is
8865 actually a primitive root: This requirement is rather pointless;
8866 a generator of the order-q subgroup is just as good, if not
8870 *) Map new X509 verification errors to alerts. Discovered and submitted by
8871 Tom Wu <tom@arcot.com>.
8874 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8875 returning non-zero before the data has been completely received
8876 when using non-blocking I/O.
8877 [Bodo Moeller; problem pointed out by John Hughes]
8879 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8880 [Ben Laurie, Lutz Jaenicke]
8882 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8883 Yoram Zahavi <YoramZ@gilian.com>).
8886 *) Add information about CygWin 1.3 and on, and preserve proper
8887 configuration for the versions before that.
8888 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8890 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8891 check whether we deal with a copy of a session and do not delete from
8892 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8893 <izhar@checkpoint.com>.
8896 *) Do not store session data into the internal session cache, if it
8897 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8898 flag is set). Proposed by Aslam <aslam@funk.com>.
8901 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8905 *) [In 0.9.6d-engine release:]
8906 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8907 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8909 *) Add the configuration target linux-s390x.
8910 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8912 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8913 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8914 variable as an indication that a ClientHello message has been
8915 received. As the flag value will be lost between multiple
8916 invocations of ssl3_accept when using non-blocking I/O, the
8917 function may not be aware that a handshake has actually taken
8918 place, thus preventing a new session from being added to the
8921 To avoid this problem, we now set s->new_session to 2 instead of
8922 using a local variable.
8923 [Lutz Jaenicke, Bodo Moeller]
8925 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8926 if the SSL_R_LENGTH_MISMATCH error is detected.
8927 [Geoff Thorpe, Bodo Moeller]
8929 *) New 'shared_ldflag' column in Configure platform table.
8932 *) Fix EVP_CIPHER_mode macro.
8933 ["Dan S. Camper" <dan@bti.net>]
8935 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8936 type, we must throw them away by setting rr->length to 0.
8937 [D P Chang <dpc@qualys.com>]
8939 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8941 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8942 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8943 worked incorrectly for those cases where range = 10..._2 and
8944 3*range is two bits longer than range.)
8947 *) Only add signing time to PKCS7 structures if it is not already
8951 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8952 OBJ_ld_ce should be OBJ_id_ce.
8953 Also some ip-pda OIDs in crypto/objects/objects.txt were
8954 incorrect (cf. RFC 3039).
8955 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8957 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8958 returns early because it has nothing to do.
8959 [Andy Schneider <andy.schneider@bjss.co.uk>]
8961 *) [In 0.9.6c-engine release:]
8962 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8963 [Andy Schneider <andy.schneider@bjss.co.uk>]
8965 *) [In 0.9.6c-engine release:]
8966 Add support for Cryptographic Appliance's keyserver technology.
8967 (Use engine 'keyclient')
8968 [Cryptographic Appliances and Geoff Thorpe]
8970 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8971 is called via tools/c89.sh because arguments have to be
8972 rearranged (all '-L' options must appear before the first object
8974 [Richard Shapiro <rshapiro@abinitio.com>]
8976 *) [In 0.9.6c-engine release:]
8977 Add support for Broadcom crypto accelerator cards, backported
8979 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8981 *) [In 0.9.6c-engine release:]
8982 Add support for SureWare crypto accelerator cards from
8983 Baltimore Technologies. (Use engine 'sureware')
8984 [Baltimore Technologies and Mark Cox]
8986 *) [In 0.9.6c-engine release:]
8987 Add support for crypto accelerator cards from Accelerated
8988 Encryption Processing, www.aep.ie. (Use engine 'aep')
8989 [AEP Inc. and Mark Cox]
8991 *) Add a configuration entry for gcc on UnixWare.
8992 [Gary Benson <gbenson@redhat.com>]
8994 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8995 messages are stored in a single piece (fixed-length part and
8996 variable-length part combined) and fix various bugs found on the way.
8999 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9000 instead. BIO_gethostbyname() does not know what timeouts are
9001 appropriate, so entries would stay in cache even when they have
9003 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9005 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9006 faced with a pathologically small ClientHello fragment that does
9007 not contain client_version: Instead of aborting with an error,
9008 simply choose the highest available protocol version (i.e.,
9009 TLS 1.0 unless it is disabled). In practice, ClientHello
9010 messages are never sent like this, but this change gives us
9011 strictly correct behaviour at least for TLS.
9014 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9015 never resets s->method to s->ctx->method when called from within
9016 one of the SSL handshake functions.
9017 [Bodo Moeller; problem pointed out by Niko Baric]
9019 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9020 (sent using the client's version number) if client_version is
9021 smaller than the protocol version in use. Also change
9022 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9023 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9024 the client will at least see that alert.
9027 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9031 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9032 client receives HelloRequest while in a handshake.
9033 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9035 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9036 should end in 'break', not 'goto end' which circumvents various
9037 cleanups done in state SSL_ST_OK. But session related stuff
9038 must be disabled for SSL_ST_OK in the case that we just sent a
9041 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9042 before just sending a HelloRequest.
9043 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9045 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9046 reveal whether illegal block cipher padding was found or a MAC
9047 verification error occurred. (Neither SSLerr() codes nor alerts
9048 are directly visible to potential attackers, but the information
9049 may leak via logfiles.)
9051 Similar changes are not required for the SSL 2.0 implementation
9052 because the number of padding bytes is sent in clear for SSL 2.0,
9053 and the extra bytes are just ignored. However ssl/s2_pkt.c
9054 failed to verify that the purported number of padding bytes is in
9058 *) Add OpenUNIX-8 support including shared libraries
9059 (Boyd Lynn Gerber <gerberb@zenez.com>).
9062 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9063 'wristwatch attack' using huge encoding parameters (cf.
9064 James H. Manger's CRYPTO 2001 paper). Note that the
9065 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9066 encoding parameters and hence was not vulnerable.
9069 *) BN_sqr() bug fix.
9070 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9072 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9073 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9074 followed by modular reduction.
9075 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9077 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9078 equivalent based on BN_pseudo_rand() instead of BN_rand().
9081 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9082 This function was broken, as the check for a new client hello message
9083 to handle SGC did not allow these large messages.
9084 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9087 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9090 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9091 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9094 *) Rework the configuration and shared library support for Tru64 Unix.
9095 The configuration part makes use of modern compiler features and
9096 still retains old compiler behavior for those that run older versions
9097 of the OS. The shared library support part includes a variant that
9098 uses the RPATH feature, and is available through the special
9099 configuration target "alpha-cc-rpath", which will never be selected
9101 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9103 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9104 with the same message size as in ssl3_get_certificate_request().
9105 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9106 messages might inadvertently be reject as too long.
9107 [Petr Lampa <lampa@fee.vutbr.cz>]
9109 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9112 *) Modified SSL library such that the verify_callback that has been set
9113 specificly for an SSL object with SSL_set_verify() is actually being
9114 used. Before the change, a verify_callback set with this function was
9115 ignored and the verify_callback() set in the SSL_CTX at the time of
9116 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9117 to allow the necessary settings.
9120 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9121 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9122 done automatically (in contradiction to the requirements of the C
9123 standard). This made problems when used from OpenSSH.
9126 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9127 dh->length and always used
9129 BN_rand_range(priv_key, dh->p).
9131 BN_rand_range() is not necessary for Diffie-Hellman, and this
9132 specific range makes Diffie-Hellman unnecessarily inefficient if
9133 dh->length (recommended exponent length) is much smaller than the
9134 length of dh->p. We could use BN_rand_range() if the order of
9135 the subgroup was stored in the DH structure, but we only have
9140 BN_rand(priv_key, l, ...)
9142 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9148 RSA_eay_public_encrypt
9149 RSA_eay_private_decrypt
9150 RSA_eay_private_encrypt (signing)
9151 RSA_eay_public_decrypt (signature verification)
9153 (default implementations for RSA_public_encrypt,
9154 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9155 always reject numbers >= n.
9158 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9159 to synchronize access to 'locking_thread'. This is necessary on
9160 systems where access to 'locking_thread' (an 'unsigned long'
9161 variable) is not atomic.
9164 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9165 *before* setting the 'crypto_lock_rand' flag. The previous code had
9166 a race condition if 0 is a valid thread ID.
9167 [Travis Vitek <vitek@roguewave.com>]
9169 *) Add support for shared libraries under Irix.
9170 [Albert Chin-A-Young <china@thewrittenword.com>]
9172 *) Add configuration option to build on Linux on both big-endian and
9174 [Ralf Baechle <ralf@uni-koblenz.de>]
9176 *) Add the possibility to create shared libraries on HP-UX.
9179 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9181 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9182 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9183 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9184 PRNG state recovery was possible based on the output of
9185 one PRNG request appropriately sized to gain knowledge on
9186 'md' followed by enough consecutive 1-byte PRNG requests
9187 to traverse all of 'state'.
9189 1. When updating 'md_local' (the current thread's copy of 'md')
9190 during PRNG output generation, hash all of the previous
9191 'md_local' value, not just the half used for PRNG output.
9193 2. Make the number of bytes from 'state' included into the hash
9194 independent from the number of PRNG bytes requested.
9196 The first measure alone would be sufficient to avoid
9197 Markku-Juhani's attack. (Actually it had never occurred
9198 to me that the half of 'md_local' used for chaining was the
9199 half from which PRNG output bytes were taken -- I had always
9200 assumed that the secret half would be used.) The second
9201 measure makes sure that additional data from 'state' is never
9202 mixed into 'md_local' in small portions; this heuristically
9203 further strengthens the PRNG.
9206 *) Fix crypto/bn/asm/mips3.s.
9209 *) When only the key is given to "enc", the IV is undefined. Print out
9210 an error message in this case.
9213 *) Handle special case when X509_NAME is empty in X509 printing routines.
9216 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9217 positive and less than q.
9220 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9221 used: it isn't thread safe and the add_lock_callback should handle
9223 [Paul Rose <Paul.Rose@bridge.com>]
9225 *) Verify that incoming data obeys the block size in
9226 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9230 [Ulf Möller, Bodo Möller]
9232 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9233 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9234 when fixing the server behaviour for backwards-compatible 'client
9235 hello' messages. (Note that the attack is impractical against
9236 SSL 3.0 and TLS 1.0 anyway because length and version checking
9237 means that the probability of guessing a valid ciphertext is
9238 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9241 Before 0.9.5, the countermeasure (hide the error by generating a
9242 random 'decryption result') did not work properly because
9243 ERR_clear_error() was missing, meaning that SSL_get_error() would
9244 detect the supposedly ignored error.
9246 Both problems are now fixed.
9249 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9250 (previously it was 1024).
9253 *) Fix for compatibility mode trust settings: ignore trust settings
9254 unless some valid trust or reject settings are present.
9257 *) Fix for blowfish EVP: its a variable length cipher.
9260 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9261 parameters in DSA public key structures and return an error in the
9262 DSA routines if parameters are absent.
9265 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9266 in the current directory if neither $RANDFILE nor $HOME was set.
9267 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9268 caused some confusion to Windows users who haven't defined $HOME.
9269 Thus RAND_file_name() is changed again: e_os.h can define a
9270 DEFAULT_HOME, which will be used if $HOME is not set.
9271 For Windows, we use "C:"; on other platforms, we still require
9272 environment variables.
9274 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9275 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9276 having multiple threads call RAND_poll() concurrently.
9279 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9280 combination of a flag and a thread ID variable.
9281 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9282 flag), *other* threads can enter ssleay_add_bytes without obeying
9283 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9284 that they do not hold after the first thread unsets add_do_not_lock).
9287 *) Change bctest again: '-x' expressions are not available in all
9291 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9293 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9294 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9296 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9297 the default extension for executables, if any. Also, make the perl
9298 scripts that use symlink() to test if it really exists and use "cp"
9299 if it doesn't. All this made OpenSSL compilable and installable in
9303 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9304 If SEQUENCE is length is indefinite just set c->slen to the total
9305 amount of data available.
9306 [Steve Henson, reported by shige@FreeBSD.org]
9307 [This change does not apply to 0.9.7.]
9309 *) Change bctest to avoid here-documents inside command substitution
9310 (workaround for FreeBSD /bin/sh bug).
9311 For compatibility with Ultrix, avoid shell functions (introduced
9312 in the bctest version that searches along $PATH).
9315 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9316 with des_encrypt() defined on some operating systems, like Solaris
9320 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9321 On the Importance of Eliminating Errors in Cryptographic
9322 Computations, J. Cryptology 14 (2001) 2, 101-119,
9323 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9326 *) MIPS assembler BIGNUM division bug fix.
9329 *) Disabled incorrect Alpha assembler code.
9332 *) Fix PKCS#7 decode routines so they correctly update the length
9333 after reading an EOC for the EXPLICIT tag.
9335 [This change does not apply to 0.9.7.]
9337 *) Fix bug in PKCS#12 key generation routines. This was triggered
9338 if a 3DES key was generated with a 0 initial byte. Include
9339 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9340 (but broken) behaviour.
9343 *) Enhance bctest to search for a working bc along $PATH and print
9345 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9347 *) Fix memory leaks in err.c: free err_data string if necessary;
9348 don't write to the wrong index in ERR_set_error_data.
9351 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9355 *) Replace rdtsc with _emit statements for VC++ version 5.
9356 [Jeremy Cooper <jeremy@baymoo.org>]
9358 *) Make it possible to reuse SSLv2 sessions.
9361 *) In copy_email() check for >= 0 as a return value for
9362 X509_NAME_get_index_by_NID() since 0 is a valid index.
9363 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9365 *) Avoid coredump with unsupported or invalid public keys by checking if
9366 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9367 PKCS7_verify() fails with non detached data.
9370 *) Don't use getenv in library functions when run as setuid/setgid.
9371 New function OPENSSL_issetugid().
9374 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9375 due to incorrect handling of multi-threading:
9377 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9379 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9381 3. Count how many times MemCheck_off() has been called so that
9382 nested use can be treated correctly. This also avoids
9383 inband-signalling in the previous code (which relied on the
9384 assumption that thread ID 0 is impossible).
9387 *) Add "-rand" option also to s_client and s_server.
9390 *) Fix CPU detection on Irix 6.x.
9391 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9392 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9394 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9397 [This change does not apply to 0.9.7.]
9399 *) Use the cached encoding of an X509_NAME structure rather than
9400 copying it. This is apparently the reason for the libsafe "errors"
9401 but the code is actually correct.
9404 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9405 Bleichenbacher's DSA attack.
9406 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9407 to be set and top=0 forces the highest bit to be set; top=-1 is new
9408 and leaves the highest bit random.
9409 [Ulf Moeller, Bodo Moeller]
9411 *) In the NCONF_...-based implementations for CONF_... queries
9412 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9413 a temporary CONF structure with the data component set to NULL
9414 (which gives segmentation faults in lh_retrieve).
9415 Instead, use NULL for the CONF pointer in CONF_get_string and
9416 CONF_get_number (which may use environment variables) and directly
9417 return NULL from CONF_get_section.
9420 *) Fix potential buffer overrun for EBCDIC.
9423 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9424 keyUsage if basicConstraints absent for a CA.
9427 *) Make SMIME_write_PKCS7() write mail header values with a format that
9428 is more generally accepted (no spaces before the semicolon), since
9429 some programs can't parse those values properly otherwise. Also make
9430 sure BIO's that break lines after each write do not create invalid
9434 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9435 macros previously used would not encode an empty SEQUENCE OF
9436 and break the signature.
9438 [This change does not apply to 0.9.7.]
9440 *) Zero the premaster secret after deriving the master secret in
9444 *) Add some EVP_add_digest_alias registrations (as found in
9445 OpenSSL_add_all_digests()) to SSL_library_init()
9446 aka OpenSSL_add_ssl_algorithms(). This provides improved
9447 compatibility with peers using X.509 certificates
9448 with unconventional AlgorithmIdentifier OIDs.
9451 *) Fix for Irix with NO_ASM.
9452 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9454 *) ./config script fixes.
9455 [Ulf Moeller, Richard Levitte]
9457 *) Fix 'openssl passwd -1'.
9460 *) Change PKCS12_key_gen_asc() so it can cope with non null
9461 terminated strings whose length is passed in the passlen
9462 parameter, for example from PEM callbacks. This was done
9463 by adding an extra length parameter to asc2uni().
9464 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9466 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9467 call failed, free the DSA structure.
9470 *) Fix to uni2asc() to cope with zero length Unicode strings.
9471 These are present in some PKCS#12 files.
9474 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9475 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9476 when writing a 32767 byte record.
9477 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9479 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9480 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9482 (RSA objects have a reference count access to which is protected
9483 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9484 so they are meant to be shared between threads.)
9485 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9486 "Reddie, Steven" <Steven.Reddie@ca.com>]
9488 *) Fix a deadlock in CRYPTO_mem_leaks().
9491 *) Use better test patterns in bntest.
9494 *) rand_win.c fix for Borland C.
9497 *) BN_rshift bugfix for n == 0.
9500 *) Add a 'bctest' script that checks for some known 'bc' bugs
9501 so that 'make test' does not abort just because 'bc' is broken.
9504 *) Store verify_result within SSL_SESSION also for client side to
9505 avoid potential security hole. (Re-used sessions on the client side
9506 always resulted in verify_result==X509_V_OK, not using the original
9507 result of the server certificate verification.)
9510 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9511 SSL3_RT_APPLICATION_DATA, return 0.
9512 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9516 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9517 releases, have been re-implemented by renaming the previous
9518 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9519 and ssl3_read_internal, respectively, and adding 'peek' parameters
9520 to them. The new ssl[23]_{read,peek} functions are calls to
9521 ssl[23]_read_internal with the 'peek' flag set appropriately.
9522 A 'peek' parameter has also been added to ssl3_read_bytes, which
9523 does the actual work for ssl3_read_internal.
9526 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9527 the method-specific "init()" handler. Also clean up ex_data after
9528 calling the method-specific "finish()" handler. Previously, this was
9529 happening the other way round.
9532 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9533 The previous value, 12, was not always sufficient for BN_mod_exp().
9536 *) Make sure that shared libraries get the internal name engine with
9537 the full version number and not just 0. This should mark the
9538 shared libraries as not backward compatible. Of course, this should
9539 be changed again when we can guarantee backward binary compatibility.
9542 *) Fix typo in get_cert_by_subject() in by_dir.c
9543 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9545 *) Rework the system to generate shared libraries:
9547 - Make note of the expected extension for the shared libraries and
9548 if there is a need for symbolic links from for example libcrypto.so.0
9549 to libcrypto.so.0.9.7. There is extended info in Configure for
9552 - Make as few rebuilds of the shared libraries as possible.
9554 - Still avoid linking the OpenSSL programs with the shared libraries.
9556 - When installing, install the shared libraries separately from the
9560 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9562 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9563 and not in SSL_clear because the latter is also used by the
9564 accept/connect functions; previously, the settings made by
9565 SSL_set_read_ahead would be lost during the handshake.
9566 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9568 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9569 Previously, it would create entries for disabled algorithms no
9573 *) Added several new manual pages for SSL_* function.
9576 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9578 *) In ssl23_get_client_hello, generate an error message when faced
9579 with an initial SSL 3.0/TLS record that is too small to contain the
9580 first two bytes of the ClientHello message, i.e. client_version.
9581 (Note that this is a pathologic case that probably has never happened
9582 in real life.) The previous approach was to use the version number
9583 from the record header as a substitute; but our protocol choice
9584 should not depend on that one because it is not authenticated
9585 by the Finished messages.
9588 *) More robust randomness gathering functions for Windows.
9589 [Jeffrey Altman <jaltman@columbia.edu>]
9591 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9592 not set then we don't setup the error code for issuer check errors
9593 to avoid possibly overwriting other errors which the callback does
9594 handle. If an application does set the flag then we assume it knows
9595 what it is doing and can handle the new informational codes
9599 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9600 a general "ANY" type, as such it should be able to decode anything
9601 including tagged types. However it didn't check the class so it would
9602 wrongly interpret tagged types in the same way as their universal
9603 counterpart and unknown types were just rejected. Changed so that the
9604 tagged and unknown types are handled in the same way as a SEQUENCE:
9605 that is the encoding is stored intact. There is also a new type
9606 "V_ASN1_OTHER" which is used when the class is not universal, in this
9607 case we have no idea what the actual type is so we just lump them all
9611 *) On VMS, stdout may very well lead to a file that is written to
9612 in a record-oriented fashion. That means that every write() will
9613 write a separate record, which will be read separately by the
9614 programs trying to read from it. This can be very confusing.
9616 The solution is to put a BIO filter in the way that will buffer
9617 text until a linefeed is reached, and then write everything a
9618 line at a time, so every record written will be an actual line,
9619 not chunks of lines and not (usually doesn't happen, but I've
9620 seen it once) several lines in one record. BIO_f_linebuffer() is
9623 Currently, it's a VMS-only method, because that's where it has
9624 been tested well enough.
9627 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9628 it can return incorrect results.
9629 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9630 but it was in 0.9.6-beta[12].)
9633 *) Disable the check for content being present when verifying detached
9634 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9635 include zero length content when signing messages.
9638 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9639 BIO_ctrl (for BIO pairs).
9642 *) Add DSO method for VMS.
9645 *) Bug fix: Montgomery multiplication could produce results with the
9649 *) Add RPM specification openssl.spec and modify it to build three
9650 packages. The default package contains applications, application
9651 documentation and run-time libraries. The devel package contains
9652 include files, static libraries and function documentation. The
9653 doc package contains the contents of the doc directory. The original
9654 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9657 *) Add a large number of documentation files for many SSL routines.
9658 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9660 *) Add a configuration entry for Sony News 4.
9661 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9663 *) Don't set the two most significant bits to one when generating a
9664 random number < q in the DSA library.
9667 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9668 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9669 the underlying transport is blocking) if a handshake took place.
9670 (The default behaviour is needed by applications such as s_client
9671 and s_server that use select() to determine when to use SSL_read;
9672 but for applications that know in advance when to expect data, it
9673 just makes things more complicated.)
9676 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9680 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9681 work better on such systems.
9682 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9684 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9685 Update PKCS12_parse() so it copies the friendlyName and the
9686 keyid to the certificates aux info.
9689 *) Fix bug in PKCS7_verify() which caused an infinite loop
9690 if there was more than one signature.
9691 [Sven Uszpelkat <su@celocom.de>]
9693 *) Major change in util/mkdef.pl to include extra information
9694 about each symbol, as well as presenting variables as well
9695 as functions. This change means that there's n more need
9696 to rebuild the .num files when some algorithms are excluded.
9699 *) Allow the verify time to be set by an application,
9700 rather than always using the current time.
9703 *) Phase 2 verify code reorganisation. The certificate
9704 verify code now looks up an issuer certificate by a
9705 number of criteria: subject name, authority key id
9706 and key usage. It also verifies self signed certificates
9707 by the same criteria. The main comparison function is
9708 X509_check_issued() which performs these checks.
9710 Lot of changes were necessary in order to support this
9711 without completely rewriting the lookup code.
9713 Authority and subject key identifier are now cached.
9715 The LHASH 'certs' is X509_STORE has now been replaced
9716 by a STACK_OF(X509_OBJECT). This is mainly because an
9717 LHASH can't store or retrieve multiple objects with
9718 the same hash value.
9720 As a result various functions (which were all internal
9721 use only) have changed to handle the new X509_STORE
9722 structure. This will break anything that messed round
9723 with X509_STORE internally.
9725 The functions X509_STORE_add_cert() now checks for an
9726 exact match, rather than just subject name.
9728 The X509_STORE API doesn't directly support the retrieval
9729 of multiple certificates matching a given criteria, however
9730 this can be worked round by performing a lookup first
9731 (which will fill the cache with candidate certificates)
9732 and then examining the cache for matches. This is probably
9733 the best we can do without throwing out X509_LOOKUP
9734 entirely (maybe later...).
9736 The X509_VERIFY_CTX structure has been enhanced considerably.
9738 All certificate lookup operations now go via a get_issuer()
9739 callback. Although this currently uses an X509_STORE it
9740 can be replaced by custom lookups. This is a simple way
9741 to bypass the X509_STORE hackery necessary to make this
9742 work and makes it possible to use more efficient techniques
9743 in future. A very simple version which uses a simple
9744 STACK for its trusted certificate store is also provided
9745 using X509_STORE_CTX_trusted_stack().
9747 The verify_cb() and verify() callbacks now have equivalents
9748 in the X509_STORE_CTX structure.
9750 X509_STORE_CTX also has a 'flags' field which can be used
9751 to customise the verify behaviour.
9754 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9755 excludes S/MIME capabilities.
9758 *) When a certificate request is read in keep a copy of the
9759 original encoding of the signed data and use it when outputting
9760 again. Signatures then use the original encoding rather than
9761 a decoded, encoded version which may cause problems if the
9762 request is improperly encoded.
9765 *) For consistency with other BIO_puts implementations, call
9766 buffer_write(b, ...) directly in buffer_puts instead of calling
9769 In BIO_puts, increment b->num_write as in BIO_write.
9770 [Peter.Sylvester@EdelWeb.fr]
9772 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9773 BN_zero, we may not return a BIGNUM with an array consisting of
9777 *) Avoid calling abort() from within the library when problems are
9778 detected, except if preprocessor symbols have been defined
9779 (such as REF_CHECK, BN_DEBUG etc.).
9782 *) New openssl application 'rsautl'. This utility can be
9783 used for low level RSA operations. DER public key
9784 BIO/fp routines also added.
9787 *) New Configure entry and patches for compiling on QNX 4.
9788 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9790 *) A demo state-machine implementation was sponsored by
9791 Nuron (http://www.nuron.com/) and is now available in
9792 demos/state_machine.
9795 *) New options added to the 'dgst' utility for signature
9796 generation and verification.
9799 *) Unrecognized PKCS#7 content types are now handled via a
9800 catch all ASN1_TYPE structure. This allows unsupported
9801 types to be stored as a "blob" and an application can
9802 encode and decode it manually.
9805 *) Fix various signed/unsigned issues to make a_strex.c
9807 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9809 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9810 length if passed a buffer. ASN1_INTEGER_to_BN failed
9811 if passed a NULL BN and its argument was negative.
9812 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9814 *) Modification to PKCS#7 encoding routines to output definite
9815 length encoding. Since currently the whole structures are in
9816 memory there's not real point in using indefinite length
9817 constructed encoding. However if OpenSSL is compiled with
9818 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9821 *) Added BIO_vprintf() and BIO_vsnprintf().
9824 *) Added more prefixes to parse for in the strings written
9825 through a logging bio, to cover all the levels that are available
9826 through syslog. The prefixes are now:
9828 PANIC, EMERG, EMR => LOG_EMERG
9829 ALERT, ALR => LOG_ALERT
9830 CRIT, CRI => LOG_CRIT
9831 ERROR, ERR => LOG_ERR
9832 WARNING, WARN, WAR => LOG_WARNING
9833 NOTICE, NOTE, NOT => LOG_NOTICE
9834 INFO, INF => LOG_INFO
9835 DEBUG, DBG => LOG_DEBUG
9837 and as before, if none of those prefixes are present at the
9838 beginning of the string, LOG_ERR is chosen.
9840 On Win32, the LOG_* levels are mapped according to this:
9842 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9843 LOG_WARNING => EVENTLOG_WARNING_TYPE
9844 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9848 *) Made it possible to reconfigure with just the configuration
9849 argument "reconf" or "reconfigure". The command line arguments
9850 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9851 and are retrieved from there when reconfiguring.
9855 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9857 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9860 *) The obj_dat.pl script was messing up the sorting of object
9861 names. The reason was that it compared the quoted version
9862 of strings as a result "OCSP" > "OCSP Signing" because
9863 " > SPACE. Changed script to store unquoted versions of
9864 names and add quotes on output. It was also omitting some
9865 names from the lookup table if they were given a default
9866 value (that is if SN is missing it is given the same
9867 value as LN and vice versa), these are now added on the
9868 grounds that if an object has a name we should be able to
9869 look it up. Finally added warning output when duplicate
9870 short or long names are found.
9873 *) Changes needed for Tandem NSK.
9874 [Scott Uroff <scott@xypro.com>]
9876 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9877 RSA_padding_check_SSLv23(), special padding was never detected
9878 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9879 version rollback attacks was not effective.
9881 In s23_clnt.c, don't use special rollback-attack detection padding
9882 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9883 client; similarly, in s23_srvr.c, don't do the rollback check if
9884 SSL 2.0 is the only protocol enabled in the server.
9887 *) Make it possible to get hexdumps of unprintable data with 'openssl
9888 asn1parse'. By implication, the functions ASN1_parse_dump() and
9889 BIO_dump_indent() are added.
9892 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9893 these print out strings and name structures based on various
9894 flags including RFC2253 support and proper handling of
9895 multibyte characters. Added options to the 'x509' utility
9896 to allow the various flags to be set.
9899 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9900 Also change the functions X509_cmp_current_time() and
9901 X509_gmtime_adj() work with an ASN1_TIME structure,
9902 this will enable certificates using GeneralizedTime in validity
9903 dates to be checked.
9906 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9907 negative public key encodings) on by default,
9908 NO_NEG_PUBKEY_BUG can be set to disable it.
9911 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9912 content octets. An i2c_ASN1_OBJECT is unnecessary because
9913 the encoding can be trivially obtained from the structure.
9916 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9917 not read locks (CRYPTO_r_[un]lock).
9920 *) A first attempt at creating official support for shared
9921 libraries through configuration. I've kept it so the
9922 default is static libraries only, and the OpenSSL programs
9923 are always statically linked for now, but there are
9924 preparations for dynamic linking in place.
9925 This has been tested on Linux and Tru64.
9928 *) Randomness polling function for Win9x, as described in:
9929 Peter Gutmann, Software Generation of Practically Strong
9933 *) Fix so PRNG is seeded in req if using an already existing
9937 *) New options to smime application. -inform and -outform
9938 allow alternative formats for the S/MIME message including
9939 PEM and DER. The -content option allows the content to be
9940 specified separately. This should allow things like Netscape
9941 form signing output easier to verify.
9944 *) Fix the ASN1 encoding of tags using the 'long form'.
9947 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9948 STRING types. These convert content octets to and from the
9949 underlying type. The actual tag and length octets are
9950 already assumed to have been read in and checked. These
9951 are needed because all other string types have virtually
9952 identical handling apart from the tag. By having versions
9953 of the ASN1 functions that just operate on content octets
9954 IMPLICIT tagging can be handled properly. It also allows
9955 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9956 and ASN1_INTEGER are identical apart from the tag.
9959 *) Change the handling of OID objects as follows:
9961 - New object identifiers are inserted in objects.txt, following
9962 the syntax given in objects.README.
9963 - objects.pl is used to process obj_mac.num and create a new
9965 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9968 This is currently kind of a hack, and the perl code in objects.pl
9969 isn't very elegant, but it works as I intended. The simplest way
9970 to check that it worked correctly is to look in obj_dat.h and
9971 check the array nid_objs and make sure the objects haven't moved
9972 around (this is important!). Additions are OK, as well as
9973 consistent name changes.
9976 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9979 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9980 The given file adds to whatever has already been seeded into the
9981 random pool through the RANDFILE configuration file option or
9982 environment variable, or the default random state file.
9985 *) mkstack.pl now sorts each macro group into lexical order.
9986 Previously the output order depended on the order the files
9987 appeared in the directory, resulting in needless rewriting
9991 *) Patches to make OpenSSL compile under Win32 again. Mostly
9992 work arounds for the VC++ problem that it treats func() as
9993 func(void). Also stripped out the parts of mkdef.pl that
9994 added extra typesafe functions: these no longer exist.
9997 *) Reorganisation of the stack code. The macros are now all
9998 collected in safestack.h . Each macro is defined in terms of
9999 a "stack macro" of the form SKM_<name>(type, a, b). The
10000 DEBUG_SAFESTACK is now handled in terms of function casts,
10001 this has the advantage of retaining type safety without the
10002 use of additional functions. If DEBUG_SAFESTACK is not defined
10003 then the non typesafe macros are used instead. Also modified the
10004 mkstack.pl script to handle the new form. Needs testing to see
10005 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10006 the default if no major problems. Similar behaviour for ASN1_SET_OF
10007 and PKCS12_STACK_OF.
10010 *) When some versions of IIS use the 'NET' form of private key the
10011 key derivation algorithm is different. Normally MD5(password) is
10012 used as a 128 bit RC4 key. In the modified case
10013 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10014 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10015 as the old Netscape_RSA functions except they have an additional
10016 'sgckey' parameter which uses the modified algorithm. Also added
10017 an -sgckey command line option to the rsa utility. Thanks to
10018 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10019 algorithm to openssl-dev.
10022 *) The evp_local.h macros were using 'c.##kname' which resulted in
10023 invalid expansion on some systems (SCO 5.0.5 for example).
10024 Corrected to 'c.kname'.
10025 [Phillip Porch <root@theporch.com>]
10027 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10028 a STACK of email addresses from a certificate or request, these look
10029 in the subject name and the subject alternative name extensions and
10030 omit any duplicate addresses.
10033 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10034 This makes DSA verification about 2 % faster.
10037 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10038 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10039 plus overhead for 1024 bit moduli).
10040 This makes exponentiations about 0.5 % faster for 1024 bit
10041 exponents (as measured by "openssl speed rsa2048").
10044 *) Rename memory handling macros to avoid conflicts with other
10046 Malloc => OPENSSL_malloc
10047 Malloc_locked => OPENSSL_malloc_locked
10048 Realloc => OPENSSL_realloc
10049 Free => OPENSSL_free
10052 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10053 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10056 *) CygWin32 support.
10057 [John Jarvie <jjarvie@newsguy.com>]
10059 *) The type-safe stack code has been rejigged. It is now only compiled
10060 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10061 by default all type-specific stack functions are "#define"d back to
10062 standard stack functions. This results in more streamlined output
10063 but retains the type-safety checking possibilities of the original
10067 *) The STACK code has been cleaned up, and certain type declarations
10068 that didn't make a lot of sense have been brought in line. This has
10069 also involved a cleanup of sorts in safestack.h to more correctly
10070 map type-safe stack functions onto their plain stack counterparts.
10071 This work has also resulted in a variety of "const"ifications of
10072 lots of the code, especially "_cmp" operations which should normally
10073 be prototyped with "const" parameters anyway.
10076 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10077 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10078 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10079 where all of 'md' is used each time the PRNG is used, but 'state'
10080 is used only indexed by a cyclic counter. As entropy may not be
10081 well distributed from the beginning, 'md' is important as a
10082 chaining variable. However, the output function chains only half
10083 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10084 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10085 in all of 'state' being rewritten, with the new values depending
10086 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10089 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10090 the handshake is continued after ssl_verify_cert_chain();
10091 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10092 can lead to 'unexplainable' connection aborts later.
10093 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10095 *) Major EVP API cipher revision.
10096 Add hooks for extra EVP features. This allows various cipher
10097 parameters to be set in the EVP interface. Support added for variable
10098 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10099 setting of RC2 and RC5 parameters.
10101 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10104 Remove lots of duplicated code from the EVP library. For example *every*
10105 cipher init() function handles the 'iv' in the same way according to the
10106 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10107 for CFB and OFB modes they zero ctx->num.
10109 New functionality allows removal of S/MIME code RC2 hack.
10111 Most of the routines have the same form and so can be declared in terms
10114 By shifting this to the top level EVP_CipherInit() it can be removed from
10115 all individual ciphers. If the cipher wants to handle IVs or keys
10116 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10119 Change lots of functions like EVP_EncryptUpdate() to now return a
10120 value: although software versions of the algorithms cannot fail
10121 any installed hardware versions can.
10124 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10125 this option is set, tolerate broken clients that send the negotiated
10126 protocol version number instead of the requested protocol version
10130 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10131 i.e. non-zero for export ciphersuites, zero otherwise.
10132 Previous versions had this flag inverted, inconsistent with
10133 rsa_tmp_cb (..._TMP_RSA_CB).
10134 [Bodo Moeller; problem reported by Amit Chopra]
10136 *) Add missing DSA library text string. Work around for some IIS
10137 key files with invalid SEQUENCE encoding.
10140 *) Add a document (doc/standards.txt) that list all kinds of standards
10141 and so on that are implemented in OpenSSL.
10144 *) Enhance c_rehash script. Old version would mishandle certificates
10145 with the same subject name hash and wouldn't handle CRLs at all.
10146 Added -fingerprint option to crl utility, to support new c_rehash
10150 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10153 *) Fix for SSL server purpose checking. Server checking was
10154 rejecting certificates which had extended key usage present
10155 but no ssl client purpose.
10156 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10158 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10159 is a little unclear about how a blank password is handled.
10160 Since the password in encoded as a BMPString with terminating
10161 double NULL a zero length password would end up as just the
10162 double NULL. However no password at all is different and is
10163 handled differently in the PKCS#12 key generation code. NS
10164 treats a blank password as zero length. MSIE treats it as no
10165 password on export: but it will try both on import. We now do
10166 the same: PKCS12_parse() tries zero length and no password if
10167 the password is set to "" or NULL (NULL is now a valid password:
10168 it wasn't before) as does the pkcs12 application.
10171 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10172 perror when PEM_read_bio_X509_REQ fails, the error message must
10173 be obtained from the error queue.
10176 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10177 it in ERR_remove_state if appropriate, and change ERR_get_state
10178 accordingly to avoid race conditions (this is necessary because
10179 thread_hash is no longer constant once set).
10182 *) Bugfix for linux-elf makefile.one.
10185 *) RSA_get_default_method() will now cause a default
10186 RSA_METHOD to be chosen if one doesn't exist already.
10187 Previously this was only set during a call to RSA_new()
10188 or RSA_new_method(NULL) meaning it was possible for
10189 RSA_get_default_method() to return NULL.
10192 *) Added native name translation to the existing DSO code
10193 that will convert (if the flag to do so is set) filenames
10194 that are sufficiently small and have no path information
10195 into a canonical native form. Eg. "blah" converted to
10196 "libblah.so" or "blah.dll" etc.
10199 *) New function ERR_error_string_n(e, buf, len) which is like
10200 ERR_error_string(e, buf), but writes at most 'len' bytes
10201 including the 0 terminator. For ERR_error_string_n, 'buf'
10203 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10205 *) CONF library reworked to become more general. A new CONF
10206 configuration file reader "class" is implemented as well as a
10207 new functions (NCONF_*, for "New CONF") to handle it. The now
10208 old CONF_* functions are still there, but are reimplemented to
10209 work in terms of the new functions. Also, a set of functions
10210 to handle the internal storage of the configuration data is
10211 provided to make it easier to write new configuration file
10212 reader "classes" (I can definitely see something reading a
10213 configuration file in XML format, for example), called _CONF_*,
10214 or "the configuration storage API"...
10216 The new configuration file reading functions are:
10218 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10219 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10221 NCONF_default, NCONF_WIN32
10223 NCONF_dump_fp, NCONF_dump_bio
10225 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10226 NCONF_new creates a new CONF object. This works in the same way
10227 as other interfaces in OpenSSL, like the BIO interface.
10228 NCONF_dump_* dump the internal storage of the configuration file,
10229 which is useful for debugging. All other functions take the same
10230 arguments as the old CONF_* functions wth the exception of the
10231 first that must be a `CONF *' instead of a `LHASH *'.
10233 To make it easer to use the new classes with the old CONF_* functions,
10234 the function CONF_set_default_method is provided.
10237 *) Add '-tls1' option to 'openssl ciphers', which was already
10238 mentioned in the documentation but had not been implemented.
10239 (This option is not yet really useful because even the additional
10240 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10243 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10244 OpenSSL-based applications) load shared libraries and bind to
10245 them in a portable way.
10246 [Geoff Thorpe, with contributions from Richard Levitte]
10248 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10250 *) Make sure _lrotl and _lrotr are only used with MSVC.
10252 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10253 (the default implementation of RAND_status).
10255 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10256 to '-clrext' (= clear extensions), as intended and documented.
10257 [Bodo Moeller; inconsistency pointed out by Michael Attili
10258 <attili@amaxo.com>]
10260 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10261 was larger than the MD block size.
10262 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10264 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10265 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10266 using the passed key: if the passed key was a private key the result
10267 of X509_print(), for example, would be to print out all the private key
10271 *) des_quad_cksum() byte order bug fix.
10272 [Ulf Möller, using the problem description in krb4-0.9.7, where
10273 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10275 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10277 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10279 *) For easily testing in shell scripts whether some command
10280 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10281 returns with exit code 0 iff no command of the given name is available.
10282 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10283 the output goes to stdout and nothing is printed to stderr.
10284 Additional arguments are always ignored.
10286 Since for each cipher there is a command of the same name,
10287 the 'no-cipher' compilation switches can be tested this way.
10289 ('openssl no-XXX' is not able to detect pseudo-commands such
10290 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10293 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10296 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10297 is set; it will be thrown away anyway because each handshake creates
10299 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10300 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10301 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10302 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10305 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10306 'Q' and 'R' lose their special meanings (quit/renegotiate).
10307 This is part of what -quiet does; unlike -quiet, -ign_eof
10308 does not suppress any output.
10311 *) Add compatibility options to the purpose and trust code. The
10312 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10313 accepts a certificate or CA, this was the previous behaviour,
10314 with all the associated security issues.
10316 X509_TRUST_COMPAT is the old trust behaviour: only and
10317 automatically trust self signed roots in certificate store. A
10318 new trust setting X509_TRUST_DEFAULT is used to specify that
10319 a purpose has no associated trust setting and it should instead
10320 use the value in the default purpose.
10323 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10324 and fix a memory leak.
10327 *) In util/mkerr.pl (which implements 'make errors'), preserve
10328 reason strings from the previous version of the .c file, as
10329 the default to have only downcase letters (and digits) in
10330 automatically generated reasons codes is not always appropriate.
10333 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10334 using strerror. Previously, ERR_reason_error_string() returned
10335 library names as reason strings for SYSerr; but SYSerr is a special
10336 case where small numbers are errno values, not library numbers.
10339 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10340 converts DSA parameters into DH parameters. (When creating parameters,
10341 DSA_generate_parameters is used.)
10344 *) Include 'length' (recommended exponent length) in C code generated
10345 by 'openssl dhparam -C'.
10348 *) The second argument to set_label in perlasm was already being used
10349 so couldn't be used as a "file scope" flag. Moved to third argument
10353 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10354 instead of RAND_bytes for encryption IVs and salts.
10357 *) Include RAND_status() into RAND_METHOD instead of implementing
10358 it only for md_rand.c Otherwise replacing the PRNG by calling
10359 RAND_set_rand_method would be impossible.
10362 *) Don't let DSA_generate_key() enter an infinite loop if the random
10363 number generation fails.
10366 *) New 'rand' application for creating pseudo-random output.
10369 *) Added configuration support for Linux/IA64
10370 [Rolf Haberrecker <rolf@suse.de>]
10372 *) Assembler module support for Mingw32.
10375 *) Shared library support for HPUX (in shlib/).
10376 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10378 *) Shared library support for Solaris gcc.
10379 [Lutz Behnke <behnke@trustcenter.de>]
10381 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10383 *) PKCS7_encrypt() was adding text MIME headers twice because they
10384 were added manually and by SMIME_crlf_copy().
10387 *) In bntest.c don't call BN_rand with zero bits argument.
10388 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10390 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10391 case was implemented. This caused BN_div_recp() to fail occasionally.
10394 *) Add an optional second argument to the set_label() in the perl
10395 assembly language builder. If this argument exists and is set
10396 to 1 it signals that the assembler should use a symbol whose
10397 scope is the entire file, not just the current function. This
10398 is needed with MASM which uses the format label:: for this scope.
10399 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10401 *) Change the ASN1 types so they are typedefs by default. Before
10402 almost all types were #define'd to ASN1_STRING which was causing
10403 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10407 *) Change names of new functions to the new get1/get0 naming
10408 convention: After 'get1', the caller owns a reference count
10409 and has to call ..._free; 'get0' returns a pointer to some
10410 data structure without incrementing reference counters.
10411 (Some of the existing 'get' functions increment a reference
10412 counter, some don't.)
10413 Similarly, 'set1' and 'add1' functions increase reference
10414 counters or duplicate objects.
10417 *) Allow for the possibility of temp RSA key generation failure:
10418 the code used to assume it always worked and crashed on failure.
10421 *) Fix potential buffer overrun problem in BIO_printf().
10422 [Ulf Möller, using public domain code by Patrick Powell; problem
10423 pointed out by David Sacerdote <das33@cornell.edu>]
10425 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10426 RAND_egd() and RAND_status(). In the command line application,
10427 the EGD socket can be specified like a seed file using RANDFILE
10431 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10432 Some CAs (e.g. Verisign) distribute certificates in this form.
10435 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10436 list to exclude them. This means that no special compilation option
10437 is needed to use anonymous DH: it just needs to be included in the
10441 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10442 EVP_MD_type. The old functionality is available in a new macro called
10443 EVP_MD_md(). Change code that uses it and update docs.
10446 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10447 where the 'void *' argument is replaced by a function pointer argument.
10448 Previously 'void *' was abused to point to functions, which works on
10449 many platforms, but is not correct. As these functions are usually
10450 called by macros defined in OpenSSL header files, most source code
10451 should work without changes.
10454 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10455 sections with information on -D... compiler switches used for
10456 compiling the library so that applications can see them. To enable
10457 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10458 must be defined. E.g.,
10459 #define OPENSSL_ALGORITHM_DEFINES
10460 #include <openssl/opensslconf.h>
10461 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10462 [Richard Levitte, Ulf and Bodo Möller]
10464 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10468 *) Change the 'other' type in certificate aux info to a STACK_OF
10469 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10470 the required ASN1 format: arbitrary types determined by an OID.
10473 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10474 argument to 'req'. This is not because the function is newer or
10475 better than others it just uses the work 'NEW' in the certificate
10476 request header lines. Some software needs this.
10479 *) Reorganise password command line arguments: now passwords can be
10480 obtained from various sources. Delete the PEM_cb function and make
10481 it the default behaviour: i.e. if the callback is NULL and the
10482 usrdata argument is not NULL interpret it as a null terminated pass
10483 phrase. If usrdata and the callback are NULL then the pass phrase
10484 is prompted for as usual.
10487 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10488 the support is automatically enabled. The resulting binaries will
10489 autodetect the card and use it if present.
10490 [Ben Laurie and Compaq Inc.]
10492 *) Work around for Netscape hang bug. This sends certificate request
10493 and server done in one record. Since this is perfectly legal in the
10494 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10495 the bugs/SSLv3 entry for more info.
10498 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10501 *) Add -rand argument to smime and pkcs12 applications and read/write
10505 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10508 *) Add command line password options to the remaining applications.
10511 *) Bug fix for BN_div_recp() for numerators with an even number of
10515 *) More tests in bntest.c, and changed test_bn output.
10518 *) ./config recognizes MacOS X now.
10521 *) Bug fix for BN_div() when the first words of num and divisor are
10522 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10525 *) Add support for various broken PKCS#8 formats, and command line
10526 options to produce them.
10529 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10530 get temporary BIGNUMs from a BN_CTX.
10533 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10537 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10538 include a #define from the old name to the new. The original intent
10539 was that statically linked binaries could for example just call
10540 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10541 link with digests. This never worked because SSLeay_add_all_digests()
10542 and SSLeay_add_all_ciphers() were in the same source file so calling
10543 one would link with the other. They are now in separate source files.
10546 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10549 *) Use a less unusual form of the Miller-Rabin primality test (it used
10550 a binary algorithm for exponentiation integrated into the Miller-Rabin
10551 loop, our standard modexp algorithms are faster).
10554 *) Support for the EBCDIC character set completed.
10555 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10557 *) Source code cleanups: use const where appropriate, eliminate casts,
10558 use void * instead of char * in lhash.
10561 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10562 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10563 this the server could overwrite ephemeral keys that the client
10567 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10568 using 50 iterations of the Rabin-Miller test.
10570 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10571 iterations of the Rabin-Miller test as required by the appendix
10572 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10573 As BN_is_prime_fasttest includes trial division, DSA parameter
10574 generation becomes much faster.
10576 This implies a change for the callback functions in DSA_is_prime
10577 and DSA_generate_parameters: The callback function is called once
10578 for each positive witness in the Rabin-Miller test, not just
10579 occasionally in the inner loop; and the parameters to the
10580 callback function now provide an iteration count for the outer
10581 loop rather than for the current invocation of the inner loop.
10582 DSA_generate_parameters additionally can call the callback
10583 function with an 'iteration count' of -1, meaning that a
10584 candidate has passed the trial division test (when q is generated
10585 from an application-provided seed, trial division is skipped).
10588 *) New function BN_is_prime_fasttest that optionally does trial
10589 division before starting the Rabin-Miller test and has
10590 an additional BN_CTX * argument (whereas BN_is_prime always
10591 has to allocate at least one BN_CTX).
10592 'callback(1, -1, cb_arg)' is called when a number has passed the
10593 trial division stage.
10596 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10600 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10603 *) New function BN_pseudo_rand().
10606 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10607 bignum version of BN_from_montgomery() with the working code from
10608 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10612 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10613 made it impossible to use the same SSL_SESSION data structure in
10614 SSL2 clients in multiple threads.
10617 *) The return value of RAND_load_file() no longer counts bytes obtained
10618 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10619 to seed the PRNG (previously an explicit byte count was required).
10620 [Ulf Möller, Bodo Möller]
10622 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10623 used (char *) instead of (void *) and had casts all over the place.
10626 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10629 *) Retain source code compatibility for BN_prime_checks macro:
10630 BN_is_prime(..., BN_prime_checks, ...) now uses
10631 BN_prime_checks_for_size to determine the appropriate number of
10632 Rabin-Miller iterations.
10635 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10636 DH_CHECK_P_NOT_SAFE_PRIME.
10637 (Check if this is true? OpenPGP calls them "strong".)
10640 *) Merge the functionality of "dh" and "gendh" programs into a new program
10641 "dhparam". The old programs are retained for now but will handle DH keys
10642 (instead of parameters) in future.
10645 *) Make the ciphers, s_server and s_client programs check the return values
10646 when a new cipher list is set.
10649 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10650 ciphers. Before when the 56bit ciphers were enabled the sorting was
10653 The syntax for the cipher sorting has been extended to support sorting by
10654 cipher-strength (using the strength_bits hard coded in the tables).
10655 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10657 Fix a bug in the cipher-command parser: when supplying a cipher command
10658 string with an "undefined" symbol (neither command nor alphanumeric
10659 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10660 an error is flagged.
10662 Due to the strength-sorting extension, the code of the
10663 ssl_create_cipher_list() function was completely rearranged. I hope that
10664 the readability was also increased :-)
10665 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10667 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10668 for the first serial number and places 2 in the serial number file. This
10669 avoids problems when the root CA is created with serial number zero and
10670 the first user certificate has the same issuer name and serial number
10674 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10675 the new code. Add documentation for this stuff.
10678 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10679 X509_*() to X509at_*() on the grounds that they don't handle X509
10680 structures and behave in an analogous way to the X509v3 functions:
10681 they shouldn't be called directly but wrapper functions should be used
10684 So we also now have some wrapper functions that call the X509at functions
10685 when passed certificate requests. (TO DO: similar things can be done with
10686 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10687 things. Some of these need some d2i or i2d and print functionality
10688 because they handle more complex structures.)
10691 *) Add missing #ifndefs that caused missing symbols when building libssl
10692 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10693 NO_RSA in ssl/s2*.c.
10694 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10696 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10697 has a return value which indicates the quality of the random data
10698 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10699 error queue. New function RAND_pseudo_bytes() generates output that is
10700 guaranteed to be unique but not unpredictable. RAND_add is like
10701 RAND_seed, but takes an extra argument for an entropy estimate
10702 (RAND_seed always assumes full entropy).
10705 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10706 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10707 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10708 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10709 false-positive rate of at most 2^-80 for random input.
10712 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10715 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10716 in the 0.9.5 release), this returns the chain
10717 from an X509_CTX structure with a dup of the stack and all
10718 the X509 reference counts upped: so the stack will exist
10719 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10722 Also make SSL_SESSION_print() print out the verify return
10726 *) Add manpage for the pkcs12 command. Also change the default
10727 behaviour so MAC iteration counts are used unless the new
10728 -nomaciter option is used. This improves file security and
10729 only older versions of MSIE (4.0 for example) need it.
10732 *) Honor the no-xxx Configure options when creating .DEF files.
10735 *) Add PKCS#10 attributes to field table: challengePassword,
10736 unstructuredName and unstructuredAddress. These are taken from
10737 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10738 international characters are used.
10740 More changes to X509_ATTRIBUTE code: allow the setting of types
10741 based on strings. Remove the 'loc' parameter when adding
10742 attributes because these will be a SET OF encoding which is sorted
10746 *) Initial changes to the 'req' utility to allow request generation
10747 automation. This will allow an application to just generate a template
10748 file containing all the field values and have req construct the
10751 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10752 used all over the place including certificate requests and PKCS#7
10753 structures. They are currently handled manually where necessary with
10754 some primitive wrappers for PKCS#7. The new functions behave in a
10755 manner analogous to the X509 extension functions: they allow
10756 attributes to be looked up by NID and added.
10758 Later something similar to the X509V3 code would be desirable to
10759 automatically handle the encoding, decoding and printing of the
10760 more complex types. The string types like challengePassword can
10761 be handled by the string table functions.
10763 Also modified the multi byte string table handling. Now there is
10764 a 'global mask' which masks out certain types. The table itself
10765 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10766 is useful when for example there is only one permissible type
10767 (as in countryName) and using the mask might result in no valid
10771 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10772 SSL_get_peer_finished to allow applications to obtain the latest
10773 Finished messages sent to the peer or expected from the peer,
10774 respectively. (SSL_get_peer_finished is usually the Finished message
10775 actually received from the peer, otherwise the protocol will be aborted.)
10777 As the Finished message are message digests of the complete handshake
10778 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10779 be used for external authentication procedures when the authentication
10780 provided by SSL/TLS is not desired or is not enough.
10783 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10784 the host supports BWX extension and if Compaq C is present on the
10785 $PATH. Just exploiting of the BWX extension results in 20-30%
10786 performance kick for some algorithms, e.g. DES and RC4 to mention
10787 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10791 *) Add support for MS "fast SGC". This is arguably a violation of the
10792 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10793 weak crypto and after checking the certificate is SGC a second one
10794 with strong crypto. MS SGC stops the first handshake after receiving
10795 the server certificate message and sends a second client hello. Since
10796 a server will typically do all the time consuming operations before
10797 expecting any further messages from the client (server key exchange
10798 is the most expensive) there is little difference between the two.
10800 To get OpenSSL to support MS SGC we have to permit a second client
10801 hello message after we have sent server done. In addition we have to
10802 reset the MAC if we do get this second client hello.
10805 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10806 if a DER encoded private key is RSA or DSA traditional format. Changed
10807 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10808 format DER encoded private key. Newer code should use PKCS#8 format which
10809 has the key type encoded in the ASN1 structure. Added DER private key
10810 support to pkcs8 application.
10813 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10814 ciphersuites has been selected (as required by the SSL 3/TLS 1
10815 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10816 is set, we interpret this as a request to violate the specification
10817 (the worst that can happen is a handshake failure, and 'correct'
10818 behaviour would result in a handshake failure anyway).
10821 *) In SSL_CTX_add_session, take into account that there might be multiple
10822 SSL_SESSION structures with the same session ID (e.g. when two threads
10823 concurrently obtain them from an external cache).
10824 The internal cache can handle only one SSL_SESSION with a given ID,
10825 so if there's a conflict, we now throw out the old one to achieve
10829 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10830 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10831 some routines that use cipher OIDs: some ciphers do not have OIDs
10832 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10836 *) Simplify the trust setting structure and code. Now we just have
10837 two sequences of OIDs for trusted and rejected settings. These will
10838 typically have values the same as the extended key usage extension
10839 and any application specific purposes.
10841 The trust checking code now has a default behaviour: it will just
10842 check for an object with the same NID as the passed id. Functions can
10843 be provided to override either the default behaviour or the behaviour
10844 for a given id. SSL client, server and email already have functions
10845 in place for compatibility: they check the NID and also return "trusted"
10846 if the certificate is self signed.
10849 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10850 traditional format into an EVP_PKEY structure.
10853 *) Add a password callback function PEM_cb() which either prompts for
10854 a password if usr_data is NULL or otherwise assumes it is a null
10855 terminated password. Allow passwords to be passed on command line
10856 environment or config files in a few more utilities.
10859 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10860 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10861 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10862 Update documentation.
10865 *) Support for ASN1 "NULL" type. This could be handled before by using
10866 ASN1_TYPE but there wasn't any function that would try to read a NULL
10867 and produce an error if it couldn't. For compatibility we also have
10868 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10869 don't allocate anything because they don't need to.
10872 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10874 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10876 *) Rebuild of the memory allocation routines used by OpenSSL code and
10877 possibly others as well. The purpose is to make an interface that
10878 provide hooks so anyone can build a separate set of allocation and
10879 deallocation routines to be used by OpenSSL, for example memory
10880 pool implementations, or something else, which was previously hard
10881 since Malloc(), Realloc() and Free() were defined as macros having
10882 the values malloc, realloc and free, respectively (except for Win32
10883 compilations). The same is provided for memory debugging code.
10884 OpenSSL already comes with functionality to find memory leaks, but
10885 this gives people a chance to debug other memory problems.
10887 With these changes, a new set of functions and macros have appeared:
10889 CRYPTO_set_mem_debug_functions() [F]
10890 CRYPTO_get_mem_debug_functions() [F]
10891 CRYPTO_dbg_set_options() [F]
10892 CRYPTO_dbg_get_options() [F]
10893 CRYPTO_malloc_debug_init() [M]
10895 The memory debug functions are NULL by default, unless the library
10896 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10897 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10898 gives the standard debugging functions that come with OpenSSL) or
10899 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10900 provided by the library user) must be used. When the standard
10901 debugging functions are used, CRYPTO_dbg_set_options can be used to
10902 request additional information:
10903 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10904 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10906 Also, things like CRYPTO_set_mem_functions will always give the
10907 expected result (the new set of functions is used for allocation
10908 and deallocation) at all times, regardless of platform and compiler
10911 To finish it up, some functions that were never use in any other
10912 way than through macros have a new API and new semantic:
10914 CRYPTO_dbg_malloc()
10915 CRYPTO_dbg_realloc()
10918 All macros of value have retained their old syntax.
10919 [Richard Levitte and Bodo Moeller]
10921 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10922 ordering of SMIMECapabilities wasn't in "strength order" and there
10923 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10927 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10928 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10929 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10931 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10932 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10933 functionality to handle multipart/signed properly) and a utility
10934 called 'smime' to call all this stuff. This is based on code I
10935 originally wrote for Celo who have kindly allowed it to be
10936 included in OpenSSL.
10939 *) Add variants des_set_key_checked and des_set_key_unchecked of
10940 des_set_key (aka des_key_sched). Global variable des_check_key
10941 decides which of these is called by des_set_key; this way
10942 des_check_key behaves as it always did, but applications and
10943 the library itself, which was buggy for des_check_key == 1,
10944 have a cleaner way to pick the version they need.
10947 *) New function PKCS12_newpass() which changes the password of a
10951 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10952 dynamic mix. In both cases the ids can be used as an index into the
10953 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10954 functions so they accept a list of the field values and the
10955 application doesn't need to directly manipulate the X509_TRUST
10959 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10963 *) Modify the way the V3 extension code looks up extensions. This now
10964 works in a similar way to the object code: we have some "standard"
10965 extensions in a static table which is searched with OBJ_bsearch()
10966 and the application can add dynamic ones if needed. The file
10967 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10968 updated whenever a new extension is added to the core code and kept
10969 in ext_nid order. There is a simple program 'tabtest.c' which checks
10970 this. New extensions are not added too often so this file can readily
10971 be maintained manually.
10973 There are two big advantages in doing things this way. The extensions
10974 can be looked up immediately and no longer need to be "added" using
10975 X509V3_add_standard_extensions(): this function now does nothing.
10976 [Side note: I get *lots* of email saying the extension code doesn't
10977 work because people forget to call this function]
10978 Also no dynamic allocation is done unless new extensions are added:
10979 so if we don't add custom extensions there is no need to call
10980 X509V3_EXT_cleanup().
10983 *) Modify enc utility's salting as follows: make salting the default. Add a
10984 magic header, so unsalted files fail gracefully instead of just decrypting
10985 to garbage. This is because not salting is a big security hole, so people
10986 should be discouraged from doing it.
10989 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10990 digest to be passed on the command line but it only used this
10991 parameter when signing a certificate. Modified so all relevant
10992 operations are affected by the digest parameter including the
10993 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10994 DSA key was used because it didn't fix the digest.
10997 *) Initial certificate chain verify code. Currently tests the untrusted
10998 certificates for consistency with the verify purpose (which is set
10999 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11001 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11002 this is because it will reject chains with invalid extensions whereas
11003 every previous version of OpenSSL and SSLeay made no checks at all.
11005 Trust code: checks the root CA for the relevant trust settings. Trust
11006 settings have an initial value consistent with the verify purpose: e.g.
11007 if the verify purpose is for SSL client use it expects the CA to be
11008 trusted for SSL client use. However the default value can be changed to
11009 permit custom trust settings: one example of this would be to only trust
11010 certificates from a specific "secure" set of CAs.
11012 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11013 which should be used for version portability: especially since the
11014 verify structure is likely to change more often now.
11016 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11017 to set them. If not set then assume SSL clients will verify SSL servers
11020 Two new options to the verify program: -untrusted allows a set of
11021 untrusted certificates to be passed in and -purpose which sets the
11022 intended purpose of the certificate. If a purpose is set then the
11023 new chain verify code is used to check extension consistency.
11026 *) Support for the authority information access extension.
11029 *) Modify RSA and DSA PEM read routines to transparently handle
11030 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11031 public keys in a format compatible with certificate
11032 SubjectPublicKeyInfo structures. Unfortunately there were already
11033 functions called *_PublicKey_* which used various odd formats so
11034 these are retained for compatibility: however the DSA variants were
11035 never in a public release so they have been deleted. Changed dsa/rsa
11036 utilities to handle the new format: note no releases ever handled public
11037 keys so we should be OK.
11039 The primary motivation for this change is to avoid the same fiasco
11040 that dogs private keys: there are several incompatible private key
11041 formats some of which are standard and some OpenSSL specific and
11042 require various evil hacks to allow partial transparent handling and
11043 even then it doesn't work with DER formats. Given the option anything
11044 other than PKCS#8 should be dumped: but the other formats have to
11045 stay in the name of compatibility.
11047 With public keys and the benefit of hindsight one standard format
11048 is used which works with EVP_PKEY, RSA or DSA structures: though
11049 it clearly returns an error if you try to read the wrong kind of key.
11051 Added a -pubkey option to the 'x509' utility to output the public key.
11052 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11053 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11054 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11055 that do the same as the EVP_PKEY_assign_*() except they up the
11056 reference count of the added key (they don't "swallow" the
11060 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11061 CRLs would fail if the file contained no certificates or no CRLs:
11062 added a new function to read in both types and return the number
11063 read: this means that if none are read it will be an error. The
11064 DER versions of the certificate and CRL reader would always fail
11065 because it isn't possible to mix certificates and CRLs in DER format
11066 without choking one or the other routine. Changed this to just read
11067 a certificate: this is the best we can do. Also modified the code
11068 in apps/verify.c to take notice of return codes: it was previously
11069 attempting to read in certificates from NULL pointers and ignoring
11070 any errors: this is one reason why the cert and CRL reader seemed
11071 to work. It doesn't check return codes from the default certificate
11072 routines: these may well fail if the certificates aren't installed.
11075 *) Code to support otherName option in GeneralName.
11078 *) First update to verify code. Change the verify utility
11079 so it warns if it is passed a self signed certificate:
11080 for consistency with the normal behaviour. X509_verify
11081 has been modified to it will now verify a self signed
11082 certificate if *exactly* the same certificate appears
11083 in the store: it was previously impossible to trust a
11084 single self signed certificate. This means that:
11085 openssl verify ss.pem
11086 now gives a warning about a self signed certificate but
11087 openssl verify -CAfile ss.pem ss.pem
11091 *) For servers, store verify_result in SSL_SESSION data structure
11092 (and add it to external session representation).
11093 This is needed when client certificate verifications fails,
11094 but an application-provided verification callback (set by
11095 SSL_CTX_set_cert_verify_callback) allows accepting the session
11096 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11097 but returns 1): When the session is reused, we have to set
11098 ssl->verify_result to the appropriate error code to avoid
11100 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11102 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11103 case in PKCS7_dataInit() where the signed PKCS7 structure
11104 didn't contain any existing data because it was being created.
11105 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11107 *) Add a salt to the key derivation routines in enc.c. This
11108 forms the first 8 bytes of the encrypted file. Also add a
11109 -S option to allow a salt to be input on the command line.
11112 *) New function X509_cmp(). Oddly enough there wasn't a function
11113 to compare two certificates. We do this by working out the SHA1
11114 hash and comparing that. X509_cmp() will be needed by the trust
11118 *) SSL_get1_session() is like SSL_get_session(), but increments
11119 the reference count in the SSL_SESSION returned.
11120 [Geoff Thorpe <geoff@eu.c2.net>]
11122 *) Fix for 'req': it was adding a null to request attributes.
11123 Also change the X509_LOOKUP and X509_INFO code to handle
11124 certificate auxiliary information.
11127 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11131 *) Add the possibility to add extra information to the memory leak
11132 detecting output, to form tracebacks, showing from where each
11133 allocation was originated: CRYPTO_push_info("constant string") adds
11134 the string plus current file name and line number to a per-thread
11135 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11136 is like calling CYRPTO_pop_info() until the stack is empty.
11137 Also updated memory leak detection code to be multi-thread-safe.
11140 *) Add options -text and -noout to pkcs7 utility and delete the
11141 encryption options which never did anything. Update docs.
11144 *) Add options to some of the utilities to allow the pass phrase
11145 to be included on either the command line (not recommended on
11146 OSes like Unix) or read from the environment. Update the
11147 manpages and fix a few bugs.
11150 *) Add a few manpages for some of the openssl commands.
11153 *) Fix the -revoke option in ca. It was freeing up memory twice,
11154 leaking and not finding already revoked certificates.
11157 *) Extensive changes to support certificate auxiliary information.
11158 This involves the use of X509_CERT_AUX structure and X509_AUX
11159 functions. An X509_AUX function such as PEM_read_X509_AUX()
11160 can still read in a certificate file in the usual way but it
11161 will also read in any additional "auxiliary information". By
11162 doing things this way a fair degree of compatibility can be
11163 retained: existing certificates can have this information added
11164 using the new 'x509' options.
11166 Current auxiliary information includes an "alias" and some trust
11167 settings. The trust settings will ultimately be used in enhanced
11168 certificate chain verification routines: currently a certificate
11169 can only be trusted if it is self signed and then it is trusted
11173 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11174 The problem was that one of the replacement routines had not been working
11175 since SSLeay releases. For now the offending routine has been replaced
11176 with non-optimised assembler. Even so, this now gives around 95%
11177 performance improvement for 1024 bit RSA signs.
11180 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11181 handling. Most clients have the effective key size in bits equal to
11182 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11183 A few however don't do this and instead use the size of the decrypted key
11184 to determine the RC2 key length and the AlgorithmIdentifier to determine
11185 the effective key length. In this case the effective key length can still
11186 be 40 bits but the key length can be 168 bits for example. This is fixed
11187 by manually forcing an RC2 key into the EVP_PKEY structure because the
11188 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11189 the key length and effective key length are equal.
11192 *) Add a bunch of functions that should simplify the creation of
11193 X509_NAME structures. Now you should be able to do:
11194 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11195 and have it automatically work out the correct field type and fill in
11196 the structures. The more adventurous can try:
11197 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11198 and it will (hopefully) work out the correct multibyte encoding.
11201 *) Change the 'req' utility to use the new field handling and multibyte
11202 copy routines. Before the DN field creation was handled in an ad hoc
11203 way in req, ca, and x509 which was rather broken and didn't support
11204 BMPStrings or UTF8Strings. Since some software doesn't implement
11205 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11206 using the dirstring_type option. See the new comment in the default
11207 openssl.cnf for more info.
11210 *) Make crypto/rand/md_rand.c more robust:
11211 - Assure unique random numbers after fork().
11212 - Make sure that concurrent threads access the global counter and
11213 md serializably so that we never lose entropy in them
11214 or use exactly the same state in multiple threads.
11215 Access to the large state is not always serializable because
11216 the additional locking could be a performance killer, and
11217 md should be large enough anyway.
11220 *) New file apps/app_rand.c with commonly needed functionality
11221 for handling the random seed file.
11223 Use the random seed file in some applications that previously did not:
11225 dsaparam -genkey (which also ignored its '-rand' option),
11228 x509 (when signing).
11229 Except on systems with /dev/urandom, it is crucial to have a random
11230 seed file at least for key creation, DSA signing, and for DH exchanges;
11231 for RSA signatures we could do without one.
11233 gendh and gendsa (unlike genrsa) used to read only the first byte
11234 of each file listed in the '-rand' option. The function as previously
11235 found in genrsa is now in app_rand.c and is used by all programs
11236 that support '-rand'.
11239 *) In RAND_write_file, use mode 0600 for creating files;
11240 don't just chmod when it may be too late.
11243 *) Report an error from X509_STORE_load_locations
11244 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11247 *) New function ASN1_mbstring_copy() this copies a string in either
11248 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11249 into an ASN1_STRING type. A mask of permissible types is passed
11250 and it chooses the "minimal" type to use or an error if not type
11254 *) Add function equivalents to the various macros in asn1.h. The old
11255 macros are retained with an M_ prefix. Code inside the library can
11256 use the M_ macros. External code (including the openssl utility)
11257 should *NOT* in order to be "shared library friendly".
11260 *) Add various functions that can check a certificate's extensions
11261 to see if it usable for various purposes such as SSL client,
11262 server or S/MIME and CAs of these types. This is currently
11263 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11264 verification. Also added a -purpose flag to x509 utility to
11265 print out all the purposes.
11268 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11272 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11273 for, obtain and decode and extension and obtain its critical flag.
11274 This allows all the necessary extension code to be handled in a
11275 single function call.
11278 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11279 platforms. See crypto/rc4/rc4_enc.c for further details.
11282 *) New -noout option to asn1parse. This causes no output to be produced
11283 its main use is when combined with -strparse and -out to extract data
11284 from a file (which may not be in ASN.1 format).
11287 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11288 when producing the local key id.
11289 [Richard Levitte <levitte@stacken.kth.se>]
11291 *) New option -dhparam in s_server. This allows a DH parameter file to be
11292 stated explicitly. If it is not stated then it tries the first server
11293 certificate file. The previous behaviour hard coded the filename
11297 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11298 a public key to be input or output. For example:
11299 openssl rsa -in key.pem -pubout -out pubkey.pem
11300 Also added necessary DSA public key functions to handle this.
11303 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11304 in the message. This was handled by allowing
11305 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11306 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11308 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11309 to the end of the strings whereas this didn't. This would cause problems
11310 if strings read with d2i_ASN1_bytes() were later modified.
11311 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11313 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11314 data and it contains EOF it will end up returning an error. This is
11315 caused by input 46 bytes long. The cause is due to the way base64
11316 BIOs find the start of base64 encoded data. They do this by trying a
11317 trial decode on each line until they find one that works. When they
11318 do a flag is set and it starts again knowing it can pass all the
11319 data directly through the decoder. Unfortunately it doesn't reset
11320 the context it uses. This means that if EOF is reached an attempt
11321 is made to pass two EOFs through the context and this causes the
11322 resulting error. This can also cause other problems as well. As is
11323 usual with these problems it takes *ages* to find and the fix is
11324 trivial: move one line.
11325 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11327 *) Ugly workaround to get s_client and s_server working under Windows. The
11328 old code wouldn't work because it needed to select() on sockets and the
11329 tty (for keypresses and to see if data could be written). Win32 only
11330 supports select() on sockets so we select() with a 1s timeout on the
11331 sockets and then see if any characters are waiting to be read, if none
11332 are present then we retry, we also assume we can always write data to
11333 the tty. This isn't nice because the code then blocks until we've
11334 received a complete line of data and it is effectively polling the
11335 keyboard at 1s intervals: however it's quite a bit better than not
11336 working at all :-) A dedicated Windows application might handle this
11337 with an event loop for example.
11340 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11341 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11342 will be called when RSA_sign() and RSA_verify() are used. This is useful
11343 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11344 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11345 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11346 This necessitated the support of an extra signature type NID_md5_sha1
11347 for SSL signatures and modifications to the SSL library to use it instead
11348 of calling RSA_public_decrypt() and RSA_private_encrypt().
11351 *) Add new -verify -CAfile and -CApath options to the crl program, these
11352 will lookup a CRL issuers certificate and verify the signature in a
11353 similar way to the verify program. Tidy up the crl program so it
11354 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11355 less strict. It will now permit CRL extensions even if it is not
11356 a V2 CRL: this will allow it to tolerate some broken CRLs.
11359 *) Initialize all non-automatic variables each time one of the openssl
11360 sub-programs is started (this is necessary as they may be started
11361 multiple times from the "OpenSSL>" prompt).
11362 [Lennart Bang, Bodo Moeller]
11364 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11365 removing all other RSA functionality (this is what NO_RSA does). This
11366 is so (for example) those in the US can disable those operations covered
11367 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11371 *) Non-copying interface to BIO pairs.
11372 (still largely untested)
11375 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11376 ASCII string. This was handled independently in various places before.
11379 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11380 UTF8 strings a character at a time.
11383 *) Use client_version from client hello to select the protocol
11384 (s23_srvr.c) and for RSA client key exchange verification
11385 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11388 *) Add various utility functions to handle SPKACs, these were previously
11389 handled by poking round in the structure internals. Added new function
11390 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11391 print, verify and generate SPKACs. Based on an original idea from
11392 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11395 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11398 *) Allow the config file extension section to be overwritten on the
11399 command line. Based on an original idea from Massimiliano Pala
11400 <madwolf@comune.modena.it>. The new option is called -extensions
11401 and can be applied to ca, req and x509. Also -reqexts to override
11402 the request extensions in req and -crlexts to override the crl extensions
11406 *) Add new feature to the SPKAC handling in ca. Now you can include
11407 the same field multiple times by preceding it by "XXXX." for example:
11410 this is the same syntax as used in the req config file.
11413 *) Allow certificate extensions to be added to certificate requests. These
11414 are specified in a 'req_extensions' option of the req section of the
11415 config file. They can be printed out with the -text option to req but
11416 are otherwise ignored at present.
11419 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11420 data read consists of only the final block it would not decrypted because
11421 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11422 A misplaced 'break' also meant the decrypted final block might not be
11423 copied until the next read.
11426 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11427 a few extra parameters to the DH structure: these will be useful if
11428 for example we want the value of 'q' or implement X9.42 DH.
11431 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11432 provides hooks that allow the default DSA functions or functions on a
11433 "per key" basis to be replaced. This allows hardware acceleration and
11434 hardware key storage to be handled without major modification to the
11435 library. Also added low level modexp hooks and CRYPTO_EX structure and
11436 associated functions.
11439 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11440 as "read only": it can't be written to and the buffer it points to will
11441 not be freed. Reading from a read only BIO is much more efficient than
11442 a normal memory BIO. This was added because there are several times when
11443 an area of memory needs to be read from a BIO. The previous method was
11444 to create a memory BIO and write the data to it, this results in two
11445 copies of the data and an O(n^2) reading algorithm. There is a new
11446 function BIO_new_mem_buf() which creates a read only memory BIO from
11447 an area of memory. Also modified the PKCS#7 routines to use read only
11451 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11452 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11453 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11454 but a retry condition occurred while trying to read the rest.
11457 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11458 NID_pkcs7_encrypted by default: this was wrong since this should almost
11459 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11460 the encrypted data type: this is a more sensible place to put it and it
11461 allows the PKCS#12 code to be tidied up that duplicated this
11465 *) Changed obj_dat.pl script so it takes its input and output files on
11466 the command line. This should avoid shell escape redirection problems
11470 *) Initial support for certificate extension requests, these are included
11471 in things like Xenroll certificate requests. Included functions to allow
11472 extensions to be obtained and added.
11475 *) -crlf option to s_client and s_server for sending newlines as
11476 CRLF (as required by many protocols).
11479 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11481 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11482 [Ralf S. Engelschall]
11484 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11485 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11487 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11491 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11492 DH parameters/keys (q is lost during that conversion, but the resulting
11493 DH parameters contain its length).
11495 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11496 much faster than DH_generate_parameters (which creates parameters
11497 where p = 2*q + 1), and also the smaller q makes DH computations
11498 much more efficient (160-bit exponentiation instead of 1024-bit
11499 exponentiation); so this provides a convenient way to support DHE
11500 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11501 utter importance to use
11502 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11504 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11505 when such DH parameters are used, because otherwise small subgroup
11506 attacks may become possible!
11509 *) Avoid memory leak in i2d_DHparams.
11512 *) Allow the -k option to be used more than once in the enc program:
11513 this allows the same encrypted message to be read by multiple recipients.
11516 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11517 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11518 it will always use the numerical form of the OID, even if it has a short
11522 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11523 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11524 otherwise bn_mod_exp was called. In the case of hardware keys for example
11525 no private key components need be present and it might store extra data
11526 in the RSA structure, which cannot be accessed from bn_mod_exp.
11527 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11528 private key operations.
11531 *) Added support for SPARC Linux.
11534 *) pem_password_cb function type incompatibly changed from
11535 typedef int pem_password_cb(char *buf, int size, int rwflag);
11537 ....(char *buf, int size, int rwflag, void *userdata);
11538 so that applications can pass data to their callbacks:
11539 The PEM[_ASN1]_{read,write}... functions and macros now take an
11540 additional void * argument, which is just handed through whenever
11541 the password callback is called.
11542 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11544 New function SSL_CTX_set_default_passwd_cb_userdata.
11546 Compatibility note: As many C implementations push function arguments
11547 onto the stack in reverse order, the new library version is likely to
11548 interoperate with programs that have been compiled with the old
11549 pem_password_cb definition (PEM_whatever takes some data that
11550 happens to be on the stack as its last argument, and the callback
11551 just ignores this garbage); but there is no guarantee whatsoever that
11554 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11555 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11556 problems not only on Windows, but also on some Unix platforms.
11557 To avoid problematic command lines, these definitions are now in an
11558 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11559 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11562 *) MIPS III/IV assembler module is reimplemented.
11565 *) More DES library cleanups: remove references to srand/rand and
11566 delete an unused file.
11569 *) Add support for the free Netwide assembler (NASM) under Win32,
11570 since not many people have MASM (ml) and it can be hard to obtain.
11571 This is currently experimental but it seems to work OK and pass all
11572 the tests. Check out INSTALL.W32 for info.
11575 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11576 without temporary keys kept an extra copy of the server key,
11577 and connections with temporary keys did not free everything in case
11581 *) New function RSA_check_key and new openssl rsa option -check
11582 for verifying the consistency of RSA keys.
11583 [Ulf Moeller, Bodo Moeller]
11585 *) Various changes to make Win32 compile work:
11586 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11587 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11588 comparison" warnings.
11589 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11592 *) Add a debugging option to PKCS#5 v2 key generation function: when
11593 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11594 derived keys are printed to stderr.
11597 *) Copy the flags in ASN1_STRING_dup().
11598 [Roman E. Pavlov <pre@mo.msk.ru>]
11600 *) The x509 application mishandled signing requests containing DSA
11601 keys when the signing key was also DSA and the parameters didn't match.
11603 It was supposed to omit the parameters when they matched the signing key:
11604 the verifying software was then supposed to automatically use the CA's
11605 parameters if they were absent from the end user certificate.
11607 Omitting parameters is no longer recommended. The test was also
11608 the wrong way round! This was probably due to unusual behaviour in
11609 EVP_cmp_parameters() which returns 1 if the parameters match.
11610 This meant that parameters were omitted when they *didn't* match and
11611 the certificate was useless. Certificates signed with 'ca' didn't have
11613 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11615 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11616 The interface is as follows:
11617 Applications can use
11618 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11619 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11620 "off" is now the default.
11621 The library internally uses
11622 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11623 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11624 to disable memory-checking temporarily.
11626 Some inconsistent states that previously were possible (and were
11627 even the default) are now avoided.
11629 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11630 with each memory chunk allocated; this is occasionally more helpful
11631 than just having a counter.
11633 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11635 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11639 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11640 which largely parallels "options", but is for changing API behaviour,
11641 whereas "options" are about protocol behaviour.
11642 Initial "mode" flags are:
11644 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11645 a single record has been written.
11646 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11647 retries use the same buffer location.
11648 (But all of the contents must be
11652 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11655 *) Fix problems with no-hmac etc.
11656 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11658 *) New functions RSA_get_default_method(), RSA_set_method() and
11659 RSA_get_method(). These allows replacement of RSA_METHODs without having
11660 to mess around with the internals of an RSA structure.
11663 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11664 Also really enable memory leak checks in openssl.c and in some
11666 [Chad C. Mulligan, Bodo Moeller]
11668 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11669 up the length of negative integers. This has now been simplified to just
11670 store the length when it is first determined and use it later, rather
11671 than trying to keep track of where data is copied and updating it to
11673 [Steve Henson, reported by Brien Wheeler
11674 <bwheeler@authentica-security.com>]
11676 *) Add a new function PKCS7_signatureVerify. This allows the verification
11677 of a PKCS#7 signature but with the signing certificate passed to the
11678 function itself. This contrasts with PKCS7_dataVerify which assumes the
11679 certificate is present in the PKCS#7 structure. This isn't always the
11680 case: certificates can be omitted from a PKCS#7 structure and be
11681 distributed by "out of band" means (such as a certificate database).
11684 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11685 function prototypes in pem.h, also change util/mkdef.pl to add the
11686 necessary function names.
11689 *) mk1mf.pl (used by Windows builds) did not properly read the
11690 options set by Configure in the top level Makefile, and Configure
11691 was not even able to write more than one option correctly.
11692 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11695 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11696 file to be loaded from a BIO or FILE pointer. The BIO version will
11697 for example allow memory BIOs to contain config info.
11700 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11701 Whoever hopes to achieve shared-library compatibility across versions
11702 must use this, not the compile-time macro.
11703 (Exercise 0.9.4: Which is the minimum library version required by
11705 Note: All this applies only to multi-threaded programs, others don't
11709 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11710 through a BIO pair triggered the default case, i.e.
11711 SSLerr(...,SSL_R_UNKNOWN_STATE).
11714 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11715 can use the SSL library even if none of the specific BIOs is
11719 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11720 for the encoded length.
11721 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11723 *) Add initial documentation of the X509V3 functions.
11726 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11727 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11728 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11729 secure PKCS#8 private key format with a high iteration count.
11732 *) Fix determination of Perl interpreter: A perl or perl5
11733 _directory_ in $PATH was also accepted as the interpreter.
11734 [Ralf S. Engelschall]
11736 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11737 wrong with it but it was very old and did things like calling
11738 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11739 unusual formatting.
11742 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11743 to use the new extension code.
11746 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11747 with macros. This should make it easier to change their form, add extra
11748 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11752 *) Add to configuration table a new entry that can specify an alternative
11753 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11754 according to Mark Crispin <MRC@Panda.COM>.
11758 *) DES CBC did not update the IV. Weird.
11761 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11762 Changing the behaviour of the former might break existing programs --
11763 where IV updating is needed, des_ncbc_encrypt can be used.
11766 *) When bntest is run from "make test" it drives bc to check its
11767 calculations, as well as internally checking them. If an internal check
11768 fails, it needs to cause bc to give a non-zero result or make test carries
11769 on without noticing the failure. Fixed.
11772 *) DES library cleanups.
11775 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11776 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11777 ciphers. NOTE: although the key derivation function has been verified
11778 against some published test vectors it has not been extensively tested
11779 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11783 *) Instead of "mkdir -p", which is not fully portable, use new
11784 Perl script "util/mkdir-p.pl".
11787 *) Rewrite the way password based encryption (PBE) is handled. It used to
11788 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11789 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11790 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11791 the 'parameter' field of the AlgorithmIdentifier is passed to the
11792 underlying key generation function so it must do its own ASN1 parsing.
11793 This has also changed the EVP_PBE_CipherInit() function which now has a
11794 'parameter' argument instead of literal salt and iteration count values
11795 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11798 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11799 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11800 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11801 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11802 value was just used as a "magic string" and not used directly its
11803 value doesn't matter.
11806 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11810 *) "linux-sparc64" configuration (ultrapenguin).
11811 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11812 "linux-sparc" configuration.
11813 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11815 *) config now generates no-xxx options for missing ciphers.
11818 *) Support the EBCDIC character set (work in progress).
11819 File ebcdic.c not yet included because it has a different license.
11820 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11822 *) Support BS2000/OSD-POSIX.
11823 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11825 *) Make callbacks for key generation use void * instead of char *.
11828 *) Make S/MIME samples compile (not yet tested).
11831 *) Additional typesafe stacks.
11834 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11838 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11840 *) New configuration variant "sco5-gcc".
11842 *) Updated some demos.
11843 [Sean O Riordain, Wade Scholine]
11845 *) Add missing BIO_free at exit of pkcs12 application.
11848 *) Fix memory leak in conf.c.
11851 *) Updates for Win32 to assembler version of MD5.
11854 *) Set #! path to perl in apps/der_chop to where we found it
11855 instead of using a fixed path.
11858 *) SHA library changes for irix64-mips4-cc.
11861 *) Improvements for VMS support.
11865 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11867 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11868 This also avoids the problems with SC4.2 and unpatched SC5.
11869 [Andy Polyakov <appro@fy.chalmers.se>]
11871 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11872 These are required because of the typesafe stack would otherwise break
11873 existing code. If old code used a structure member which used to be STACK
11874 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11875 sk_num or sk_value it would produce an error because the num, data members
11876 are not present in STACK_OF. Now it just produces a warning. sk_set
11877 replaces the old method of assigning a value to sk_value
11878 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11879 that does this will no longer work (and should use sk_set instead) but
11880 this could be regarded as a "questionable" behaviour anyway.
11883 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11884 correctly handle encrypted S/MIME data.
11887 *) Change type of various DES function arguments from des_cblock
11888 (which means, in function argument declarations, pointer to char)
11889 to des_cblock * (meaning pointer to array with 8 char elements),
11890 which allows the compiler to do more typechecking; it was like
11891 that back in SSLeay, but with lots of ugly casts.
11893 Introduce new type const_des_cblock.
11896 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11897 problems: find RecipientInfo structure that matches recipient certificate
11898 and initialise the ASN1 structures properly based on passed cipher.
11901 *) Belatedly make the BN tests actually check the results.
11904 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11905 to and from BNs: it was completely broken. New compilation option
11906 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11907 key elements as negative integers.
11910 *) Reorganize and speed up MD5.
11911 [Andy Polyakov <appro@fy.chalmers.se>]
11914 [Richard Levitte <richard@levitte.org>]
11916 *) New option -out to asn1parse to allow the parsed structure to be
11917 output to a file. This is most useful when combined with the -strparse
11918 option to examine the output of things like OCTET STRINGS.
11921 *) Make SSL library a little more fool-proof by not requiring any longer
11922 that SSL_set_{accept,connect}_state be called before
11923 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11924 in many applications because usually everything *appeared* to work as
11925 intended anyway -- now it really works as intended).
11928 *) Move openssl.cnf out of lib/.
11931 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11932 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11933 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11934 [Ralf S. Engelschall]
11936 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11937 handle PKCS#7 enveloped data properly.
11938 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11940 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11941 copying pointers. The cert_st handling is changed by this in
11942 various ways (and thus what used to be known as ctx->default_cert
11943 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11944 any longer when s->cert does not give us what we need).
11945 ssl_cert_instantiate becomes obsolete by this change.
11946 As soon as we've got the new code right (possibly it already is?),
11947 we have solved a couple of bugs of the earlier code where s->cert
11948 was used as if it could not have been shared with other SSL structures.
11950 Note that using the SSL API in certain dirty ways now will result
11951 in different behaviour than observed with earlier library versions:
11952 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11953 does not influence s as it used to.
11955 In order to clean up things more thoroughly, inside SSL_SESSION
11956 we don't use CERT any longer, but a new structure SESS_CERT
11957 that holds per-session data (if available); currently, this is
11958 the peer's certificate chain and, for clients, the server's certificate
11959 and temporary key. CERT holds only those values that can have
11960 meaningful defaults in an SSL_CTX.
11963 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11964 from the internal representation. Various PKCS#7 fixes: remove some
11965 evil casts and set the enc_dig_alg field properly based on the signing
11969 *) Allow PKCS#12 password to be set from the command line or the
11970 environment. Let 'ca' get its config file name from the environment
11971 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11975 *) Allow certificate policies extension to use an IA5STRING for the
11976 organization field. This is contrary to the PKIX definition but
11977 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11981 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11982 without disallowing inline assembler and the like for non-pedantic builds.
11985 *) Support Borland C++ builder.
11986 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11988 *) Support Mingw32.
11991 *) SHA-1 cleanups and performance enhancements.
11992 [Andy Polyakov <appro@fy.chalmers.se>]
11994 *) Sparc v8plus assembler for the bignum library.
11995 [Andy Polyakov <appro@fy.chalmers.se>]
11997 *) Accept any -xxx and +xxx compiler options in Configure.
12000 *) Update HPUX configuration.
12003 *) Add missing sk_<type>_unshift() function to safestack.h
12004 [Ralf S. Engelschall]
12006 *) New function SSL_CTX_use_certificate_chain_file that sets the
12007 "extra_cert"s in addition to the certificate. (This makes sense
12008 only for "PEM" format files, as chains as a whole are not
12012 *) Support verify_depth from the SSL API.
12013 x509_vfy.c had what can be considered an off-by-one-error:
12014 Its depth (which was not part of the external interface)
12015 was actually counting the number of certificates in a chain;
12016 now it really counts the depth.
12019 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12020 instead of X509err, which often resulted in confusing error
12021 messages since the error codes are not globally unique
12022 (e.g. an alleged error in ssl3_accept when a certificate
12023 didn't match the private key).
12025 *) New function SSL_CTX_set_session_id_context that allows to set a default
12026 value (so that you don't need SSL_set_session_id_context for each
12027 connection using the SSL_CTX).
12030 *) OAEP decoding bug fix.
12033 *) Support INSTALL_PREFIX for package builders, as proposed by
12037 *) New Configure options "threads" and "no-threads". For systems
12038 where the proper compiler options are known (currently Solaris
12039 and Linux), "threads" is the default.
12042 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12045 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12046 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12047 such as /usr/local/bin.
12050 *) "make linux-shared" to build shared libraries.
12051 [Niels Poppe <niels@netbox.org>]
12053 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12056 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12057 extension adding in x509 utility.
12060 *) Remove NOPROTO sections and error code comments.
12063 *) Partial rewrite of the DEF file generator to now parse the ANSI
12067 *) New Configure options --prefix=DIR and --openssldir=DIR.
12070 *) Complete rewrite of the error code script(s). It is all now handled
12071 by one script at the top level which handles error code gathering,
12072 header rewriting and C source file generation. It should be much better
12073 than the old method: it now uses a modified version of Ulf's parser to
12074 read the ANSI prototypes in all header files (thus the old K&R definitions
12075 aren't needed for error creation any more) and do a better job of
12076 translating function codes into names. The old 'ASN1 error code imbedded
12077 in a comment' is no longer necessary and it doesn't use .err files which
12078 have now been deleted. Also the error code call doesn't have to appear all
12079 on one line (which resulted in some large lines...).
12082 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12085 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12086 0 (which usually indicates a closed connection), but continue reading.
12089 *) Fix some race conditions.
12092 *) Add support for CRL distribution points extension. Add Certificate
12093 Policies and CRL distribution points documentation.
12096 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12099 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12100 8 of keying material. Merlin has also confirmed interop with this fix
12101 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12102 [Merlin Hughes <merlin@baltimore.ie>]
12104 *) Fix lots of warnings.
12105 [Richard Levitte <levitte@stacken.kth.se>]
12107 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12108 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12109 [Richard Levitte <levitte@stacken.kth.se>]
12111 *) Fix problems with sizeof(long) == 8.
12112 [Andy Polyakov <appro@fy.chalmers.se>]
12114 *) Change functions to ANSI C.
12117 *) Fix typos in error codes.
12118 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12120 *) Remove defunct assembler files from Configure.
12123 *) SPARC v8 assembler BIGNUM implementation.
12124 [Andy Polyakov <appro@fy.chalmers.se>]
12126 *) Support for Certificate Policies extension: both print and set.
12127 Various additions to support the r2i method this uses.
12130 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12131 return a const string when you are expecting an allocated buffer.
12134 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12135 types DirectoryString and DisplayText.
12138 *) Add code to allow r2i extensions to access the configuration database,
12139 add an LHASH database driver and add several ctx helper functions.
12142 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12143 fail when they extended the size of a BIGNUM.
12146 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12147 support typesafe stack.
12150 *) Fix typo in SSL_[gs]et_options().
12151 [Nils Frostberg <nils@medcom.se>]
12153 *) Delete various functions and files that belonged to the (now obsolete)
12154 old X509V3 handling code.
12157 *) New Configure option "rsaref".
12160 *) Don't auto-generate pem.h.
12163 *) Introduce type-safe ASN.1 SETs.
12166 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12167 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12169 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12170 that links with OpenSSL (well at least cause lots of warnings), but fear
12171 not: the conversion is trivial, and it eliminates loads of evil casts. A
12172 few STACKed things have been converted already. Feel free to convert more.
12173 In the fullness of time, I'll do away with the STACK type altogether.
12176 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12177 specified in <certfile> by updating the entry in the index.txt file.
12178 This way one no longer has to edit the index.txt file manually for
12179 revoking a certificate. The -revoke option does the gory details now.
12180 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12182 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12183 `-text' option at all and this way the `-noout -text' combination was
12184 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12185 [Ralf S. Engelschall]
12187 *) Make sure a corresponding plain text error message exists for the
12188 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12189 verify callback function determined that a certificate was revoked.
12190 [Ralf S. Engelschall]
12192 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12193 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12194 all available ciphers including rc5, which was forgotten until now.
12195 In order to let the testing shell script know which algorithms
12196 are available, a new (up to now undocumented) command
12197 "openssl list-cipher-commands" is used.
12200 *) Bugfix: s_client occasionally would sleep in select() when
12201 it should have checked SSL_pending() first.
12204 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12205 the raw DSA values prior to ASN.1 encoding.
12208 *) Tweaks to Configure
12209 [Niels Poppe <niels@netbox.org>]
12211 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12215 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12218 *) New config option to avoid instructions that are illegal on the 80386.
12219 The default code is faster, but requires at least a 486.
12222 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12223 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12224 same as SSL2_VERSION anyway.
12227 *) New "-showcerts" option for s_client.
12230 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12231 application. Various cleanups and fixes.
12234 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12235 modify error routines to work internally. Add error codes and PBE init
12236 to library startup routines.
12239 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12240 packing functions to asn1 and evp. Changed function names and error
12241 codes along the way.
12244 *) PKCS12 integration: and so it begins... First of several patches to
12245 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12246 objects to objects.h
12249 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12250 and display support for Thawte strong extranet extension.
12253 *) Add LinuxPPC support.
12254 [Jeff Dubrule <igor@pobox.org>]
12256 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12257 bn_div_words in alpha.s.
12258 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12260 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12261 OAEP isn't supported when OpenSSL is built with RSAref.
12262 [Ulf Moeller <ulf@fitug.de>]
12264 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12265 so they no longer are missing under -DNOPROTO.
12266 [Soren S. Jorvang <soren@t.dk>]
12269 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12271 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12272 doesn't work when the session is reused. Coming soon!
12275 *) Fix a security hole, that allows sessions to be reused in the wrong
12276 context thus bypassing client cert protection! All software that uses
12277 client certs and session caches in multiple contexts NEEDS PATCHING to
12278 allow session reuse! A fuller solution is in the works.
12279 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12281 *) Some more source tree cleanups (removed obsolete files
12282 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12283 permission on "config" script to be executable) and a fix for the INSTALL
12285 [Ulf Moeller <ulf@fitug.de>]
12287 *) Remove some legacy and erroneous uses of malloc, free instead of
12289 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12291 *) Make rsa_oaep_test return non-zero on error.
12292 [Ulf Moeller <ulf@fitug.de>]
12294 *) Add support for native Solaris shared libraries. Configure
12295 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12296 if someone would make that last step automatic.
12297 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12299 *) ctx_size was not built with the right compiler during "make links". Fixed.
12302 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12303 except NULL ciphers". This means the default cipher list will no longer
12304 enable NULL ciphers. They need to be specifically enabled e.g. with
12305 the string "DEFAULT:eNULL".
12308 *) Fix to RSA private encryption routines: if p < q then it would
12309 occasionally produce an invalid result. This will only happen with
12310 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12313 *) Be less restrictive and allow also `perl util/perlpath.pl
12314 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12315 because this way one can also use an interpreter named `perl5' (which is
12316 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12317 installed as `perl').
12318 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12320 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12321 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12323 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12324 advapi32.lib to Win32 build and change the pem test comparison
12325 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12326 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12327 and crypto/des/ede_cbcm_enc.c.
12330 *) DES quad checksum was broken on big-endian architectures. Fixed.
12333 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12334 Win32 test batch file so it (might) work again. The Win32 test batch file
12335 is horrible: I feel ill....
12338 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12339 in e_os.h. Audit of header files to check ANSI and non ANSI
12340 sections: 10 functions were absent from non ANSI section and not exported
12341 from Windows DLLs. Fixed up libeay.num for new functions.
12344 *) Make `openssl version' output lines consistent.
12345 [Ralf S. Engelschall]
12347 *) Fix Win32 symbol export lists for BIO functions: Added
12348 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12349 to ms/libeay{16,32}.def.
12350 [Ralf S. Engelschall]
12352 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12353 fine under Unix and passes some trivial tests I've now added. But the
12354 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12355 added to make sure no one expects that this stuff really works in the
12356 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12357 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12359 [Ralf S. Engelschall]
12361 *) Fix the generation of two part addresses in perl.
12362 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12364 *) Add config entry for Linux on MIPS.
12365 [John Tobey <jtobey@channel1.com>]
12367 *) Make links whenever Configure is run, unless we are on Windoze.
12370 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12371 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12375 *) Add a useful kludge to allow package maintainers to specify compiler and
12376 other platforms details on the command line without having to patch the
12377 Configure script everytime: One now can use ``perl Configure
12378 <id>:<details>'', i.e. platform ids are allowed to have details appended
12379 to them (separated by colons). This is treated as there would be a static
12380 pre-configured entry in Configure's %table under key <id> with value
12381 <details> and ``perl Configure <id>'' is called. So, when you want to
12382 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12383 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12384 now, which overrides the FreeBSD-elf entry on-the-fly.
12385 [Ralf S. Engelschall]
12387 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12390 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12391 on the `perl Configure ...' command line. This way one can compile
12392 OpenSSL libraries with Position Independent Code (PIC) which is needed
12393 for linking it into DSOs.
12394 [Ralf S. Engelschall]
12396 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12400 *) Cleaned up the LICENSE document: The official contact for any license
12401 questions now is the OpenSSL core team under openssl-core@openssl.org.
12402 And add a paragraph about the dual-license situation to make sure people
12403 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12404 to the OpenSSL toolkit.
12405 [Ralf S. Engelschall]
12407 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12408 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12409 Additionally cleaned up the `make links' target: Remove unnecessary
12410 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12411 to speed processing and no longer clutter the display with confusing
12412 stuff. Instead only the actually done links are displayed.
12413 [Ralf S. Engelschall]
12415 *) Permit null encryption ciphersuites, used for authentication only. It used
12416 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12417 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12421 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12422 signed attributes when verifying signatures (this would break them),
12423 the detached data encoding was wrong and public keys obtained using
12424 X509_get_pubkey() weren't freed.
12427 *) Add text documentation for the BUFFER functions. Also added a work around
12428 to a Win95 console bug. This was triggered by the password read stuff: the
12429 last character typed gets carried over to the next fread(). If you were
12430 generating a new cert request using 'req' for example then the last
12431 character of the passphrase would be CR which would then enter the first
12435 *) Added the new `Includes OpenSSL Cryptography Software' button as
12436 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12437 button and can be used by applications based on OpenSSL to show the
12438 relationship to the OpenSSL project.
12439 [Ralf S. Engelschall]
12441 *) Remove confusing variables in function signatures in files
12442 ssl/ssl_lib.c and ssl/ssl.h.
12443 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12445 *) Don't install bss_file.c under PREFIX/include/
12446 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12448 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12449 functions that return function pointers and has support for NT specific
12450 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12451 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12452 unsigned to signed types: this was killing the Win32 compile.
12455 *) Add new certificate file to stack functions,
12456 SSL_add_dir_cert_subjects_to_stack() and
12457 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12458 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12459 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12460 This means that Apache-SSL and similar packages don't have to mess around
12461 to add as many CAs as they want to the preferred list.
12464 *) Experiment with doxygen documentation. Currently only partially applied to
12466 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12467 openssl.doxy as the configuration file.
12470 *) Get rid of remaining C++-style comments which strict C compilers hate.
12471 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12473 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12474 compiled in by default: it has problems with large keys.
12477 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12478 DH private keys and/or callback functions which directly correspond to
12479 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12480 is needed for applications which have to configure certificates on a
12481 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12483 For the RSA certificate situation is makes no difference, but
12484 for the DSA certificate situation this fixes the "no shared cipher"
12485 problem where the OpenSSL cipher selection procedure failed because the
12486 temporary keys were not overtaken from the context and the API provided
12487 no way to reconfigure them.
12488 The new functions now let applications reconfigure the stuff and they
12489 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12490 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12491 non-public-API function ssl_cert_instantiate() is used as a helper
12492 function and also to reduce code redundancy inside ssl_rsa.c.
12493 [Ralf S. Engelschall]
12495 *) Move s_server -dcert and -dkey options out of the undocumented feature
12496 area because they are useful for the DSA situation and should be
12497 recognized by the users.
12498 [Ralf S. Engelschall]
12500 *) Fix the cipher decision scheme for export ciphers: the export bits are
12501 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12502 SSL_EXP_MASK. So, the original variable has to be used instead of the
12503 already masked variable.
12504 [Richard Levitte <levitte@stacken.kth.se>]
12506 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12507 [Richard Levitte <levitte@stacken.kth.se>]
12509 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12510 from `int' to `unsigned int' because it's a length and initialized by
12511 EVP_DigestFinal() which expects an `unsigned int *'.
12512 [Richard Levitte <levitte@stacken.kth.se>]
12514 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12515 script. Instead use the usual Shell->Perl transition trick.
12516 [Ralf S. Engelschall]
12518 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12519 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12520 -noout -modulus' as it's already the case for `openssl rsa -noout
12521 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12522 currently the public key is printed (a decision which was already done by
12523 `openssl dsa -modulus' in the past) which serves a similar purpose.
12524 Additionally the NO_RSA no longer completely removes the whole -modulus
12525 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12527 [Ralf S. Engelschall]
12529 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12530 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12531 [Arne Ansper <arne@ats.cyber.ee>]
12533 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12534 to be added. Now both 'req' and 'ca' can use new objects defined in the
12538 *) Add cool BIO that does syslog (or event log on NT).
12539 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12541 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12542 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12543 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12544 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12547 *) Add preliminary config info for new extension code.
12550 *) Make RSA_NO_PADDING really use no padding.
12551 [Ulf Moeller <ulf@fitug.de>]
12553 *) Generate errors when private/public key check is done.
12556 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12557 for some CRL extensions and new objects added.
12560 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12561 key usage extension and fuller support for authority key id.
12564 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12565 padding method for RSA, which is recommended for new applications in PKCS
12566 #1 v2.0 (RFC 2437, October 1998).
12567 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12568 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12569 against Bleichbacher's attack on RSA.
12570 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12573 *) Updates to the new SSL compression code
12574 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12576 *) Fix so that the version number in the master secret, when passed
12577 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12578 (because the server will not accept higher), that the version number
12579 is 0x03,0x01, not 0x03,0x00
12580 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12582 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12583 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12584 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12587 *) Support for RAW extensions where an arbitrary extension can be
12588 created by including its DER encoding. See apps/openssl.cnf for
12592 *) Make sure latest Perl versions don't interpret some generated C array
12593 code as Perl array code in the crypto/err/err_genc.pl script.
12594 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12596 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12597 not many people have the assembler. Various Win32 compilation fixes and
12598 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12599 build instructions.
12602 *) Modify configure script 'Configure' to automatically create crypto/date.h
12603 file under Win32 and also build pem.h from pem.org. New script
12604 util/mkfiles.pl to create the MINFO file on environments that can't do a
12605 'make files': perl util/mkfiles.pl >MINFO should work.
12608 *) Major rework of DES function declarations, in the pursuit of correctness
12609 and purity. As a result, many evil casts evaporated, and some weirdness,
12610 too. You may find this causes warnings in your code. Zapping your evil
12611 casts will probably fix them. Mostly.
12614 *) Fix for a typo in asn1.h. Bug fix to object creation script
12615 obj_dat.pl. It considered a zero in an object definition to mean
12616 "end of object": none of the objects in objects.h have any zeros
12617 so it wasn't spotted.
12618 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12620 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12621 Masking (CBCM). In the absence of test vectors, the best I have been able
12622 to do is check that the decrypt undoes the encrypt, so far. Send me test
12623 vectors if you have them.
12626 *) Correct calculation of key length for export ciphers (too much space was
12627 allocated for null ciphers). This has not been tested!
12630 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12631 message is now correct (it understands "crypto" and "ssl" on its
12632 command line). There is also now an "update" option. This will update
12633 the util/ssleay.num and util/libeay.num files with any new functions.
12635 perl util/mkdef.pl crypto ssl update
12636 it will update them.
12639 *) Overhauled the Perl interface (perl/*):
12640 - ported BN stuff to OpenSSL's different BN library
12641 - made the perl/ source tree CVS-aware
12642 - renamed the package from SSLeay to OpenSSL (the files still contain
12643 their history because I've copied them in the repository)
12644 - removed obsolete files (the test scripts will be replaced
12645 by better Test::Harness variants in the future)
12646 [Ralf S. Engelschall]
12648 *) First cut for a very conservative source tree cleanup:
12649 1. merge various obsolete readme texts into doc/ssleay.txt
12650 where we collect the old documents and readme texts.
12651 2. remove the first part of files where I'm already sure that we no
12652 longer need them because of three reasons: either they are just temporary
12653 files which were left by Eric or they are preserved original files where
12654 I've verified that the diff is also available in the CVS via "cvs diff
12655 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12656 the crypto/md/ stuff).
12657 [Ralf S. Engelschall]
12659 *) More extension code. Incomplete support for subject and issuer alt
12660 name, issuer and authority key id. Change the i2v function parameters
12661 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12662 what that's for :-) Fix to ASN1 macro which messed up
12663 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12666 *) Preliminary support for ENUMERATED type. This is largely copied from the
12670 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12671 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12673 *) Make sure `make rehash' target really finds the `openssl' program.
12674 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12676 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12677 like to hear about it if this slows down other processors.
12680 *) Add CygWin32 platform information to Configure script.
12681 [Alan Batie <batie@aahz.jf.intel.com>]
12683 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12684 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12686 *) New program nseq to manipulate netscape certificate sequences
12689 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12693 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12694 but the BN code had some problems that would cause failures when
12695 doing certificate verification and some other functions.
12696 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12698 *) Add ASN1 and PEM code to support netscape certificate sequences.
12701 *) Add ASN1 and PEM code to support netscape certificate sequences.
12704 *) Add several PKIX and private extended key usage OIDs.
12707 *) Modify the 'ca' program to handle the new extension code. Modify
12708 openssl.cnf for new extension format, add comments.
12711 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12712 and add a sample to openssl.cnf so req -x509 now adds appropriate
12716 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12717 error code, add initial support to X509_print() and x509 application.
12720 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12721 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12722 stuff is currently isolated and isn't even compiled yet.
12725 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12726 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12727 Removed the versions check from X509 routines when loading extensions:
12728 this allows certain broken certificates that don't set the version
12729 properly to be processed.
12732 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12733 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12734 can still be regenerated with "make depend".
12737 *) Spelling mistake in C version of CAST-128.
12738 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12740 *) Changes to the error generation code. The perl script err-code.pl
12741 now reads in the old error codes and retains the old numbers, only
12742 adding new ones if necessary. It also only changes the .err files if new
12743 codes are added. The makefiles have been modified to only insert errors
12744 when needed (to avoid needlessly modifying header files). This is done
12745 by only inserting errors if the .err file is newer than the auto generated
12746 C file. To rebuild all the error codes from scratch (the old behaviour)
12747 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12748 or delete all the .err files.
12751 *) CAST-128 was incorrectly implemented for short keys. The C version has
12752 been fixed, but is untested. The assembler versions are also fixed, but
12753 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12754 to regenerate it if needed.
12755 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12756 Hagino <itojun@kame.net>]
12758 *) File was opened incorrectly in randfile.c.
12759 [Ulf Möller <ulf@fitug.de>]
12761 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12762 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12763 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12764 al: it's just almost always a UTCTime. Note this patch adds new error
12765 codes so do a "make errors" if there are problems.
12768 *) Correct Linux 1 recognition in config.
12769 [Ulf Möller <ulf@fitug.de>]
12771 *) Remove pointless MD5 hash when using DSA keys in ca.
12772 [Anonymous <nobody@replay.com>]
12774 *) Generate an error if given an empty string as a cert directory. Also
12775 generate an error if handed NULL (previously returned 0 to indicate an
12776 error, but didn't set one).
12777 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12779 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12782 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12783 parameters. This was causing a warning which killed off the Win32 compile.
12786 *) Remove C++ style comments from crypto/bn/bn_local.h.
12787 [Neil Costigan <neil.costigan@celocom.com>]
12789 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12790 based on a text string, looking up short and long names and finally
12791 "dot" format. The "dot" format stuff didn't work. Added new function
12792 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12793 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12794 OID is not part of the table.
12797 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12798 X509_LOOKUP_by_alias().
12801 *) Sort openssl functions by name.
12804 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12805 encryption from sample DSA keys (in case anyone is interested the password
12809 *) Make _all_ *_free functions accept a NULL pointer.
12810 [Frans Heymans <fheymans@isaserver.be>]
12812 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12814 [Anonymous <nobody@replay.com>]
12816 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12817 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12819 *) Don't blow it for numeric -newkey arguments to apps/req.
12820 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12822 *) Temp key "for export" tests were wrong in s3_srvr.c.
12823 [Anonymous <nobody@replay.com>]
12825 *) Add prototype for temp key callback functions
12826 SSL_CTX_set_tmp_{rsa,dh}_callback().
12829 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12830 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12833 *) X509_name_add_entry() freed the wrong thing after an error.
12834 [Arne Ansper <arne@ats.cyber.ee>]
12836 *) rsa_eay.c would attempt to free a NULL context.
12837 [Arne Ansper <arne@ats.cyber.ee>]
12839 *) BIO_s_socket() had a broken should_retry() on Windoze.
12840 [Arne Ansper <arne@ats.cyber.ee>]
12842 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12843 [Arne Ansper <arne@ats.cyber.ee>]
12845 *) Make sure the already existing X509_STORE->depth variable is initialized
12846 in X509_STORE_new(), but document the fact that this variable is still
12847 unused in the certificate verification process.
12848 [Ralf S. Engelschall]
12850 *) Fix the various library and apps files to free up pkeys obtained from
12851 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12854 *) Fix reference counting in X509_PUBKEY_get(). This makes
12855 demos/maurice/example2.c work, amongst others, probably.
12856 [Steve Henson and Ben Laurie]
12858 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12859 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12860 are no longer created. This way we have a single and consistent command
12861 line interface `openssl <command>', similar to `cvs <command>'.
12862 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12864 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12865 BIT STRING wrapper always have zero unused bits.
12868 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12871 *) Make the top-level INSTALL documentation easier to understand.
12874 *) Makefiles updated to exit if an error occurs in a sub-directory
12875 make (including if user presses ^C) [Paul Sutton]
12877 *) Make Montgomery context stuff explicit in RSA data structure.
12880 *) Fix build order of pem and err to allow for generated pem.h.
12883 *) Fix renumbering bug in X509_NAME_delete_entry().
12886 *) Enhanced the err-ins.pl script so it makes the error library number
12887 global and can add a library name. This is needed for external ASN1 and
12888 other error libraries.
12891 *) Fixed sk_insert which never worked properly.
12894 *) Fix ASN1 macros so they can handle indefinite length constructed
12895 EXPLICIT tags. Some non standard certificates use these: they can now
12899 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12900 into a single doc/ssleay.txt bundle. This way the information is still
12901 preserved but no longer messes up this directory. Now it's new room for
12902 the new set of documentation files.
12903 [Ralf S. Engelschall]
12905 *) SETs were incorrectly DER encoded. This was a major pain, because they
12906 shared code with SEQUENCEs, which aren't coded the same. This means that
12907 almost everything to do with SETs or SEQUENCEs has either changed name or
12908 number of arguments.
12909 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12911 *) Fix test data to work with the above.
12914 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12915 was already fixed by Eric for 0.9.1 it seems.
12916 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12918 *) Autodetect FreeBSD3.
12921 *) Fix various bugs in Configure. This affects the following platforms:
12925 unixware-2.0-pentium
12929 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12930 before they are needed.
12933 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12937 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12939 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12940 changed SSLeay to OpenSSL in version strings.
12941 [Ralf S. Engelschall]
12943 *) Some fixups to the top-level documents.
12946 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12947 because the symlink to include/ was missing.
12948 [Ralf S. Engelschall]
12950 *) Incorporated the popular no-RSA/DSA-only patches
12951 which allow to compile a RSA-free SSLeay.
12952 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12954 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12955 when "ssleay" is still not found.
12956 [Ralf S. Engelschall]
12958 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12959 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12961 *) Updated the README file.
12962 [Ralf S. Engelschall]
12964 *) Added various .cvsignore files in the CVS repository subdirs
12965 to make a "cvs update" really silent.
12966 [Ralf S. Engelschall]
12968 *) Recompiled the error-definition header files and added
12969 missing symbols to the Win32 linker tables.
12970 [Ralf S. Engelschall]
12972 *) Cleaned up the top-level documents;
12973 o new files: CHANGES and LICENSE
12974 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12975 o merged COPYRIGHT into LICENSE
12976 o removed obsolete TODO file
12977 o renamed MICROSOFT to INSTALL.W32
12978 [Ralf S. Engelschall]
12980 *) Removed dummy files from the 0.9.1b source tree:
12981 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12982 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12983 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12984 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12985 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12986 [Ralf S. Engelschall]
12988 *) Added various platform portability fixes.
12991 *) The Genesis of the OpenSSL rpject:
12992 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12993 Young and Tim J. Hudson created while they were working for C2Net until
12995 [The OpenSSL Project]
12998 Changes between 0.9.0b and 0.9.1b [not released]
13000 *) Updated a few CA certificates under certs/
13003 *) Changed some BIGNUM api stuff.
13006 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13007 DGUX x86, Linux Alpha, etc.
13010 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13011 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13015 *) Add -strparse option to asn1pars program which parses nested
13017 [Dr Stephen Henson <shenson@bigfoot.com>]
13019 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13022 *) DSA fix for "ca" program.
13025 *) Added "-genkey" option to "dsaparam" program.
13028 *) Added RIPE MD160 (rmd160) message digest.
13031 *) Added -a (all) option to "ssleay version" command.
13034 *) Added PLATFORM define which is the id given to Configure.
13037 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13040 *) Extended the ASN.1 parser routines.
13043 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13046 *) Added a BN_CTX to the BN library.
13049 *) Fixed the weak key values in DES library
13052 *) Changed API in EVP library for cipher aliases.
13055 *) Added support for RC2/64bit cipher.
13058 *) Converted the lhash library to the crypto/mem.c functions.
13061 *) Added more recognized ASN.1 object ids.
13064 *) Added more RSA padding checks for SSL/TLS.
13067 *) Added BIO proxy/filter functionality.
13070 *) Added extra_certs to SSL_CTX which can be used
13071 send extra CA certificates to the client in the CA cert chain sending
13072 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13075 *) Now Fortezza is denied in the authentication phase because
13076 this is key exchange mechanism is not supported by SSLeay at all.
13079 *) Additional PKCS1 checks.
13082 *) Support the string "TLSv1" for all TLS v1 ciphers.
13085 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13086 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13089 *) Fixed a few memory leaks.
13092 *) Fixed various code and comment typos.
13095 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13096 bytes sent in the client random.
13097 [Edward Bishop <ebishop@spyglass.com>]