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 1.1.1b and 1.1.1c [xx XXX xxxx]
14 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
16 *) Added SCA hardening for modular field inversion in EC_GROUP through
17 a new dedicated field_inv() pointer in EC_METHOD.
18 This also addresses a leakage affecting conversions from projective
19 to affine coordinates.
20 [Billy Bob Brumley, Nicola Tuveri]
22 *) Change the info callback signals for the start and end of a post-handshake
23 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
24 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
25 confused by this and assume that a TLSv1.2 renegotiation has started. This
26 can break KeyUpdate handling. Instead we no longer signal the start and end
27 of a post handshake message exchange (although the messages themselves are
28 still signalled). This could break some applications that were expecting
29 the old signals. However without this KeyUpdate is not usable for many
33 *) Fix a bug in the computation of the endpoint-pair shared secret used
34 by DTLS over SCTP. This breaks interoperability with older versions
35 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
36 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
37 interoperability with such broken implementations. However, enabling
38 this switch breaks interoperability with correct implementations.
40 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
41 re-used X509_PUBKEY object if the second PUBKEY is malformed.
44 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
47 *) Remove the 'dist' target and add a tarball building script. The
48 'dist' target has fallen out of use, and it shouldn't be
49 necessary to configure just to create a source distribution.
52 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
54 *) Timing vulnerability in DSA signature generation
56 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
57 timing side channel attack. An attacker could use variations in the signing
58 algorithm to recover the private key.
60 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
64 *) Timing vulnerability in ECDSA signature generation
66 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
67 timing side channel attack. An attacker could use variations in the signing
68 algorithm to recover the private key.
70 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
74 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
75 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
76 are retained for backwards compatibility.
79 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
80 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
81 of two gigabytes and the error handling improved.
83 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
84 categorized as a normal bug, not a security issue, because the DRBG reseeds
85 automatically and is fully functional even without additional randomness
86 provided by the application.
88 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
90 *) Add a new ClientHello callback. Provides a callback interface that gives
91 the application the ability to adjust the nascent SSL object at the
92 earliest stage of ClientHello processing, immediately after extensions have
93 been collected but before they have been processed. In particular, this
94 callback can adjust the supported TLS versions in response to the contents
98 *) Add SM2 base algorithm support.
101 *) s390x assembly pack: add (improved) hardware-support for the following
102 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
103 aes-cfb/cfb8, aes-ecb.
106 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
107 parameter is no longer accepted, as it leads to a corrupt table. NULL
108 pem_str is reserved for alias entries only.
111 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
112 step for prime curves. The new implementation is based on formulae from
113 differential addition-and-doubling in homogeneous projective coordinates
114 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
115 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
116 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
117 to work in projective coordinates.
118 [Billy Bob Brumley, Nicola Tuveri]
120 *) Change generating and checking of primes so that the error rate of not
121 being prime depends on the intended use based on the size of the input.
122 For larger primes this will result in more rounds of Miller-Rabin.
123 The maximal error rate for primes with more than 1080 bits is lowered
125 [Kurt Roeckx, Annie Yousar]
127 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
130 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
131 moving between systems, and to avoid confusion when a Windows build is
132 done with mingw vs with MSVC. For POSIX installs, there's still a
133 symlink or copy named 'tsget' to avoid that confusion as well.
136 *) Revert blinding in ECDSA sign and instead make problematic addition
137 length-invariant. Switch even to fixed-length Montgomery multiplication.
140 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
141 step for binary curves. The new implementation is based on formulae from
142 differential addition-and-doubling in mixed Lopez-Dahab projective
143 coordinates, modified to independently blind the operands.
144 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
146 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
147 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
148 EC_METHODs to implement their own specialized "ladder step", to take
149 advantage of more favorable coordinate systems or more efficient
150 differential addition-and-doubling algorithms.
151 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
153 *) Modified the random device based seed sources to keep the relevant
154 file descriptors open rather than reopening them on each access.
155 This allows such sources to operate in a chroot() jail without
156 the associated device nodes being available. This behaviour can be
157 controlled using RAND_keep_random_devices_open().
160 *) Numerous side-channel attack mitigations have been applied. This may have
161 performance impacts for some algorithms for the benefit of improved
162 security. Specific changes are noted in this change log by their respective
166 *) AIX shared library support overhaul. Switch to AIX "natural" way of
167 handling shared libraries, which means collecting shared objects of
168 different versions and bitnesses in one common archive. This allows to
169 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
170 doesn't affect the way 3rd party applications are linked, only how
171 multi-version installation is managed.
174 *) Make ec_group_do_inverse_ord() more robust and available to other
175 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
176 mitigations are applied to the fallback BN_mod_inverse().
177 When using this function rather than BN_mod_inverse() directly, new
178 EC cryptosystem implementations are then safer-by-default.
181 *) Add coordinate blinding for EC_POINT and implement projective
182 coordinate blinding for generic prime curves as a countermeasure to
183 chosen point SCA attacks.
184 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
186 *) Add blinding to ECDSA and DSA signatures to protect against side channel
187 attacks discovered by Keegan Ryan (NCC Group).
190 *) Enforce checking in the pkeyutl command line app to ensure that the input
191 length does not exceed the maximum supported digest length when performing
192 a sign, verify or verifyrecover operation.
195 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
196 I/O in combination with something like select() or poll() will hang. This
197 can be turned off again using SSL_CTX_clear_mode().
198 Many applications do not properly handle non-application data records, and
199 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
200 around the problems in those applications, but can also break some.
201 It's recommended to read the manpages about SSL_read(), SSL_write(),
202 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
203 SSL_CTX_set_read_ahead() again.
206 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
207 now allow empty (zero character) pass phrases.
210 *) Apply blinding to binary field modular inversion and remove patent
211 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
214 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
215 binary and prime elliptic curves.
218 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
219 constant time fixed point multiplication.
222 *) Revise elliptic curve scalar multiplication with timing attack
223 defenses: ec_wNAF_mul redirects to a constant time implementation
224 when computing fixed point and variable point multiplication (which
225 in OpenSSL are mostly used with secret scalars in keygen, sign,
226 ECDH derive operations).
227 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
230 *) Updated CONTRIBUTING
233 *) Updated DRBG / RAND to request nonce and additional low entropy
234 randomness from the system.
235 [Matthias St. Pierre]
237 *) Updated 'openssl rehash' to use OpenSSL consistent default.
240 *) Moved the load of the ssl_conf module to libcrypto, which helps
241 loading engines that libssl uses before libssl is initialised.
244 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
247 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
248 [Ingo Schwarze, Rich Salz]
250 *) Added output of accepting IP address and port for 'openssl s_server'
253 *) Added a new API for TLSv1.3 ciphersuites:
254 SSL_CTX_set_ciphersuites()
255 SSL_set_ciphersuites()
258 *) Memory allocation failures consistenly add an error to the error
262 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
263 in libcrypto when run as setuid/setgid.
266 *) Load any config file by default when libssl is used.
269 *) Added new public header file <openssl/rand_drbg.h> and documentation
270 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
271 [Matthias St. Pierre]
273 *) QNX support removed (cannot find contributors to get their approval
274 for the license change).
277 *) TLSv1.3 replay protection for early data has been implemented. See the
278 SSL_read_early_data() man page for further details.
281 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
282 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
283 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
284 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
285 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
286 configuration has been separated out. See the ciphers man page or the
287 SSL_CTX_set_ciphersuites() man page for more information.
290 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
291 in responder mode now supports the new "-multi" option, which
292 spawns the specified number of child processes to handle OCSP
293 requests. The "-timeout" option now also limits the OCSP
294 responder's patience to wait to receive the full client request
295 on a newly accepted connection. Child processes are respawned
296 as needed, and the CA index file is automatically reloaded
297 when changed. This makes it possible to run the "ocsp" responder
298 as a long-running service, making the OpenSSL CA somewhat more
299 feature-complete. In this mode, most diagnostic messages logged
300 after entering the event loop are logged via syslog(3) rather than
304 *) Added support for X448 and Ed448. Heavily based on original work by
308 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
309 objects loaded. This adds the functions OSSL_STORE_expect() and
310 OSSL_STORE_find() as well as needed tools to construct searches and
311 get the search data out of them.
314 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
315 version of OpenSSL should review their configuration settings to ensure
316 that they are still appropriate for TLSv1.3. For further information see:
317 https://wiki.openssl.org/index.php/TLS1.3
320 *) Grand redesign of the OpenSSL random generator
322 The default RAND method now utilizes an AES-CTR DRBG according to
323 NIST standard SP 800-90Ar1. The new random generator is essentially
324 a port of the default random generator from the OpenSSL FIPS 2.0
325 object module. It is a hybrid deterministic random bit generator
326 using an AES-CTR bit stream and which seeds and reseeds itself
327 automatically using trusted system entropy sources.
329 Some of its new features are:
330 o Support for multiple DRBG instances with seed chaining.
331 o The default RAND method makes use of a DRBG.
332 o There is a public and private DRBG instance.
333 o The DRBG instances are fork-safe.
334 o Keep all global DRBG instances on the secure heap if it is enabled.
335 o The public and private DRBG instance are per thread for lock free
337 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
339 *) Changed Configure so it only says what it does and doesn't dump
340 so much data. Instead, ./configdata.pm should be used as a script
341 to display all sorts of configuration data.
344 *) Added processing of "make variables" to Configure.
347 *) Added SHA512/224 and SHA512/256 algorithm support.
350 *) The last traces of Netware support, first removed in 1.1.0, have
354 *) Get rid of Makefile.shared, and in the process, make the processing
355 of certain files (rc.obj, or the .def/.map/.opt files produced from
356 the ordinal files) more visible and hopefully easier to trace and
357 debug (or make silent).
360 *) Make it possible to have environment variable assignments as
361 arguments to config / Configure.
364 *) Add multi-prime RSA (RFC 8017) support.
367 *) Add SM3 implemented according to GB/T 32905-2016
368 [ Jack Lloyd <jack.lloyd@ribose.com>,
369 Ronald Tse <ronald.tse@ribose.com>,
370 Erick Borsboom <erick.borsboom@ribose.com> ]
372 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
373 as documented in RFC6066.
374 Based on a patch from Tomasz Moń
375 [Filipe Raimundo da Silva]
377 *) Add SM4 implemented according to GB/T 32907-2016.
378 [ Jack Lloyd <jack.lloyd@ribose.com>,
379 Ronald Tse <ronald.tse@ribose.com>,
380 Erick Borsboom <erick.borsboom@ribose.com> ]
382 *) Reimplement -newreq-nodes and ERR_error_string_n; the
383 original author does not agree with the license change.
386 *) Add ARIA AEAD TLS support.
389 *) Some macro definitions to support VS6 have been removed. Visual
390 Studio 6 has not worked since 1.1.0
393 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
394 without clearing the errors.
397 *) Add "atfork" functions. If building on a system that without
398 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
399 requirements. The RAND facility now uses/requires this.
405 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
406 not possible to disable entirely. However, it's still possible to
407 disable the console reading UI method, UI_OpenSSL() (use UI_null()
410 To disable, configure with 'no-ui-console'. 'no-ui' is still
411 possible to use as an alias. Check at compile time with the
412 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
413 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
416 *) Add a STORE module, which implements a uniform and URI based reader of
417 stores that can contain keys, certificates, CRLs and numerous other
418 objects. The main API is loosely based on a few stdio functions,
419 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
420 OSSL_STORE_error and OSSL_STORE_close.
421 The implementation uses backends called "loaders" to implement arbitrary
422 URI schemes. There is one built in "loader" for the 'file' scheme.
425 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
426 then adjusted to work on FreeBSD 8.4 as well.
427 Enable by configuring with 'enable-devcryptoeng'. This is done by default
428 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
431 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
432 util/mkerr.pl, which is adapted to allow those prefixes, leading to
433 error code calls like this:
435 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
437 With this change, we claim the namespaces OSSL and OPENSSL in a manner
438 that can be encoded in C. For the foreseeable future, this will only
440 [Richard Levitte and Tim Hudson]
442 *) Removed BSD cryptodev engine.
445 *) Add a build target 'build_all_generated', to build all generated files
446 and only that. This can be used to prepare everything that requires
447 things like perl for a system that lacks perl and then move everything
448 to that system and do the rest of the build there.
451 *) In the UI interface, make it possible to duplicate the user data. This
452 can be used by engines that need to retain the data for a longer time
453 than just the call where this user data is passed.
456 *) Ignore the '-named_curve auto' value for compatibility of applications
458 [Tomas Mraz <tmraz@fedoraproject.org>]
460 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
461 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
462 alerts across multiple records (some of which could be empty). In practice
463 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
464 prohibts this altogether and other libraries (BoringSSL, NSS) do not
465 support this at all. Supporting it adds significant complexity to the
466 record layer, and its removal is unlikely to cause inter-operability
470 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
471 with Z. These are meant to replace LONG and ZLONG and to be size safe.
472 The use of LONG and ZLONG is discouraged and scheduled for deprecation
476 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
477 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
478 [Richard Levitte, Andy Polyakov]
480 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
484 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
485 platform rather than 'mingw'.
488 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
489 success if they are asked to add an object which already exists
490 in the store. This change cascades to other functions which load
491 certificates and CRLs.
494 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
495 facilitate stack unwinding even from assembly subroutines.
498 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
499 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
502 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
503 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
504 which is the minimum version we support.
507 *) Certificate time validation (X509_cmp_time) enforces stricter
508 compliance with RFC 5280. Fractional seconds and timezone offsets
509 are no longer allowed.
512 *) Add support for ARIA
515 *) s_client will now send the Server Name Indication (SNI) extension by
516 default unless the new "-noservername" option is used. The server name is
517 based on the host provided to the "-connect" option unless overridden by
521 *) Add support for SipHash
524 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
525 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
526 prevent issues where no progress is being made and the peer continually
527 sends unrecognised record types, using up resources processing them.
530 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
531 using the algorithm defined in
532 https://www.akkadia.org/drepper/SHA-crypt.txt
535 *) Heartbeat support has been removed; the ABI is changed for now.
536 [Richard Levitte, Rich Salz]
538 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
541 *) The RSA "null" method, which was partially supported to avoid patent
542 issues, has been replaced to always returns NULL.
546 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
548 *) Client DoS due to large DH parameter
550 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
551 malicious server can send a very large prime value to the client. This will
552 cause the client to spend an unreasonably long period of time generating a
553 key for this prime resulting in a hang until the client has finished. This
554 could be exploited in a Denial Of Service attack.
556 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
560 *) Cache timing vulnerability in RSA Key Generation
562 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
563 a cache timing side channel attack. An attacker with sufficient access to
564 mount cache timing attacks during the RSA key generation process could
565 recover the private key.
567 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
568 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
572 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
573 parameter is no longer accepted, as it leads to a corrupt table. NULL
574 pem_str is reserved for alias entries only.
577 *) Revert blinding in ECDSA sign and instead make problematic addition
578 length-invariant. Switch even to fixed-length Montgomery multiplication.
581 *) Change generating and checking of primes so that the error rate of not
582 being prime depends on the intended use based on the size of the input.
583 For larger primes this will result in more rounds of Miller-Rabin.
584 The maximal error rate for primes with more than 1080 bits is lowered
586 [Kurt Roeckx, Annie Yousar]
588 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
591 *) Add blinding to ECDSA and DSA signatures to protect against side channel
592 attacks discovered by Keegan Ryan (NCC Group).
595 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
596 now allow empty (zero character) pass phrases.
599 *) Certificate time validation (X509_cmp_time) enforces stricter
600 compliance with RFC 5280. Fractional seconds and timezone offsets
601 are no longer allowed.
604 *) Fixed a text canonicalisation bug in CMS
606 Where a CMS detached signature is used with text content the text goes
607 through a canonicalisation process first prior to signing or verifying a
608 signature. This process strips trailing space at the end of lines, converts
609 line terminators to CRLF and removes additional trailing line terminators
610 at the end of a file. A bug in the canonicalisation process meant that
611 some characters, such as form-feed, were incorrectly treated as whitespace
612 and removed. This is contrary to the specification (RFC5485). This fix
613 could mean that detached text data signed with an earlier version of
614 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
615 signed with a fixed OpenSSL may fail to verify with an earlier version of
616 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
617 and use the "-binary" flag (for the "cms" command line application) or set
618 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
621 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
623 *) Constructed ASN.1 types with a recursive definition could exceed the stack
625 Constructed ASN.1 types with a recursive definition (such as can be found
626 in PKCS7) could eventually exceed the stack given malicious input with
627 excessive recursion. This could result in a Denial Of Service attack. There
628 are no such structures used within SSL/TLS that come from untrusted sources
629 so this is considered safe.
631 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
636 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
638 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
639 effectively reduced to only comparing the least significant bit of each
640 byte. This allows an attacker to forge messages that would be considered as
641 authenticated in an amount of tries lower than that guaranteed by the
642 security claims of the scheme. The module can only be compiled by the
643 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
645 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
650 *) Add a build target 'build_all_generated', to build all generated files
651 and only that. This can be used to prepare everything that requires
652 things like perl for a system that lacks perl and then move everything
653 to that system and do the rest of the build there.
656 *) Backport SSL_OP_NO_RENGOTIATION
658 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
659 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
660 changes this is no longer possible in 1.1.0. Therefore the new
661 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
662 1.1.0 to provide equivalent functionality.
664 Note that if an application built against 1.1.0h headers (or above) is run
665 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
666 accepted but nothing will happen, i.e. renegotiation will not be prevented.
669 *) Removed the OS390-Unix config target. It relied on a script that doesn't
673 *) rsaz_1024_mul_avx2 overflow bug on x86_64
675 There is an overflow bug in the AVX2 Montgomery multiplication procedure
676 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
677 Analysis suggests that attacks against RSA and DSA as a result of this
678 defect would be very difficult to perform and are not believed likely.
679 Attacks against DH1024 are considered just feasible, because most of the
680 work necessary to deduce information about a private key may be performed
681 offline. The amount of resources required for such an attack would be
682 significant. However, for an attack on TLS to be meaningful, the server
683 would have to share the DH1024 private key among multiple clients, which is
684 no longer an option since CVE-2016-0701.
686 This only affects processors that support the AVX2 but not ADX extensions
687 like Intel Haswell (4th generation).
689 This issue was reported to OpenSSL by David Benjamin (Google). The issue
690 was originally found via the OSS-Fuzz project.
694 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
696 *) bn_sqrx8x_internal carry bug on x86_64
698 There is a carry propagating bug in the x86_64 Montgomery squaring
699 procedure. No EC algorithms are affected. Analysis suggests that attacks
700 against RSA and DSA as a result of this defect would be very difficult to
701 perform and are not believed likely. Attacks against DH are considered just
702 feasible (although very difficult) because most of the work necessary to
703 deduce information about a private key may be performed offline. The amount
704 of resources required for such an attack would be very significant and
705 likely only accessible to a limited number of attackers. An attacker would
706 additionally need online access to an unpatched system using the target
707 private key in a scenario with persistent DH parameters and a private
708 key that is shared between multiple clients.
710 This only affects processors that support the BMI1, BMI2 and ADX extensions
711 like Intel Broadwell (5th generation) and later or AMD Ryzen.
713 This issue was reported to OpenSSL by the OSS-Fuzz project.
717 *) Malformed X.509 IPAddressFamily could cause OOB read
719 If an X.509 certificate has a malformed IPAddressFamily extension,
720 OpenSSL could do a one-byte buffer overread. The most likely result
721 would be an erroneous display of the certificate in text format.
723 This issue was reported to OpenSSL by the OSS-Fuzz project.
727 Changes between 1.1.0e and 1.1.0f [25 May 2017]
729 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
730 platform rather than 'mingw'.
733 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
734 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
735 which is the minimum version we support.
738 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
740 *) Encrypt-Then-Mac renegotiation crash
742 During a renegotiation handshake if the Encrypt-Then-Mac extension is
743 negotiated where it was not in the original handshake (or vice-versa) then
744 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
745 and servers are affected.
747 This issue was reported to OpenSSL by Joe Orton (Red Hat).
751 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
753 *) Truncated packet could crash via OOB read
755 If one side of an SSL/TLS path is running on a 32-bit host and a specific
756 cipher is being used, then a truncated packet can cause that host to
757 perform an out-of-bounds read, usually resulting in a crash.
759 This issue was reported to OpenSSL by Robert Święcki of Google.
763 *) Bad (EC)DHE parameters cause a client crash
765 If a malicious server supplies bad parameters for a DHE or ECDHE key
766 exchange then this can result in the client attempting to dereference a
767 NULL pointer leading to a client crash. This could be exploited in a Denial
770 This issue was reported to OpenSSL by Guido Vranken.
774 *) BN_mod_exp may produce incorrect results on x86_64
776 There is a carry propagating bug in the x86_64 Montgomery squaring
777 procedure. No EC algorithms are affected. Analysis suggests that attacks
778 against RSA and DSA as a result of this defect would be very difficult to
779 perform and are not believed likely. Attacks against DH are considered just
780 feasible (although very difficult) because most of the work necessary to
781 deduce information about a private key may be performed offline. The amount
782 of resources required for such an attack would be very significant and
783 likely only accessible to a limited number of attackers. An attacker would
784 additionally need online access to an unpatched system using the target
785 private key in a scenario with persistent DH parameters and a private
786 key that is shared between multiple clients. For example this can occur by
787 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
788 similar to CVE-2015-3193 but must be treated as a separate problem.
790 This issue was reported to OpenSSL by the OSS-Fuzz project.
794 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
796 *) ChaCha20/Poly1305 heap-buffer-overflow
798 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
799 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
800 crash. This issue is not considered to be exploitable beyond a DoS.
802 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
806 *) CMS Null dereference
808 Applications parsing invalid CMS structures can crash with a NULL pointer
809 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
810 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
811 structure callback if an attempt is made to free certain invalid encodings.
812 Only CHOICE structures using a callback which do not handle NULL value are
815 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
819 *) Montgomery multiplication may produce incorrect results
821 There is a carry propagating bug in the Broadwell-specific Montgomery
822 multiplication procedure that handles input lengths divisible by, but
823 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
824 and DH private keys are impossible. This is because the subroutine in
825 question is not used in operations with the private key itself and an input
826 of the attacker's direct choice. Otherwise the bug can manifest itself as
827 transient authentication and key negotiation failures or reproducible
828 erroneous outcome of public-key operations with specially crafted input.
829 Among EC algorithms only Brainpool P-512 curves are affected and one
830 presumably can attack ECDH key negotiation. Impact was not analyzed in
831 detail, because pre-requisites for attack are considered unlikely. Namely
832 multiple clients have to choose the curve in question and the server has to
833 share the private key among them, neither of which is default behaviour.
834 Even then only clients that chose the curve will be affected.
836 This issue was publicly reported as transient failures and was not
837 initially recognized as a security issue. Thanks to Richard Morgan for
838 providing reproducible case.
842 *) Removed automatic addition of RPATH in shared libraries and executables,
843 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
846 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
848 *) Fix Use After Free for large message sizes
850 The patch applied to address CVE-2016-6307 resulted in an issue where if a
851 message larger than approx 16k is received then the underlying buffer to
852 store the incoming message is reallocated and moved. Unfortunately a
853 dangling pointer to the old location is left which results in an attempt to
854 write to the previously freed location. This is likely to result in a
855 crash, however it could potentially lead to execution of arbitrary code.
857 This issue only affects OpenSSL 1.1.0a.
859 This issue was reported to OpenSSL by Robert Święcki.
863 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
865 *) OCSP Status Request extension unbounded memory growth
867 A malicious client can send an excessively large OCSP Status Request
868 extension. If that client continually requests renegotiation, sending a
869 large OCSP Status Request extension each time, then there will be unbounded
870 memory growth on the server. This will eventually lead to a Denial Of
871 Service attack through memory exhaustion. Servers with a default
872 configuration are vulnerable even if they do not support OCSP. Builds using
873 the "no-ocsp" build time option are not affected.
875 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
879 *) SSL_peek() hang on empty record
881 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
882 sends an empty record. This could be exploited by a malicious peer in a
883 Denial Of Service attack.
885 This issue was reported to OpenSSL by Alex Gaynor.
889 *) Excessive allocation of memory in tls_get_message_header() and
890 dtls1_preprocess_fragment()
892 A (D)TLS message includes 3 bytes for its length in the header for the
893 message. This would allow for messages up to 16Mb in length. Messages of
894 this length are excessive and OpenSSL includes a check to ensure that a
895 peer is sending reasonably sized messages in order to avoid too much memory
896 being consumed to service a connection. A flaw in the logic of version
897 1.1.0 means that memory for the message is allocated too early, prior to
898 the excessive message length check. Due to way memory is allocated in
899 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
900 to service a connection. This could lead to a Denial of Service through
901 memory exhaustion. However, the excessive message length check still takes
902 place, and this would cause the connection to immediately fail. Assuming
903 that the application calls SSL_free() on the failed connection in a timely
904 manner then the 21Mb of allocated memory will then be immediately freed
905 again. Therefore the excessive memory allocation will be transitory in
906 nature. This then means that there is only a security impact if:
908 1) The application does not call SSL_free() in a timely manner in the event
909 that the connection fails
911 2) The application is working in a constrained environment where there is
912 very little free memory
914 3) The attacker initiates multiple connection attempts such that there are
915 multiple connections in a state where memory has been allocated for the
916 connection; SSL_free() has not yet been called; and there is insufficient
917 memory to service the multiple requests.
919 Except in the instance of (1) above any Denial Of Service is likely to be
920 transitory because as soon as the connection fails the memory is
921 subsequently freed again in the SSL_free() call. However there is an
922 increased risk during this period of application crashes due to the lack of
923 memory - which would then mean a more serious Denial of Service.
925 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
926 (CVE-2016-6307 and CVE-2016-6308)
929 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
930 had to be removed. Primary reason is that vendor assembler can't
931 assemble our modules with -KPIC flag. As result it, assembly
932 support, was not even available as option. But its lack means
933 lack of side-channel resistant code, which is incompatible with
934 security by todays standards. Fortunately gcc is readily available
935 prepackaged option, which we firmly point at...
938 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
940 *) Windows command-line tool supports UTF-8 opt-in option for arguments
941 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
942 (to any value) allows Windows user to access PKCS#12 file generated
943 with Windows CryptoAPI and protected with non-ASCII password, as well
944 as files generated under UTF-8 locale on Linux also protected with
948 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
949 have been disabled by default and removed from DEFAULT, just like RC4.
950 See the RC4 item below to re-enable both.
953 *) The method for finding the storage location for the Windows RAND seed file
954 has changed. First we check %RANDFILE%. If that is not set then we check
955 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
956 all else fails we fall back to C:\.
959 *) The EVP_EncryptUpdate() function has had its return type changed from void
960 to int. A return of 0 indicates and error while a return of 1 indicates
964 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
965 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
966 off the constant time implementation for RSA, DSA and DH have been made
967 no-ops and deprecated.
970 *) Windows RAND implementation was simplified to only get entropy by
971 calling CryptGenRandom(). Various other RAND-related tickets
973 [Joseph Wylie Yandle, Rich Salz]
975 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
976 and OPENSSL_LH_, respectively. The old names are available
977 with API compatibility. They new names are now completely documented.
980 *) Unify TYPE_up_ref(obj) methods signature.
981 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
982 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
983 int (instead of void) like all others TYPE_up_ref() methods.
984 So now these methods also check the return value of CRYPTO_atomic_add(),
985 and the validity of object reference counter.
986 [fdasilvayy@gmail.com]
988 *) With Windows Visual Studio builds, the .pdb files are installed
989 alongside the installed libraries and executables. For a static
990 library installation, ossl_static.pdb is the associate compiler
991 generated .pdb file to be used when linking programs.
994 *) Remove openssl.spec. Packaging files belong with the packagers.
997 *) Automatic Darwin/OSX configuration has had a refresh, it will now
998 recognise x86_64 architectures automatically. You can still decide
999 to build for a different bitness with the environment variable
1000 KERNEL_BITS (can be 32 or 64), for example:
1002 KERNEL_BITS=32 ./config
1006 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1007 256 bit AES and HMAC with SHA256.
1010 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1013 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1016 *) To enable users to have their own config files and build file templates,
1017 Configure looks in the directory indicated by the environment variable
1018 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1019 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1020 name and is used as is.
1023 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1024 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1025 X509_CERT_FILE_CTX was removed.
1028 *) "shared" builds are now the default. To create only static libraries use
1029 the "no-shared" Configure option.
1032 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1033 All of these option have not worked for some while and are fundamental
1037 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1038 global cleanup functions are no longer required because they are handled
1039 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1040 Explicitly de-initing can cause problems (e.g. where a library that uses
1041 OpenSSL de-inits, but an application is still using it). The affected
1042 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1043 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1044 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1045 COMP_zlib_cleanup().
1048 *) --strict-warnings no longer enables runtime debugging options
1049 such as REF_DEBUG. Instead, debug options are automatically
1050 enabled with '--debug' builds.
1051 [Andy Polyakov, Emilia Käsper]
1053 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1054 have been moved out of the public header files. New functions for managing
1055 these have been added.
1058 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1059 objects have been moved out of the public header files. New
1060 functions for managing these have been added.
1063 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1064 have been moved out of the public header files. New functions for managing
1065 these have been added.
1068 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1069 moved out of the public header files. New functions for managing these
1073 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1076 *) Removed the mk1mf build scripts.
1079 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1080 it is always safe to #include a header now.
1083 *) Removed the aged BC-32 config and all its supporting scripts
1086 *) Removed support for Ultrix, Netware, and OS/2.
1089 *) Add support for HKDF.
1090 [Alessandro Ghedini]
1092 *) Add support for blake2b and blake2s
1095 *) Added support for "pipelining". Ciphers that have the
1096 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1097 encryptions/decryptions simultaneously. There are currently no built-in
1098 ciphers with this property but the expectation is that engines will be able
1099 to offer it to significantly improve throughput. Support has been extended
1100 into libssl so that multiple records for a single connection can be
1101 processed in one go (for >=TLS 1.1).
1104 *) Added the AFALG engine. This is an async capable engine which is able to
1105 offload work to the Linux kernel. In this initial version it only supports
1106 AES128-CBC. The kernel must be version 4.1.0 or greater.
1109 *) OpenSSL now uses a new threading API. It is no longer necessary to
1110 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1111 are two supported threading models: pthreads and windows threads. It is
1112 also possible to configure OpenSSL at compile time for "no-threads". The
1113 old threading API should no longer be used. The functions have been
1114 replaced with "no-op" compatibility macros.
1115 [Alessandro Ghedini, Matt Caswell]
1117 *) Modify behavior of ALPN to invoke callback after SNI/servername
1118 callback, such that updates to the SSL_CTX affect ALPN.
1121 *) Add SSL_CIPHER queries for authentication and key-exchange.
1124 *) Changes to the DEFAULT cipherlist:
1125 - Prefer (EC)DHE handshakes over plain RSA.
1126 - Prefer AEAD ciphers over legacy ciphers.
1127 - Prefer ECDSA over RSA when both certificates are available.
1128 - Prefer TLSv1.2 ciphers/PRF.
1129 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1133 *) Change the ECC default curve list to be this, in order: x25519,
1134 secp256r1, secp521r1, secp384r1.
1137 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1138 disabled by default. They can be re-enabled using the
1139 enable-weak-ssl-ciphers option to Configure.
1142 *) If the server has ALPN configured, but supports no protocols that the
1143 client advertises, send a fatal "no_application_protocol" alert.
1144 This behaviour is SHALL in RFC 7301, though it isn't universally
1145 implemented by other servers.
1148 *) Add X25519 support.
1149 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1150 for public and private key encoding using the format documented in
1151 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1152 key generation and key derivation.
1154 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1158 *) Deprecate SRP_VBASE_get_by_user.
1159 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1160 In order to fix an unavoidable memory leak (CVE-2016-0798),
1161 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1162 seed, even if the seed is configured.
1164 Users should use SRP_VBASE_get1_by_user instead. Note that in
1165 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1166 also that even though configuring the SRP seed attempts to hide
1167 invalid usernames by continuing the handshake with fake
1168 credentials, this behaviour is not constant time and no strong
1169 guarantees are made that the handshake is indistinguishable from
1170 that of a valid user.
1173 *) Configuration change; it's now possible to build dynamic engines
1174 without having to build shared libraries and vice versa. This
1175 only applies to the engines in engines/, those in crypto/engine/
1176 will always be built into libcrypto (i.e. "static").
1178 Building dynamic engines is enabled by default; to disable, use
1179 the configuration option "disable-dynamic-engine".
1181 The only requirements for building dynamic engines are the
1182 presence of the DSO module and building with position independent
1183 code, so they will also automatically be disabled if configuring
1184 with "disable-dso" or "disable-pic".
1186 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1187 are also taken away from openssl/opensslconf.h, as they are
1191 *) Configuration change; if there is a known flag to compile
1192 position independent code, it will always be applied on the
1193 libcrypto and libssl object files, and never on the application
1194 object files. This means other libraries that use routines from
1195 libcrypto / libssl can be made into shared libraries regardless
1196 of how OpenSSL was configured.
1198 If this isn't desirable, the configuration options "disable-pic"
1199 or "no-pic" can be used to disable the use of PIC. This will
1200 also disable building shared libraries and dynamic engines.
1203 *) Removed JPAKE code. It was experimental and has no wide use.
1206 *) The INSTALL_PREFIX Makefile variable has been renamed to
1207 DESTDIR. That makes for less confusion on what this variable
1208 is for. Also, the configuration option --install_prefix is
1212 *) Heartbeat for TLS has been removed and is disabled by default
1213 for DTLS; configure with enable-heartbeats. Code that uses the
1214 old #define's might need to be updated.
1215 [Emilia Käsper, Rich Salz]
1217 *) Rename REF_CHECK to REF_DEBUG.
1220 *) New "unified" build system
1222 The "unified" build system is aimed to be a common system for all
1223 platforms we support. With it comes new support for VMS.
1225 This system builds supports building in a different directory tree
1226 than the source tree. It produces one Makefile (for unix family
1227 or lookalikes), or one descrip.mms (for VMS).
1229 The source of information to make the Makefile / descrip.mms is
1230 small files called 'build.info', holding the necessary
1231 information for each directory with source to compile, and a
1232 template in Configurations, like unix-Makefile.tmpl or
1235 With this change, the library names were also renamed on Windows
1236 and on VMS. They now have names that are closer to the standard
1237 on Unix, and include the major version number, and in certain
1238 cases, the architecture they are built for. See "Notes on shared
1239 libraries" in INSTALL.
1241 We rely heavily on the perl module Text::Template.
1244 *) Added support for auto-initialisation and de-initialisation of the library.
1245 OpenSSL no longer requires explicit init or deinit routines to be called,
1246 except in certain circumstances. See the OPENSSL_init_crypto() and
1247 OPENSSL_init_ssl() man pages for further information.
1250 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1251 "peer" argument is now expected to be a BIO_ADDR object.
1253 *) Rewrite of BIO networking library. The BIO library lacked consistent
1254 support of IPv6, and adding it required some more extensive
1255 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1256 which hold all types of addresses and chains of address information.
1257 It also introduces a new API, with functions like BIO_socket,
1258 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1259 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1260 have been adapted accordingly.
1263 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1267 *) CRIME protection: disable compression by default, even if OpenSSL is
1268 compiled with zlib enabled. Applications can still enable compression
1269 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1270 using the SSL_CONF library to configure compression.
1273 *) The signature of the session callback configured with
1274 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1275 was explicitly marked as 'const unsigned char*' instead of
1279 *) Always DPURIFY. Remove the use of uninitialized memory in the
1280 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1283 *) Removed many obsolete configuration items, including
1284 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1285 MD2_CHAR, MD2_INT, MD2_LONG
1287 IDEA_SHORT, IDEA_LONG
1288 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1289 [Rich Salz, with advice from Andy Polyakov]
1291 *) Many BN internals have been moved to an internal header file.
1292 [Rich Salz with help from Andy Polyakov]
1294 *) Configuration and writing out the results from it has changed.
1295 Files such as Makefile include/openssl/opensslconf.h and are now
1296 produced through general templates, such as Makefile.in and
1297 crypto/opensslconf.h.in and some help from the perl module
1300 Also, the center of configuration information is no longer
1301 Makefile. Instead, Configure produces a perl module in
1302 configdata.pm which holds most of the config data (in the hash
1303 table %config), the target data that comes from the target
1304 configuration in one of the Configurations/*.conf files (in
1308 *) To clarify their intended purposes, the Configure options
1309 --prefix and --openssldir change their semantics, and become more
1310 straightforward and less interdependent.
1312 --prefix shall be used exclusively to give the location INSTALLTOP
1313 where programs, scripts, libraries, include files and manuals are
1314 going to be installed. The default is now /usr/local.
1316 --openssldir shall be used exclusively to give the default
1317 location OPENSSLDIR where certificates, private keys, CRLs are
1318 managed. This is also where the default openssl.cnf gets
1320 If the directory given with this option is a relative path, the
1321 values of both the --prefix value and the --openssldir value will
1322 be combined to become OPENSSLDIR.
1323 The default for --openssldir is INSTALLTOP/ssl.
1325 Anyone who uses --openssldir to specify where OpenSSL is to be
1326 installed MUST change to use --prefix instead.
1329 *) The GOST engine was out of date and therefore it has been removed. An up
1330 to date GOST engine is now being maintained in an external repository.
1331 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1332 support for GOST ciphersuites (these are only activated if a GOST engine
1336 *) EGD is no longer supported by default; use enable-egd when
1338 [Ben Kaduk and Rich Salz]
1340 *) The distribution now has Makefile.in files, which are used to
1341 create Makefile's when Configure is run. *Configure must be run
1342 before trying to build now.*
1345 *) The return value for SSL_CIPHER_description() for error conditions
1349 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1351 Obtaining and performing DNSSEC validation of TLSA records is
1352 the application's responsibility. The application provides
1353 the TLSA records of its choice to OpenSSL, and these are then
1354 used to authenticate the peer.
1356 The TLSA records need not even come from DNS. They can, for
1357 example, be used to implement local end-entity certificate or
1358 trust-anchor "pinning", where the "pin" data takes the form
1359 of TLSA records, which can augment or replace verification
1360 based on the usual WebPKI public certification authorities.
1363 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1364 continues to support deprecated interfaces in default builds.
1365 However, applications are strongly advised to compile their
1366 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1367 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1368 or the 1.1.0 releases.
1370 In environments in which all applications have been ported to
1371 not use any deprecated interfaces OpenSSL's Configure script
1372 should be used with the --api=1.1.0 option to entirely remove
1373 support for the deprecated features from the library and
1374 unconditionally disable them in the installed headers.
1375 Essentially the same effect can be achieved with the "no-deprecated"
1376 argument to Configure, except that this will always restrict
1377 the build to just the latest API, rather than a fixed API
1380 As applications are ported to future revisions of the API,
1381 they should update their compile-time OPENSSL_API_COMPAT define
1382 accordingly, but in most cases should be able to continue to
1383 compile with later releases.
1385 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1386 0x10000000L and 0x00908000L, respectively. However those
1387 versions did not support the OPENSSL_API_COMPAT feature, and
1388 so applications are not typically tested for explicit support
1389 of just the undeprecated features of either release.
1392 *) Add support for setting the minimum and maximum supported protocol.
1393 It can bet set via the SSL_set_min_proto_version() and
1394 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1395 MaxProtocol. It's recommended to use the new APIs to disable
1396 protocols instead of disabling individual protocols using
1397 SSL_set_options() or SSL_CONF's Protocol. This change also
1398 removes support for disabling TLS 1.2 in the OpenSSL TLS
1399 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1402 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1405 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1406 and integrates ECDSA and ECDH functionality into EC. Implementations can
1407 now redirect key generation and no longer need to convert to or from
1410 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1411 include the ec.h header file instead.
1414 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1415 ciphers who are no longer supported and drops support the ephemeral RSA key
1416 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1419 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1420 opaque. For HMAC_CTX, the following constructors and destructors
1423 HMAC_CTX *HMAC_CTX_new(void);
1424 void HMAC_CTX_free(HMAC_CTX *ctx);
1426 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1427 destroy such methods has been added. See EVP_MD_meth_new(3) and
1428 EVP_CIPHER_meth_new(3) for documentation.
1431 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1432 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1433 EVP_MD_CTX_reset() should be called instead to reinitialise
1434 an already created structure.
1435 2) For consistency with the majority of our object creators and
1436 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1437 EVP_MD_CTX_(new|free). The old names are retained as macros
1438 for deprecated builds.
1441 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1442 cryptographic operations to be performed asynchronously as long as an
1443 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1444 further details. Libssl has also had this capability integrated with the
1445 introduction of the new mode SSL_MODE_ASYNC and associated error
1446 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1447 pages. This work was developed in partnership with Intel Corp.
1450 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1451 always enabled now. If you want to disable the support you should
1452 exclude it using the list of supported ciphers. This also means that the
1453 "-no_ecdhe" option has been removed from s_server.
1456 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1457 SSL_{CTX_}set1_curves() which can set a list.
1460 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1461 curve you want to support using SSL_{CTX_}set1_curves().
1464 *) State machine rewrite. The state machine code has been significantly
1465 refactored in order to remove much duplication of code and solve issues
1466 with the old code (see ssl/statem/README for further details). This change
1467 does have some associated API changes. Notably the SSL_state() function
1468 has been removed and replaced by SSL_get_state which now returns an
1469 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1470 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1474 *) All instances of the string "ssleay" in the public API were replaced
1475 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1476 Some error codes related to internal RSA_eay API's were renamed.
1479 *) The demo files in crypto/threads were moved to demo/threads.
1482 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1484 [Matt Caswell, Rich Salz]
1486 *) New ASN.1 embed macro.
1488 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1489 structure is not allocated: it is part of the parent. That is instead of
1497 This reduces memory fragmentation and make it impossible to accidentally
1498 set a mandatory field to NULL.
1500 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1501 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1502 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1506 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1509 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1510 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1511 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1512 DES and RC4 ciphersuites.
1515 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1516 This changes the decoding behaviour for some invalid messages,
1517 though the change is mostly in the more lenient direction, and
1518 legacy behaviour is preserved as much as possible.
1521 *) Fix no-stdio build.
1522 [ David Woodhouse <David.Woodhouse@intel.com> and also
1523 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1525 *) New testing framework
1526 The testing framework has been largely rewritten and is now using
1527 perl and the perl modules Test::Harness and an extended variant of
1528 Test::More called OpenSSL::Test to do its work. All test scripts in
1529 test/ have been rewritten into test recipes, and all direct calls to
1530 executables in test/Makefile have become individual recipes using the
1531 simplified testing OpenSSL::Test::Simple.
1533 For documentation on our testing modules, do:
1535 perldoc test/testlib/OpenSSL/Test/Simple.pm
1536 perldoc test/testlib/OpenSSL/Test.pm
1540 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1541 are used; the latter aborts on memory leaks (usually checked on exit).
1542 Some undocumented "set malloc, etc., hooks" functions were removed
1543 and others were changed. All are now documented.
1546 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1548 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1550 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1551 from RFC4279, RFC4785, RFC5487, RFC5489.
1553 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1554 original RSA_PSK patch.
1557 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1558 era flag was never set throughout the codebase (only read). Also removed
1559 SSL3_FLAGS_POP_BUFFER which was only used if
1560 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1563 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1564 to be "oneline" instead of "compat".
1567 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1568 not aware of clients that still exhibit this bug, and the workaround
1569 hasn't been working properly for a while.
1572 *) The return type of BIO_number_read() and BIO_number_written() as well as
1573 the corresponding num_read and num_write members in the BIO structure has
1574 changed from unsigned long to uint64_t. On platforms where an unsigned
1575 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1579 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1580 OpenSSL without support for them. It also means that maintaining
1581 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1582 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1585 *) Removed support for the two export grade static DH ciphersuites
1586 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1587 were newly added (along with a number of other static DH ciphersuites) to
1588 1.0.2. However the two export ones have *never* worked since they were
1589 introduced. It seems strange in any case to be adding new export
1590 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1593 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1594 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1595 and turned into macros which simply call the new preferred function names
1596 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1597 should use the new names instead. Also as part of this change the ssl23.h
1598 header file has been removed.
1601 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1602 code and the associated standard is no longer considered fit-for-purpose.
1605 *) RT2547 was closed. When generating a private key, try to make the
1606 output file readable only by the owner. This behavior change might
1607 be noticeable when interacting with other software.
1609 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1613 *) Added HTTP GET support to the ocsp command.
1616 *) Changed default digest for the dgst and enc commands from MD5 to
1620 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1623 *) Added support for TLS extended master secret from
1624 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1625 initial patch which was a great help during development.
1628 *) All libssl internal structures have been removed from the public header
1629 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1630 now redundant). Users should not attempt to access internal structures
1631 directly. Instead they should use the provided API functions.
1634 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1635 Access to deprecated functions can be re-enabled by running config with
1636 "enable-deprecated". In addition applications wishing to use deprecated
1637 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1638 will, by default, disable some transitive includes that previously existed
1639 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1642 *) Added support for OCB mode. OpenSSL has been granted a patent license
1643 compatible with the OpenSSL license for use of OCB. Details are available
1644 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1645 for OCB can be removed by calling config with no-ocb.
1648 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1649 compatible client hello.
1652 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1653 done while fixing the error code for the key-too-small case.
1654 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1656 *) CA.sh has been removed; use CA.pl instead.
1659 *) Removed old DES API.
1662 *) Remove various unsupported platforms:
1668 Sinix/ReliantUNIX RM400
1673 16-bit platforms such as WIN16
1676 *) Clean up OPENSSL_NO_xxx #define's
1677 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1678 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1679 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1680 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1681 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1682 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1683 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1684 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1685 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1686 Remove MS_STATIC; it's a relic from platforms <32 bits.
1689 *) Cleaned up dead code
1690 Remove all but one '#ifdef undef' which is to be looked at.
1693 *) Clean up calling of xxx_free routines.
1694 Just like free(), fix most of the xxx_free routines to accept
1695 NULL. Remove the non-null checks from callers. Save much code.
1698 *) Add secure heap for storage of private keys (when possible).
1699 Add BIO_s_secmem(), CBIGNUM, etc.
1700 Contributed by Akamai Technologies under our Corporate CLA.
1703 *) Experimental support for a new, fast, unbiased prime candidate generator,
1704 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1705 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1707 *) New output format NSS in the sess_id command line tool. This allows
1708 exporting the session id and the master key in NSS keylog format.
1709 [Martin Kaiser <martin@kaiser.cx>]
1711 *) Harmonize version and its documentation. -f flag is used to display
1713 [mancha <mancha1@zoho.com>]
1715 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1716 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1717 [mancha <mancha1@zoho.com>]
1719 *) Fix some double frees. These are not thought to be exploitable.
1720 [mancha <mancha1@zoho.com>]
1722 *) A missing bounds check in the handling of the TLS heartbeat extension
1723 can be used to reveal up to 64k of memory to a connected client or
1726 Thanks for Neel Mehta of Google Security for discovering this bug and to
1727 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1728 preparing the fix (CVE-2014-0160)
1729 [Adam Langley, Bodo Moeller]
1731 *) Fix for the attack described in the paper "Recovering OpenSSL
1732 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1733 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1734 http://eprint.iacr.org/2014/140
1736 Thanks to Yuval Yarom and Naomi Benger for discovering this
1737 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1738 [Yuval Yarom and Naomi Benger]
1740 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1741 this fixes a limitation in previous versions of OpenSSL.
1744 *) Experimental encrypt-then-mac support.
1746 Experimental support for encrypt then mac from
1747 draft-gutmann-tls-encrypt-then-mac-02.txt
1749 To enable it set the appropriate extension number (0x42 for the test
1750 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1752 For non-compliant peers (i.e. just about everything) this should have no
1755 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1759 *) Add EVP support for key wrapping algorithms, to avoid problems with
1760 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1761 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1762 algorithms and include tests cases.
1765 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1769 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1770 MGF1 digest and OAEP label.
1773 *) Make openssl verify return errors.
1774 [Chris Palmer <palmer@google.com> and Ben Laurie]
1776 *) New function ASN1_TIME_diff to calculate the difference between two
1777 ASN1_TIME structures or one structure and the current time.
1780 *) Update fips_test_suite to support multiple command line options. New
1781 test to induce all self test errors in sequence and check expected
1785 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1786 sign or verify all in one operation.
1789 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1790 test programs and fips_test_suite. Includes functionality to parse
1791 the minimal script output of fipsalgest.pl directly.
1794 *) Add authorisation parameter to FIPS_module_mode_set().
1797 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1800 *) Use separate DRBG fields for internal and external flags. New function
1801 FIPS_drbg_health_check() to perform on demand health checking. Add
1802 generation tests to fips_test_suite with reduced health check interval to
1803 demonstrate periodic health checking. Add "nodh" option to
1804 fips_test_suite to skip very slow DH test.
1807 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1811 *) More extensive health check for DRBG checking many more failure modes.
1812 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1813 combination: call this in fips_test_suite.
1816 *) Add support for canonical generation of DSA parameter 'g'. See
1819 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1820 POST to handle HMAC cases.
1823 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1824 to return numerical and string versions of the FIPS module number.
1827 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1828 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1829 outside the validated module in the FIPS capable OpenSSL.
1832 *) Minor change to DRBG entropy callback semantics. In some cases
1833 there is no multiple of the block length between min_len and
1834 max_len. Allow the callback to return more than max_len bytes
1835 of entropy but discard any extra: it is the callback's responsibility
1836 to ensure that the extra data discarded does not impact the
1837 requested amount of entropy.
1840 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1841 information in FIPS186-3, SP800-57 and SP800-131A.
1844 *) CCM support via EVP. Interface is very similar to GCM case except we
1845 must supply all data in one chunk (i.e. no update, final) and the
1846 message length must be supplied if AAD is used. Add algorithm test
1850 *) Initial version of POST overhaul. Add POST callback to allow the status
1851 of POST to be monitored and/or failures induced. Modify fips_test_suite
1852 to use callback. Always run all selftests even if one fails.
1855 *) XTS support including algorithm test driver in the fips_gcmtest program.
1856 Note: this does increase the maximum key length from 32 to 64 bytes but
1857 there should be no binary compatibility issues as existing applications
1858 will never use XTS mode.
1861 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1862 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1863 performs algorithm blocking for unapproved PRNG types. Also do not
1864 set PRNG type in FIPS_mode_set(): leave this to the application.
1865 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1866 the standard OpenSSL PRNG: set additional data to a date time vector.
1869 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1870 This shouldn't present any incompatibility problems because applications
1871 shouldn't be using these directly and any that are will need to rethink
1872 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1875 *) Extensive self tests and health checking required by SP800-90 DRBG.
1876 Remove strength parameter from FIPS_drbg_instantiate and always
1877 instantiate at maximum supported strength.
1880 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1883 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1886 *) New function DH_compute_key_padded() to compute a DH key and pad with
1887 leading zeroes if needed: this complies with SP800-56A et al.
1890 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1891 anything, incomplete, subject to change and largely untested at present.
1894 *) Modify fipscanisteronly build option to only build the necessary object
1895 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1898 *) Add experimental option FIPSSYMS to give all symbols in
1899 fipscanister.o and FIPS or fips prefix. This will avoid
1900 conflicts with future versions of OpenSSL. Add perl script
1901 util/fipsas.pl to preprocess assembly language source files
1902 and rename any affected symbols.
1905 *) Add selftest checks and algorithm block of non-fips algorithms in
1906 FIPS mode. Remove DES2 from selftests.
1909 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1910 return internal method without any ENGINE dependencies. Add new
1911 tiny fips sign and verify functions.
1914 *) New build option no-ec2m to disable characteristic 2 code.
1917 *) New build option "fipscanisteronly". This only builds fipscanister.o
1918 and (currently) associated fips utilities. Uses the file Makefile.fips
1919 instead of Makefile.org as the prototype.
1922 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1923 Update fips_gcmtest to use IV generator.
1926 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1927 setting output buffer to NULL. The *Final function must be
1928 called although it will not retrieve any additional data. The tag
1929 can be set or retrieved with a ctrl. The IV length is by default 12
1930 bytes (96 bits) but can be set to an alternative value. If the IV
1931 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1935 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1936 underlying do_cipher function handles all cipher semantics itself
1937 including padding and finalisation. This is useful if (for example)
1938 an ENGINE cipher handles block padding itself. The behaviour of
1939 do_cipher is subtly changed if this flag is set: the return value
1940 is the number of characters written to the output buffer (zero is
1941 no longer an error code) or a negative error code. Also if the
1942 input buffer is NULL and length 0 finalisation should be performed.
1945 *) If a candidate issuer certificate is already part of the constructed
1946 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1949 *) Improve forward-security support: add functions
1951 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1952 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1954 for use by SSL/TLS servers; the callback function will be called whenever a
1955 new session is created, and gets to decide whether the session may be
1956 cached to make it resumable (return 0) or not (return 1). (As by the
1957 SSL/TLS protocol specifications, the session_id sent by the server will be
1958 empty to indicate that the session is not resumable; also, the server will
1959 not generate RFC 4507 (RFC 5077) session tickets.)
1961 A simple reasonable callback implementation is to return is_forward_secure.
1962 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1963 by the SSL/TLS server library, indicating whether it can provide forward
1965 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1967 *) New -verify_name option in command line utilities to set verification
1971 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1972 Add CMAC pkey methods.
1975 *) Experimental renegotiation in s_server -www mode. If the client
1976 browses /reneg connection is renegotiated. If /renegcert it is
1977 renegotiated requesting a certificate.
1980 *) Add an "external" session cache for debugging purposes to s_server. This
1981 should help trace issues which normally are only apparent in deployed
1982 multi-process servers.
1985 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1986 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1987 BIO_set_cipher() and some obscure PEM functions were changed so they
1988 can now return an error. The RAND changes required a change to the
1989 RAND_METHOD structure.
1992 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1993 a gcc attribute to warn if the result of a function is ignored. This
1994 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1995 whose return value is often ignored.
1998 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1999 These allow SCTs (signed certificate timestamps) to be requested and
2000 validated when establishing a connection.
2001 [Rob Percival <robpercival@google.com>]
2003 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2005 *) Prevent padding oracle in AES-NI CBC MAC check
2007 A MITM attacker can use a padding oracle attack to decrypt traffic
2008 when the connection uses an AES CBC cipher and the server support
2011 This issue was introduced as part of the fix for Lucky 13 padding
2012 attack (CVE-2013-0169). The padding check was rewritten to be in
2013 constant time by making sure that always the same bytes are read and
2014 compared against either the MAC or padding bytes. But it no longer
2015 checked that there was enough data to have both the MAC and padding
2018 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2022 *) Fix EVP_EncodeUpdate overflow
2024 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2025 Base64 encoding of binary data. If an attacker is able to supply very large
2026 amounts of input data then a length check can overflow resulting in a heap
2029 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2030 the PEM_write_bio* family of functions. These are mainly used within the
2031 OpenSSL command line applications, so any application which processes data
2032 from an untrusted source and outputs it as a PEM file should be considered
2033 vulnerable to this issue. User applications that call these APIs directly
2034 with large amounts of untrusted data may also be vulnerable.
2036 This issue was reported by Guido Vranken.
2040 *) Fix EVP_EncryptUpdate overflow
2042 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2043 is able to supply very large amounts of input data after a previous call to
2044 EVP_EncryptUpdate() with a partial block then a length check can overflow
2045 resulting in a heap corruption. Following an analysis of all OpenSSL
2046 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2047 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2048 the first called function after an EVP_EncryptInit(), and therefore that
2049 specific call must be safe. The second form is where the length passed to
2050 EVP_EncryptUpdate() can be seen from the code to be some small value and
2051 therefore there is no possibility of an overflow. Since all instances are
2052 one of these two forms, it is believed that there can be no overflows in
2053 internal code due to this problem. It should be noted that
2054 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2055 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2056 of these calls have also been analysed too and it is believed there are no
2057 instances in internal usage where an overflow could occur.
2059 This issue was reported by Guido Vranken.
2063 *) Prevent ASN.1 BIO excessive memory allocation
2065 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2066 a short invalid encoding can cause allocation of large amounts of memory
2067 potentially consuming excessive resources or exhausting memory.
2069 Any application parsing untrusted data through d2i BIO functions is
2070 affected. The memory based functions such as d2i_X509() are *not* affected.
2071 Since the memory based functions are used by the TLS library, TLS
2072 applications are not affected.
2074 This issue was reported by Brian Carpenter.
2080 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2081 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2082 in arbitrary stack data being returned in the buffer.
2084 This issue was reported by Guido Vranken.
2088 *) Modify behavior of ALPN to invoke callback after SNI/servername
2089 callback, such that updates to the SSL_CTX affect ALPN.
2092 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2096 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2097 methods are enabled and ssl2 is disabled the methods return NULL.
2100 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2102 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2103 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2104 provide any "EXPORT" or "LOW" strength ciphers.
2107 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2108 is by default disabled at build-time. Builds that are not configured with
2109 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2110 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2111 will need to explicitly call either of:
2113 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2115 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2117 as appropriate. Even if either of those is used, or the application
2118 explicitly uses the version-specific SSLv2_method() or its client and
2119 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2120 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2121 ciphers, and SSLv2 56-bit DES are no longer available.
2125 *) Fix a double-free in DSA code
2127 A double free bug was discovered when OpenSSL parses malformed DSA private
2128 keys and could lead to a DoS attack or memory corruption for applications
2129 that receive DSA private keys from untrusted sources. This scenario is
2132 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2137 *) Disable SRP fake user seed to address a server memory leak.
2139 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2141 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2142 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2143 was changed to ignore the "fake user" SRP seed, even if the seed
2146 Users should use SRP_VBASE_get1_by_user instead. Note that in
2147 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2148 also that even though configuring the SRP seed attempts to hide
2149 invalid usernames by continuing the handshake with fake
2150 credentials, this behaviour is not constant time and no strong
2151 guarantees are made that the handshake is indistinguishable from
2152 that of a valid user.
2156 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2158 In the BN_hex2bn function the number of hex digits is calculated using an
2159 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2160 large values of |i| this can result in |bn_expand| not allocating any
2161 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2162 field as NULL leading to a subsequent NULL ptr deref. For very large values
2163 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2164 In this case memory is allocated to the internal BIGNUM data field, but it
2165 is insufficiently sized leading to heap corruption. A similar issue exists
2166 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2167 is ever called by user applications with very large untrusted hex/dec data.
2168 This is anticipated to be a rare occurrence.
2170 All OpenSSL internal usage of these functions use data that is not expected
2171 to be untrusted, e.g. config file data or application command line
2172 arguments. If user developed applications generate config file data based
2173 on untrusted data then it is possible that this could also lead to security
2174 consequences. This is also anticipated to be rare.
2176 This issue was reported to OpenSSL by Guido Vranken.
2180 *) Fix memory issues in BIO_*printf functions
2182 The internal |fmtstr| function used in processing a "%s" format string in
2183 the BIO_*printf functions could overflow while calculating the length of a
2184 string and cause an OOB read when printing very long strings.
2186 Additionally the internal |doapr_outch| function can attempt to write to an
2187 OOB memory location (at an offset from the NULL pointer) in the event of a
2188 memory allocation failure. In 1.0.2 and below this could be caused where
2189 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2190 could be in processing a very long "%s" format string. Memory leaks can
2193 The first issue may mask the second issue dependent on compiler behaviour.
2194 These problems could enable attacks where large amounts of untrusted data
2195 is passed to the BIO_*printf functions. If applications use these functions
2196 in this way then they could be vulnerable. OpenSSL itself uses these
2197 functions when printing out human-readable dumps of ASN.1 data. Therefore
2198 applications that print this data could be vulnerable if the data is from
2199 untrusted sources. OpenSSL command line applications could also be
2200 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2201 as command line arguments.
2203 Libssl is not considered directly vulnerable. Additionally certificates etc
2204 received via remote connections via libssl are also unlikely to be able to
2205 trigger these issues because of message size limits enforced within libssl.
2207 This issue was reported to OpenSSL Guido Vranken.
2211 *) Side channel attack on modular exponentiation
2213 A side-channel attack was found which makes use of cache-bank conflicts on
2214 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2215 of RSA keys. The ability to exploit this issue is limited as it relies on
2216 an attacker who has control of code in a thread running on the same
2217 hyper-threaded core as the victim thread which is performing decryptions.
2219 This issue was reported to OpenSSL by Yuval Yarom, The University of
2220 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2221 Nadia Heninger, University of Pennsylvania with more information at
2222 http://cachebleed.info.
2226 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2227 if no keysize is specified with default_bits. This fixes an
2228 omission in an earlier change that changed all RSA/DSA key generation
2229 apps to use 2048 bits by default.
2232 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2233 *) DH small subgroups
2235 Historically OpenSSL only ever generated DH parameters based on "safe"
2236 primes. More recently (in version 1.0.2) support was provided for
2237 generating X9.42 style parameter files such as those required for RFC 5114
2238 support. The primes used in such files may not be "safe". Where an
2239 application is using DH configured with parameters based on primes that are
2240 not "safe" then an attacker could use this fact to find a peer's private
2241 DH exponent. This attack requires that the attacker complete multiple
2242 handshakes in which the peer uses the same private DH exponent. For example
2243 this could be used to discover a TLS server's private DH exponent if it's
2244 reusing the private DH exponent or it's using a static DH ciphersuite.
2246 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2247 TLS. It is not on by default. If the option is not set then the server
2248 reuses the same private DH exponent for the life of the server process and
2249 would be vulnerable to this attack. It is believed that many popular
2250 applications do set this option and would therefore not be at risk.
2252 The fix for this issue adds an additional check where a "q" parameter is
2253 available (as is the case in X9.42 based parameters). This detects the
2254 only known attack, and is the only possible defense for static DH
2255 ciphersuites. This could have some performance impact.
2257 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2258 default and cannot be disabled. This could have some performance impact.
2260 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2264 *) SSLv2 doesn't block disabled ciphers
2266 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2267 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2268 been disabled, provided that the SSLv2 protocol was not also disabled via
2271 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2272 and Sebastian Schinzel.
2276 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2278 *) BN_mod_exp may produce incorrect results on x86_64
2280 There is a carry propagating bug in the x86_64 Montgomery squaring
2281 procedure. No EC algorithms are affected. Analysis suggests that attacks
2282 against RSA and DSA as a result of this defect would be very difficult to
2283 perform and are not believed likely. Attacks against DH are considered just
2284 feasible (although very difficult) because most of the work necessary to
2285 deduce information about a private key may be performed offline. The amount
2286 of resources required for such an attack would be very significant and
2287 likely only accessible to a limited number of attackers. An attacker would
2288 additionally need online access to an unpatched system using the target
2289 private key in a scenario with persistent DH parameters and a private
2290 key that is shared between multiple clients. For example this can occur by
2291 default in OpenSSL DHE based SSL/TLS ciphersuites.
2293 This issue was reported to OpenSSL by Hanno Böck.
2297 *) Certificate verify crash with missing PSS parameter
2299 The signature verification routines will crash with a NULL pointer
2300 dereference if presented with an ASN.1 signature using the RSA PSS
2301 algorithm and absent mask generation function parameter. Since these
2302 routines are used to verify certificate signature algorithms this can be
2303 used to crash any certificate verification operation and exploited in a
2304 DoS attack. Any application which performs certificate verification is
2305 vulnerable including OpenSSL clients and servers which enable client
2308 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2312 *) X509_ATTRIBUTE memory leak
2314 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2315 memory. This structure is used by the PKCS#7 and CMS routines so any
2316 application which reads PKCS#7 or CMS data from untrusted sources is
2317 affected. SSL/TLS is not affected.
2319 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2324 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2325 This changes the decoding behaviour for some invalid messages,
2326 though the change is mostly in the more lenient direction, and
2327 legacy behaviour is preserved as much as possible.
2330 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2332 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2334 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2336 *) Alternate chains certificate forgery
2338 During certificate verification, OpenSSL will attempt to find an
2339 alternative certificate chain if the first attempt to build such a chain
2340 fails. An error in the implementation of this logic can mean that an
2341 attacker could cause certain checks on untrusted certificates to be
2342 bypassed, such as the CA flag, enabling them to use a valid leaf
2343 certificate to act as a CA and "issue" an invalid certificate.
2345 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2349 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2351 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2352 incompatibility in the handling of HMAC. The previous ABI has now been
2356 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2358 *) Malformed ECParameters causes infinite loop
2360 When processing an ECParameters structure OpenSSL enters an infinite loop
2361 if the curve specified is over a specially malformed binary polynomial
2364 This can be used to perform denial of service against any
2365 system which processes public keys, certificate requests or
2366 certificates. This includes TLS clients and TLS servers with
2367 client authentication enabled.
2369 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2373 *) Exploitable out-of-bounds read in X509_cmp_time
2375 X509_cmp_time does not properly check the length of the ASN1_TIME
2376 string and can read a few bytes out of bounds. In addition,
2377 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2380 An attacker can use this to craft malformed certificates and CRLs of
2381 various sizes and potentially cause a segmentation fault, resulting in
2382 a DoS on applications that verify certificates or CRLs. TLS clients
2383 that verify CRLs are affected. TLS clients and servers with client
2384 authentication enabled may be affected if they use custom verification
2387 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2388 independently by Hanno Böck.
2392 *) PKCS7 crash with missing EnvelopedContent
2394 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2395 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2396 with missing content and trigger a NULL pointer dereference on parsing.
2398 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2399 structures from untrusted sources are affected. OpenSSL clients and
2400 servers are not affected.
2402 This issue was reported to OpenSSL by Michal Zalewski (Google).
2406 *) CMS verify infinite loop with unknown hash function
2408 When verifying a signedData message the CMS code can enter an infinite loop
2409 if presented with an unknown hash function OID. This can be used to perform
2410 denial of service against any system which verifies signedData messages using
2412 This issue was reported to OpenSSL by Johannes Bauer.
2416 *) Race condition handling NewSessionTicket
2418 If a NewSessionTicket is received by a multi-threaded client when attempting to
2419 reuse a previous ticket then a race condition can occur potentially leading to
2420 a double free of the ticket data.
2424 *) Only support 256-bit or stronger elliptic curves with the
2425 'ecdh_auto' setting (server) or by default (client). Of supported
2426 curves, prefer P-256 (both).
2429 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2431 *) ClientHello sigalgs DoS fix
2433 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2434 invalid signature algorithms extension a NULL pointer dereference will
2435 occur. This can be exploited in a DoS attack against the server.
2437 This issue was was reported to OpenSSL by David Ramos of Stanford
2440 [Stephen Henson and Matt Caswell]
2442 *) Multiblock corrupted pointer fix
2444 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2445 feature only applies on 64 bit x86 architecture platforms that support AES
2446 NI instructions. A defect in the implementation of "multiblock" can cause
2447 OpenSSL's internal write buffer to become incorrectly set to NULL when
2448 using non-blocking IO. Typically, when the user application is using a
2449 socket BIO for writing, this will only result in a failed connection.
2450 However if some other BIO is used then it is likely that a segmentation
2451 fault will be triggered, thus enabling a potential DoS attack.
2453 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2457 *) Segmentation fault in DTLSv1_listen fix
2459 The DTLSv1_listen function is intended to be stateless and processes the
2460 initial ClientHello from many peers. It is common for user code to loop
2461 over the call to DTLSv1_listen until a valid ClientHello is received with
2462 an associated cookie. A defect in the implementation of DTLSv1_listen means
2463 that state is preserved in the SSL object from one invocation to the next
2464 that can lead to a segmentation fault. Errors processing the initial
2465 ClientHello can trigger this scenario. An example of such an error could be
2466 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2469 This issue was reported to OpenSSL by Per Allansson.
2473 *) Segmentation fault in ASN1_TYPE_cmp fix
2475 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2476 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2477 certificate signature algorithm consistency this can be used to crash any
2478 certificate verification operation and exploited in a DoS attack. Any
2479 application which performs certificate verification is vulnerable including
2480 OpenSSL clients and servers which enable client authentication.
2484 *) Segmentation fault for invalid PSS parameters fix
2486 The signature verification routines will crash with a NULL pointer
2487 dereference if presented with an ASN.1 signature using the RSA PSS
2488 algorithm and invalid parameters. Since these routines are used to verify
2489 certificate signature algorithms this can be used to crash any
2490 certificate verification operation and exploited in a DoS attack. Any
2491 application which performs certificate verification is vulnerable including
2492 OpenSSL clients and servers which enable client authentication.
2494 This issue was was reported to OpenSSL by Brian Carpenter.
2498 *) ASN.1 structure reuse memory corruption fix
2500 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2501 memory corruption via an invalid write. Such reuse is and has been
2502 strongly discouraged and is believed to be rare.
2504 Applications that parse structures containing CHOICE or ANY DEFINED BY
2505 components may be affected. Certificate parsing (d2i_X509 and related
2506 functions) are however not affected. OpenSSL clients and servers are
2511 *) PKCS7 NULL pointer dereferences fix
2513 The PKCS#7 parsing code does not handle missing outer ContentInfo
2514 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2515 missing content and trigger a NULL pointer dereference on parsing.
2517 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2518 otherwise parse PKCS#7 structures from untrusted sources are
2519 affected. OpenSSL clients and servers are not affected.
2521 This issue was reported to OpenSSL by Michal Zalewski (Google).
2525 *) DoS via reachable assert in SSLv2 servers fix
2527 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2528 servers that both support SSLv2 and enable export cipher suites by sending
2529 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2531 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2532 (OpenSSL development team).
2536 *) Empty CKE with client auth and DHE fix
2538 If client auth is used then a server can seg fault in the event of a DHE
2539 ciphersuite being selected and a zero length ClientKeyExchange message
2540 being sent by the client. This could be exploited in a DoS attack.
2544 *) Handshake with unseeded PRNG fix
2546 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2547 with an unseeded PRNG. The conditions are:
2548 - The client is on a platform where the PRNG has not been seeded
2549 automatically, and the user has not seeded manually
2550 - A protocol specific client method version has been used (i.e. not
2551 SSL_client_methodv23)
2552 - A ciphersuite is used that does not require additional random data from
2553 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2555 If the handshake succeeds then the client random that has been used will
2556 have been generated from a PRNG with insufficient entropy and therefore the
2557 output may be predictable.
2559 For example using the following command with an unseeded openssl will
2560 succeed on an unpatched platform:
2562 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2566 *) Use After Free following d2i_ECPrivatekey error fix
2568 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2569 could cause a use after free condition. This, in turn, could cause a double
2570 free in several private key parsing functions (such as d2i_PrivateKey
2571 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2572 for applications that receive EC private keys from untrusted
2573 sources. This scenario is considered rare.
2575 This issue was discovered by the BoringSSL project and fixed in their
2580 *) X509_to_X509_REQ NULL pointer deref fix
2582 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2583 the certificate key is invalid. This function is rarely used in practice.
2585 This issue was discovered by Brian Carpenter.
2589 *) Removed the export ciphers from the DEFAULT ciphers
2592 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2594 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2595 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2596 So far those who have to target multiple platforms would compromise
2597 and argue that binary targeting say ARMv5 would still execute on
2598 ARMv8. "Universal" build resolves this compromise by providing
2599 near-optimal performance even on newer platforms.
2602 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2603 (other platforms pending).
2604 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2606 *) Add support for the SignedCertificateTimestampList certificate and
2607 OCSP response extensions from RFC6962.
2610 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2611 for corner cases. (Certain input points at infinity could lead to
2612 bogus results, with non-infinity inputs mapped to infinity too.)
2615 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2616 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2617 common cases are optimized and there still is room for further
2618 improvements. Vector Permutation AES for Altivec is also added.
2621 *) Add support for little-endian ppc64 Linux target.
2622 [Marcelo Cerri (IBM)]
2624 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2625 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2626 are optimized and there still is room for further improvements.
2627 Both 32- and 64-bit modes are supported.
2628 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2630 *) Improved ARMv7 NEON support.
2633 *) Support for SPARC Architecture 2011 crypto extensions, first
2634 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2635 SHA256/512, MD5, GHASH and modular exponentiation.
2636 [Andy Polyakov, David Miller]
2638 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2640 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2642 *) Support for new and upcoming Intel processors, including AVX2,
2643 BMI and SHA ISA extensions. This includes additional "stitched"
2644 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2647 This work was sponsored by Intel Corp.
2650 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2651 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2652 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2655 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2656 this fixes a limitation in previous versions of OpenSSL.
2659 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2660 MGF1 digest and OAEP label.
2663 *) Add EVP support for key wrapping algorithms, to avoid problems with
2664 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2665 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2666 algorithms and include tests cases.
2669 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2671 [Douglas E. Engert, Steve Henson]
2673 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2674 difference in days and seconds between two tm or ASN1_TIME structures.
2677 *) Add -rev test option to s_server to just reverse order of characters
2678 received by client and send back to server. Also prints an abbreviated
2679 summary of the connection parameters.
2682 *) New option -brief for s_client and s_server to print out a brief summary
2683 of connection parameters.
2686 *) Add callbacks for arbitrary TLS extensions.
2687 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2689 *) New option -crl_download in several openssl utilities to download CRLs
2690 from CRLDP extension in certificates.
2693 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2696 *) New function X509_CRL_diff to generate a delta CRL from the difference
2697 of two full CRLs. Add support to "crl" utility.
2700 *) New functions to set lookup_crls function and to retrieve
2701 X509_STORE from X509_STORE_CTX.
2704 *) Print out deprecated issuer and subject unique ID fields in
2708 *) Extend OCSP I/O functions so they can be used for simple general purpose
2709 HTTP as well as OCSP. New wrapper function which can be used to download
2710 CRLs using the OCSP API.
2713 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2716 *) SSL_CONF* functions. These provide a common framework for application
2717 configuration using configuration files or command lines.
2720 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2721 message callback and prints the results. Needs compile time option
2722 "enable-ssl-trace". New options to s_client and s_server to enable
2726 *) New ctrl and macro to retrieve supported points extensions.
2727 Print out extension in s_server and s_client.
2730 *) New functions to retrieve certificate signature and signature
2734 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2738 *) New Suite B modes for TLS code. These use and enforce the requirements
2739 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2740 only use Suite B curves. The Suite B modes can be set by using the
2741 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2744 *) New chain verification flags for Suite B levels of security. Check
2745 algorithms are acceptable when flags are set in X509_verify_cert.
2748 *) Make tls1_check_chain return a set of flags indicating checks passed
2749 by a certificate chain. Add additional tests to handle client
2750 certificates: checks for matching certificate type and issuer name
2754 *) If an attempt is made to use a signature algorithm not in the peer
2755 preference list abort the handshake. If client has no suitable
2756 signature algorithms in response to a certificate request do not
2757 use the certificate.
2760 *) If server EC tmp key is not in client preference list abort handshake.
2763 *) Add support for certificate stores in CERT structure. This makes it
2764 possible to have different stores per SSL structure or one store in
2765 the parent SSL_CTX. Include distinct stores for certificate chain
2766 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2767 to build and store a certificate chain in CERT structure: returning
2768 an error if the chain cannot be built: this will allow applications
2769 to test if a chain is correctly configured.
2771 Note: if the CERT based stores are not set then the parent SSL_CTX
2772 store is used to retain compatibility with existing behaviour.
2776 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2777 mask based on the current session, check mask when sending client
2778 hello and checking the requested ciphersuite.
2781 *) New ctrls to retrieve and set certificate types in a certificate
2782 request message. Print out received values in s_client. If certificate
2783 types is not set with custom values set sensible values based on
2784 supported signature algorithms.
2787 *) Support for distinct client and server supported signature algorithms.
2790 *) Add certificate callback. If set this is called whenever a certificate
2791 is required by client or server. An application can decide which
2792 certificate chain to present based on arbitrary criteria: for example
2793 supported signature algorithms. Add very simple example to s_server.
2794 This fixes many of the problems and restrictions of the existing client
2795 certificate callback: for example you can now clear an existing
2796 certificate and specify the whole chain.
2799 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2800 the certificate can be used for (if anything). Set valid_flags field
2801 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2802 to have similar checks in it.
2804 Add new "cert_flags" field to CERT structure and include a "strict mode".
2805 This enforces some TLS certificate requirements (such as only permitting
2806 certificate signature algorithms contained in the supported algorithms
2807 extension) which some implementations ignore: this option should be used
2808 with caution as it could cause interoperability issues.
2811 *) Update and tidy signature algorithm extension processing. Work out
2812 shared signature algorithms based on preferences and peer algorithms
2813 and print them out in s_client and s_server. Abort handshake if no
2814 shared signature algorithms.
2817 *) Add new functions to allow customised supported signature algorithms
2818 for SSL and SSL_CTX structures. Add options to s_client and s_server
2822 *) New function SSL_certs_clear() to delete all references to certificates
2823 from an SSL structure. Before this once a certificate had been added
2824 it couldn't be removed.
2827 *) Integrate hostname, email address and IP address checking with certificate
2828 verification. New verify options supporting checking in openssl utility.
2831 *) Fixes and wildcard matching support to hostname and email checking
2832 functions. Add manual page.
2833 [Florian Weimer (Red Hat Product Security Team)]
2835 *) New functions to check a hostname email or IP address against a
2836 certificate. Add options x509 utility to print results of checks against
2840 *) Fix OCSP checking.
2841 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2843 *) Initial experimental support for explicitly trusted non-root CAs.
2844 OpenSSL still tries to build a complete chain to a root but if an
2845 intermediate CA has a trust setting included that is used. The first
2846 setting is used: whether to trust (e.g., -addtrust option to the x509
2850 *) Add -trusted_first option which attempts to find certificates in the
2851 trusted store even if an untrusted chain is also supplied.
2854 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2855 platform support for Linux and Android.
2858 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2861 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2862 When in FIPS mode the approved implementations are used as normal,
2863 when not in FIPS mode the internal unapproved versions are used instead.
2864 This means that the FIPS capable OpenSSL isn't forced to use the
2865 (often lower performance) FIPS implementations outside FIPS mode.
2868 *) Transparently support X9.42 DH parameters when calling
2869 PEM_read_bio_DHparameters. This means existing applications can handle
2870 the new parameter format automatically.
2873 *) Initial experimental support for X9.42 DH parameter format: mainly
2874 to support use of 'q' parameter for RFC5114 parameters.
2877 *) Add DH parameters from RFC5114 including test data to dhtest.
2880 *) Support for automatic EC temporary key parameter selection. If enabled
2881 the most preferred EC parameters are automatically used instead of
2882 hardcoded fixed parameters. Now a server just has to call:
2883 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2884 support ECDH and use the most appropriate parameters.
2887 *) Enhance and tidy EC curve and point format TLS extension code. Use
2888 static structures instead of allocation if default values are used.
2889 New ctrls to set curves we wish to support and to retrieve shared curves.
2890 Print out shared curves in s_server. New options to s_server and s_client
2891 to set list of supported curves.
2894 *) New ctrls to retrieve supported signature algorithms and
2895 supported curve values as an array of NIDs. Extend openssl utility
2896 to print out received values.
2899 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2900 between NIDs and the more common NIST names such as "P-256". Enhance
2901 ecparam utility and ECC method to recognise the NIST names for curves.
2904 *) Enhance SSL/TLS certificate chain handling to support different
2905 chains for each certificate instead of one chain in the parent SSL_CTX.
2908 *) Support for fixed DH ciphersuite client authentication: where both
2909 server and client use DH certificates with common parameters.
2912 *) Support for fixed DH ciphersuites: those requiring DH server
2916 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2918 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2919 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2920 X509_CINF_get_signature were reverted post internal team review.
2922 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2924 *) Build fixes for the Windows and OpenVMS platforms
2925 [Matt Caswell and Richard Levitte]
2927 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2929 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2930 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2931 dereference. This could lead to a Denial Of Service attack. Thanks to
2932 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2936 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2937 dtls1_buffer_record function under certain conditions. In particular this
2938 could occur if an attacker sent repeated DTLS records with the same
2939 sequence number but for the next epoch. The memory leak could be exploited
2940 by an attacker in a Denial of Service attack through memory exhaustion.
2941 Thanks to Chris Mueller for reporting this issue.
2945 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2946 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2947 method would be set to NULL which could later result in a NULL pointer
2948 dereference. Thanks to Frank Schmirler for reporting this issue.
2952 *) Abort handshake if server key exchange message is omitted for ephemeral
2955 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2956 reporting this issue.
2960 *) Remove non-export ephemeral RSA code on client and server. This code
2961 violated the TLS standard by allowing the use of temporary RSA keys in
2962 non-export ciphersuites and could be used by a server to effectively
2963 downgrade the RSA key length used to a value smaller than the server
2964 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2965 INRIA or reporting this issue.
2969 *) Fixed issue where DH client certificates are accepted without verification.
2970 An OpenSSL server will accept a DH certificate for client authentication
2971 without the certificate verify message. This effectively allows a client to
2972 authenticate without the use of a private key. This only affects servers
2973 which trust a client certificate authority which issues certificates
2974 containing DH keys: these are extremely rare and hardly ever encountered.
2975 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2980 *) Ensure that the session ID context of an SSL is updated when its
2981 SSL_CTX is updated via SSL_set_SSL_CTX.
2983 The session ID context is typically set from the parent SSL_CTX,
2984 and can vary with the CTX.
2987 *) Fix various certificate fingerprint issues.
2989 By using non-DER or invalid encodings outside the signed portion of a
2990 certificate the fingerprint can be changed without breaking the signature.
2991 Although no details of the signed portion of the certificate can be changed
2992 this can cause problems with some applications: e.g. those using the
2993 certificate fingerprint for blacklists.
2995 1. Reject signatures with non zero unused bits.
2997 If the BIT STRING containing the signature has non zero unused bits reject
2998 the signature. All current signature algorithms require zero unused bits.
3000 2. Check certificate algorithm consistency.
3002 Check the AlgorithmIdentifier inside TBS matches the one in the
3003 certificate signature. NB: this will result in signature failure
3004 errors for some broken certificates.
3006 Thanks to Konrad Kraszewski from Google for reporting this issue.
3008 3. Check DSA/ECDSA signatures use DER.
3010 Re-encode DSA/ECDSA signatures and compare with the original received
3011 signature. Return an error if there is a mismatch.
3013 This will reject various cases including garbage after signature
3014 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3015 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3016 (negative or with leading zeroes).
3018 Further analysis was conducted and fixes were developed by Stephen Henson
3019 of the OpenSSL core team.
3024 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3025 results on some platforms, including x86_64. This bug occurs at random
3026 with a very low probability, and is not known to be exploitable in any
3027 way, though its exact impact is difficult to determine. Thanks to Pieter
3028 Wuille (Blockstream) who reported this issue and also suggested an initial
3029 fix. Further analysis was conducted by the OpenSSL development team and
3030 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3031 the OpenSSL core team.
3035 *) Do not resume sessions on the server if the negotiated protocol
3036 version does not match the session's version. Resuming with a different
3037 version, while not strictly forbidden by the RFC, is of questionable
3038 sanity and breaks all known clients.
3039 [David Benjamin, Emilia Käsper]
3041 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3042 early CCS messages during renegotiation. (Note that because
3043 renegotiation is encrypted, this early CCS was not exploitable.)
3046 *) Tighten client-side session ticket handling during renegotiation:
3047 ensure that the client only accepts a session ticket if the server sends
3048 the extension anew in the ServerHello. Previously, a TLS client would
3049 reuse the old extension state and thus accept a session ticket if one was
3050 announced in the initial ServerHello.
3052 Similarly, ensure that the client requires a session ticket if one
3053 was advertised in the ServerHello. Previously, a TLS client would
3054 ignore a missing NewSessionTicket message.
3057 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3059 *) SRTP Memory Leak.
3061 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3062 sends a carefully crafted handshake message, to cause OpenSSL to fail
3063 to free up to 64k of memory causing a memory leak. This could be
3064 exploited in a Denial Of Service attack. This issue affects OpenSSL
3065 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3066 whether SRTP is used or configured. Implementations of OpenSSL that
3067 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3069 The fix was developed by the OpenSSL team.
3073 *) Session Ticket Memory Leak.
3075 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3076 integrity of that ticket is first verified. In the event of a session
3077 ticket integrity check failing, OpenSSL will fail to free memory
3078 causing a memory leak. By sending a large number of invalid session
3079 tickets an attacker could exploit this issue in a Denial Of Service
3084 *) Build option no-ssl3 is incomplete.
3086 When OpenSSL is configured with "no-ssl3" as a build option, servers
3087 could accept and complete a SSL 3.0 handshake, and clients could be
3088 configured to send them.
3090 [Akamai and the OpenSSL team]
3092 *) Add support for TLS_FALLBACK_SCSV.
3093 Client applications doing fallback retries should call
3094 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3096 [Adam Langley, Bodo Moeller]
3098 *) Add additional DigestInfo checks.
3100 Re-encode DigestInto in DER and check against the original when
3101 verifying RSA signature: this will reject any improperly encoded
3102 DigestInfo structures.
3104 Note: this is a precautionary measure and no attacks are currently known.
3108 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3110 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3111 SRP code can be overrun an internal buffer. Add sanity check that
3112 g, A, B < N to SRP code.
3114 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3115 Group for discovering this issue.
3119 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3120 TLS 1.0 instead of higher protocol versions when the ClientHello message
3121 is badly fragmented. This allows a man-in-the-middle attacker to force a
3122 downgrade to TLS 1.0 even if both the server and the client support a
3123 higher protocol version, by modifying the client's TLS records.
3125 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3126 researching this issue.
3130 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3131 to a denial of service attack. A malicious server can crash the client
3132 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3133 ciphersuite and sending carefully crafted handshake messages.
3135 Thanks to Felix Gröbert (Google) for discovering and researching this
3140 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3141 to leak memory. This can be exploited through a Denial of Service attack.
3142 Thanks to Adam Langley for discovering and researching this issue.
3146 *) An attacker can force openssl to consume large amounts of memory whilst
3147 processing DTLS handshake messages. This can be exploited through a
3148 Denial of Service attack.
3149 Thanks to Adam Langley for discovering and researching this issue.
3153 *) An attacker can force an error condition which causes openssl to crash
3154 whilst processing DTLS packets due to memory being freed twice. This
3155 can be exploited through a Denial of Service attack.
3156 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3161 *) If a multithreaded client connects to a malicious server using a resumed
3162 session and the server sends an ec point format extension it could write
3163 up to 255 bytes to freed memory.
3165 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3170 *) A malicious server can crash an OpenSSL client with a null pointer
3171 dereference (read) by specifying an SRP ciphersuite even though it was not
3172 properly negotiated with the client. This can be exploited through a
3173 Denial of Service attack.
3175 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3176 discovering and researching this issue.
3180 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3181 X509_name_oneline, X509_name_print_ex et al. to leak some information
3182 from the stack. Applications may be affected if they echo pretty printing
3183 output to the attacker.
3185 Thanks to Ivan Fratric (Google) for discovering this issue.
3187 [Emilia Käsper, and Steve Henson]
3189 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3190 for corner cases. (Certain input points at infinity could lead to
3191 bogus results, with non-infinity inputs mapped to infinity too.)
3194 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3196 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3197 handshake can force the use of weak keying material in OpenSSL
3198 SSL/TLS clients and servers.
3200 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3201 researching this issue. (CVE-2014-0224)
3202 [KIKUCHI Masashi, Steve Henson]
3204 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3205 OpenSSL DTLS client the code can be made to recurse eventually crashing
3208 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3210 [Imre Rad, Steve Henson]
3212 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3213 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3214 client or server. This is potentially exploitable to run arbitrary
3215 code on a vulnerable client or server.
3217 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3218 [Jüri Aedla, Steve Henson]
3220 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3221 are subject to a denial of service attack.
3223 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3224 this issue. (CVE-2014-3470)
3225 [Felix Gröbert, Ivan Fratric, Steve Henson]
3227 *) Harmonize version and its documentation. -f flag is used to display
3229 [mancha <mancha1@zoho.com>]
3231 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3232 in i2d_ECPrivateKey.
3233 [mancha <mancha1@zoho.com>]
3235 *) Fix some double frees. These are not thought to be exploitable.
3236 [mancha <mancha1@zoho.com>]
3238 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3240 *) A missing bounds check in the handling of the TLS heartbeat extension
3241 can be used to reveal up to 64k of memory to a connected client or
3244 Thanks for Neel Mehta of Google Security for discovering this bug and to
3245 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3246 preparing the fix (CVE-2014-0160)
3247 [Adam Langley, Bodo Moeller]
3249 *) Fix for the attack described in the paper "Recovering OpenSSL
3250 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3251 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3252 http://eprint.iacr.org/2014/140
3254 Thanks to Yuval Yarom and Naomi Benger for discovering this
3255 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3256 [Yuval Yarom and Naomi Benger]
3258 *) TLS pad extension: draft-agl-tls-padding-03
3260 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3261 TLS client Hello record length value would otherwise be > 255 and
3262 less that 512 pad with a dummy extension containing zeroes so it
3263 is at least 512 bytes long.
3265 [Adam Langley, Steve Henson]
3267 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3269 *) Fix for TLS record tampering bug. A carefully crafted invalid
3270 handshake could crash OpenSSL with a NULL pointer exception.
3271 Thanks to Anton Johansson for reporting this issues.
3274 *) Keep original DTLS digest and encryption contexts in retransmission
3275 structures so we can use the previous session parameters if they need
3276 to be resent. (CVE-2013-6450)
3279 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3280 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3281 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3282 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3283 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3284 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3285 [Rob Stradling, Adam Langley]
3287 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3289 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3290 supporting platforms or when small records were transferred.
3291 [Andy Polyakov, Steve Henson]
3293 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3295 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3297 This addresses the flaw in CBC record processing discovered by
3298 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3299 at: http://www.isg.rhul.ac.uk/tls/
3301 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3302 Security Group at Royal Holloway, University of London
3303 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3304 Emilia Käsper for the initial patch.
3306 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3308 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3309 ciphersuites which can be exploited in a denial of service attack.
3310 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3311 and detecting this bug and to Wolfgang Ettlinger
3312 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3316 *) Return an error when checking OCSP signatures when key is NULL.
3317 This fixes a DoS attack. (CVE-2013-0166)
3320 *) Make openssl verify return errors.
3321 [Chris Palmer <palmer@google.com> and Ben Laurie]
3323 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3324 the right response is stapled. Also change SSL_get_certificate()
3325 so it returns the certificate actually sent.
3326 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3327 [Rob Stradling <rob.stradling@comodo.com>]
3329 *) Fix possible deadlock when decoding public keys.
3332 *) Don't use TLS 1.0 record version number in initial client hello
3336 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3338 *) Sanity check record length before skipping explicit IV in TLS
3339 1.2, 1.1 and DTLS to fix DoS attack.
3341 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3342 fuzzing as a service testing platform.
3346 *) Initialise tkeylen properly when encrypting CMS messages.
3347 Thanks to Solar Designer of Openwall for reporting this issue.
3350 *) In FIPS mode don't try to use composite ciphers as they are not
3354 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3356 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3357 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3358 mean any application compiled against OpenSSL 1.0.0 headers setting
3359 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3360 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3361 0x10000000L Any application which was previously compiled against
3362 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3363 will need to be recompiled as a result. Letting be results in
3364 inability to disable specifically TLS 1.1 and in client context,
3365 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3368 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3369 disable just protocol X, but all protocols above X *if* there are
3370 protocols *below* X still enabled. In more practical terms it means
3371 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3372 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3373 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3377 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3379 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3380 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3381 in CRYPTO_realloc_clean.
3383 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3384 issue and to Adam Langley <agl@chromium.org> for fixing it.
3386 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3388 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3391 *) Workarounds for some broken servers that "hang" if a client hello
3392 record length exceeds 255 bytes.
3394 1. Do not use record version number > TLS 1.0 in initial client
3395 hello: some (but not all) hanging servers will now work.
3396 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3397 the number of ciphers sent in the client hello. This should be
3398 set to an even number, such as 50, for example by passing:
3399 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3400 Most broken servers should now work.
3401 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3402 TLS 1.2 client support entirely.
3405 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3408 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3410 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3411 STRING form instead of a DigestInfo.
3414 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3415 and the RSA_sign/RSA_verify functions. This was made more apparent when
3416 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3417 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3418 the correct format in RSA_verify so both forms transparently work.
3421 *) Some servers which support TLS 1.0 can choke if we initially indicate
3422 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3423 encrypted premaster secret. As a workaround use the maximum permitted
3424 client version in client hello, this should keep such servers happy
3425 and still work with previous versions of OpenSSL.
3428 *) Add support for TLS/DTLS heartbeats.
3429 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3431 *) Add support for SCTP.
3432 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3434 *) Improved PRNG seeding for VOS.
3435 [Paul Green <Paul.Green@stratus.com>]
3437 *) Extensive assembler packs updates, most notably:
3439 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3440 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3441 - x86_64: bit-sliced AES implementation;
3442 - ARM: NEON support, contemporary platforms optimizations;
3443 - s390x: z196 support;
3444 - *: GHASH and GF(2^m) multiplication implementations;
3448 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3449 (removal of unnecessary code)
3450 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3452 *) Add TLS key material exporter from RFC 5705.
3455 *) Add DTLS-SRTP negotiation from RFC 5764.
3458 *) Add Next Protocol Negotiation,
3459 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3460 disabled with a no-npn flag to config or Configure. Code donated
3462 [Adam Langley <agl@google.com> and Ben Laurie]
3464 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3465 NIST-P256, NIST-P521, with constant-time single point multiplication on
3466 typical inputs. Compiler support for the nonstandard type __uint128_t is
3467 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3468 Code made available under Apache License version 2.0.
3470 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3471 line to include this in your build of OpenSSL, and run "make depend" (or
3472 "make update"). This enables the following EC_METHODs:
3474 EC_GFp_nistp224_method()
3475 EC_GFp_nistp256_method()
3476 EC_GFp_nistp521_method()
3478 EC_GROUP_new_by_curve_name() will automatically use these (while
3479 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3481 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3483 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3484 all platforms. Move ssize_t definition from e_os.h to the public
3485 header file e_os2.h as it now appears in public header file cms.h
3488 *) New -sigopt option to the ca, req and x509 utilities. Additional
3489 signature parameters can be passed using this option and in
3493 *) Add RSA PSS signing function. This will generate and set the
3494 appropriate AlgorithmIdentifiers for PSS based on those in the
3495 corresponding EVP_MD_CTX structure. No application support yet.
3498 *) Support for companion algorithm specific ASN1 signing routines.
3499 New function ASN1_item_sign_ctx() signs a pre-initialised
3500 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3501 the appropriate parameters.
3504 *) Add new algorithm specific ASN1 verification initialisation function
3505 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3506 handling will be the same no matter what EVP_PKEY_METHOD is used.
3507 Add a PSS handler to support verification of PSS signatures: checked
3508 against a number of sample certificates.
3511 *) Add signature printing for PSS. Add PSS OIDs.
3512 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3514 *) Add algorithm specific signature printing. An individual ASN1 method
3515 can now print out signatures instead of the standard hex dump.
3517 More complex signatures (e.g. PSS) can print out more meaningful
3518 information. Include DSA version that prints out the signature
3522 *) Password based recipient info support for CMS library: implementing
3526 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3527 neatly separates the code into cipher and PBE sections and is required
3528 for some algorithms that split PBES2 into separate pieces (such as
3529 password based CMS).
3532 *) Session-handling fixes:
3533 - Fix handling of connections that are resuming with a session ID,
3534 but also support Session Tickets.
3535 - Fix a bug that suppressed issuing of a new ticket if the client
3536 presented a ticket with an expired session.
3537 - Try to set the ticket lifetime hint to something reasonable.
3538 - Make tickets shorter by excluding irrelevant information.
3539 - On the client side, don't ignore renewed tickets.
3540 [Adam Langley, Bodo Moeller (Google)]
3542 *) Fix PSK session representation.
3545 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3547 This work was sponsored by Intel.
3550 *) Add GCM support to TLS library. Some custom code is needed to split
3551 the IV between the fixed (from PRF) and explicit (from TLS record)
3552 portions. This adds all GCM ciphersuites supported by RFC5288 and
3553 RFC5289. Generalise some AES* cipherstrings to include GCM and
3554 add a special AESGCM string for GCM only.
3557 *) Expand range of ctrls for AES GCM. Permit setting invocation
3558 field on decrypt and retrieval of invocation field only on encrypt.
3561 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3562 As required by RFC5289 these ciphersuites cannot be used if for
3563 versions of TLS earlier than 1.2.
3566 *) For FIPS capable OpenSSL interpret a NULL default public key method
3567 as unset and return the appropriate default but do *not* set the default.
3568 This means we can return the appropriate method in applications that
3569 switch between FIPS and non-FIPS modes.
3572 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3573 ENGINE is used then we cannot handle that in the FIPS module so we
3574 keep original code iff non-FIPS operations are allowed.
3577 *) Add -attime option to openssl utilities.
3578 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3580 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3583 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3584 FIPS EC methods unconditionally for now.
3587 *) New build option no-ec2m to disable characteristic 2 code.
3590 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3591 all cases can be covered as some introduce binary incompatibilities.
3594 *) Redirect RSA operations to FIPS module including keygen,
3595 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3598 *) Add similar low level API blocking to ciphers.
3601 *) Low level digest APIs are not approved in FIPS mode: any attempt
3602 to use these will cause a fatal error. Applications that *really* want
3603 to use them can use the private_* version instead.
3606 *) Redirect cipher operations to FIPS module for FIPS builds.
3609 *) Redirect digest operations to FIPS module for FIPS builds.
3612 *) Update build system to add "fips" flag which will link in fipscanister.o
3613 for static and shared library builds embedding a signature if needed.
3616 *) Output TLS supported curves in preference order instead of numerical
3617 order. This is currently hardcoded for the highest order curves first.
3618 This should be configurable so applications can judge speed vs strength.
3621 *) Add TLS v1.2 server support for client authentication.
3624 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3628 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3629 FIPS modules versions.
3632 *) Add TLS v1.2 client side support for client authentication. Keep cache
3633 of handshake records longer as we don't know the hash algorithm to use
3634 until after the certificate request message is received.
3637 *) Initial TLS v1.2 client support. Add a default signature algorithms
3638 extension including all the algorithms we support. Parse new signature
3639 format in client key exchange. Relax some ECC signing restrictions for
3640 TLS v1.2 as indicated in RFC5246.
3643 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3644 to new signature format when needed using client digest preference.
3645 All server ciphersuites should now work correctly in TLS v1.2. No client
3646 support yet and no support for client certificates.
3649 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3650 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3651 ciphersuites. At present only RSA key exchange ciphersuites work with
3652 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3653 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3654 and version checking.
3657 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3658 with this defined it will not be affected by any changes to ssl internal
3659 structures. Add several utility functions to allow openssl application
3660 to work with OPENSSL_NO_SSL_INTERN defined.
3663 *) A long standing patch to add support for SRP from EdelWeb (Peter
3664 Sylvester and Christophe Renou) was integrated.
3665 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3666 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3669 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3672 *) Permit abbreviated handshakes when renegotiating using the function
3673 SSL_renegotiate_abbreviated().
3674 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3676 *) Add call to ENGINE_register_all_complete() to
3677 ENGINE_load_builtin_engines(), so some implementations get used
3678 automatically instead of needing explicit application support.
3681 *) Add support for TLS key exporter as described in RFC5705.
3682 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3684 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3685 a few changes are required:
3687 Add SSL_OP_NO_TLSv1_1 flag.
3688 Add TLSv1_1 methods.
3689 Update version checking logic to handle version 1.1.
3690 Add explicit IV handling (ported from DTLS code).
3691 Add command line options to s_client/s_server.
3694 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3696 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3697 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3698 content decryption and always return the same error. Note: this attack
3699 needs on average 2^20 messages so it only affects automated senders. The
3700 old behaviour can be re-enabled in the CMS code by setting the
3701 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3702 an MMA defence is not necessary.
3703 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3704 this issue. (CVE-2012-0884)
3707 *) Fix CVE-2011-4619: make sure we really are receiving a
3708 client hello before rejecting multiple SGC restarts. Thanks to
3709 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3712 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3714 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3715 Thanks to Antonio Martin, Enterprise Secure Access Research and
3716 Development, Cisco Systems, Inc. for discovering this bug and
3717 preparing a fix. (CVE-2012-0050)
3720 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3722 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3723 of the Vaudenay padding oracle attack on CBC mode encryption
3724 which enables an efficient plaintext recovery attack against
3725 the OpenSSL implementation of DTLS. Their attack exploits timing
3726 differences arising during decryption processing. A research
3727 paper describing this attack can be found at:
3728 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3729 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3730 Security Group at Royal Holloway, University of London
3731 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3732 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3733 for preparing the fix. (CVE-2011-4108)
3734 [Robin Seggelmann, Michael Tuexen]
3736 *) Clear bytes used for block padding of SSL 3.0 records.
3738 [Adam Langley (Google)]
3740 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3741 Kadianakis <desnacked@gmail.com> for discovering this issue and
3742 Adam Langley for preparing the fix. (CVE-2011-4619)
3743 [Adam Langley (Google)]
3745 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3746 [Andrey Kulikov <amdeich@gmail.com>]
3748 *) Prevent malformed RFC3779 data triggering an assertion failure.
3749 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3750 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3751 [Rob Austein <sra@hactrn.net>]
3753 *) Improved PRNG seeding for VOS.
3754 [Paul Green <Paul.Green@stratus.com>]
3756 *) Fix ssl_ciph.c set-up race.
3757 [Adam Langley (Google)]
3759 *) Fix spurious failures in ecdsatest.c.
3760 [Emilia Käsper (Google)]
3762 *) Fix the BIO_f_buffer() implementation (which was mixing different
3763 interpretations of the '..._len' fields).
3764 [Adam Langley (Google)]
3766 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3767 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3768 threads won't reuse the same blinding coefficients.
3770 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3771 lock to call BN_BLINDING_invert_ex, and avoids one use of
3772 BN_BLINDING_update for each BN_BLINDING structure (previously,
3773 the last update always remained unused).
3774 [Emilia Käsper (Google)]
3776 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3777 [Bob Buckholz (Google)]
3779 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3781 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3782 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3783 [Kaspar Brand <ossl@velox.ch>]
3785 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3786 for multi-threaded use of ECDH. (CVE-2011-3210)
3787 [Adam Langley (Google)]
3789 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3792 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3793 signature public key algorithm by using OID xref utilities instead.
3794 Before this you could only use some ECC ciphersuites with SHA1 only.
3797 *) Add protection against ECDSA timing attacks as mentioned in the paper
3798 by Billy Bob Brumley and Nicola Tuveri, see:
3800 http://eprint.iacr.org/2011/232.pdf
3802 [Billy Bob Brumley and Nicola Tuveri]
3804 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3806 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3807 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3809 *) Fix bug in string printing code: if *any* escaping is enabled we must
3810 escape the escape character (backslash) or the resulting string is
3814 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3816 *) Disable code workaround for ancient and obsolete Netscape browsers
3817 and servers: an attacker can use it in a ciphersuite downgrade attack.
3818 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3821 *) Fixed J-PAKE implementation error, originally discovered by
3822 Sebastien Martini, further info and confirmation from Stefan
3823 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3826 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3828 *) Fix extension code to avoid race conditions which can result in a buffer
3829 overrun vulnerability: resumed sessions must not be modified as they can
3830 be shared by multiple threads. CVE-2010-3864
3833 *) Fix WIN32 build system to correctly link an ENGINE directory into
3837 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3839 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3841 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3843 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3845 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3846 context. The operation can be customised via the ctrl mechanism in
3847 case ENGINEs want to include additional functionality.
3850 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3853 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3854 output hashes compatible with older versions of OpenSSL.
3855 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3857 *) Fix compression algorithm handling: if resuming a session use the
3858 compression algorithm of the resumed session instead of determining
3859 it from client hello again. Don't allow server to change algorithm.
3862 *) Add load_crls() function to apps tidying load_certs() too. Add option
3863 to verify utility to allow additional CRLs to be included.
3866 *) Update OCSP request code to permit adding custom headers to the request:
3867 some responders need this.
3870 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3872 [Julia Lawall <julia@diku.dk>]
3874 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3875 needlessly dereferenced structures, used obsolete functions and
3876 didn't handle all updated verify codes correctly.
3879 *) Disable MD2 in the default configuration.
3882 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3883 indicate the initial BIO being pushed or popped. This makes it possible
3884 to determine whether the BIO is the one explicitly called or as a result
3885 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3886 it handles reference counts correctly and doesn't zero out the I/O bio
3887 when it is not being explicitly popped. WARNING: applications which
3888 included workarounds for the old buggy behaviour will need to be modified
3889 or they could free up already freed BIOs.
3892 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3893 renaming to all platforms (within the 0.9.8 branch, this was
3894 done conditionally on Netware platforms to avoid a name clash).
3895 [Guenter <lists@gknw.net>]
3897 *) Add ECDHE and PSK support to DTLS.
3898 [Michael Tuexen <tuexen@fh-muenster.de>]
3900 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3904 *) Add "missing" function EVP_MD_flags() (without this the only way to
3905 retrieve a digest flags is by accessing the structure directly. Update
3906 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3907 or cipher is registered as in the "from" argument. Print out all
3908 registered digests in the dgst usage message instead of manually
3909 attempting to work them out.
3912 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3913 this allows the use of compression and extensions. Change default cipher
3914 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3915 by default unless an application cipher string requests it.
3918 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3919 key ids to find matching certificates and keys but some PKCS#12 files
3920 don't follow the (somewhat unwritten) rules and this strategy fails.
3921 Now just gather all certificates together and the first private key
3922 then look for the first certificate that matches the key.
3925 *) Support use of registered digest and cipher names for dgst and cipher
3926 commands instead of having to add each one as a special case. So now
3933 openssl dgst -sha256 foo
3935 and this works for ENGINE based algorithms too.
3939 *) Update Gost ENGINE to support parameter files.
3940 [Victor B. Wagner <vitus@cryptocom.ru>]
3942 *) Support GeneralizedTime in ca utility.
3943 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3945 *) Enhance the hash format used for certificate directory links. The new
3946 form uses the canonical encoding (meaning equivalent names will work
3947 even if they aren't identical) and uses SHA1 instead of MD5. This form
3948 is incompatible with the older format and as a result c_rehash should
3949 be used to rebuild symbolic links.
3952 *) Make PKCS#8 the default write format for private keys, replacing the
3953 traditional format. This form is standardised, more secure and doesn't
3954 include an implicit MD5 dependency.
3957 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3958 committed to OpenSSL should pass this lot as a minimum.
3961 *) Add session ticket override functionality for use by EAP-FAST.
3962 [Jouni Malinen <j@w1.fi>]
3964 *) Modify HMAC functions to return a value. Since these can be implemented
3965 in an ENGINE errors can occur.
3968 *) Type-checked OBJ_bsearch_ex.
3971 *) Type-checked OBJ_bsearch. Also some constification necessitated
3972 by type-checking. Still to come: TXT_DB, bsearch(?),
3973 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3977 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3978 seconds to a tm structure directly, instead of going through OS
3979 specific date routines. This avoids any issues with OS routines such
3980 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3981 and X509_time_adj_ex() to cover the extended range. The existing
3982 X509_time_adj() is still usable and will no longer have any date issues.
3985 *) Delta CRL support. New use deltas option which will attempt to locate
3986 and search any appropriate delta CRLs available.
3988 This work was sponsored by Google.
3991 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3992 code and add additional score elements. Validate alternate CRL paths
3993 as part of the CRL checking and indicate a new error "CRL path validation
3994 error" in this case. Applications wanting additional details can use
3995 the verify callback and check the new "parent" field. If this is not
3996 NULL CRL path validation is taking place. Existing applications won't
3997 see this because it requires extended CRL support which is off by
4000 This work was sponsored by Google.
4003 *) Support for freshest CRL extension.
4005 This work was sponsored by Google.
4008 *) Initial indirect CRL support. Currently only supported in the CRLs
4009 passed directly and not via lookup. Process certificate issuer
4010 CRL entry extension and lookup CRL entries by bother issuer name
4011 and serial number. Check and process CRL issuer entry in IDP extension.
4013 This work was sponsored by Google.
4016 *) Add support for distinct certificate and CRL paths. The CRL issuer
4017 certificate is validated separately in this case. Only enabled if
4018 an extended CRL support flag is set: this flag will enable additional
4019 CRL functionality in future.
4021 This work was sponsored by Google.
4024 *) Add support for policy mappings extension.
4026 This work was sponsored by Google.
4029 *) Fixes to pathlength constraint, self issued certificate handling,
4030 policy processing to align with RFC3280 and PKITS tests.
4032 This work was sponsored by Google.
4035 *) Support for name constraints certificate extension. DN, email, DNS
4036 and URI types are currently supported.
4038 This work was sponsored by Google.
4041 *) To cater for systems that provide a pointer-based thread ID rather
4042 than numeric, deprecate the current numeric thread ID mechanism and
4043 replace it with a structure and associated callback type. This
4044 mechanism allows a numeric "hash" to be extracted from a thread ID in
4045 either case, and on platforms where pointers are larger than 'long',
4046 mixing is done to help ensure the numeric 'hash' is usable even if it
4047 can't be guaranteed unique. The default mechanism is to use "&errno"
4048 as a pointer-based thread ID to distinguish between threads.
4050 Applications that want to provide their own thread IDs should now use
4051 CRYPTO_THREADID_set_callback() to register a callback that will call
4052 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4054 Note that ERR_remove_state() is now deprecated, because it is tied
4055 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4056 to free the current thread's error state should be replaced by
4057 ERR_remove_thread_state(NULL).
4059 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4060 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4061 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4062 application was previously providing a numeric thread callback that
4063 was inappropriate for distinguishing threads, then uniqueness might
4064 have been obtained with &errno that happened immediately in the
4065 intermediate development versions of OpenSSL; this is no longer the
4066 case, the numeric thread callback will now override the automatic use
4068 [Geoff Thorpe, with help from Bodo Moeller]
4070 *) Initial support for different CRL issuing certificates. This covers a
4071 simple case where the self issued certificates in the chain exist and
4072 the real CRL issuer is higher in the existing chain.
4074 This work was sponsored by Google.
4077 *) Removed effectively defunct crypto/store from the build.
4080 *) Revamp of STACK to provide stronger type-checking. Still to come:
4081 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4082 ASN1_STRING, CONF_VALUE.
4085 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4086 RAM on SSL connections. This option can save about 34k per idle SSL.
4089 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4090 STACK, TXT_DB, bsearch, qsort.
4093 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4094 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4095 support for data, signedData, compressedData, digestedData and
4096 encryptedData, envelopedData types included. Scripts to check against
4097 RFC4134 examples draft and interop and consistency checks of many
4098 content types and variants.
4101 *) Add options to enc utility to support use of zlib compression BIO.
4104 *) Extend mk1mf to support importing of options and assembly language
4105 files from Configure script, currently only included in VC-WIN32.
4106 The assembly language rules can now optionally generate the source
4107 files from the associated perl scripts.
4110 *) Implement remaining functionality needed to support GOST ciphersuites.
4111 Interop testing has been performed using CryptoPro implementations.
4112 [Victor B. Wagner <vitus@cryptocom.ru>]
4114 *) s390x assembler pack.
4117 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4121 *) Implement Opaque PRF Input TLS extension as specified in
4122 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4123 official specification yet and no extension type assignment by
4124 IANA exists, this extension (for now) will have to be explicitly
4125 enabled when building OpenSSL by providing the extension number
4126 to use. For example, specify an option
4128 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4130 to the "config" or "Configure" script to enable the extension,
4131 assuming extension number 0x9527 (which is a completely arbitrary
4132 and unofficial assignment based on the MD5 hash of the Internet
4133 Draft). Note that by doing so, you potentially lose
4134 interoperability with other TLS implementations since these might
4135 be using the same extension number for other purposes.
4137 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4138 opaque PRF input value to use in the handshake. This will create
4139 an internal copy of the length-'len' string at 'src', and will
4140 return non-zero for success.
4142 To get more control and flexibility, provide a callback function
4145 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4146 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4150 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4153 Callback function 'cb' will be called in handshakes, and is
4154 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4155 Argument 'arg' is for application purposes (the value as given to
4156 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4157 be provided to the callback function). The callback function
4158 has to return non-zero to report success: usually 1 to use opaque
4159 PRF input just if possible, or 2 to enforce use of the opaque PRF
4160 input. In the latter case, the library will abort the handshake
4161 if opaque PRF input is not successfully negotiated.
4163 Arguments 'peerinput' and 'len' given to the callback function
4164 will always be NULL and 0 in the case of a client. A server will
4165 see the client's opaque PRF input through these variables if
4166 available (NULL and 0 otherwise). Note that if the server
4167 provides an opaque PRF input, the length must be the same as the
4168 length of the client's opaque PRF input.
4170 Note that the callback function will only be called when creating
4171 a new session (session resumption can resume whatever was
4172 previously negotiated), and will not be called in SSL 2.0
4173 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4174 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4175 for applications that need to enforce opaque PRF input.
4179 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4182 [Victor B. Wagner <vitus@cryptocom.ru>]
4184 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4185 RFC4507bis. The encrypted ticket format is an encrypted encoded
4186 SSL_SESSION structure, that way new session features are automatically
4189 If a client application caches session in an SSL_SESSION structure
4190 support is transparent because tickets are now stored in the encoded
4193 The SSL_CTX structure automatically generates keys for ticket
4194 protection in servers so again support should be possible
4195 with no application modification.
4197 If a client or server wishes to disable RFC4507 support then the option
4198 SSL_OP_NO_TICKET can be set.
4200 Add a TLS extension debugging callback to allow the contents of any client
4201 or server extensions to be examined.
4203 This work was sponsored by Google.
4206 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4207 OpenSSL should now compile cleanly on gcc 4.2
4208 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4210 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4211 support including streaming MAC support: this is required for GOST
4212 ciphersuite support.
4213 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4215 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4216 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4217 to output in BER and PEM format.
4220 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4221 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4222 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4223 ENGINE support for HMAC keys which are unextractable. New -mac and
4224 -macopt options to dgst utility.
4227 *) New option -sigopt to dgst utility. Update dgst to use
4228 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4229 alternative signing parameters such as X9.31 or PSS in the dgst
4233 *) Change ssl_cipher_apply_rule(), the internal function that does
4234 the work each time a ciphersuite string requests enabling
4235 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4236 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4237 the order of disabled ciphersuites such that those ciphersuites
4238 that most recently went from enabled to disabled not only stay
4239 in order with respect to each other, but also have higher priority
4240 than other disabled ciphersuites the next time ciphersuites are
4243 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4244 the same ciphersuites as with "HIGH" alone, but in a specific
4245 order where the PSK ciphersuites come first (since they are the
4246 most recently disabled ciphersuites when "HIGH" is parsed).
4248 Also, change ssl_create_cipher_list() (using this new
4249 functionality) such that between otherwise identical
4250 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4254 *) Change ssl_create_cipher_list() so that it automatically
4255 arranges the ciphersuites in reasonable order before starting
4256 to process the rule string. Thus, the definition for "DEFAULT"
4257 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4258 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4259 This makes it much easier to arrive at a reasonable default order
4260 in applications for which anonymous ciphers are OK (meaning
4261 that you can't actually use DEFAULT).
4262 [Bodo Moeller; suggested by Victor Duchovni]
4264 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4265 processing) into multiple integers instead of setting
4266 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4267 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4268 (These masks as well as the individual bit definitions are hidden
4269 away into the non-exported interface ssl/ssl_locl.h, so this
4270 change to the definition of the SSL_CIPHER structure shouldn't
4271 affect applications.) This give us more bits for each of these
4272 categories, so there is no longer a need to coagulate AES128 and
4273 AES256 into a single algorithm bit, and to coagulate Camellia128
4274 and Camellia256 into a single algorithm bit, which has led to all
4277 Thus, among other things, the kludge introduced in 0.9.7m and
4278 0.9.8e for masking out AES256 independently of AES128 or masking
4279 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4281 With the change, we also introduce new ciphersuite aliases that
4282 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4286 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4287 Use the leftmost N bytes of the signature input if the input is
4288 larger than the prime q (with N being the size in bytes of q).
4291 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4292 it yet and it is largely untested.
4295 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4298 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4299 some compilers (gcc 4.2 and later) reject their use. Safestack is
4300 reimplemented. Update ASN1 to avoid use of legacy functions.
4303 *) Win32/64 targets are linked with Winsock2.
4306 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4307 to external functions. This can be used to increase CRL handling
4308 efficiency especially when CRLs are very large by (for example) storing
4309 the CRL revoked certificates in a database.
4312 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4313 new CRLs added to a directory can be used. New command line option
4314 -verify_return_error to s_client and s_server. This causes real errors
4315 to be returned by the verify callback instead of carrying on no matter
4316 what. This reflects the way a "real world" verify callback would behave.
4319 *) GOST engine, supporting several GOST algorithms and public key formats.
4320 Kindly donated by Cryptocom.
4323 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4324 partitioned by DP are handled but no indirect CRL or reason partitioning
4325 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4326 selected via a scoring technique which handles IDP and AKID in CRLs.
4329 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4330 will ultimately be used for all verify operations: this will remove the
4331 X509_STORE dependency on certificate verification and allow alternative
4332 lookup methods. X509_STORE based implementations of these two callbacks.
4335 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4336 Modify get_crl() to find a valid (unexpired) CRL if possible.
4339 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4340 this would be called X509_CRL_cmp() but that name is already used by
4341 a function that just compares CRL issuer names. Cache several CRL
4342 extensions in X509_CRL structure and cache CRLDP in X509.
4345 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4346 this maps equivalent X509_NAME structures into a consistent structure.
4347 Name comparison can then be performed rapidly using memcmp().
4350 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4354 *) Allow digests to supply their own micalg string for S/MIME type using
4355 the ctrl EVP_MD_CTRL_MICALG.
4358 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4359 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4360 ctrl. It can then customise the structure before and/or after signing
4364 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4365 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4366 to free up any added signature OIDs.
4369 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4370 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4371 digest and cipher tables. New options added to openssl utility:
4372 list-message-digest-algorithms and list-cipher-algorithms.
4375 *) Change the array representation of binary polynomials: the list
4376 of degrees of non-zero coefficients is now terminated with -1.
4377 Previously it was terminated with 0, which was also part of the
4378 value; thus, the array representation was not applicable to
4379 polynomials where t^0 has coefficient zero. This change makes
4380 the array representation useful in a more general context.
4383 *) Various modifications and fixes to SSL/TLS cipher string
4384 handling. For ECC, the code now distinguishes between fixed ECDH
4385 with RSA certificates on the one hand and with ECDSA certificates
4386 on the other hand, since these are separate ciphersuites. The
4387 unused code for Fortezza ciphersuites has been removed.
4389 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4390 (not "ECDHE"). For consistency with the code for DH
4391 certificates, use of ECDH certificates is now considered ECDH
4392 authentication, not RSA or ECDSA authentication (the latter is
4393 merely the CA's signing algorithm and not actively used in the
4396 The temporary ciphersuite alias "ECCdraft" is no longer
4397 available, and ECC ciphersuites are no longer excluded from "ALL"
4398 and "DEFAULT". The following aliases now exist for RFC 4492
4399 ciphersuites, most of these by analogy with the DH case:
4401 kECDHr - ECDH cert, signed with RSA
4402 kECDHe - ECDH cert, signed with ECDSA
4403 kECDH - ECDH cert (signed with either RSA or ECDSA)
4404 kEECDH - ephemeral ECDH
4405 ECDH - ECDH cert or ephemeral ECDH
4411 AECDH - anonymous ECDH
4412 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4416 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4417 Use correct micalg parameters depending on digest(s) in signed message.
4420 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4421 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4424 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4425 an engine to register a method. Add ENGINE lookups for methods and
4426 functional reference processing.
4429 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4430 EVP_{Sign,Verify}* which allow an application to customise the signature
4434 *) New -resign option to smime utility. This adds one or more signers
4435 to an existing PKCS#7 signedData structure. Also -md option to use an
4436 alternative message digest algorithm for signing.
4439 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4440 create PKCS7 structures containing multiple signers. Update smime
4441 application to support multiple signers.
4444 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4448 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4449 Reorganize PBE internals to lookup from a static table using NIDs,
4450 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4451 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4452 PRF which will be automatically used with PBES2.
4455 *) Replace the algorithm specific calls to generate keys in "req" with the
4459 *) Update PKCS#7 enveloped data routines to use new API. This is now
4460 supported by any public key method supporting the encrypt operation. A
4461 ctrl is added to allow the public key algorithm to examine or modify
4462 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4466 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4467 a default digest type to use. In most cases this will be SHA1 but some
4468 algorithms (such as GOST) need to specify an alternative digest. The
4469 return value indicates how strong the preference is 1 means optional and
4470 2 is mandatory (that is it is the only supported type). Modify
4471 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4472 use the default md. Update openssl utilities to use the default digest
4473 type for signing if it is not explicitly indicated.
4476 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4477 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4478 signing method from the key type. This effectively removes the link
4479 between digests and public key types.
4482 *) Add an OID cross reference table and utility functions. Its purpose is to
4483 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4484 rsaEncryption. This will allow some of the algorithm specific hackery
4485 needed to use the correct OID to be removed.
4488 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4489 structures for PKCS7_sign(). They are now set up by the relevant public
4493 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4496 *) Add support for key derivation (agreement) in the API, DH method and
4500 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4501 public and private key formats. As a side effect these add additional
4502 command line functionality not previously available: DSA signatures can be
4503 generated and verified using pkeyutl and DH key support and generation in
4508 [Oliver Tappe <zooey@hirschkaefer.de>]
4510 *) New make target "install_html_docs" installs HTML renditions of the
4512 [Oliver Tappe <zooey@hirschkaefer.de>]
4514 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4515 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4516 support key and parameter generation and add initial key generation
4517 functionality for RSA.
4520 *) Add functions for main EVP_PKEY_method operations. The undocumented
4521 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4522 EVP_PKEY_{encrypt,decrypt}_old.
4525 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4526 key API, doesn't do much yet.
4529 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4530 public key algorithms. New option to openssl utility:
4531 "list-public-key-algorithms" to print out info.
4534 *) Implement the Supported Elliptic Curves Extension for
4535 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4538 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4539 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4542 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4543 utilities such as rsa, dsa, dsaparam etc except they process any key
4547 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4548 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4549 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4553 *) Initial support for pluggable public key ASN1.
4554 De-spaghettify the public key ASN1 handling. Move public and private
4555 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4556 algorithm specific handling to a single module within the relevant
4557 algorithm directory. Add functions to allow (near) opaque processing
4558 of public and private key structures.
4561 *) Implement the Supported Point Formats Extension for
4562 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4565 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4566 for the psk identity [hint] and the psk callback functions to the
4567 SSL_SESSION, SSL and SSL_CTX structure.
4570 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4574 SSL_CTX_use_psk_identity_hint
4575 SSL_get_psk_identity_hint
4576 SSL_get_psk_identity
4577 SSL_use_psk_identity_hint
4579 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4581 *) Add RFC 3161 compliant time stamp request creation, response generation
4582 and response verification functionality.
4583 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4585 *) Add initial support for TLS extensions, specifically for the server_name
4586 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4587 have new members for a host name. The SSL data structure has an
4588 additional member SSL_CTX *initial_ctx so that new sessions can be
4589 stored in that context to allow for session resumption, even after the
4590 SSL has been switched to a new SSL_CTX in reaction to a client's
4591 server_name extension.
4593 New functions (subject to change):
4595 SSL_get_servername()
4596 SSL_get_servername_type()
4599 New CTRL codes and macros (subject to change):
4601 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4602 - SSL_CTX_set_tlsext_servername_callback()
4603 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4604 - SSL_CTX_set_tlsext_servername_arg()
4605 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4607 openssl s_client has a new '-servername ...' option.
4609 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4610 '-key2 ...', '-servername_fatal' (subject to change). This allows
4611 testing the HostName extension for a specific single host name ('-cert'
4612 and '-key' remain fallbacks for handshakes without HostName
4613 negotiation). If the unrecognized_name alert has to be sent, this by
4614 default is a warning; it becomes fatal with the '-servername_fatal'
4617 [Peter Sylvester, Remy Allais, Christophe Renou]
4619 *) Whirlpool hash implementation is added.
4622 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4623 bn(64,32). Because of instruction set limitations it doesn't have
4624 any negative impact on performance. This was done mostly in order
4625 to make it possible to share assembler modules, such as bn_mul_mont
4626 implementations, between 32- and 64-bit builds without hassle.
4629 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4630 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4634 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4635 dedicated Montgomery multiplication procedure, is introduced.
4636 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4637 "64-bit" performance on certain 32-bit targets.
4640 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4641 in SSL structures. New SSL ctrl to set maximum send fragment size.
4642 Save memory by setting the I/O buffer sizes dynamically instead of
4643 using the maximum available value.
4646 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4647 in addition to the text details.
4650 *) Very, very preliminary EXPERIMENTAL support for printing of general
4651 ASN1 structures. This currently produces rather ugly output and doesn't
4652 handle several customised structures at all.
4655 *) Integrated support for PVK file format and some related formats such
4656 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4657 these in the 'rsa' and 'dsa' utilities.
4660 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4663 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4664 place for the (very old) "NETSCAPE" format certificates which are now
4665 handled using new ASN1 code equivalents.
4668 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4669 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4670 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4673 *) Modify CRL distribution points extension code to print out previously
4674 unsupported fields. Enhance extension setting code to allow setting of
4678 *) Add print and set support for Issuing Distribution Point CRL extension.
4681 *) Change 'Configure' script to enable Camellia by default.
4684 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4686 *) When rejecting SSL/TLS records due to an incorrect version number, never
4687 update s->server with a new major version number. As of
4688 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4689 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4690 the previous behavior could result in a read attempt at NULL when
4691 receiving specific incorrect SSL/TLS records once record payload
4692 protection is active. (CVE-2010-0740)
4693 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4695 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4696 could be crashed if the relevant tables were not present (e.g. chrooted).
4697 [Tomas Hoger <thoger@redhat.com>]
4699 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4701 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4702 [Martin Olsson, Neel Mehta]
4704 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4705 accommodate for stack sorting, always a write lock!).
4708 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4709 excessive delays in the RAND_poll(): over a minute. As a workaround
4710 include a time check in the inner Heap32Next loop too.
4713 *) The code that handled flushing of data in SSL/TLS originally used the
4714 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4715 the problem outlined in PR#1949. The fix suggested there however can
4716 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4717 of Apache). So instead simplify the code to flush unconditionally.
4718 This should be fine since flushing with no data to flush is a no op.
4721 *) Handle TLS versions 2.0 and later properly and correctly use the
4722 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4723 off ancient servers have a habit of sticking around for a while...
4726 *) Modify compression code so it frees up structures without using the
4727 ex_data callbacks. This works around a problem where some applications
4728 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4729 restarting) then use compression (e.g. SSL with compression) later.
4730 This results in significant per-connection memory leaks and
4731 has caused some security issues including CVE-2008-1678 and
4735 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4736 change when encrypting or decrypting.
4739 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4740 connect and renegotiate with servers which do not support RI.
4741 Until RI is more widely deployed this option is enabled by default.
4744 *) Add "missing" ssl ctrls to clear options and mode.
4747 *) If client attempts to renegotiate and doesn't support RI respond with
4748 a no_renegotiation alert as required by RFC5746. Some renegotiating
4749 TLS clients will continue a connection gracefully when they receive
4750 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4751 waiting for a server hello which it will never receive. Now we treat a
4752 received no_renegotiation alert as a fatal error. This is because
4753 applications requesting a renegotiation might well expect it to succeed
4754 and would have no code in place to handle the server denying it so the
4755 only safe thing to do is to terminate the connection.
4758 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4759 peer supports secure renegotiation and 0 otherwise. Print out peer
4760 renegotiation support in s_client/s_server.
4763 *) Replace the highly broken and deprecated SPKAC certification method with
4764 the updated NID creation version. This should correctly handle UTF8.
4767 *) Implement RFC5746. Re-enable renegotiation but require the extension
4768 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4769 turns out to be a bad idea. It has been replaced by
4770 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4771 SSL_CTX_set_options(). This is really not recommended unless you
4772 know what you are doing.
4773 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4775 *) Fixes to stateless session resumption handling. Use initial_ctx when
4776 issuing and attempting to decrypt tickets in case it has changed during
4777 servername handling. Use a non-zero length session ID when attempting
4778 stateless session resumption: this makes it possible to determine if
4779 a resumption has occurred immediately after receiving server hello
4780 (several places in OpenSSL subtly assume this) instead of later in
4784 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4785 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4786 fixes for a few places where the return code is not checked
4788 [Julia Lawall <julia@diku.dk>]
4790 *) Add --strict-warnings option to Configure script to include devteam
4791 warnings in other configurations.
4794 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4795 makes it possible to install openssl libraries in locations which
4796 have names other than "lib", for example "/usr/lib64" which some
4798 [Steve Henson, based on patch from Jeremy Utley]
4800 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4801 X690 8.9.12 and can produce some misleading textual output of OIDs.
4802 [Steve Henson, reported by Dan Kaminsky]
4804 *) Delete MD2 from algorithm tables. This follows the recommendation in
4805 several standards that it is not used in new applications due to
4806 several cryptographic weaknesses. For binary compatibility reasons
4807 the MD2 API is still compiled in by default.
4810 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4814 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4815 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4817 [Guenter <lists@gknw.net>]
4819 *) Fix the server certificate chain building code to use X509_verify_cert(),
4820 it used to have an ad-hoc builder which was unable to cope with anything
4821 other than a simple chain.
4822 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4824 *) Don't check self signed certificate signatures in X509_verify_cert()
4825 by default (a flag can override this): it just wastes time without
4826 adding any security. As a useful side effect self signed root CAs
4827 with non-FIPS digests are now usable in FIPS mode.
4830 *) In dtls1_process_out_of_seq_message() the check if the current message
4831 is already buffered was missing. For every new message was memory
4832 allocated, allowing an attacker to perform an denial of service attack
4833 with sending out of seq handshake messages until there is no memory
4834 left. Additionally every future message was buffered, even if the
4835 sequence number made no sense and would be part of another handshake.
4836 So only messages with sequence numbers less than 10 in advance will be
4837 buffered. (CVE-2009-1378)
4838 [Robin Seggelmann, discovered by Daniel Mentz]
4840 *) Records are buffered if they arrive with a future epoch to be
4841 processed after finishing the corresponding handshake. There is
4842 currently no limitation to this buffer allowing an attacker to perform
4843 a DOS attack with sending records with future epochs until there is no
4844 memory left. This patch adds the pqueue_size() function to determine
4845 the size of a buffer and limits the record buffer to 100 entries.
4847 [Robin Seggelmann, discovered by Daniel Mentz]
4849 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4850 parent structure is freed. (CVE-2009-1379)
4853 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4854 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4857 [Ilya O. <vrghost@gmail.com>]
4859 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4861 *) Disable renegotiation completely - this fixes a severe security
4862 problem (CVE-2009-3555) at the cost of breaking all
4863 renegotiation. Renegotiation can be re-enabled by setting
4864 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4865 run-time. This is really not recommended unless you know what
4869 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4871 *) Don't set val to NULL when freeing up structures, it is freed up by
4872 underlying code. If sizeof(void *) > sizeof(long) this can result in
4873 zeroing past the valid field. (CVE-2009-0789)
4874 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4876 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4877 checked correctly. This would allow some invalid signed attributes to
4878 appear to verify correctly. (CVE-2009-0591)
4879 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4881 *) Reject UniversalString and BMPString types with invalid lengths. This
4882 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4883 a legal length. (CVE-2009-0590)
4886 *) Set S/MIME signing as the default purpose rather than setting it
4887 unconditionally. This allows applications to override it at the store
4891 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4892 to handle some structures.
4895 *) Improve efficiency of mem_gets: don't search whole buffer each time
4897 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4899 *) New -hex option for openssl rand.
4902 *) Print out UTF8String and NumericString when parsing ASN1.
4905 *) Support NumericString type for name components.
4908 *) Allow CC in the environment to override the automatically chosen
4909 compiler. Note that nothing is done to ensure flags work with the
4913 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4915 *) Properly check EVP_VerifyFinal() and similar return values
4917 [Ben Laurie, Bodo Moeller, Google Security Team]
4919 *) Enable TLS extensions by default.
4922 *) Allow the CHIL engine to be loaded, whether the application is
4923 multithreaded or not. (This does not release the developer from the
4924 obligation to set up the dynamic locking callbacks.)
4925 [Sander Temme <sander@temme.net>]
4927 *) Use correct exit code if there is an error in dgst command.
4928 [Steve Henson; problem pointed out by Roland Dirlewanger]
4930 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4931 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4934 *) Add experimental JPAKE support, including demo authentication in
4935 s_client and s_server.
4938 *) Set the comparison function in v3_addr_canonize().
4939 [Rob Austein <sra@hactrn.net>]
4941 *) Add support for XMPP STARTTLS in s_client.
4942 [Philip Paeps <philip@freebsd.org>]
4944 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4945 to ensure that even with this option, only ciphersuites in the
4946 server's preference list will be accepted. (Note that the option
4947 applies only when resuming a session, so the earlier behavior was
4948 just about the algorithm choice for symmetric cryptography.)
4951 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4953 *) Fix NULL pointer dereference if a DTLS server received
4954 ChangeCipherSpec as first record (CVE-2009-1386).
4957 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4958 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4961 *) The fix in 0.9.8c that supposedly got rid of unsafe
4962 double-checked locking was incomplete for RSA blinding,
4963 addressing just one layer of what turns out to have been
4964 doubly unsafe triple-checked locking.
4966 So now fix this for real by retiring the MONT_HELPER macro
4967 in crypto/rsa/rsa_eay.c.
4969 [Bodo Moeller; problem pointed out by Marius Schilder]
4971 *) Various precautionary measures:
4973 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4975 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4976 (NB: This would require knowledge of the secret session ticket key
4977 to exploit, in which case you'd be SOL either way.)
4979 - Change bn_nist.c so that it will properly handle input BIGNUMs
4980 outside the expected range.
4982 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4985 [Neel Mehta, Bodo Moeller]
4987 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4988 the load fails. Useful for distros.
4989 [Ben Laurie and the FreeBSD team]
4991 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4994 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4997 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4999 This work was sponsored by Logica.
5002 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5003 keystores. Support for SSL/TLS client authentication too.
5004 Not compiled unless enable-capieng specified to Configure.
5006 This work was sponsored by Logica.
5009 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5010 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5011 attribute creation routines such as certificate requests and PKCS#12
5015 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5017 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5018 handshake which could lead to a client crash as found using the
5019 Codenomicon TLS test suite (CVE-2008-1672)
5020 [Steve Henson, Mark Cox]
5022 *) Fix double free in TLS server name extensions which could lead to
5023 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5026 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5028 Clear the error queue to ensure that error entries left from
5029 older function calls do not interfere with the correct operation.
5030 [Lutz Jaenicke, Erik de Castro Lopo]
5032 *) Remove root CA certificates of commercial CAs:
5034 The OpenSSL project does not recommend any specific CA and does not
5035 have any policy with respect to including or excluding any CA.
5036 Therefore it does not make any sense to ship an arbitrary selection
5037 of root CA certificates with the OpenSSL software.
5040 *) RSA OAEP patches to fix two separate invalid memory reads.
5041 The first one involves inputs when 'lzero' is greater than
5042 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5043 before the beginning of from). The second one involves inputs where
5044 the 'db' section contains nothing but zeroes (there is a one-byte
5045 invalid read after the end of 'db').
5046 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5048 *) Partial backport from 0.9.9-dev:
5050 Introduce bn_mul_mont (dedicated Montgomery multiplication
5051 procedure) as a candidate for BIGNUM assembler implementation.
5052 While 0.9.9-dev uses assembler for various architectures, only
5053 x86_64 is available by default here in the 0.9.8 branch, and
5054 32-bit x86 is available through a compile-time setting.
5056 To try the 32-bit x86 assembler implementation, use Configure
5057 option "enable-montasm" (which exists only for this backport).
5059 As "enable-montasm" for 32-bit x86 disclaims code stability
5060 anyway, in this constellation we activate additional code
5061 backported from 0.9.9-dev for further performance improvements,
5062 namely BN_from_montgomery_word. (To enable this otherwise,
5063 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5065 [Andy Polyakov (backport partially by Bodo Moeller)]
5067 *) Add TLS session ticket callback. This allows an application to set
5068 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5069 values. This is useful for key rollover for example where several key
5070 sets may exist with different names.
5073 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5074 This was broken until now in 0.9.8 releases, such that the only way
5075 a registered ENGINE could be used (assuming it initialises
5076 successfully on the host) was to explicitly set it as the default
5077 for the relevant algorithms. This is in contradiction with 0.9.7
5078 behaviour and the documentation. With this fix, when an ENGINE is
5079 registered into a given algorithm's table of implementations, the
5080 'uptodate' flag is reset so that auto-discovery will be used next
5081 time a new context for that algorithm attempts to select an
5083 [Ian Lister (tweaked by Geoff Thorpe)]
5085 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5086 implementation in the following ways:
5088 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5091 Lack of BER streaming support means one pass streaming processing is
5092 only supported if data is detached: setting the streaming flag is
5093 ignored for embedded content.
5095 CMS support is disabled by default and must be explicitly enabled
5096 with the enable-cms configuration option.
5099 *) Update the GMP engine glue to do direct copies between BIGNUM and
5100 mpz_t when openssl and GMP use the same limb size. Otherwise the
5101 existing "conversion via a text string export" trick is still used.
5102 [Paul Sheer <paulsheer@gmail.com>]
5104 *) Zlib compression BIO. This is a filter BIO which compressed and
5105 uncompresses any data passed through it.
5108 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5109 RFC3394 compatible AES key wrapping.
5112 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5113 sets string data without copying. X509_ALGOR_set0() and
5114 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5115 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5116 from an X509_ATTRIBUTE structure optionally checking it occurs only
5117 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5121 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5122 to get the expected BN_FLG_CONSTTIME behavior.
5123 [Bodo Moeller (Google)]
5127 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5128 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5129 - added some more tests to do_tests.pl
5130 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5131 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5132 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5133 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5134 - various changes to netware.pl to enable gcc-cross builds on Win32
5136 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5137 - various changes to fix missing prototype warnings
5138 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5139 - added AES, WHIRLPOOL and CPUID assembler code to build files
5140 - added missing AES assembler make rules to mk1mf.pl
5141 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5142 [Guenter Knauf <eflash@gmx.net>]
5144 *) Implement certificate status request TLS extension defined in RFC3546.
5145 A client can set the appropriate parameters and receive the encoded
5146 OCSP response via a callback. A server can query the supplied parameters
5147 and set the encoded OCSP response in the callback. Add simplified examples
5148 to s_client and s_server.
5151 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5153 *) Fix various bugs:
5154 + Binary incompatibility of ssl_ctx_st structure
5155 + DTLS interoperation with non-compliant servers
5156 + Don't call get_session_cb() without proposed session
5157 + Fix ia64 assembler code
5158 [Andy Polyakov, Steve Henson]
5160 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5162 *) DTLS Handshake overhaul. There were longstanding issues with
5163 OpenSSL DTLS implementation, which were making it impossible for
5164 RFC 4347 compliant client to communicate with OpenSSL server.
5165 Unfortunately just fixing these incompatibilities would "cut off"
5166 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5167 server keeps tolerating non RFC compliant syntax. The opposite is
5168 not true, 0.9.8f client can not communicate with earlier server.
5169 This update even addresses CVE-2007-4995.
5172 *) Changes to avoid need for function casts in OpenSSL: some compilers
5173 (gcc 4.2 and later) reject their use.
5174 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5177 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5178 RFC4507bis. The encrypted ticket format is an encrypted encoded
5179 SSL_SESSION structure, that way new session features are automatically
5182 If a client application caches session in an SSL_SESSION structure
5183 support is transparent because tickets are now stored in the encoded
5186 The SSL_CTX structure automatically generates keys for ticket
5187 protection in servers so again support should be possible
5188 with no application modification.
5190 If a client or server wishes to disable RFC4507 support then the option
5191 SSL_OP_NO_TICKET can be set.
5193 Add a TLS extension debugging callback to allow the contents of any client
5194 or server extensions to be examined.
5196 This work was sponsored by Google.
5199 *) Add initial support for TLS extensions, specifically for the server_name
5200 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5201 have new members for a host name. The SSL data structure has an
5202 additional member SSL_CTX *initial_ctx so that new sessions can be
5203 stored in that context to allow for session resumption, even after the
5204 SSL has been switched to a new SSL_CTX in reaction to a client's
5205 server_name extension.
5207 New functions (subject to change):
5209 SSL_get_servername()
5210 SSL_get_servername_type()
5213 New CTRL codes and macros (subject to change):
5215 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5216 - SSL_CTX_set_tlsext_servername_callback()
5217 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5218 - SSL_CTX_set_tlsext_servername_arg()
5219 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5221 openssl s_client has a new '-servername ...' option.
5223 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5224 '-key2 ...', '-servername_fatal' (subject to change). This allows
5225 testing the HostName extension for a specific single host name ('-cert'
5226 and '-key' remain fallbacks for handshakes without HostName
5227 negotiation). If the unrecognized_name alert has to be sent, this by
5228 default is a warning; it becomes fatal with the '-servername_fatal'
5231 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5233 *) Add AES and SSE2 assembly language support to VC++ build.
5236 *) Mitigate attack on final subtraction in Montgomery reduction.
5239 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5240 (which previously caused an internal error).
5243 *) Squeeze another 10% out of IGE mode when in != out.
5246 *) AES IGE mode speedup.
5247 [Dean Gaudet (Google)]
5249 *) Add the Korean symmetric 128-bit cipher SEED (see
5250 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5251 add SEED ciphersuites from RFC 4162:
5253 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5254 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5255 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5256 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5258 To minimize changes between patchlevels in the OpenSSL 0.9.8
5259 series, SEED remains excluded from compilation unless OpenSSL
5260 is configured with 'enable-seed'.
5261 [KISA, Bodo Moeller]
5263 *) Mitigate branch prediction attacks, which can be practical if a
5264 single processor is shared, allowing a spy process to extract
5265 information. For detailed background information, see
5266 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5267 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5268 and Necessary Software Countermeasures"). The core of the change
5269 are new versions BN_div_no_branch() and
5270 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5271 respectively, which are slower, but avoid the security-relevant
5272 conditional branches. These are automatically called by BN_div()
5273 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5274 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5275 remove a conditional branch.
5277 BN_FLG_CONSTTIME is the new name for the previous
5278 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5279 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5280 in the exponent causes BN_mod_exp_mont() to use the alternative
5281 implementation in BN_mod_exp_mont_consttime().) The old name
5282 remains as a deprecated alias.
5284 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5285 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5286 constant-time implementations for more than just exponentiation.
5287 Here too the old name is kept as a deprecated alias.
5289 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5290 the BN_BLINDING structure gets an independent copy of the
5291 modulus. This means that the previous "BIGNUM *m" argument to
5292 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5293 essentially becomes "const BIGNUM *m", although we can't actually
5294 change this in the header file before 0.9.9. It allows
5295 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5296 enable BN_FLG_CONSTTIME.
5298 [Matthew D Wood (Intel Corp)]
5300 *) In the SSL/TLS server implementation, be strict about session ID
5301 context matching (which matters if an application uses a single
5302 external cache for different purposes). Previously,
5303 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5304 set. This did ensure strict client verification, but meant that,
5305 with applications using a single external cache for quite
5306 different requirements, clients could circumvent ciphersuite
5307 restrictions for a given session ID context by starting a session
5308 in a different context.
5311 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5312 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5313 authentication-only ciphersuites.
5316 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5317 not complete and could lead to a possible single byte overflow
5318 (CVE-2007-5135) [Ben Laurie]
5320 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5322 *) Since AES128 and AES256 (and similarly Camellia128 and
5323 Camellia256) share a single mask bit in the logic of
5324 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5325 kludge to work properly if AES128 is available and AES256 isn't
5326 (or if Camellia128 is available and Camellia256 isn't).
5329 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5330 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5331 When a point or a seed is encoded in a BIT STRING, we need to
5332 prevent the removal of trailing zero bits to get the proper DER
5333 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5334 of a NamedBitList, for which trailing 0 bits need to be removed.)
5337 *) Have SSL/TLS server implementation tolerate "mismatched" record
5338 protocol version while receiving ClientHello even if the
5339 ClientHello is fragmented. (The server can't insist on the
5340 particular protocol version it has chosen before the ServerHello
5341 message has informed the client about his choice.)
5344 *) Add RFC 3779 support.
5345 [Rob Austein for ARIN, Ben Laurie]
5347 *) Load error codes if they are not already present instead of using a
5348 static variable. This allows them to be cleanly unloaded and reloaded.
5349 Improve header file function name parsing.
5352 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5353 or CAPABILITY handshake as required by RFCs.
5356 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5358 *) Introduce limits to prevent malicious keys being able to
5359 cause a denial of service. (CVE-2006-2940)
5360 [Steve Henson, Bodo Moeller]
5362 *) Fix ASN.1 parsing of certain invalid structures that can result
5363 in a denial of service. (CVE-2006-2937) [Steve Henson]
5365 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5366 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5368 *) Fix SSL client code which could crash if connecting to a
5369 malicious SSLv2 server. (CVE-2006-4343)
5370 [Tavis Ormandy and Will Drewry, Google Security Team]
5372 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5373 match only those. Before that, "AES256-SHA" would be interpreted
5374 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5375 the same strength classification in 0.9.7h) as we currently only
5376 have a single AES bit in the ciphersuite description bitmap.
5377 That change, however, also applied to ciphersuite strings such as
5378 "RC4-MD5" that intentionally matched multiple ciphersuites --
5379 namely, SSL 2.0 ciphersuites in addition to the more common ones
5380 from SSL 3.0/TLS 1.0.
5382 So we change the selection algorithm again: Naming an explicit
5383 ciphersuite selects this one ciphersuite, and any other similar
5384 ciphersuite (same bitmap) from *other* protocol versions.
5385 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5386 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5388 Since SSL 2.0 does not have any ciphersuites for which the
5389 128/256 bit distinction would be relevant, this works for now.
5390 The proper fix will be to use different bits for AES128 and
5391 AES256, which would have avoided the problems from the beginning;
5392 however, bits are scarce, so we can only do this in a new release
5393 (not just a patchlevel) when we can change the SSL_CIPHER
5394 definition to split the single 'unsigned long mask' bitmap into
5395 multiple values to extend the available space.
5399 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5401 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5402 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5404 *) Add AES IGE and biIGE modes.
5407 *) Change the Unix randomness entropy gathering to use poll() when
5408 possible instead of select(), since the latter has some
5409 undesirable limitations.
5410 [Darryl Miles via Richard Levitte and Bodo Moeller]
5412 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5413 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5414 cannot be implicitly activated as part of, e.g., the "AES" alias.
5415 However, please upgrade to OpenSSL 0.9.9[-dev] for
5416 non-experimental use of the ECC ciphersuites to get TLS extension
5417 support, which is required for curve and point format negotiation
5418 to avoid potential handshake problems.
5421 *) Disable rogue ciphersuites:
5423 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5424 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5425 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5427 The latter two were purportedly from
5428 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5431 Also deactivate the remaining ciphersuites from
5432 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5433 unofficial, and the ID has long expired.
5436 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5437 dual-core machines) and other potential thread-safety issues.
5440 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5441 versions), which is now available for royalty-free use
5442 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5443 Also, add Camellia TLS ciphersuites from RFC 4132.
5445 To minimize changes between patchlevels in the OpenSSL 0.9.8
5446 series, Camellia remains excluded from compilation unless OpenSSL
5447 is configured with 'enable-camellia'.
5450 *) Disable the padding bug check when compression is in use. The padding
5451 bug check assumes the first packet is of even length, this is not
5452 necessarily true if compression is enabled and can result in false
5453 positives causing handshake failure. The actual bug test is ancient
5454 code so it is hoped that implementations will either have fixed it by
5455 now or any which still have the bug do not support compression.
5458 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5460 *) When applying a cipher rule check to see if string match is an explicit
5461 cipher suite and only match that one cipher suite if it is.
5464 *) Link in manifests for VC++ if needed.
5465 [Austin Ziegler <halostatue@gmail.com>]
5467 *) Update support for ECC-based TLS ciphersuites according to
5468 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5469 TLS extensions, which are supported starting with the 0.9.9
5470 branch, not in the OpenSSL 0.9.8 branch).
5473 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5474 opaque EVP_CIPHER_CTX handling.
5477 *) Fixes and enhancements to zlib compression code. We now only use
5478 "zlib1.dll" and use the default __cdecl calling convention on Win32
5479 to conform with the standards mentioned here:
5480 http://www.zlib.net/DLL_FAQ.txt
5481 Static zlib linking now works on Windows and the new --with-zlib-include
5482 --with-zlib-lib options to Configure can be used to supply the location
5483 of the headers and library. Gracefully handle case where zlib library
5487 *) Several fixes and enhancements to the OID generation code. The old code
5488 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5489 handle numbers larger than ULONG_MAX, truncated printing and had a
5490 non standard OBJ_obj2txt() behaviour.
5493 *) Add support for building of engines under engine/ as shared libraries
5494 under VC++ build system.
5497 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5498 Hopefully, we will not see any false combination of paths any more.
5501 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5503 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5504 (part of SSL_OP_ALL). This option used to disable the
5505 countermeasure against man-in-the-middle protocol-version
5506 rollback in the SSL 2.0 server implementation, which is a bad
5507 idea. (CVE-2005-2969)
5509 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5510 for Information Security, National Institute of Advanced Industrial
5511 Science and Technology [AIST], Japan)]
5513 *) Add two function to clear and return the verify parameter flags.
5516 *) Keep cipherlists sorted in the source instead of sorting them at
5517 runtime, thus removing the need for a lock.
5520 *) Avoid some small subgroup attacks in Diffie-Hellman.
5521 [Nick Mathewson and Ben Laurie]
5523 *) Add functions for well-known primes.
5526 *) Extended Windows CE support.
5527 [Satoshi Nakamura and Andy Polyakov]
5529 *) Initialize SSL_METHOD structures at compile time instead of during
5530 runtime, thus removing the need for a lock.
5533 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5534 attempting to decrypt each encrypted key in turn. Add support to
5538 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5540 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5543 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5546 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5547 key into the same file any more.
5550 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5553 *) Add -utf8 command line and config file option to 'ca'.
5554 [Stefan <stf@udoma.org]
5556 *) Removed the macro des_crypt(), as it seems to conflict with some
5557 libraries. Use DES_crypt().
5560 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5561 involves renaming the source and generated shared-libs for
5562 both. The engines will accept the corrected or legacy ids
5563 ('ncipher' and '4758_cca' respectively) when binding. NB,
5564 this only applies when building 'shared'.
5565 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5567 *) Add attribute functions to EVP_PKEY structure. Modify
5568 PKCS12_create() to recognize a CSP name attribute and
5569 use it. Make -CSP option work again in pkcs12 utility.
5572 *) Add new functionality to the bn blinding code:
5573 - automatic re-creation of the BN_BLINDING parameters after
5574 a fixed number of uses (currently 32)
5575 - add new function for parameter creation
5576 - introduce flags to control the update behaviour of the
5577 BN_BLINDING parameters
5578 - hide BN_BLINDING structure
5579 Add a second BN_BLINDING slot to the RSA structure to improve
5580 performance when a single RSA object is shared among several
5584 *) Add support for DTLS.
5585 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5587 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5588 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5591 *) Remove buggy and incomplete DH cert support from
5592 ssl/ssl_rsa.c and ssl/s3_both.c
5595 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5596 the apps/openssl applications.
5599 *) Compile clean with "-Wall -Wmissing-prototypes
5600 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5601 DEBUG_SAFESTACK must also be set.
5604 *) Change ./Configure so that certain algorithms can be disabled by default.
5605 The new counterpiece to "no-xxx" is "enable-xxx".
5607 The patented RC5 and MDC2 algorithms will now be disabled unless
5608 "enable-rc5" and "enable-mdc2", respectively, are specified.
5610 (IDEA remains enabled despite being patented. This is because IDEA
5611 is frequently required for interoperability, and there is no license
5612 fee for non-commercial use. As before, "no-idea" can be used to
5613 avoid this algorithm.)
5617 *) Add processing of proxy certificates (see RFC 3820). This work was
5618 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5619 EGEE (Enabling Grids for E-science in Europe).
5622 *) RC4 performance overhaul on modern architectures/implementations, such
5623 as Intel P4, IA-64 and AMD64.
5626 *) New utility extract-section.pl. This can be used specify an alternative
5627 section number in a pod file instead of having to treat each file as
5628 a separate case in Makefile. This can be done by adding two lines to the
5631 =for comment openssl_section:XXX
5633 The blank line is mandatory.
5637 *) New arguments -certform, -keyform and -pass for s_client and s_server
5638 to allow alternative format key and certificate files and passphrase
5642 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5643 update associated structures and add various utility functions.
5645 Add new policy related verify parameters, include policy checking in
5646 standard verify code. Enhance 'smime' application with extra parameters
5647 to support policy checking and print out.
5650 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5651 Nehemiah processors. These extensions support AES encryption in hardware
5652 as well as RNG (though RNG support is currently disabled).
5653 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5655 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5658 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5659 [Andy Polyakov and a number of other people]
5661 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5662 implementation contributed by IBM.
5663 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5665 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5666 exponent rather than 'unsigned long'. There is a corresponding change to
5667 the new 'rsa_keygen' element of the RSA_METHOD structure.
5668 [Jelte Jansen, Geoff Thorpe]
5670 *) Functionality for creating the initial serial number file is now
5671 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5673 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5674 number file to 1, which is bound to cause problems. To avoid
5675 the problems while respecting compatibility between different 0.9.7
5676 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5677 CA.pl for serial number initialization. With the new release 0.9.8,
5678 we can fix the problem directly in the 'ca' utility.)
5681 *) Reduced header interdependencies by declaring more opaque objects in
5682 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5683 give fewer recursive includes, which could break lazy source code - so
5684 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5685 developers should define this symbol when building and using openssl to
5686 ensure they track the recommended behaviour, interfaces, [etc], but
5687 backwards-compatible behaviour prevails when this isn't defined.
5690 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5693 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5694 This will generate a random key of the appropriate length based on the
5695 cipher context. The EVP_CIPHER can provide its own random key generation
5696 routine to support keys of a specific form. This is used in the des and
5697 3des routines to generate a key of the correct parity. Update S/MIME
5698 code to use new functions and hence generate correct parity DES keys.
5699 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5700 valid (weak or incorrect parity).
5703 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5704 as looking them up. This is useful when the verified structure may contain
5705 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5706 present unless the new PKCS7_NO_CRL flag is asserted.
5709 *) Extend ASN1 oid configuration module. It now additionally accepts the
5712 shortName = some long name, 1.2.3.4
5715 *) Reimplemented the BN_CTX implementation. There is now no more static
5716 limitation on the number of variables it can handle nor the depth of the
5717 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5718 information can now expand as required, and rather than having a single
5719 static array of bignums, BN_CTX now uses a linked-list of such arrays
5720 allowing it to expand on demand whilst maintaining the usefulness of
5721 BN_CTX's "bundling".
5724 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5725 to allow all RSA operations to function using a single BN_CTX.
5728 *) Preliminary support for certificate policy evaluation and checking. This
5729 is initially intended to pass the tests outlined in "Conformance Testing
5730 of Relying Party Client Certificate Path Processing Logic" v1.07.
5733 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5734 remained unused and not that useful. A variety of other little bignum
5735 tweaks and fixes have also been made continuing on from the audit (see
5739 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5740 associated ASN1, EVP and SSL functions and old ASN1 macros.
5743 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5744 and this should never fail. So the return value from the use of
5745 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5746 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5749 *) BN_CTX_get() should return zero-valued bignums, providing the same
5750 initialised value as BN_new().
5751 [Geoff Thorpe, suggested by Ulf Möller]
5753 *) Support for inhibitAnyPolicy certificate extension.
5756 *) An audit of the BIGNUM code is underway, for which debugging code is
5757 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5758 is considered valid when processing BIGNUMs, and causes execution to
5759 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5760 further steps are taken to deliberately pollute unused data in BIGNUM
5761 structures to try and expose faulty code further on. For now, openssl will
5762 (in its default mode of operation) continue to tolerate the inconsistent
5763 forms that it has tolerated in the past, but authors and packagers should
5764 consider trying openssl and their own applications when compiled with
5765 these debugging symbols defined. It will help highlight potential bugs in
5766 their own code, and will improve the test coverage for OpenSSL itself. At
5767 some point, these tighter rules will become openssl's default to improve
5768 maintainability, though the assert()s and other overheads will remain only
5769 in debugging configurations. See bn.h for more details.
5770 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5772 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5773 that can only be obtained through BN_CTX_new() (which implicitly
5774 initialises it). The presence of this function only made it possible
5775 to overwrite an existing structure (and cause memory leaks).
5778 *) Because of the callback-based approach for implementing LHASH as a
5779 template type, lh_insert() adds opaque objects to hash-tables and
5780 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5781 to clean up those corresponding objects before destroying the hash table
5782 (and losing the object pointers). So some over-zealous constifications in
5783 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5784 objects as "const" and the lh_doall[_arg] callback wrappers are not
5785 prototyped to have "const" restrictions on the object pointers they are
5786 given (and so aren't required to cast them away any more).
5789 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5790 (speed) prefers to use its own implementation. The two implementations
5791 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5792 its object type properly exposed (MS_TM) instead of casting to/from "char
5793 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5794 aren't necessarily the greatest nomenclatures - but this is what was used
5795 internally to the implementation so I've used that for now.
5798 *) Ensure that deprecated functions do not get compiled when
5799 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5800 the self-tests were still using deprecated key-generation functions so
5801 these have been updated also.
5804 *) Reorganise PKCS#7 code to separate the digest location functionality
5805 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5806 New function PKCS7_set_digest() to set the digest type for PKCS#7
5807 digestedData type. Add additional code to correctly generate the
5808 digestedData type and add support for this type in PKCS7 initialization
5812 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5813 structure of type "other".
5816 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5817 sure the loop does correctly stop and breaking ("division by zero")
5818 modulus operations are not performed. The (pre-generated) prime
5819 table crypto/bn/bn_prime.h was already correct, but it could not be
5820 re-generated on some platforms because of the "division by zero"
5821 situation in the script.
5822 [Ralf S. Engelschall]
5824 *) Update support for ECC-based TLS ciphersuites according to
5825 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5826 SHA-1 now is only used for "small" curves (where the
5827 representation of a field element takes up to 24 bytes); for
5828 larger curves, the field element resulting from ECDH is directly
5829 used as premaster secret.
5830 [Douglas Stebila (Sun Microsystems Laboratories)]
5832 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5833 curve secp160r1 to the tests.
5834 [Douglas Stebila (Sun Microsystems Laboratories)]
5836 *) Add the possibility to load symbols globally with DSO.
5837 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5839 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5840 control of the error stack.
5843 *) Add support for STORE in ENGINE.
5846 *) Add the STORE type. The intention is to provide a common interface
5847 to certificate and key stores, be they simple file-based stores, or
5848 HSM-type store, or LDAP stores, or...
5849 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5852 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5853 pass a list of arguments to any function as well as provide a way
5854 for a function to pass data back to the caller.
5857 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5858 works like BUF_strdup() but can be used to duplicate a portion of
5859 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5863 *) Add the function sk_find_ex() which works like sk_find(), but will
5864 return an index to an element even if an exact match couldn't be
5865 found. The index is guaranteed to point at the element where the
5866 searched-for key would be inserted to preserve sorting order.
5869 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5870 takes an extra flags argument for optional functionality. Currently,
5871 the following flags are defined:
5873 OBJ_BSEARCH_VALUE_ON_NOMATCH
5874 This one gets OBJ_bsearch_ex() to return a pointer to the first
5875 element where the comparing function returns a negative or zero
5878 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5879 This one gets OBJ_bsearch_ex() to return a pointer to the first
5880 element where the comparing function returns zero. This is useful
5881 if there are more than one element where the comparing function
5885 *) Make it possible to create self-signed certificates with 'openssl ca'
5886 in such a way that the self-signed certificate becomes part of the
5887 CA database and uses the same mechanisms for serial number generation
5888 as all other certificate signing. The new flag '-selfsign' enables
5889 this functionality. Adapt CA.sh and CA.pl.in.
5892 *) Add functionality to check the public key of a certificate request
5893 against a given private. This is useful to check that a certificate
5894 request can be signed by that key (self-signing).
5897 *) Make it possible to have multiple active certificates with the same
5898 subject in the CA index file. This is done only if the keyword
5899 'unique_subject' is set to 'no' in the main CA section (default
5900 if 'CA_default') of the configuration file. The value is saved
5901 with the database itself in a separate index attribute file,
5902 named like the index file with '.attr' appended to the name.
5905 *) Generate multi-valued AVAs using '+' notation in config files for
5909 *) Support for nameConstraints certificate extension.
5912 *) Support for policyConstraints certificate extension.
5915 *) Support for policyMappings certificate extension.
5918 *) Make sure the default DSA_METHOD implementation only uses its
5919 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5920 and change its own handlers to be NULL so as to remove unnecessary
5921 indirection. This lets alternative implementations fallback to the
5922 default implementation more easily.
5925 *) Support for directoryName in GeneralName related extensions
5929 *) Make it possible to link applications using Makefile.shared.
5930 Make that possible even when linking against static libraries!
5933 *) Support for single pass processing for S/MIME signing. This now
5934 means that S/MIME signing can be done from a pipe, in addition
5935 cleartext signing (multipart/signed type) is effectively streaming
5936 and the signed data does not need to be all held in memory.
5938 This is done with a new flag PKCS7_STREAM. When this flag is set
5939 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5940 is done after the data is output (and digests calculated) in
5941 SMIME_write_PKCS7().
5944 *) Add full support for -rpath/-R, both in shared libraries and
5945 applications, at least on the platforms where it's known how
5949 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5950 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5951 will now compute a table of multiples of the generator that
5952 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5953 faster (notably in the case of a single point multiplication,
5954 scalar * generator).
5955 [Nils Larsch, Bodo Moeller]
5957 *) IPv6 support for certificate extensions. The various extensions
5958 which use the IP:a.b.c.d can now take IPv6 addresses using the
5959 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5963 *) Added an ENGINE that implements RSA by performing private key
5964 exponentiations with the GMP library. The conversions to and from
5965 GMP's mpz_t format aren't optimised nor are any montgomery forms
5966 cached, and on x86 it appears OpenSSL's own performance has caught up.
5967 However there are likely to be other architectures where GMP could
5968 provide a boost. This ENGINE is not built in by default, but it can be
5969 specified at Configure time and should be accompanied by the necessary
5970 linker additions, eg;
5971 ./config -DOPENSSL_USE_GMP -lgmp
5974 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5975 testing availability of engines with "-t" - the old behaviour is
5976 produced by increasing the feature's verbosity with "-tt".
5979 *) ECDSA routines: under certain error conditions uninitialized BN objects
5980 could be freed. Solution: make sure initialization is performed early
5981 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5985 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5986 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5987 software implementations. For DSA and DH, parameter generation can
5988 also be overridden by providing the appropriate method callbacks.
5991 *) Change the "progress" mechanism used in key-generation and
5992 primality testing to functions that take a new BN_GENCB pointer in
5993 place of callback/argument pairs. The new API functions have "_ex"
5994 postfixes and the older functions are reimplemented as wrappers for
5995 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5996 declarations of the old functions to help (graceful) attempts to
5997 migrate to the new functions. Also, the new key-generation API
5998 functions operate on a caller-supplied key-structure and return
5999 success/failure rather than returning a key or NULL - this is to
6000 help make "keygen" another member function of RSA_METHOD etc.
6002 Example for using the new callback interface:
6004 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6008 BN_GENCB_set(&my_cb, my_callback, my_arg);
6010 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6011 /* For the meaning of a, b in calls to my_callback(), see the
6012 * documentation of the function that calls the callback.
6013 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6014 * my_callback should return 1 if it wants BN_is_prime_ex()
6015 * to continue, or 0 to stop.
6020 *) Change the ZLIB compression method to be stateful, and make it
6021 available to TLS with the number defined in
6022 draft-ietf-tls-compression-04.txt.
6025 *) Add the ASN.1 structures and functions for CertificatePair, which
6026 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6028 CertificatePair ::= SEQUENCE {
6029 forward [0] Certificate OPTIONAL,
6030 reverse [1] Certificate OPTIONAL,
6031 -- at least one of the pair shall be present -- }
6033 Also implement the PEM functions to read and write certificate
6034 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6036 This needed to be defined, mostly for the sake of the LDAP
6037 attribute crossCertificatePair, but may prove useful elsewhere as
6041 *) Make it possible to inhibit symlinking of shared libraries in
6042 Makefile.shared, for Cygwin's sake.
6045 *) Extend the BIGNUM API by creating a function
6046 void BN_set_negative(BIGNUM *a, int neg);
6047 and a macro that behave like
6048 int BN_is_negative(const BIGNUM *a);
6050 to avoid the need to access 'a->neg' directly in applications.
6053 *) Implement fast modular reduction for pseudo-Mersenne primes
6054 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6055 EC_GROUP_new_curve_GFp() will now automatically use this
6057 [Nils Larsch <nla@trustcenter.de>]
6059 *) Add new lock type (CRYPTO_LOCK_BN).
6062 *) Change the ENGINE framework to automatically load engines
6063 dynamically from specific directories unless they could be
6064 found to already be built in or loaded. Move all the
6065 current engines except for the cryptodev one to a new
6067 The engines in engines/ are built as shared libraries if
6068 the "shared" options was given to ./Configure or ./config.
6069 Otherwise, they are inserted in libcrypto.a.
6070 /usr/local/ssl/engines is the default directory for dynamic
6071 engines, but that can be overridden at configure time through
6072 the usual use of --prefix and/or --openssldir, and at run
6073 time with the environment variable OPENSSL_ENGINES.
6074 [Geoff Thorpe and Richard Levitte]
6076 *) Add Makefile.shared, a helper makefile to build shared
6077 libraries. Adapt Makefile.org.
6080 *) Add version info to Win32 DLLs.
6081 [Peter 'Luna' Runestig" <peter@runestig.com>]
6083 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6084 can be added using this API to created arbitrary PKCS#12
6085 files while avoiding the low level API.
6087 New options to PKCS12_create(), key or cert can be NULL and
6088 will then be omitted from the output file. The encryption
6089 algorithm NIDs can be set to -1 for no encryption, the mac
6090 iteration count can be set to 0 to omit the mac.
6092 Enhance pkcs12 utility by making the -nokeys and -nocerts
6093 options work when creating a PKCS#12 file. New option -nomac
6094 to omit the mac, NONE can be set for an encryption algorithm.
6095 New code is modified to use the enhanced PKCS12_create()
6096 instead of the low level API.
6099 *) Extend ASN1 encoder to support indefinite length constructed
6100 encoding. This can output sequences tags and octet strings in
6101 this form. Modify pk7_asn1.c to support indefinite length
6102 encoding. This is experimental and needs additional code to
6103 be useful, such as an ASN1 bio and some enhanced streaming
6106 Extend template encode functionality so that tagging is passed
6107 down to the template encoder.
6110 *) Let 'openssl req' fail if an argument to '-newkey' is not
6111 recognized instead of using RSA as a default.
6114 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6115 As these are not official, they are not included in "ALL";
6116 the "ECCdraft" ciphersuite group alias can be used to select them.
6117 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6119 *) Add ECDH engine support.
6120 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6122 *) Add ECDH in new directory crypto/ecdh/.
6123 [Douglas Stebila (Sun Microsystems Laboratories)]
6125 *) Let BN_rand_range() abort with an error after 100 iterations
6126 without success (which indicates a broken PRNG).
6129 *) Change BN_mod_sqrt() so that it verifies that the input value
6130 is really the square of the return value. (Previously,
6131 BN_mod_sqrt would show GIGO behaviour.)
6134 *) Add named elliptic curves over binary fields from X9.62, SECG,
6135 and WAP/WTLS; add OIDs that were still missing.
6137 [Sheueling Chang Shantz and Douglas Stebila
6138 (Sun Microsystems Laboratories)]
6140 *) Extend the EC library for elliptic curves over binary fields
6141 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6144 EC_GF2m_simple_method
6148 EC_GROUP_new_curve_GF2m
6149 EC_GROUP_set_curve_GF2m
6150 EC_GROUP_get_curve_GF2m
6151 EC_POINT_set_affine_coordinates_GF2m
6152 EC_POINT_get_affine_coordinates_GF2m
6153 EC_POINT_set_compressed_coordinates_GF2m
6155 Point compression for binary fields is disabled by default for
6156 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6159 As binary polynomials are represented as BIGNUMs, various members
6160 of the EC_GROUP and EC_POINT data structures can be shared
6161 between the implementations for prime fields and binary fields;
6162 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6163 are essentially identical to their ..._GFp counterparts.
6164 (For simplicity, the '..._GFp' prefix has been dropped from
6165 various internal method names.)
6167 An internal 'field_div' method (similar to 'field_mul' and
6168 'field_sqr') has been added; this is used only for binary fields.
6170 [Sheueling Chang Shantz and Douglas Stebila
6171 (Sun Microsystems Laboratories)]
6173 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6174 through methods ('mul', 'precompute_mult').
6176 The generic implementations (now internally called 'ec_wNAF_mul'
6177 and 'ec_wNAF_precomputed_mult') remain the default if these
6178 methods are undefined.
6180 [Sheueling Chang Shantz and Douglas Stebila
6181 (Sun Microsystems Laboratories)]
6183 *) New function EC_GROUP_get_degree, which is defined through
6184 EC_METHOD. For curves over prime fields, this returns the bit
6185 length of the modulus.
6187 [Sheueling Chang Shantz and Douglas Stebila
6188 (Sun Microsystems Laboratories)]
6190 *) New functions EC_GROUP_dup, EC_POINT_dup.
6191 (These simply call ..._new and ..._copy).
6193 [Sheueling Chang Shantz and Douglas Stebila
6194 (Sun Microsystems Laboratories)]
6196 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6197 Polynomials are represented as BIGNUMs (where the sign bit is not
6198 used) in the following functions [macros]:
6201 BN_GF2m_sub [= BN_GF2m_add]
6202 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6203 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6204 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6206 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6207 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6208 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6209 BN_GF2m_cmp [= BN_ucmp]
6211 (Note that only the 'mod' functions are actually for fields GF(2^m).
6212 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6214 For some functions, an the irreducible polynomial defining a
6215 field can be given as an 'unsigned int[]' with strictly
6216 decreasing elements giving the indices of those bits that are set;
6217 i.e., p[] represents the polynomial
6218 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6220 p[0] > p[1] > ... > p[k] = 0.
6221 This applies to the following functions:
6226 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6227 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6229 BN_GF2m_mod_sqrt_arr
6230 BN_GF2m_mod_solve_quad_arr
6234 Conversion can be performed by the following functions:
6239 bntest.c has additional tests for binary polynomial arithmetic.
6241 Two implementations for BN_GF2m_mod_div() are available.
6242 The default algorithm simply uses BN_GF2m_mod_inv() and
6243 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6244 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6245 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6247 [Sheueling Chang Shantz and Douglas Stebila
6248 (Sun Microsystems Laboratories)]
6250 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6251 functionality is disabled at compile-time.
6252 [Douglas Stebila <douglas.stebila@sun.com>]
6254 *) Change default behaviour of 'openssl asn1parse' so that more
6255 information is visible when viewing, e.g., a certificate:
6257 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6258 mode the content of non-printable OCTET STRINGs is output in a
6259 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6260 avoid the appearance of a printable string.
6261 [Nils Larsch <nla@trustcenter.de>]
6263 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6265 EC_GROUP_set_asn1_flag()
6266 EC_GROUP_get_asn1_flag()
6267 EC_GROUP_set_point_conversion_form()
6268 EC_GROUP_get_point_conversion_form()
6269 These control ASN1 encoding details:
6270 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6271 has been set to OPENSSL_EC_NAMED_CURVE.
6272 - Points are encoded in uncompressed form by default; options for
6273 asn1_for are as for point2oct, namely
6274 POINT_CONVERSION_COMPRESSED
6275 POINT_CONVERSION_UNCOMPRESSED
6276 POINT_CONVERSION_HYBRID
6278 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6281 EC_GROUP_get0_seed()
6282 EC_GROUP_get_seed_len()
6283 This is used only for ASN1 purposes (so far).
6284 [Nils Larsch <nla@trustcenter.de>]
6286 *) Add 'field_type' member to EC_METHOD, which holds the NID
6287 of the appropriate field type OID. The new function
6288 EC_METHOD_get_field_type() returns this value.
6289 [Nils Larsch <nla@trustcenter.de>]
6294 EC_POINT_point2hex()
6295 EC_POINT_hex2point()
6296 providing useful interfaces to EC_POINT_point2oct() and
6297 EC_POINT_oct2point().
6298 [Nils Larsch <nla@trustcenter.de>]
6300 *) Change internals of the EC library so that the functions
6301 EC_GROUP_set_generator()
6302 EC_GROUP_get_generator()
6303 EC_GROUP_get_order()
6304 EC_GROUP_get_cofactor()
6305 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6306 to methods, which would lead to unnecessary code duplication when
6307 adding different types of curves.
6308 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6310 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6311 arithmetic, and such that modified wNAFs are generated
6312 (which avoid length expansion in many cases).
6315 *) Add a function EC_GROUP_check_discriminant() (defined via
6316 EC_METHOD) that verifies that the curve discriminant is non-zero.
6318 Add a function EC_GROUP_check() that makes some sanity tests
6319 on a EC_GROUP, its generator and order. This includes
6320 EC_GROUP_check_discriminant().
6321 [Nils Larsch <nla@trustcenter.de>]
6323 *) Add ECDSA in new directory crypto/ecdsa/.
6325 Add applications 'openssl ecparam' and 'openssl ecdsa'
6326 (these are based on 'openssl dsaparam' and 'openssl dsa').
6328 ECDSA support is also included in various other files across the
6329 library. Most notably,
6330 - 'openssl req' now has a '-newkey ecdsa:file' option;
6331 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6332 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6333 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6334 them suitable for ECDSA where domain parameters must be
6335 extracted before the specific public key;
6336 - ECDSA engine support has been added.
6337 [Nils Larsch <nla@trustcenter.de>]
6339 *) Include some named elliptic curves, and add OIDs from X9.62,
6340 SECG, and WAP/WTLS. Each curve can be obtained from the new
6342 EC_GROUP_new_by_curve_name(),
6343 and the list of available named curves can be obtained with
6344 EC_get_builtin_curves().
6345 Also add a 'curve_name' member to EC_GROUP objects, which can be
6347 EC_GROUP_set_curve_name()
6348 EC_GROUP_get_curve_name()
6349 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6351 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6352 was actually never needed) and in BN_mul(). The removal in BN_mul()
6353 required a small change in bn_mul_part_recursive() and the addition
6354 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6355 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6356 bn_sub_words() and bn_add_words() except they take arrays with
6360 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6362 *) Cleanse PEM buffers before freeing them since they may contain
6364 [Benjamin Bennett <ben@psc.edu>]
6366 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6367 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6368 authentication-only ciphersuites.
6371 *) Since AES128 and AES256 share a single mask bit in the logic of
6372 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6373 kludge to work properly if AES128 is available and AES256 isn't.
6376 *) Expand security boundary to match 1.1.1 module.
6379 *) Remove redundant features: hash file source, editing of test vectors
6380 modify fipsld to use external fips_premain.c signature.
6383 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6384 run algorithm test programs.
6387 *) Make algorithm test programs more tolerant of whitespace.
6390 *) Have SSL/TLS server implementation tolerate "mismatched" record
6391 protocol version while receiving ClientHello even if the
6392 ClientHello is fragmented. (The server can't insist on the
6393 particular protocol version it has chosen before the ServerHello
6394 message has informed the client about his choice.)
6397 *) Load error codes if they are not already present instead of using a
6398 static variable. This allows them to be cleanly unloaded and reloaded.
6401 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6403 *) Introduce limits to prevent malicious keys being able to
6404 cause a denial of service. (CVE-2006-2940)
6405 [Steve Henson, Bodo Moeller]
6407 *) Fix ASN.1 parsing of certain invalid structures that can result
6408 in a denial of service. (CVE-2006-2937) [Steve Henson]
6410 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6411 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6413 *) Fix SSL client code which could crash if connecting to a
6414 malicious SSLv2 server. (CVE-2006-4343)
6415 [Tavis Ormandy and Will Drewry, Google Security Team]
6417 *) Change ciphersuite string processing so that an explicit
6418 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6419 will no longer include "AES128-SHA"), and any other similar
6420 ciphersuite (same bitmap) from *other* protocol versions (so that
6421 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6422 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6423 changes from 0.9.8b and 0.9.8d.
6426 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6428 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6429 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6431 *) Change the Unix randomness entropy gathering to use poll() when
6432 possible instead of select(), since the latter has some
6433 undesirable limitations.
6434 [Darryl Miles via Richard Levitte and Bodo Moeller]
6436 *) Disable rogue ciphersuites:
6438 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6439 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6440 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6442 The latter two were purportedly from
6443 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6446 Also deactivate the remaining ciphersuites from
6447 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6448 unofficial, and the ID has long expired.
6451 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6452 dual-core machines) and other potential thread-safety issues.
6455 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6457 *) Adapt fipsld and the build system to link against the validated FIPS
6458 module in FIPS mode.
6461 *) Fixes for VC++ 2005 build under Windows.
6464 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6465 from a Windows bash shell such as MSYS. It is autodetected from the
6466 "config" script when run from a VC++ environment. Modify standard VC++
6467 build to use fipscanister.o from the GNU make build.
6470 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6472 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6473 The value now differs depending on if you build for FIPS or not.
6474 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6475 safely run with a non-FIPSed libcrypto, as it may crash because of
6476 the difference induced by this change.
6479 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6481 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6482 (part of SSL_OP_ALL). This option used to disable the
6483 countermeasure against man-in-the-middle protocol-version
6484 rollback in the SSL 2.0 server implementation, which is a bad
6485 idea. (CVE-2005-2969)
6487 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6488 for Information Security, National Institute of Advanced Industrial
6489 Science and Technology [AIST], Japan)]
6491 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6492 mainly for FIPS compliance and not fully integrated at this stage.
6495 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6496 the exponentiation using a fixed-length exponent. (Otherwise,
6497 the information leaked through timing could expose the secret key
6498 after many signatures; cf. Bleichenbacher's attack on DSA with
6502 *) Make a new fixed-window mod_exp implementation the default for
6503 RSA, DSA, and DH private-key operations so that the sequence of
6504 squares and multiplies and the memory access pattern are
6505 independent of the particular secret key. This will mitigate
6506 cache-timing and potential related attacks.
6508 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6509 and this is automatically used by BN_mod_exp_mont() if the new flag
6510 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6511 will use this BN flag for private exponents unless the flag
6512 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6513 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6515 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6517 *) Change the client implementation for SSLv23_method() and
6518 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6519 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6520 (Previously, the SSL 2.0 backwards compatible Client Hello
6521 message format would be used even with SSL_OP_NO_SSLv2.)
6524 *) Add support for smime-type MIME parameter in S/MIME messages which some
6528 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6529 a threadsafe manner. Modify rsa code to use new function and add calls
6530 to dsa and dh code (which had race conditions before).
6533 *) Include the fixed error library code in the C error file definitions
6534 instead of fixing them up at runtime. This keeps the error code
6535 structures constant.
6538 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6540 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6543 *) Fixes for newer kerberos headers. NB: the casts are needed because
6544 the 'length' field is signed on one version and unsigned on another
6545 with no (?) obvious way to tell the difference, without these VC++
6546 complains. Also the "definition" of FAR (blank) is no longer included
6547 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6548 some needed definitions.
6551 *) Undo Cygwin change.
6554 *) Added support for proxy certificates according to RFC 3820.
6555 Because they may be a security thread to unaware applications,
6556 they must be explicitly allowed in run-time. See
6557 docs/HOWTO/proxy_certificates.txt for further information.
6560 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6562 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6563 server and client random values. Previously
6564 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6565 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6567 This change has negligible security impact because:
6569 1. Server and client random values still have 24 bytes of pseudo random
6572 2. Server and client random values are sent in the clear in the initial
6575 3. The master secret is derived using the premaster secret (48 bytes in
6576 size for static RSA ciphersuites) as well as client server and random
6579 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6582 [Stephen Henson, reported by UK NISCC]
6584 *) Use Windows randomness collection on Cygwin.
6587 *) Fix hang in EGD/PRNGD query when communication socket is closed
6588 prematurely by EGD/PRNGD.
6589 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6591 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6594 *) Back-port of selected performance improvements from development
6595 branch, as well as improved support for PowerPC platforms.
6598 *) Add lots of checks for memory allocation failure, error codes to indicate
6599 failure and freeing up memory if a failure occurs.
6600 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6602 *) Add new -passin argument to dgst.
6605 *) Perform some character comparisons of different types in X509_NAME_cmp:
6606 this is needed for some certificates that re-encode DNs into UTF8Strings
6607 (in violation of RFC3280) and can't or won't issue name rollover
6611 *) Make an explicit check during certificate validation to see that
6612 the CA setting in each certificate on the chain is correct. As a
6613 side effect always do the following basic checks on extensions,
6614 not just when there's an associated purpose to the check:
6616 - if there is an unhandled critical extension (unless the user
6617 has chosen to ignore this fault)
6618 - if the path length has been exceeded (if one is set at all)
6619 - that certain extensions fit the associated purpose (if one has
6623 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6625 *) Avoid a race condition when CRLs are checked in a multi threaded
6626 environment. This would happen due to the reordering of the revoked
6627 entries during signature checking and serial number lookup. Now the
6628 encoding is cached and the serial number sort performed under a lock.
6629 Add new STACK function sk_is_sorted().
6632 *) Add Delta CRL to the extension code.
6635 *) Various fixes to s3_pkt.c so alerts are sent properly.
6636 [David Holmes <d.holmes@f5.com>]
6638 *) Reduce the chances of duplicate issuer name and serial numbers (in
6639 violation of RFC3280) using the OpenSSL certificate creation utilities.
6640 This is done by creating a random 64 bit value for the initial serial
6641 number when a serial number file is created or when a self signed
6642 certificate is created using 'openssl req -x509'. The initial serial
6643 number file is created using 'openssl x509 -next_serial' in CA.pl
6644 rather than being initialized to 1.
6647 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6649 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6650 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6651 [Joe Orton, Steve Henson]
6653 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6655 [Joe Orton, Steve Henson]
6657 *) Make it possible to have multiple active certificates with the same
6658 subject in the CA index file. This is done only if the keyword
6659 'unique_subject' is set to 'no' in the main CA section (default
6660 if 'CA_default') of the configuration file. The value is saved
6661 with the database itself in a separate index attribute file,
6662 named like the index file with '.attr' appended to the name.
6665 *) X509 verify fixes. Disable broken certificate workarounds when
6666 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6667 keyUsage extension present. Don't accept CRLs with unhandled critical
6668 extensions: since verify currently doesn't process CRL extensions this
6669 rejects a CRL with *any* critical extensions. Add new verify error codes
6673 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6674 A clarification of RFC2560 will require the use of OCTET STRINGs and
6675 some implementations cannot handle the current raw format. Since OpenSSL
6676 copies and compares OCSP nonces as opaque blobs without any attempt at
6677 parsing them this should not create any compatibility issues.
6680 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6681 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6682 this HMAC (and other) operations are several times slower than OpenSSL
6686 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6687 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6689 *) Use the correct content when signing type "other".
6692 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6694 *) Fix various bugs revealed by running the NISCC test suite:
6696 Stop out of bounds reads in the ASN1 code when presented with
6697 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6699 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6701 If verify callback ignores invalid public key errors don't try to check
6702 certificate signature with the NULL public key.
6706 *) New -ignore_err option in ocsp application to stop the server
6707 exiting on the first error in a request.
6710 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6711 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6715 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6716 extra data after the compression methods not only for TLS 1.0
6717 but also for SSL 3.0 (as required by the specification).
6718 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6720 *) Change X509_certificate_type() to mark the key as exported/exportable
6721 when it's 512 *bits* long, not 512 bytes.
6724 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6725 blocks during encryption.
6728 *) Various fixes to base64 BIO and non blocking I/O. On write
6729 flushes were not handled properly if the BIO retried. On read
6730 data was not being buffered properly and had various logic bugs.
6731 This also affects blocking I/O when the data being decoded is a
6735 *) Various S/MIME bugfixes and compatibility changes:
6736 output correct application/pkcs7 MIME type if
6737 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6738 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6739 of files as .eml work). Correctly handle very long lines in MIME
6743 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6745 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6746 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6747 a protocol version number mismatch like a decryption error
6748 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6751 *) Turn on RSA blinding by default in the default implementation
6752 to avoid a timing attack. Applications that don't want it can call
6753 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6754 They would be ill-advised to do so in most cases.
6755 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6757 *) Change RSA blinding code so that it works when the PRNG is not
6758 seeded (in this case, the secret RSA exponent is abused as
6759 an unpredictable seed -- if it is not unpredictable, there
6760 is no point in blinding anyway). Make RSA blinding thread-safe
6761 by remembering the creator's thread ID in rsa->blinding and
6762 having all other threads use local one-time blinding factors
6763 (this requires more computation than sharing rsa->blinding, but
6764 avoids excessive locking; and if an RSA object is not shared
6765 between threads, blinding will still be very fast).
6768 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6769 ENGINE as defaults for all supported algorithms irrespective of
6770 the 'flags' parameter. 'flags' is now honoured, so applications
6771 should make sure they are passing it correctly.
6774 *) Target "mingw" now allows native Windows code to be generated in
6775 the Cygwin environment as well as with the MinGW compiler.
6778 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6780 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6781 via timing by performing a MAC computation even if incorrect
6782 block cipher padding has been found. This is a countermeasure
6783 against active attacks where the attacker has to distinguish
6784 between bad padding and a MAC verification error. (CVE-2003-0078)
6786 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6787 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6788 Martin Vuagnoux (EPFL, Ilion)]
6790 *) Make the no-err option work as intended. The intention with no-err
6791 is not to have the whole error stack handling routines removed from
6792 libcrypto, it's only intended to remove all the function name and
6793 reason texts, thereby removing some of the footprint that may not
6794 be interesting if those errors aren't displayed anyway.
6796 NOTE: it's still possible for any application or module to have it's
6797 own set of error texts inserted. The routines are there, just not
6798 used by default when no-err is given.
6801 *) Add support for FreeBSD on IA64.
6802 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6804 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6805 Kerberos function mit_des_cbc_cksum(). Before this change,
6806 the value returned by DES_cbc_cksum() was like the one from
6807 mit_des_cbc_cksum(), except the bytes were swapped.
6808 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6810 *) Allow an application to disable the automatic SSL chain building.
6811 Before this a rather primitive chain build was always performed in
6812 ssl3_output_cert_chain(): an application had no way to send the
6813 correct chain if the automatic operation produced an incorrect result.
6815 Now the chain builder is disabled if either:
6817 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6819 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6821 The reasoning behind this is that an application would not want the
6822 auto chain building to take place if extra chain certificates are
6823 present and it might also want a means of sending no additional
6824 certificates (for example the chain has two certificates and the
6828 *) Add the possibility to build without the ENGINE framework.
6829 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6831 *) Under Win32 gmtime() can return NULL: check return value in
6832 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6835 *) DSA routines: under certain error conditions uninitialized BN objects
6836 could be freed. Solution: make sure initialization is performed early
6837 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6838 Nils Larsch <nla@trustcenter.de> via PR#459)
6841 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6842 checked on reconnect on the client side, therefore session resumption
6843 could still fail with a "ssl session id is different" error. This
6844 behaviour is masked when SSL_OP_ALL is used due to
6845 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6846 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6847 followup to PR #377.
6850 *) IA-32 assembler support enhancements: unified ELF targets, support
6851 for SCO/Caldera platforms, fix for Cygwin shared build.
6854 *) Add support for FreeBSD on sparc64. As a consequence, support for
6855 FreeBSD on non-x86 processors is separate from x86 processors on
6856 the config script, much like the NetBSD support.
6857 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6859 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6861 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6864 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6865 code (06) was taken as the first octet of the session ID and the last
6866 octet was ignored consequently. As a result SSLv2 client side session
6867 caching could not have worked due to the session ID mismatch between
6869 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6873 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6874 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6878 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6879 seems that in spite of existing for more than a year, many application
6880 author have done nothing to provide the necessary callbacks, which
6881 means that this particular engine will not work properly anywhere.
6882 This is a very unfortunate situation which forces us, in the name
6883 of usability, to give the hw_ncipher.c a static lock, which is part
6885 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6886 appear in 0.9.8 or later. We EXPECT application authors to have
6887 dealt properly with this when 0.9.8 is released (unless we actually
6888 make such changes in the libcrypto locking code that changes will
6889 have to be made anyway).
6892 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6893 octets have been read, EOF or an error occurs. Without this change
6894 some truncated ASN1 structures will not produce an error.
6897 *) Disable Heimdal support, since it hasn't been fully implemented.
6898 Still give the possibility to force the use of Heimdal, but with
6899 warnings and a request that patches get sent to openssl-dev.
6902 *) Add the VC-CE target, introduce the WINCE sysname, and add
6903 INSTALL.WCE and appropriate conditionals to make it build.
6904 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6906 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6907 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6908 edit numbers of the version.
6909 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6911 *) Introduce safe string copy and catenation functions
6912 (BUF_strlcpy() and BUF_strlcat()).
6913 [Ben Laurie (CHATS) and Richard Levitte]
6915 *) Avoid using fixed-size buffers for one-line DNs.
6916 [Ben Laurie (CHATS)]
6918 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6919 resizing buffers containing secrets, and use where appropriate.
6920 [Ben Laurie (CHATS)]
6922 *) Avoid using fixed size buffers for configuration file location.
6923 [Ben Laurie (CHATS)]
6925 *) Avoid filename truncation for various CA files.
6926 [Ben Laurie (CHATS)]
6928 *) Use sizeof in preference to magic numbers.
6929 [Ben Laurie (CHATS)]
6931 *) Avoid filename truncation in cert requests.
6932 [Ben Laurie (CHATS)]
6934 *) Add assertions to check for (supposedly impossible) buffer
6936 [Ben Laurie (CHATS)]
6938 *) Don't cache truncated DNS entries in the local cache (this could
6939 potentially lead to a spoofing attack).
6940 [Ben Laurie (CHATS)]
6942 *) Fix various buffers to be large enough for hex/decimal
6943 representations in a platform independent manner.
6944 [Ben Laurie (CHATS)]
6946 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6947 resizing buffers containing secrets, and use where appropriate.
6948 [Ben Laurie (CHATS)]
6950 *) Add BIO_indent() to avoid much slightly worrying code to do
6952 [Ben Laurie (CHATS)]
6954 *) Convert sprintf()/BIO_puts() to BIO_printf().
6955 [Ben Laurie (CHATS)]
6957 *) buffer_gets() could terminate with the buffer only half
6959 [Ben Laurie (CHATS)]
6961 *) Add assertions to prevent user-supplied crypto functions from
6962 overflowing internal buffers by having large block sizes, etc.
6963 [Ben Laurie (CHATS)]
6965 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6967 [Ben Laurie (CHATS)]
6969 *) Eliminate unused copy of key in RC4.
6970 [Ben Laurie (CHATS)]
6972 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6973 [Ben Laurie (CHATS)]
6975 *) Fix off-by-one error in EGD path.
6976 [Ben Laurie (CHATS)]
6978 *) If RANDFILE path is too long, ignore instead of truncating.
6979 [Ben Laurie (CHATS)]
6981 *) Eliminate unused and incorrectly sized X.509 structure
6983 [Ben Laurie (CHATS)]
6985 *) Eliminate unused and dangerous function knumber().
6986 [Ben Laurie (CHATS)]
6988 *) Eliminate unused and dangerous structure, KSSL_ERR.
6989 [Ben Laurie (CHATS)]
6991 *) Protect against overlong session ID context length in an encoded
6992 session object. Since these are local, this does not appear to be
6994 [Ben Laurie (CHATS)]
6996 *) Change from security patch (see 0.9.6e below) that did not affect
6997 the 0.9.6 release series:
6999 Remote buffer overflow in SSL3 protocol - an attacker could
7000 supply an oversized master key in Kerberos-enabled versions.
7002 [Ben Laurie (CHATS)]
7004 *) Change the SSL kerb5 codes to match RFC 2712.
7007 *) Make -nameopt work fully for req and add -reqopt switch.
7008 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7010 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7011 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7013 *) Make sure tests can be performed even if the corresponding algorithms
7014 have been removed entirely. This was also the last step to make
7015 OpenSSL compilable with DJGPP under all reasonable conditions.
7016 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7018 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7019 to allow version independent disabling of normally unselected ciphers,
7020 which may be activated as a side-effect of selecting a single cipher.
7022 (E.g., cipher list string "RSA" enables ciphersuites that are left
7023 out of "ALL" because they do not provide symmetric encryption.
7024 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7025 [Lutz Jaenicke, Bodo Moeller]
7027 *) Add appropriate support for separate platform-dependent build
7028 directories. The recommended way to make a platform-dependent
7029 build directory is the following (tested on Linux), maybe with
7032 # Place yourself outside of the OpenSSL source tree. In
7033 # this example, the environment variable OPENSSL_SOURCE
7034 # is assumed to contain the absolute OpenSSL source directory.
7035 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7036 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7037 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7038 mkdir -p `dirname $F`
7039 ln -s $OPENSSL_SOURCE/$F $F
7042 To be absolutely sure not to disturb the source tree, a "make clean"
7043 is a good thing. If it isn't successful, don't worry about it,
7044 it probably means the source directory is very clean.
7047 *) Make sure any ENGINE control commands make local copies of string
7048 pointers passed to them whenever necessary. Otherwise it is possible
7049 the caller may have overwritten (or deallocated) the original string
7050 data when a later ENGINE operation tries to use the stored values.
7051 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7053 *) Improve diagnostics in file reading and command-line digests.
7054 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7056 *) Add AES modes CFB and OFB to the object database. Correct an
7057 error in AES-CFB decryption.
7060 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7061 allows existing EVP_CIPHER_CTX structures to be reused after
7062 calling EVP_*Final(). This behaviour is used by encryption
7063 BIOs and some applications. This has the side effect that
7064 applications must explicitly clean up cipher contexts with
7065 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7068 *) Check the values of dna and dnb in bn_mul_recursive before calling
7069 bn_mul_comba (a non zero value means the a or b arrays do not contain
7070 n2 elements) and fallback to bn_mul_normal if either is not zero.
7073 *) Fix escaping of non-ASCII characters when using the -subj option
7074 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7077 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7078 form for "surname", serialNumber has no short form.
7079 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7080 therefore remove "mail" short name for "internet 7".
7081 The OID for unique identifiers in X509 certificates is
7082 x500UniqueIdentifier, not uniqueIdentifier.
7083 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7086 *) Add an "init" command to the ENGINE config module and auto initialize
7087 ENGINEs. Without any "init" command the ENGINE will be initialized
7088 after all ctrl commands have been executed on it. If init=1 the
7089 ENGINE is initialized at that point (ctrls before that point are run
7090 on the uninitialized ENGINE and after on the initialized one). If
7091 init=0 then the ENGINE will not be initialized at all.
7094 *) Fix the 'app_verify_callback' interface so that the user-defined
7095 argument is actually passed to the callback: In the
7096 SSL_CTX_set_cert_verify_callback() prototype, the callback
7097 declaration has been changed from
7100 int (*cb)(X509_STORE_CTX *,void *);
7101 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7102 i=s->ctx->app_verify_callback(&ctx)
7103 has been changed into
7104 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7106 To update applications using SSL_CTX_set_cert_verify_callback(),
7107 a dummy argument can be added to their callback functions.
7108 [D. K. Smetters <smetters@parc.xerox.com>]
7110 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7111 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7113 *) Add and OPENSSL_LOAD_CONF define which will cause
7114 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7115 This allows older applications to transparently support certain
7116 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7117 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7118 load the config file and OPENSSL_add_all_algorithms_conf() which will
7119 always load it have also been added.
7122 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7123 Adjust NIDs and EVP layer.
7124 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7126 *) Config modules support in openssl utility.
7128 Most commands now load modules from the config file,
7129 though in a few (such as version) this isn't done
7130 because it couldn't be used for anything.
7132 In the case of ca and req the config file used is
7133 the same as the utility itself: that is the -config
7134 command line option can be used to specify an
7138 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7139 use "openssl_conf" if filename is NULL use default openssl config file.
7142 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7143 config section name. Add a new flag to tolerate a missing config file
7144 and move code to CONF_modules_load_file().
7147 *) Support for crypto accelerator cards from Accelerated Encryption
7148 Processing, www.aep.ie. (Use engine 'aep')
7149 The support was copied from 0.9.6c [engine] and adapted/corrected
7150 to work with the new engine framework.
7151 [AEP Inc. and Richard Levitte]
7153 *) Support for SureWare crypto accelerator cards from Baltimore
7154 Technologies. (Use engine 'sureware')
7155 The support was copied from 0.9.6c [engine] and adapted
7156 to work with the new engine framework.
7159 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7160 make the newer ENGINE framework commands for the CHIL engine work.
7161 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7163 *) Make it possible to produce shared libraries on ReliantUNIX.
7164 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7166 *) Add the configuration target debug-linux-ppro.
7167 Make 'openssl rsa' use the general key loading routines
7168 implemented in apps.c, and make those routines able to
7169 handle the key format FORMAT_NETSCAPE and the variant
7171 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7173 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7174 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7176 *) Add -keyform to rsautl, and document -engine.
7177 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7179 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7180 BIO_R_NO_SUCH_FILE error code rather than the generic
7181 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7184 *) Add new functions
7186 ERR_peek_last_error_line
7187 ERR_peek_last_error_line_data.
7188 These are similar to
7191 ERR_peek_error_line_data,
7192 but report on the latest error recorded rather than the first one
7193 still in the error queue.
7194 [Ben Laurie, Bodo Moeller]
7196 *) default_algorithms option in ENGINE config module. This allows things
7198 default_algorithms = ALL
7199 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7202 *) Preliminary ENGINE config module.
7205 *) New experimental application configuration code.
7208 *) Change the AES code to follow the same name structure as all other
7209 symmetric ciphers, and behave the same way. Move everything to
7210 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7211 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7213 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7214 [Ben Laurie and Theo de Raadt]
7216 *) Add option to output public keys in req command.
7217 [Massimiliano Pala madwolf@openca.org]
7219 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7220 (up to about 10% better than before for P-192 and P-224).
7223 *) New functions/macros
7225 SSL_CTX_set_msg_callback(ctx, cb)
7226 SSL_CTX_set_msg_callback_arg(ctx, arg)
7227 SSL_set_msg_callback(ssl, cb)
7228 SSL_set_msg_callback_arg(ssl, arg)
7230 to request calling a callback function
7232 void cb(int write_p, int version, int content_type,
7233 const void *buf, size_t len, SSL *ssl, void *arg)
7235 whenever a protocol message has been completely received
7236 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7237 protocol version according to which the SSL library interprets
7238 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7239 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7240 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7241 specification (change_cipher_spec(20), alert(21), handshake(22)).
7242 'buf' and 'len' point to the actual message, 'ssl' to the
7243 SSL object, and 'arg' is the application-defined value set by
7244 SSL[_CTX]_set_msg_callback_arg().
7246 'openssl s_client' and 'openssl s_server' have new '-msg' options
7247 to enable a callback that displays all protocol messages.
7250 *) Change the shared library support so shared libraries are built as
7251 soon as the corresponding static library is finished, and thereby get
7252 openssl and the test programs linked against the shared library.
7253 This still only happens when the keyword "shard" has been given to
7254 the configuration scripts.
7256 NOTE: shared library support is still an experimental thing, and
7257 backward binary compatibility is still not guaranteed.
7258 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7260 *) Add support for Subject Information Access extension.
7261 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7263 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7264 additional bytes when new memory had to be allocated, not just
7265 when reusing an existing buffer.
7268 *) New command line and configuration option 'utf8' for the req command.
7269 This allows field values to be specified as UTF8 strings.
7272 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7273 runs for the former and machine-readable output for the latter.
7276 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7277 of the e-mail address in the DN (i.e., it will go into a certificate
7278 extension only). The new configuration file option 'email_in_dn = no'
7279 has the same effect.
7280 [Massimiliano Pala madwolf@openca.org]
7282 *) Change all functions with names starting with des_ to be starting
7283 with DES_ instead. Add wrappers that are compatible with libdes,
7284 but are named _ossl_old_des_*. Finally, add macros that map the
7285 des_* symbols to the corresponding _ossl_old_des_* if libdes
7286 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7287 desired, the des_* symbols will be mapped to DES_*, with one
7290 Since we provide two compatibility mappings, the user needs to
7291 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7292 compatibility is desired. The default (i.e., when that macro
7293 isn't defined) is OpenSSL 0.9.6c compatibility.
7295 There are also macros that enable and disable the support of old
7296 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7297 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7298 are defined, the default will apply: to support the old des routines.
7300 In either case, one must include openssl/des.h to get the correct
7301 definitions. Do not try to just include openssl/des_old.h, that
7304 NOTE: This is a major break of an old API into a new one. Software
7305 authors are encouraged to switch to the DES_ style functions. Some
7306 time in the future, des_old.h and the libdes compatibility functions
7307 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7308 default), and then completely removed.
7311 *) Test for certificates which contain unsupported critical extensions.
7312 If such a certificate is found during a verify operation it is
7313 rejected by default: this behaviour can be overridden by either
7314 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7315 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7316 X509_supported_extension() has also been added which returns 1 if a
7317 particular extension is supported.
7320 *) Modify the behaviour of EVP cipher functions in similar way to digests
7321 to retain compatibility with existing code.
7324 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7325 compatibility with existing code. In particular the 'ctx' parameter does
7326 not have to be to be initialized before the call to EVP_DigestInit() and
7327 it is tidied up after a call to EVP_DigestFinal(). New function
7328 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7329 EVP_MD_CTX_copy() changed to not require the destination to be
7330 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7331 requires the destination to be valid.
7333 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7334 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7337 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7338 so that complete 'Handshake' protocol structures are kept in memory
7339 instead of overwriting 'msg_type' and 'length' with 'body' data.
7342 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7343 [Massimo Santin via Richard Levitte]
7345 *) Major restructuring to the underlying ENGINE code. This includes
7346 reduction of linker bloat, separation of pure "ENGINE" manipulation
7347 (initialisation, etc) from functionality dealing with implementations
7348 of specific crypto interfaces. This change also introduces integrated
7349 support for symmetric ciphers and digest implementations - so ENGINEs
7350 can now accelerate these by providing EVP_CIPHER and EVP_MD
7351 implementations of their own. This is detailed in crypto/engine/README
7352 as it couldn't be adequately described here. However, there are a few
7353 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7354 were changed in the original introduction of ENGINE code have now
7355 reverted back - the hooking from this code to ENGINE is now a good
7356 deal more passive and at run-time, operations deal directly with
7357 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7358 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7359 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7360 they were not being used by the framework as there is no concept of a
7361 BIGNUM_METHOD and they could not be generalised to the new
7362 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7363 ENGINE_cpy() has been removed as it cannot be consistently defined in
7367 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7370 *) Change mkdef.pl to sort symbols that get the same entry number,
7371 and make sure the automatically generated functions ERR_load_*
7372 become part of libeay.num as well.
7375 *) New function SSL_renegotiate_pending(). This returns true once
7376 renegotiation has been requested (either SSL_renegotiate() call
7377 or HelloRequest/ClientHello received from the peer) and becomes
7378 false once a handshake has been completed.
7379 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7380 sends a HelloRequest, but does not ensure that a handshake takes
7381 place. SSL_renegotiate_pending() is useful for checking if the
7382 client has followed the request.)
7385 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7386 By default, clients may request session resumption even during
7387 renegotiation (if session ID contexts permit); with this option,
7388 session resumption is possible only in the first handshake.
7390 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7391 more bits available for options that should not be part of
7392 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7395 *) Add some demos for certificate and certificate request creation.
7398 *) Make maximum certificate chain size accepted from the peer application
7399 settable (SSL*_get/set_max_cert_list()), as proposed by
7400 "Douglas E. Engert" <deengert@anl.gov>.
7403 *) Add support for shared libraries for Unixware-7
7404 (Boyd Lynn Gerber <gerberb@zenez.com>).
7407 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7408 be done prior to destruction. Use this to unload error strings from
7409 ENGINEs that load their own error strings. NB: This adds two new API
7410 functions to "get" and "set" this destroy handler in an ENGINE.
7413 *) Alter all existing ENGINE implementations (except "openssl" and
7414 "openbsd") to dynamically instantiate their own error strings. This
7415 makes them more flexible to be built both as statically-linked ENGINEs
7416 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7417 Also, add stub code to each that makes building them as self-contained
7418 shared-libraries easier (see README.ENGINE).
7421 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7422 implementations into applications that are completely implemented in
7423 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7424 commands that can be used to configure what shared-library to load and
7425 to control aspects of the way it is handled. Also, made an update to
7426 the README.ENGINE file that brings its information up-to-date and
7427 provides some information and instructions on the "dynamic" ENGINE
7428 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7431 *) Make it possible to unload ranges of ERR strings with a new
7432 "ERR_unload_strings" function.
7435 *) Add a copy() function to EVP_MD.
7438 *) Make EVP_MD routines take a context pointer instead of just the
7439 md_data void pointer.
7442 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7443 that the digest can only process a single chunk of data
7444 (typically because it is provided by a piece of
7445 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7446 is only going to provide a single chunk of data, and hence the
7447 framework needn't accumulate the data for oneshot drivers.
7450 *) As with "ERR", make it possible to replace the underlying "ex_data"
7451 functions. This change also alters the storage and management of global
7452 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7453 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7454 index counters. The API functions that use this state have been changed
7455 to take a "class_index" rather than pointers to the class's local STACK
7456 and counter, and there is now an API function to dynamically create new
7457 classes. This centralisation allows us to (a) plug a lot of the
7458 thread-safety problems that existed, and (b) makes it possible to clean
7459 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7460 such data would previously have always leaked in application code and
7461 workarounds were in place to make the memory debugging turn a blind eye
7462 to it. Application code that doesn't use this new function will still
7463 leak as before, but their memory debugging output will announce it now
7464 rather than letting it slide.
7466 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7467 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7468 has a return value to indicate success or failure.
7471 *) Make it possible to replace the underlying "ERR" functions such that the
7472 global state (2 LHASH tables and 2 locks) is only used by the "default"
7473 implementation. This change also adds two functions to "get" and "set"
7474 the implementation prior to it being automatically set the first time
7475 any other ERR function takes place. Ie. an application can call "get",
7476 pass the return value to a module it has just loaded, and that module
7477 can call its own "set" function using that value. This means the
7478 module's "ERR" operations will use (and modify) the error state in the
7479 application and not in its own statically linked copy of OpenSSL code.
7482 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7483 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7484 the operation, and provides a more encapsulated way for external code
7485 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7486 to use these functions rather than manually incrementing the counts.
7488 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7491 *) Add EVP test program.
7494 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7497 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7498 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7499 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7500 These allow a CRL to be built without having to access X509_CRL fields
7501 directly. Modify 'ca' application to use new functions.
7504 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7505 bug workarounds. Rollback attack detection is a security feature.
7506 The problem will only arise on OpenSSL servers when TLSv1 is not
7507 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7508 Software authors not wanting to support TLSv1 will have special reasons
7509 for their choice and can explicitly enable this option.
7510 [Bodo Moeller, Lutz Jaenicke]
7512 *) Rationalise EVP so it can be extended: don't include a union of
7513 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7514 (similar to those existing for EVP_CIPHER_CTX).
7519 EVP_MD_CTX_init(&md); /* new function call */
7520 EVP_DigestInit(&md, EVP_sha1());
7521 EVP_DigestUpdate(&md, in, len);
7522 EVP_DigestFinal(&md, out, NULL);
7523 EVP_MD_CTX_cleanup(&md); /* new function call */
7527 *) Make DES key schedule conform to the usual scheme, as well as
7528 correcting its structure. This means that calls to DES functions
7529 now have to pass a pointer to a des_key_schedule instead of a
7530 plain des_key_schedule (which was actually always a pointer
7533 des_key_schedule ks;
7535 des_set_key_checked(..., &ks);
7536 des_ncbc_encrypt(..., &ks, ...);
7538 (Note that a later change renames 'des_...' into 'DES_...'.)
7541 *) Initial reduction of linker bloat: the use of some functions, such as
7542 PEM causes large amounts of unused functions to be linked in due to
7543 poor organisation. For example pem_all.c contains every PEM function
7544 which has a knock on effect of linking in large amounts of (unused)
7545 ASN1 code. Grouping together similar functions and splitting unrelated
7546 functions prevents this.
7549 *) Cleanup of EVP macros.
7552 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7553 correct _ecb suffix.
7556 *) Add initial OCSP responder support to ocsp application. The
7557 revocation information is handled using the text based index
7558 use by the ca application. The responder can either handle
7559 requests generated internally, supplied in files (for example
7560 via a CGI script) or using an internal minimal server.
7563 *) Add configuration choices to get zlib compression for TLS.
7566 *) Changes to Kerberos SSL for RFC 2712 compliance:
7567 1. Implemented real KerberosWrapper, instead of just using
7568 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7569 2. Implemented optional authenticator field of KerberosWrapper.
7571 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7572 and authenticator structs; see crypto/krb5/.
7574 Generalized Kerberos calls to support multiple Kerberos libraries.
7575 [Vern Staats <staatsvr@asc.hpc.mil>,
7576 Jeffrey Altman <jaltman@columbia.edu>
7577 via Richard Levitte]
7579 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7580 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7581 values for each of the key sizes rather than having just
7582 parameters (and 'speed' generating keys each time).
7585 *) Speed up EVP routines.
7588 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7589 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7590 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7591 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7593 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7594 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7595 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7598 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7600 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7603 *) Added the OS2-EMX target.
7604 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7606 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7607 to support NCONF routines in extension code. New function CONF_set_nconf()
7608 to allow functions which take an NCONF to also handle the old LHASH
7609 structure: this means that the old CONF compatible routines can be
7610 retained (in particular wrt extensions) without having to duplicate the
7611 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7614 *) Enhance the general user interface with mechanisms for inner control
7615 and with possibilities to have yes/no kind of prompts.
7618 *) Change all calls to low level digest routines in the library and
7619 applications to use EVP. Add missing calls to HMAC_cleanup() and
7620 don't assume HMAC_CTX can be copied using memcpy().
7621 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7623 *) Add the possibility to control engines through control names but with
7624 arbitrary arguments instead of just a string.
7625 Change the key loaders to take a UI_METHOD instead of a callback
7626 function pointer. NOTE: this breaks binary compatibility with earlier
7627 versions of OpenSSL [engine].
7628 Adapt the nCipher code for these new conditions and add a card insertion
7632 *) Enhance the general user interface with mechanisms to better support
7633 dialog box interfaces, application-defined prompts, the possibility
7634 to use defaults (for example default passwords from somewhere else)
7635 and interrupts/cancellations.
7638 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7639 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7642 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7643 tidy up some unnecessarily weird code in 'sk_new()').
7644 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7646 *) Change the key loading routines for ENGINEs to use the same kind
7647 callback (pem_password_cb) as all other routines that need this
7651 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7652 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7653 than this minimum value is recommended.
7656 *) New random seeder for OpenVMS, using the system process statistics
7657 that are easily reachable.
7660 *) Windows apparently can't transparently handle global
7661 variables defined in DLLs. Initialisations such as:
7663 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7665 won't compile. This is used by the any applications that need to
7666 declare their own ASN1 modules. This was fixed by adding the option
7667 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7668 needed for static libraries under Win32.
7671 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7672 setting of purpose and trust fields. New X509_STORE trust and
7673 purpose functions and tidy up setting in other SSL functions.
7676 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7677 structure. These are inherited by X509_STORE_CTX when it is
7678 initialised. This allows various defaults to be set in the
7679 X509_STORE structure (such as flags for CRL checking and custom
7680 purpose or trust settings) for functions which only use X509_STORE_CTX
7681 internally such as S/MIME.
7683 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7684 trust settings if they are not set in X509_STORE. This allows X509_STORE
7685 purposes and trust (in S/MIME for example) to override any set by default.
7687 Add command line options for CRL checking to smime, s_client and s_server
7691 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7692 are set then the CRL is looked up in the X509_STORE structure and
7693 its validity and signature checked, then if the certificate is found
7694 in the CRL the verify fails with a revoked error.
7696 Various new CRL related callbacks added to X509_STORE_CTX structure.
7698 Command line options added to 'verify' application to support this.
7700 This needs some additional work, such as being able to handle multiple
7701 CRLs with different times, extension based lookup (rather than just
7702 by subject name) and ultimately more complete V2 CRL extension
7706 *) Add a general user interface API (crypto/ui/). This is designed
7707 to replace things like des_read_password and friends (backward
7708 compatibility functions using this new API are provided).
7709 The purpose is to remove prompting functions from the DES code
7710 section as well as provide for prompting through dialog boxes in
7711 a window system and the like.
7714 *) Add "ex_data" support to ENGINE so implementations can add state at a
7715 per-structure level rather than having to store it globally.
7718 *) Make it possible for ENGINE structures to be copied when retrieved by
7719 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7720 This causes the "original" ENGINE structure to act like a template,
7721 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7722 operational state can be localised to each ENGINE structure, despite the
7723 fact they all share the same "methods". New ENGINE structures returned in
7724 this case have no functional references and the return value is the single
7725 structural reference. This matches the single structural reference returned
7726 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7730 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7731 needs to match any other type at all we need to manually clear the
7735 *) Changes to the "openssl engine" utility to include;
7736 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7737 about an ENGINE's available control commands.
7738 - executing control commands from command line arguments using the
7739 '-pre' and '-post' switches. '-post' is only used if '-t' is
7740 specified and the ENGINE is successfully initialised. The syntax for
7741 the individual commands are colon-separated, for example;
7742 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7745 *) New dynamic control command support for ENGINEs. ENGINEs can now
7746 declare their own commands (numbers), names (strings), descriptions,
7747 and input types for run-time discovery by calling applications. A
7748 subset of these commands are implicitly classed as "executable"
7749 depending on their input type, and only these can be invoked through
7750 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7751 can be based on user input, config files, etc). The distinction is
7752 that "executable" commands cannot return anything other than a boolean
7753 result and can only support numeric or string input, whereas some
7754 discoverable commands may only be for direct use through
7755 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7756 pointers, or other custom uses. The "executable" commands are to
7757 support parameterisations of ENGINE behaviour that can be
7758 unambiguously defined by ENGINEs and used consistently across any
7759 OpenSSL-based application. Commands have been added to all the
7760 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7761 control over shared-library paths without source code alterations.
7764 *) Changed all ENGINE implementations to dynamically allocate their
7765 ENGINEs rather than declaring them statically. Apart from this being
7766 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7767 this also allows the implementations to compile without using the
7768 internal engine_int.h header.
7771 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7772 'const' value. Any code that should be able to modify a RAND_METHOD
7773 should already have non-const pointers to it (ie. they should only
7774 modify their own ones).
7777 *) Made a variety of little tweaks to the ENGINE code.
7778 - "atalla" and "ubsec" string definitions were moved from header files
7779 to C code. "nuron" string definitions were placed in variables
7780 rather than hard-coded - allowing parameterisation of these values
7781 later on via ctrl() commands.
7782 - Removed unused "#if 0"'d code.
7783 - Fixed engine list iteration code so it uses ENGINE_free() to release
7784 structural references.
7785 - Constified the RAND_METHOD element of ENGINE structures.
7786 - Constified various get/set functions as appropriate and added
7787 missing functions (including a catch-all ENGINE_cpy that duplicates
7788 all ENGINE values onto a new ENGINE except reference counts/state).
7789 - Removed NULL parameter checks in get/set functions. Setting a method
7790 or function to NULL is a way of cancelling out a previously set
7791 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7792 and doesn't justify the extra error symbols and code.
7793 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7794 flags from engine_int.h to engine.h.
7795 - Changed prototypes for ENGINE handler functions (init(), finish(),
7796 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7799 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7800 to the algorithm using long division. The binary algorithm can be
7801 used only if the modulus is odd. On 32-bit systems, it is faster
7802 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7803 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7804 up to 450 bits. In 64-bit environments, the binary algorithm
7805 appears to be advantageous for much longer moduli; here we use it
7806 for moduli up to 2048 bits.
7809 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7810 could not support the combine flag in choice fields.
7813 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7814 extensions from a certificate request to the certificate.
7817 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7818 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7819 file: this allows the display of the certificate about to be
7820 signed to be customised, to allow certain fields to be included
7821 or excluded and extension details. The old system didn't display
7822 multicharacter strings properly, omitted fields not in the policy
7823 and couldn't display additional details such as extensions.
7826 *) Function EC_POINTs_mul for multiple scalar multiplication
7827 of an arbitrary number of elliptic curve points
7828 \sum scalars[i]*points[i],
7829 optionally including the generator defined for the EC_GROUP:
7830 scalar*generator + \sum scalars[i]*points[i].
7832 EC_POINT_mul is a simple wrapper function for the typical case
7833 that the point list has just one item (besides the optional
7837 *) First EC_METHODs for curves over GF(p):
7839 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7840 operations and provides various method functions that can also
7841 operate with faster implementations of modular arithmetic.
7843 EC_GFp_mont_method() reuses most functions that are part of
7844 EC_GFp_simple_method, but uses Montgomery arithmetic.
7846 [Bodo Moeller; point addition and point doubling
7847 implementation directly derived from source code provided by
7848 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7850 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7851 crypto/ec/ec_lib.c):
7853 Curves are EC_GROUP objects (with an optional group generator)
7854 based on EC_METHODs that are built into the library.
7856 Points are EC_POINT objects based on EC_GROUP objects.
7858 Most of the framework would be able to handle curves over arbitrary
7859 finite fields, but as there are no obvious types for fields other
7860 than GF(p), some functions are limited to that for now.
7863 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7864 that the file contains a complete HTTP response.
7867 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7868 change the def and num file printf format specifier from "%-40sXXX"
7869 to "%-39s XXX". The latter will always guarantee a space after the
7870 field while the former will cause them to run together if the field
7871 is 40 of more characters long.
7874 *) Constify the cipher and digest 'method' functions and structures
7875 and modify related functions to take constant EVP_MD and EVP_CIPHER
7879 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7880 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7883 *) Modify EVP_Digest*() routines so they now return values. Although the
7884 internal software routines can never fail additional hardware versions
7888 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7890 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7891 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7894 ERR_R_NESTED_ASN1_ERROR
7896 ERR_R_MISSING_ASN1_EOS
7897 were 4 .. 9, conflicting with
7898 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7900 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7901 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7903 Add new error code 'ERR_R_INTERNAL_ERROR'.
7906 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7910 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7911 sets the subject name for a new request or supersedes the
7912 subject name in a given request. Formats that can be parsed are
7913 'CN=Some Name, OU=myOU, C=IT'
7915 'CN=Some Name/OU=myOU/C=IT'.
7917 Add options '-batch' and '-verbose' to 'openssl req'.
7918 [Massimiliano Pala <madwolf@hackmasters.net>]
7920 *) Introduce the possibility to access global variables through
7921 functions on platform were that's the best way to handle exporting
7922 global variables in shared libraries. To enable this functionality,
7923 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7924 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7925 is normally done by Configure or something similar).
7927 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7928 in the source file (foo.c) like this:
7930 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7931 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7933 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7934 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7936 OPENSSL_DECLARE_GLOBAL(int,foo);
7937 #define foo OPENSSL_GLOBAL_REF(foo)
7938 OPENSSL_DECLARE_GLOBAL(double,bar);
7939 #define bar OPENSSL_GLOBAL_REF(bar)
7941 The #defines are very important, and therefore so is including the
7942 header file everywhere where the defined globals are used.
7944 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7945 of ASN.1 items, but that structure is a bit different.
7947 The largest change is in util/mkdef.pl which has been enhanced with
7948 better and easier to understand logic to choose which symbols should
7949 go into the Windows .def files as well as a number of fixes and code
7950 cleanup (among others, algorithm keywords are now sorted
7951 lexicographically to avoid constant rewrites).
7954 *) In BN_div() keep a copy of the sign of 'num' before writing the
7955 result to 'rm' because if rm==num the value will be overwritten
7956 and produce the wrong result if 'num' is negative: this caused
7957 problems with BN_mod() and BN_nnmod().
7960 *) Function OCSP_request_verify(). This checks the signature on an
7961 OCSP request and verifies the signer certificate. The signer
7962 certificate is just checked for a generic purpose and OCSP request
7966 *) Add OCSP_check_validity() function to check the validity of OCSP
7967 responses. OCSP responses are prepared in real time and may only
7968 be a few seconds old. Simply checking that the current time lies
7969 between thisUpdate and nextUpdate max reject otherwise valid responses
7970 caused by either OCSP responder or client clock inaccuracy. Instead
7971 we allow thisUpdate and nextUpdate to fall within a certain period of
7972 the current time. The age of the response can also optionally be
7973 checked. Two new options -validity_period and -status_age added to
7977 *) If signature or public key algorithm is unrecognized print out its
7978 OID rather that just UNKNOWN.
7981 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7982 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7983 ID to be generated from the issuer certificate alone which can then be
7984 passed to OCSP_id_issuer_cmp().
7987 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7988 ASN1 modules to export functions returning ASN1_ITEM pointers
7989 instead of the ASN1_ITEM structures themselves. This adds several
7990 new macros which allow the underlying ASN1 function/structure to
7991 be accessed transparently. As a result code should not use ASN1_ITEM
7992 references directly (such as &X509_it) but instead use the relevant
7993 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7994 use of the new ASN1 code on platforms where exporting structures
7995 is problematical (for example in shared libraries) but exporting
7996 functions returning pointers to structures is not.
7999 *) Add support for overriding the generation of SSL/TLS session IDs.
8000 These callbacks can be registered either in an SSL_CTX or per SSL.
8001 The purpose of this is to allow applications to control, if they wish,
8002 the arbitrary values chosen for use as session IDs, particularly as it
8003 can be useful for session caching in multiple-server environments. A
8004 command-line switch for testing this (and any client code that wishes
8005 to use such a feature) has been added to "s_server".
8006 [Geoff Thorpe, Lutz Jaenicke]
8008 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8009 of the form '#if defined(...) || defined(...) || ...' and
8010 '#if !defined(...) && !defined(...) && ...'. This also avoids
8011 the growing number of special cases it was previously handling.
8014 *) Make all configuration macros available for application by making
8015 sure they are available in opensslconf.h, by giving them names starting
8016 with "OPENSSL_" to avoid conflicts with other packages and by making
8017 sure e_os2.h will cover all platform-specific cases together with
8019 Additionally, it is now possible to define configuration/platform-
8020 specific names (called "system identities"). In the C code, these
8021 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8022 macro with the name beginning with "OPENSSL_SYS_", which is determined
8023 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8027 *) New option -set_serial to 'req' and 'x509' this allows the serial
8028 number to use to be specified on the command line. Previously self
8029 signed certificates were hard coded with serial number 0 and the
8030 CA options of 'x509' had to use a serial number in a file which was
8034 *) New options to 'ca' utility to support V2 CRL entry extensions.
8035 Currently CRL reason, invalidity date and hold instruction are
8036 supported. Add new CRL extensions to V3 code and some new objects.
8039 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8040 disable standard block padding (aka PKCS#5 padding) in the EVP
8041 API, which was previously mandatory. This means that the data is
8042 not padded in any way and so the total length much be a multiple
8043 of the block size, otherwise an error occurs.
8046 *) Initial (incomplete) OCSP SSL support.
8049 *) New function OCSP_parse_url(). This splits up a URL into its host,
8050 port and path components: primarily to parse OCSP URLs. New -url
8051 option to ocsp utility.
8054 *) New nonce behavior. The return value of OCSP_check_nonce() now
8055 reflects the various checks performed. Applications can decide
8056 whether to tolerate certain situations such as an absent nonce
8057 in a response when one was present in a request: the ocsp application
8058 just prints out a warning. New function OCSP_add1_basic_nonce()
8059 this is to allow responders to include a nonce in a response even if
8060 the request is nonce-less.
8063 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8064 skipped when using openssl x509 multiple times on a single input file,
8065 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8068 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8069 set string type: to handle setting ASN1_TIME structures. Fix ca
8070 utility to correctly initialize revocation date of CRLs.
8073 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8074 the clients preferred ciphersuites and rather use its own preferences.
8075 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8076 Internet Explorer by ensuring unchanged hash method during stepup.
8077 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8080 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8081 to aes and add a new 'exist' option to print out symbols that don't
8085 *) Additional options to ocsp utility to allow flags to be set and
8086 additional certificates supplied.
8089 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8090 OCSP client a number of certificate to only verify the response
8094 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8095 handle the new API. Currently only ECB, CBC modes supported. Add new
8098 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8099 Encryption Standard (AES) Ciphersuites for Transport Layer
8100 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8101 not enabled by default and were not part of the "ALL" ciphersuite
8102 alias because they were not yet official; they could be
8103 explicitly requested by specifying the "AESdraft" ciphersuite
8104 group alias. In the final release of OpenSSL 0.9.7, the group
8105 alias is called "AES" and is part of "ALL".)
8106 [Ben Laurie, Steve Henson, Bodo Moeller]
8108 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8109 request to response.
8112 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8113 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8114 extract information from a certificate request. OCSP_response_create()
8115 creates a response and optionally adds a basic response structure.
8116 OCSP_basic_add1_status() adds a complete single response to a basic
8117 response and returns the OCSP_SINGLERESP structure just added (to allow
8118 extensions to be included for example). OCSP_basic_add1_cert() adds a
8119 certificate to a basic response and OCSP_basic_sign() signs a basic
8120 response with various flags. New helper functions ASN1_TIME_check()
8121 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8122 (converts ASN1_TIME to GeneralizedTime).
8125 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8126 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8127 structure from a certificate. X509_pubkey_digest() digests the public_key
8128 contents: this is used in various key identifiers.
8131 *) Make sk_sort() tolerate a NULL argument.
8132 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8134 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8135 passed by the function are trusted implicitly. If any of them signed the
8136 response then it is assumed to be valid and is not verified.
8139 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8140 to data. This was previously part of the PKCS7 ASN1 code. This
8141 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8142 [Steve Henson, reported by Kenneth R. Robinette
8143 <support@securenetterm.com>]
8145 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8146 routines: without these tracing memory leaks is very painful.
8147 Fix leaks in PKCS12 and PKCS7 routines.
8150 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8151 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8152 effectively meant GeneralizedTime would never be used. Now it
8153 is initialised to -1 but X509_time_adj() now has to check the value
8154 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8155 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8156 [Steve Henson, reported by Kenneth R. Robinette
8157 <support@securenetterm.com>]
8159 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8160 result in a zero length in the ASN1_INTEGER structure which was
8161 not consistent with the structure when d2i_ASN1_INTEGER() was used
8162 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8163 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8164 where it did not print out a minus for negative ASN1_INTEGER.
8167 *) Add summary printout to ocsp utility. The various functions which
8168 convert status values to strings have been renamed to:
8169 OCSP_response_status_str(), OCSP_cert_status_str() and
8170 OCSP_crl_reason_str() and are no longer static. New options
8171 to verify nonce values and to disable verification. OCSP response
8172 printout format cleaned up.
8175 *) Add additional OCSP certificate checks. These are those specified
8176 in RFC2560. This consists of two separate checks: the CA of the
8177 certificate being checked must either be the OCSP signer certificate
8178 or the issuer of the OCSP signer certificate. In the latter case the
8179 OCSP signer certificate must contain the OCSP signing extended key
8180 usage. This check is performed by attempting to match the OCSP
8181 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8182 in the OCSP_CERTID structures of the response.
8185 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8186 and related routines. This uses the standard OpenSSL certificate
8187 verify routines to perform initial checks (just CA validity) and
8188 to obtain the certificate chain. Then additional checks will be
8189 performed on the chain. Currently the root CA is checked to see
8190 if it is explicitly trusted for OCSP signing. This is used to set
8191 a root CA as a global signing root: that is any certificate that
8192 chains to that CA is an acceptable OCSP signing certificate.
8195 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8196 extensions from a separate configuration file.
8197 As when reading extensions from the main configuration file,
8198 the '-extensions ...' option may be used for specifying the
8200 [Massimiliano Pala <madwolf@comune.modena.it>]
8202 *) New OCSP utility. Allows OCSP requests to be generated or
8203 read. The request can be sent to a responder and the output
8204 parsed, outputed or printed in text form. Not complete yet:
8205 still needs to check the OCSP response validity.
8208 *) New subcommands for 'openssl ca':
8209 'openssl ca -status <serial>' prints the status of the cert with
8210 the given serial number (according to the index file).
8211 'openssl ca -updatedb' updates the expiry status of certificates
8213 [Massimiliano Pala <madwolf@comune.modena.it>]
8215 *) New '-newreq-nodes' command option to CA.pl. This is like
8216 '-newreq', but calls 'openssl req' with the '-nodes' option
8217 so that the resulting key is not encrypted.
8218 [Damien Miller <djm@mindrot.org>]
8220 *) New configuration for the GNU Hurd.
8221 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8223 *) Initial code to implement OCSP basic response verify. This
8224 is currently incomplete. Currently just finds the signer's
8225 certificate and verifies the signature on the response.
8228 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8229 value of OPENSSLDIR. This is available via the new '-d' option
8230 to 'openssl version', and is also included in 'openssl version -a'.
8233 *) Allowing defining memory allocation callbacks that will be given
8234 file name and line number information in additional arguments
8235 (a const char* and an int). The basic functionality remains, as
8236 well as the original possibility to just replace malloc(),
8237 realloc() and free() by functions that do not know about these
8238 additional arguments. To register and find out the current
8239 settings for extended allocation functions, the following
8240 functions are provided:
8242 CRYPTO_set_mem_ex_functions
8243 CRYPTO_set_locked_mem_ex_functions
8244 CRYPTO_get_mem_ex_functions
8245 CRYPTO_get_locked_mem_ex_functions
8247 These work the same way as CRYPTO_set_mem_functions and friends.
8248 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8249 extended allocation function is enabled.
8250 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8251 a conventional allocation function is enabled.
8252 [Richard Levitte, Bodo Moeller]
8254 *) Finish off removing the remaining LHASH function pointer casts.
8255 There should no longer be any prototype-casting required when using
8256 the LHASH abstraction, and any casts that remain are "bugs". See
8257 the callback types and macros at the head of lhash.h for details
8258 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8261 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8262 If /dev/[u]random devices are not available or do not return enough
8263 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8265 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8266 /etc/entropy will be queried once each in this sequence, querying stops
8267 when enough entropy was collected without querying more sockets.
8270 *) Change the Unix RAND_poll() variant to be able to poll several
8271 random devices, as specified by DEVRANDOM, until a sufficient amount
8272 of data has been collected. We spend at most 10 ms on each file
8273 (select timeout) and read in non-blocking mode. DEVRANDOM now
8274 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8275 (previously it was just the string "/dev/urandom"), so on typical
8276 platforms the 10 ms delay will never occur.
8277 Also separate out the Unix variant to its own file, rand_unix.c.
8278 For VMS, there's a currently-empty rand_vms.c.
8281 *) Move OCSP client related routines to ocsp_cl.c. These
8282 provide utility functions which an application needing
8283 to issue a request to an OCSP responder and analyse the
8284 response will typically need: as opposed to those which an
8285 OCSP responder itself would need which will be added later.
8287 OCSP_request_sign() signs an OCSP request with an API similar
8288 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8289 response. OCSP_response_get1_basic() extracts basic response
8290 from response. OCSP_resp_find_status(): finds and extracts status
8291 information from an OCSP_CERTID structure (which will be created
8292 when the request structure is built). These are built from lower
8293 level functions which work on OCSP_SINGLERESP structures but
8294 won't normally be used unless the application wishes to examine
8295 extensions in the OCSP response for example.
8297 Replace nonce routines with a pair of functions.
8298 OCSP_request_add1_nonce() adds a nonce value and optionally
8299 generates a random value. OCSP_check_nonce() checks the
8300 validity of the nonce in an OCSP response.
8303 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8304 This doesn't copy the supplied OCSP_CERTID and avoids the
8305 need to free up the newly created id. Change return type
8306 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8307 This can then be used to add extensions to the request.
8308 Deleted OCSP_request_new(), since most of its functionality
8309 is now in OCSP_REQUEST_new() (and the case insensitive name
8310 clash) apart from the ability to set the request name which
8311 will be added elsewhere.
8314 *) Update OCSP API. Remove obsolete extensions argument from
8315 various functions. Extensions are now handled using the new
8316 OCSP extension code. New simple OCSP HTTP function which
8317 can be used to send requests and parse the response.
8320 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8321 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8322 uses the special reorder version of SET OF to sort the attributes
8323 and reorder them to match the encoded order. This resolves a long
8324 standing problem: a verify on a PKCS7 structure just after signing
8325 it used to fail because the attribute order did not match the
8326 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8327 it uses the received order. This is necessary to tolerate some broken
8328 software that does not order SET OF. This is handled by encoding
8329 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8330 to produce the required SET OF.
8333 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8334 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8335 files to get correct declarations of the ASN.1 item variables.
8338 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8339 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8340 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8341 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8342 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8343 ASN1_ITEM and no wrapper functions.
8346 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8347 replace the old function pointer based I/O routines. Change most of
8348 the *_d2i_bio() and *_d2i_fp() functions to use these.
8351 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8352 lines, recognize more "algorithms" that can be deselected, and make
8353 it complain about algorithm deselection that isn't recognised.
8356 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8357 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8358 to use new functions. Add NO_ASN1_OLD which can be set to remove
8359 some old style ASN1 functions: this can be used to determine if old
8360 code will still work when these eventually go away.
8363 *) New extension functions for OCSP structures, these follow the
8364 same conventions as certificates and CRLs.
8367 *) New function X509V3_add1_i2d(). This automatically encodes and
8368 adds an extension. Its behaviour can be customised with various
8369 flags to append, replace or delete. Various wrappers added for
8370 certificates and CRLs.
8373 *) Fix to avoid calling the underlying ASN1 print routine when
8374 an extension cannot be parsed. Correct a typo in the
8375 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8378 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8379 entries for variables.
8382 *) Add functionality to apps/openssl.c for detecting locking
8383 problems: As the program is single-threaded, all we have
8384 to do is register a locking callback using an array for
8385 storing which locks are currently held by the program.
8388 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8389 SSL_get_ex_data_X509_STORE_idx(), which is used in
8390 ssl_verify_cert_chain() and thus can be called at any time
8391 during TLS/SSL handshakes so that thread-safety is essential.
8392 Unfortunately, the ex_data design is not at all suited
8393 for multi-threaded use, so it probably should be abolished.
8396 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8397 [Broadcom, tweaked and integrated by Geoff Thorpe]
8399 *) Move common extension printing code to new function
8400 X509V3_print_extensions(). Reorganise OCSP print routines and
8401 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8404 *) New function X509_signature_print() to remove duplication in some
8408 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8409 set (this was treated exactly the same as SET OF previously). This
8410 is used to reorder the STACK representing the structure to match the
8411 encoding. This will be used to get round a problem where a PKCS7
8412 structure which was signed could not be verified because the STACK
8413 order did not reflect the encoded order.
8416 *) Reimplement the OCSP ASN1 module using the new code.
8419 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8420 for its ASN1 operations. The old style function pointers still exist
8421 for now but they will eventually go away.
8424 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8425 completely replaces the old ASN1 functionality with a table driven
8426 encoder and decoder which interprets an ASN1_ITEM structure describing
8427 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8428 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8429 has also been converted to the new form.
8432 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8433 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8434 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8435 for negative moduli.
8438 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8439 of not touching the result's sign bit.
8442 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8446 *) Changed the LHASH code to use prototypes for callbacks, and created
8447 macros to declare and implement thin (optionally static) functions
8448 that provide type-safety and avoid function pointer casting for the
8449 type-specific callbacks.
8452 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8454 [Veers Staats <staatsvr@asc.hpc.mil>,
8455 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8457 *) Reformat the FAQ so the different questions and answers can be divided
8458 in sections depending on the subject.
8461 *) Have the zlib compression code load ZLIB.DLL dynamically under
8465 *) New function BN_mod_sqrt for computing square roots modulo a prime
8466 (using the probabilistic Tonelli-Shanks algorithm unless
8467 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8468 be handled deterministically).
8469 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8471 *) Make BN_mod_inverse faster by explicitly handling small quotients
8472 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8473 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8476 *) New function BN_kronecker.
8479 *) Fix BN_gcd so that it works on negative inputs; the result is
8480 positive unless both parameters are zero.
8481 Previously something reasonably close to an infinite loop was
8482 possible because numbers could be growing instead of shrinking
8483 in the implementation of Euclid's algorithm.
8486 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8487 sign of the number in question.
8489 Fix BN_is_word(a,w) to work correctly for w == 0.
8491 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8492 because its test if the absolute value of 'a' equals 'w'.
8493 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8494 it exists mostly for use in the implementations of BN_is_zero(),
8495 BN_is_one(), and BN_is_word().
8498 *) New function BN_swap.
8501 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8502 the exponentiation functions are more likely to produce reasonable
8503 results on negative inputs.
8506 *) Change BN_mod_mul so that the result is always non-negative.
8507 Previously, it could be negative if one of the factors was negative;
8508 I don't think anyone really wanted that behaviour.
8511 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8512 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8513 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8514 and add new functions:
8523 BN_mod_lshift1_quick
8527 These functions always generate non-negative results.
8529 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8530 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8532 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8533 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8534 be reduced modulo m.
8535 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8538 The following entry accidentally appeared in the CHANGES file
8539 distributed with OpenSSL 0.9.7. The modifications described in
8540 it do *not* apply to OpenSSL 0.9.7.
8542 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8543 was actually never needed) and in BN_mul(). The removal in BN_mul()
8544 required a small change in bn_mul_part_recursive() and the addition
8545 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8546 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8547 bn_sub_words() and bn_add_words() except they take arrays with
8552 *) In 'openssl passwd', verify passwords read from the terminal
8553 unless the '-salt' option is used (which usually means that
8554 verification would just waste user's time since the resulting
8555 hash is going to be compared with some given password hash)
8556 or the new '-noverify' option is used.
8558 This is an incompatible change, but it does not affect
8559 non-interactive use of 'openssl passwd' (passwords on the command
8560 line, '-stdin' option, '-in ...' option) and thus should not
8564 *) Remove all references to RSAref, since there's no more need for it.
8567 *) Make DSO load along a path given through an environment variable
8568 (SHLIB_PATH) with shl_load().
8571 *) Constify the ENGINE code as a result of BIGNUM constification.
8572 Also constify the RSA code and most things related to it. In a
8573 few places, most notable in the depth of the ASN.1 code, ugly
8574 casts back to non-const were required (to be solved at a later
8578 *) Make it so the openssl application has all engines loaded by default.
8581 *) Constify the BIGNUM routines a little more.
8584 *) Add the following functions:
8586 ENGINE_load_cswift()
8588 ENGINE_load_atalla()
8590 ENGINE_load_builtin_engines()
8592 That way, an application can itself choose if external engines that
8593 are built-in in OpenSSL shall ever be used or not. The benefit is
8594 that applications won't have to be linked with libdl or other dso
8595 libraries unless it's really needed.
8597 Changed 'openssl engine' to load all engines on demand.
8598 Changed the engine header files to avoid the duplication of some
8599 declarations (they differed!).
8602 *) 'openssl engine' can now list capabilities.
8605 *) Better error reporting in 'openssl engine'.
8608 *) Never call load_dh_param(NULL) in s_server.
8611 *) Add engine application. It can currently list engines by name and
8612 identity, and test if they are actually available.
8615 *) Improve RPM specification file by forcing symbolic linking and making
8616 sure the installed documentation is also owned by root.root.
8617 [Damien Miller <djm@mindrot.org>]
8619 *) Give the OpenSSL applications more possibilities to make use of
8620 keys (public as well as private) handled by engines.
8623 *) Add OCSP code that comes from CertCo.
8626 *) Add VMS support for the Rijndael code.
8629 *) Added untested support for Nuron crypto accelerator.
8632 *) Add support for external cryptographic devices. This code was
8633 previously distributed separately as the "engine" branch.
8634 [Geoff Thorpe, Richard Levitte]
8636 *) Rework the filename-translation in the DSO code. It is now possible to
8637 have far greater control over how a "name" is turned into a filename
8638 depending on the operating environment and any oddities about the
8639 different shared library filenames on each system.
8642 *) Support threads on FreeBSD-elf in Configure.
8645 *) Fix for SHA1 assembly problem with MASM: it produces
8646 warnings about corrupt line number information when assembling
8647 with debugging information. This is caused by the overlapping
8649 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8652 NCONF_get_number() has no error checking at all. As a replacement,
8653 NCONF_get_number_e() is defined (_e for "error checking") and is
8654 promoted strongly. The old NCONF_get_number is kept around for
8655 binary backward compatibility.
8656 Make it possible for methods to load from something other than a BIO,
8657 by providing a function pointer that is given a name instead of a BIO.
8658 For example, this could be used to load configuration data from an
8662 *) Fix for non blocking accept BIOs. Added new I/O special reason
8663 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8664 with non blocking I/O was not possible because no retry code was
8665 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8669 *) Added the beginnings of Rijndael support.
8672 *) Fix for bug in DirectoryString mask setting. Add support for
8673 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8674 to allow certificate printing to more controllable, additional
8675 'certopt' option to 'x509' to allow new printing options to be
8679 *) Clean old EAY MD5 hack from e_os.h.
8682 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8684 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8685 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8686 [Joe Orton, Steve Henson]
8688 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8690 *) Fix additional bug revealed by the NISCC test suite:
8692 Stop bug triggering large recursion when presented with
8693 certain ASN.1 tags (CVE-2003-0851)
8696 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8698 *) Fix various bugs revealed by running the NISCC test suite:
8700 Stop out of bounds reads in the ASN1 code when presented with
8701 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8703 If verify callback ignores invalid public key errors don't try to check
8704 certificate signature with the NULL public key.
8708 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8709 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8713 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8714 extra data after the compression methods not only for TLS 1.0
8715 but also for SSL 3.0 (as required by the specification).
8716 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8718 *) Change X509_certificate_type() to mark the key as exported/exportable
8719 when it's 512 *bits* long, not 512 bytes.
8722 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8724 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8725 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8726 a protocol version number mismatch like a decryption error
8727 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8730 *) Turn on RSA blinding by default in the default implementation
8731 to avoid a timing attack. Applications that don't want it can call
8732 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8733 They would be ill-advised to do so in most cases.
8734 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8736 *) Change RSA blinding code so that it works when the PRNG is not
8737 seeded (in this case, the secret RSA exponent is abused as
8738 an unpredictable seed -- if it is not unpredictable, there
8739 is no point in blinding anyway). Make RSA blinding thread-safe
8740 by remembering the creator's thread ID in rsa->blinding and
8741 having all other threads use local one-time blinding factors
8742 (this requires more computation than sharing rsa->blinding, but
8743 avoids excessive locking; and if an RSA object is not shared
8744 between threads, blinding will still be very fast).
8747 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8749 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8750 via timing by performing a MAC computation even if incorrect
8751 block cipher padding has been found. This is a countermeasure
8752 against active attacks where the attacker has to distinguish
8753 between bad padding and a MAC verification error. (CVE-2003-0078)
8755 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8756 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8757 Martin Vuagnoux (EPFL, Ilion)]
8759 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8761 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8762 memory from it's contents. This is done with a counter that will
8763 place alternating values in each byte. This can be used to solve
8764 two issues: 1) the removal of calls to memset() by highly optimizing
8765 compilers, and 2) cleansing with other values than 0, since those can
8766 be read through on certain media, for example a swap space on disk.
8769 *) Bugfix: client side session caching did not work with external caching,
8770 because the session->cipher setting was not restored when reloading
8771 from the external cache. This problem was masked, when
8772 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8773 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8776 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8777 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8778 [Zeev Lieber <zeev-l@yahoo.com>]
8780 *) Undo an undocumented change introduced in 0.9.6e which caused
8781 repeated calls to OpenSSL_add_all_ciphers() and
8782 OpenSSL_add_all_digests() to be ignored, even after calling
8786 *) Change the default configuration reader to deal with last line not
8787 being properly terminated.
8790 *) Change X509_NAME_cmp() so it applies the special rules on handling
8791 DN values that are of type PrintableString, as well as RDNs of type
8792 emailAddress where the value has the type ia5String.
8793 [stefank@valicert.com via Richard Levitte]
8795 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8796 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8797 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8798 the bitwise-OR of the two for use by the majority of applications
8799 wanting this behaviour, and update the docs. The documented
8800 behaviour and actual behaviour were inconsistent and had been
8801 changing anyway, so this is more a bug-fix than a behavioural
8803 [Geoff Thorpe, diagnosed by Nadav Har'El]
8805 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8806 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8809 *) Fix initialization code race conditions in
8810 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8811 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8812 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8813 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8814 ssl2_get_cipher_by_char(),
8815 ssl3_get_cipher_by_char().
8816 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8818 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8819 the cached sessions are flushed, as the remove_cb() might use ex_data
8820 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8821 (see [openssl.org #212]).
8822 [Geoff Thorpe, Lutz Jaenicke]
8824 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8825 length, instead of the encoding length to d2i_ASN1_OBJECT.
8828 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8830 *) [In 0.9.6g-engine release:]
8831 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8832 [Lynn Gazis <lgazis@rainbow.com>]
8834 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8836 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8837 and get fix the header length calculation.
8838 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8839 Alon Kantor <alonk@checkpoint.com> (and others),
8842 *) Use proper error handling instead of 'assertions' in buffer
8843 overflow checks added in 0.9.6e. This prevents DoS (the
8844 assertions could call abort()).
8845 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8847 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8849 *) Add various sanity checks to asn1_get_length() to reject
8850 the ASN1 length bytes if they exceed sizeof(long), will appear
8851 negative or the content length exceeds the length of the
8853 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8855 *) Fix cipher selection routines: ciphers without encryption had no flags
8856 for the cipher strength set and where therefore not handled correctly
8857 by the selection routines (PR #130).
8860 *) Fix EVP_dsa_sha macro.
8864 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8865 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8866 that was added in OpenSSL 0.9.6d.
8868 As the countermeasure turned out to be incompatible with some
8869 broken SSL implementations, the new option is part of SSL_OP_ALL.
8870 SSL_OP_ALL is usually employed when compatibility with weird SSL
8871 implementations is desired (e.g. '-bugs' option to 's_client' and
8872 's_server'), so the new option is automatically set in many
8876 *) Changes in security patch:
8878 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8879 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8880 Air Force Materiel Command, USAF, under agreement number
8883 *) Add various sanity checks to asn1_get_length() to reject
8884 the ASN1 length bytes if they exceed sizeof(long), will appear
8885 negative or the content length exceeds the length of the
8886 supplied buffer. (CVE-2002-0659)
8887 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8889 *) Assertions for various potential buffer overflows, not known to
8891 [Ben Laurie (CHATS)]
8893 *) Various temporary buffers to hold ASCII versions of integers were
8894 too small for 64 bit platforms. (CVE-2002-0655)
8895 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8897 *) Remote buffer overflow in SSL3 protocol - an attacker could
8898 supply an oversized session ID to a client. (CVE-2002-0656)
8899 [Ben Laurie (CHATS)]
8901 *) Remote buffer overflow in SSL2 protocol - an attacker could
8902 supply an oversized client master key. (CVE-2002-0656)
8903 [Ben Laurie (CHATS)]
8905 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8907 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8908 encoded as NULL) with id-dsa-with-sha1.
8909 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8911 *) Check various X509_...() return values in apps/req.c.
8912 [Nils Larsch <nla@trustcenter.de>]
8914 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8915 an end-of-file condition would erroneously be flagged, when the CRLF
8916 was just at the end of a processed block. The bug was discovered when
8917 processing data through a buffering memory BIO handing the data to a
8918 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8919 <ptsekov@syntrex.com> and Nedelcho Stanev.
8922 *) Implement a countermeasure against a vulnerability recently found
8923 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8924 before application data chunks to avoid the use of known IVs
8925 with data potentially chosen by the attacker.
8928 *) Fix length checks in ssl3_get_client_hello().
8931 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8932 to prevent ssl3_read_internal() from incorrectly assuming that
8933 ssl3_read_bytes() found application data while handshake
8934 processing was enabled when in fact s->s3->in_read_app_data was
8935 merely automatically cleared during the initial handshake.
8936 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8938 *) Fix object definitions for Private and Enterprise: they were not
8939 recognized in their shortname (=lowercase) representation. Extend
8940 obj_dat.pl to issue an error when using undefined keywords instead
8941 of silently ignoring the problem (Svenning Sorensen
8942 <sss@sss.dnsalias.net>).
8945 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8946 generators, i.e. generators other than 2 and 5. (Previously, the
8947 code did not properly initialise the 'add' and 'rem' values to
8948 BN_generate_prime().)
8950 In the new general case, we do not insist that 'generator' is
8951 actually a primitive root: This requirement is rather pointless;
8952 a generator of the order-q subgroup is just as good, if not
8956 *) Map new X509 verification errors to alerts. Discovered and submitted by
8957 Tom Wu <tom@arcot.com>.
8960 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8961 returning non-zero before the data has been completely received
8962 when using non-blocking I/O.
8963 [Bodo Moeller; problem pointed out by John Hughes]
8965 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8966 [Ben Laurie, Lutz Jaenicke]
8968 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8969 Yoram Zahavi <YoramZ@gilian.com>).
8972 *) Add information about CygWin 1.3 and on, and preserve proper
8973 configuration for the versions before that.
8974 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8976 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8977 check whether we deal with a copy of a session and do not delete from
8978 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8979 <izhar@checkpoint.com>.
8982 *) Do not store session data into the internal session cache, if it
8983 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8984 flag is set). Proposed by Aslam <aslam@funk.com>.
8987 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8991 *) [In 0.9.6d-engine release:]
8992 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8993 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8995 *) Add the configuration target linux-s390x.
8996 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8998 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8999 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9000 variable as an indication that a ClientHello message has been
9001 received. As the flag value will be lost between multiple
9002 invocations of ssl3_accept when using non-blocking I/O, the
9003 function may not be aware that a handshake has actually taken
9004 place, thus preventing a new session from being added to the
9007 To avoid this problem, we now set s->new_session to 2 instead of
9008 using a local variable.
9009 [Lutz Jaenicke, Bodo Moeller]
9011 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9012 if the SSL_R_LENGTH_MISMATCH error is detected.
9013 [Geoff Thorpe, Bodo Moeller]
9015 *) New 'shared_ldflag' column in Configure platform table.
9018 *) Fix EVP_CIPHER_mode macro.
9019 ["Dan S. Camper" <dan@bti.net>]
9021 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9022 type, we must throw them away by setting rr->length to 0.
9023 [D P Chang <dpc@qualys.com>]
9025 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9027 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9028 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9029 worked incorrectly for those cases where range = 10..._2 and
9030 3*range is two bits longer than range.)
9033 *) Only add signing time to PKCS7 structures if it is not already
9037 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9038 OBJ_ld_ce should be OBJ_id_ce.
9039 Also some ip-pda OIDs in crypto/objects/objects.txt were
9040 incorrect (cf. RFC 3039).
9041 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9043 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9044 returns early because it has nothing to do.
9045 [Andy Schneider <andy.schneider@bjss.co.uk>]
9047 *) [In 0.9.6c-engine release:]
9048 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9049 [Andy Schneider <andy.schneider@bjss.co.uk>]
9051 *) [In 0.9.6c-engine release:]
9052 Add support for Cryptographic Appliance's keyserver technology.
9053 (Use engine 'keyclient')
9054 [Cryptographic Appliances and Geoff Thorpe]
9056 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9057 is called via tools/c89.sh because arguments have to be
9058 rearranged (all '-L' options must appear before the first object
9060 [Richard Shapiro <rshapiro@abinitio.com>]
9062 *) [In 0.9.6c-engine release:]
9063 Add support for Broadcom crypto accelerator cards, backported
9065 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9067 *) [In 0.9.6c-engine release:]
9068 Add support for SureWare crypto accelerator cards from
9069 Baltimore Technologies. (Use engine 'sureware')
9070 [Baltimore Technologies and Mark Cox]
9072 *) [In 0.9.6c-engine release:]
9073 Add support for crypto accelerator cards from Accelerated
9074 Encryption Processing, www.aep.ie. (Use engine 'aep')
9075 [AEP Inc. and Mark Cox]
9077 *) Add a configuration entry for gcc on UnixWare.
9078 [Gary Benson <gbenson@redhat.com>]
9080 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9081 messages are stored in a single piece (fixed-length part and
9082 variable-length part combined) and fix various bugs found on the way.
9085 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9086 instead. BIO_gethostbyname() does not know what timeouts are
9087 appropriate, so entries would stay in cache even when they have
9089 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9091 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9092 faced with a pathologically small ClientHello fragment that does
9093 not contain client_version: Instead of aborting with an error,
9094 simply choose the highest available protocol version (i.e.,
9095 TLS 1.0 unless it is disabled). In practice, ClientHello
9096 messages are never sent like this, but this change gives us
9097 strictly correct behaviour at least for TLS.
9100 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9101 never resets s->method to s->ctx->method when called from within
9102 one of the SSL handshake functions.
9103 [Bodo Moeller; problem pointed out by Niko Baric]
9105 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9106 (sent using the client's version number) if client_version is
9107 smaller than the protocol version in use. Also change
9108 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9109 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9110 the client will at least see that alert.
9113 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9117 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9118 client receives HelloRequest while in a handshake.
9119 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9121 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9122 should end in 'break', not 'goto end' which circumvents various
9123 cleanups done in state SSL_ST_OK. But session related stuff
9124 must be disabled for SSL_ST_OK in the case that we just sent a
9127 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9128 before just sending a HelloRequest.
9129 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9131 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9132 reveal whether illegal block cipher padding was found or a MAC
9133 verification error occurred. (Neither SSLerr() codes nor alerts
9134 are directly visible to potential attackers, but the information
9135 may leak via logfiles.)
9137 Similar changes are not required for the SSL 2.0 implementation
9138 because the number of padding bytes is sent in clear for SSL 2.0,
9139 and the extra bytes are just ignored. However ssl/s2_pkt.c
9140 failed to verify that the purported number of padding bytes is in
9144 *) Add OpenUNIX-8 support including shared libraries
9145 (Boyd Lynn Gerber <gerberb@zenez.com>).
9148 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9149 'wristwatch attack' using huge encoding parameters (cf.
9150 James H. Manger's CRYPTO 2001 paper). Note that the
9151 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9152 encoding parameters and hence was not vulnerable.
9155 *) BN_sqr() bug fix.
9156 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9158 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9159 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9160 followed by modular reduction.
9161 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9163 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9164 equivalent based on BN_pseudo_rand() instead of BN_rand().
9167 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9168 This function was broken, as the check for a new client hello message
9169 to handle SGC did not allow these large messages.
9170 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9173 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9176 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9177 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9180 *) Rework the configuration and shared library support for Tru64 Unix.
9181 The configuration part makes use of modern compiler features and
9182 still retains old compiler behavior for those that run older versions
9183 of the OS. The shared library support part includes a variant that
9184 uses the RPATH feature, and is available through the special
9185 configuration target "alpha-cc-rpath", which will never be selected
9187 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9189 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9190 with the same message size as in ssl3_get_certificate_request().
9191 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9192 messages might inadvertently be reject as too long.
9193 [Petr Lampa <lampa@fee.vutbr.cz>]
9195 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9198 *) Modified SSL library such that the verify_callback that has been set
9199 specificly for an SSL object with SSL_set_verify() is actually being
9200 used. Before the change, a verify_callback set with this function was
9201 ignored and the verify_callback() set in the SSL_CTX at the time of
9202 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9203 to allow the necessary settings.
9206 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9207 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9208 done automatically (in contradiction to the requirements of the C
9209 standard). This made problems when used from OpenSSH.
9212 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9213 dh->length and always used
9215 BN_rand_range(priv_key, dh->p).
9217 BN_rand_range() is not necessary for Diffie-Hellman, and this
9218 specific range makes Diffie-Hellman unnecessarily inefficient if
9219 dh->length (recommended exponent length) is much smaller than the
9220 length of dh->p. We could use BN_rand_range() if the order of
9221 the subgroup was stored in the DH structure, but we only have
9226 BN_rand(priv_key, l, ...)
9228 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9234 RSA_eay_public_encrypt
9235 RSA_eay_private_decrypt
9236 RSA_eay_private_encrypt (signing)
9237 RSA_eay_public_decrypt (signature verification)
9239 (default implementations for RSA_public_encrypt,
9240 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9241 always reject numbers >= n.
9244 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9245 to synchronize access to 'locking_thread'. This is necessary on
9246 systems where access to 'locking_thread' (an 'unsigned long'
9247 variable) is not atomic.
9250 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9251 *before* setting the 'crypto_lock_rand' flag. The previous code had
9252 a race condition if 0 is a valid thread ID.
9253 [Travis Vitek <vitek@roguewave.com>]
9255 *) Add support for shared libraries under Irix.
9256 [Albert Chin-A-Young <china@thewrittenword.com>]
9258 *) Add configuration option to build on Linux on both big-endian and
9260 [Ralf Baechle <ralf@uni-koblenz.de>]
9262 *) Add the possibility to create shared libraries on HP-UX.
9265 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9267 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9268 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9269 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9270 PRNG state recovery was possible based on the output of
9271 one PRNG request appropriately sized to gain knowledge on
9272 'md' followed by enough consecutive 1-byte PRNG requests
9273 to traverse all of 'state'.
9275 1. When updating 'md_local' (the current thread's copy of 'md')
9276 during PRNG output generation, hash all of the previous
9277 'md_local' value, not just the half used for PRNG output.
9279 2. Make the number of bytes from 'state' included into the hash
9280 independent from the number of PRNG bytes requested.
9282 The first measure alone would be sufficient to avoid
9283 Markku-Juhani's attack. (Actually it had never occurred
9284 to me that the half of 'md_local' used for chaining was the
9285 half from which PRNG output bytes were taken -- I had always
9286 assumed that the secret half would be used.) The second
9287 measure makes sure that additional data from 'state' is never
9288 mixed into 'md_local' in small portions; this heuristically
9289 further strengthens the PRNG.
9292 *) Fix crypto/bn/asm/mips3.s.
9295 *) When only the key is given to "enc", the IV is undefined. Print out
9296 an error message in this case.
9299 *) Handle special case when X509_NAME is empty in X509 printing routines.
9302 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9303 positive and less than q.
9306 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9307 used: it isn't thread safe and the add_lock_callback should handle
9309 [Paul Rose <Paul.Rose@bridge.com>]
9311 *) Verify that incoming data obeys the block size in
9312 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9316 [Ulf Möller, Bodo Möller]
9318 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9319 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9320 when fixing the server behaviour for backwards-compatible 'client
9321 hello' messages. (Note that the attack is impractical against
9322 SSL 3.0 and TLS 1.0 anyway because length and version checking
9323 means that the probability of guessing a valid ciphertext is
9324 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9327 Before 0.9.5, the countermeasure (hide the error by generating a
9328 random 'decryption result') did not work properly because
9329 ERR_clear_error() was missing, meaning that SSL_get_error() would
9330 detect the supposedly ignored error.
9332 Both problems are now fixed.
9335 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9336 (previously it was 1024).
9339 *) Fix for compatibility mode trust settings: ignore trust settings
9340 unless some valid trust or reject settings are present.
9343 *) Fix for blowfish EVP: its a variable length cipher.
9346 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9347 parameters in DSA public key structures and return an error in the
9348 DSA routines if parameters are absent.
9351 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9352 in the current directory if neither $RANDFILE nor $HOME was set.
9353 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9354 caused some confusion to Windows users who haven't defined $HOME.
9355 Thus RAND_file_name() is changed again: e_os.h can define a
9356 DEFAULT_HOME, which will be used if $HOME is not set.
9357 For Windows, we use "C:"; on other platforms, we still require
9358 environment variables.
9360 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9361 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9362 having multiple threads call RAND_poll() concurrently.
9365 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9366 combination of a flag and a thread ID variable.
9367 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9368 flag), *other* threads can enter ssleay_add_bytes without obeying
9369 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9370 that they do not hold after the first thread unsets add_do_not_lock).
9373 *) Change bctest again: '-x' expressions are not available in all
9377 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9379 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9380 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9382 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9383 the default extension for executables, if any. Also, make the perl
9384 scripts that use symlink() to test if it really exists and use "cp"
9385 if it doesn't. All this made OpenSSL compilable and installable in
9389 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9390 If SEQUENCE is length is indefinite just set c->slen to the total
9391 amount of data available.
9392 [Steve Henson, reported by shige@FreeBSD.org]
9393 [This change does not apply to 0.9.7.]
9395 *) Change bctest to avoid here-documents inside command substitution
9396 (workaround for FreeBSD /bin/sh bug).
9397 For compatibility with Ultrix, avoid shell functions (introduced
9398 in the bctest version that searches along $PATH).
9401 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9402 with des_encrypt() defined on some operating systems, like Solaris
9406 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9407 On the Importance of Eliminating Errors in Cryptographic
9408 Computations, J. Cryptology 14 (2001) 2, 101-119,
9409 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9412 *) MIPS assembler BIGNUM division bug fix.
9415 *) Disabled incorrect Alpha assembler code.
9418 *) Fix PKCS#7 decode routines so they correctly update the length
9419 after reading an EOC for the EXPLICIT tag.
9421 [This change does not apply to 0.9.7.]
9423 *) Fix bug in PKCS#12 key generation routines. This was triggered
9424 if a 3DES key was generated with a 0 initial byte. Include
9425 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9426 (but broken) behaviour.
9429 *) Enhance bctest to search for a working bc along $PATH and print
9431 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9433 *) Fix memory leaks in err.c: free err_data string if necessary;
9434 don't write to the wrong index in ERR_set_error_data.
9437 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9441 *) Replace rdtsc with _emit statements for VC++ version 5.
9442 [Jeremy Cooper <jeremy@baymoo.org>]
9444 *) Make it possible to reuse SSLv2 sessions.
9447 *) In copy_email() check for >= 0 as a return value for
9448 X509_NAME_get_index_by_NID() since 0 is a valid index.
9449 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9451 *) Avoid coredump with unsupported or invalid public keys by checking if
9452 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9453 PKCS7_verify() fails with non detached data.
9456 *) Don't use getenv in library functions when run as setuid/setgid.
9457 New function OPENSSL_issetugid().
9460 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9461 due to incorrect handling of multi-threading:
9463 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9465 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9467 3. Count how many times MemCheck_off() has been called so that
9468 nested use can be treated correctly. This also avoids
9469 inband-signalling in the previous code (which relied on the
9470 assumption that thread ID 0 is impossible).
9473 *) Add "-rand" option also to s_client and s_server.
9476 *) Fix CPU detection on Irix 6.x.
9477 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9478 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9480 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9483 [This change does not apply to 0.9.7.]
9485 *) Use the cached encoding of an X509_NAME structure rather than
9486 copying it. This is apparently the reason for the libsafe "errors"
9487 but the code is actually correct.
9490 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9491 Bleichenbacher's DSA attack.
9492 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9493 to be set and top=0 forces the highest bit to be set; top=-1 is new
9494 and leaves the highest bit random.
9495 [Ulf Moeller, Bodo Moeller]
9497 *) In the NCONF_...-based implementations for CONF_... queries
9498 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9499 a temporary CONF structure with the data component set to NULL
9500 (which gives segmentation faults in lh_retrieve).
9501 Instead, use NULL for the CONF pointer in CONF_get_string and
9502 CONF_get_number (which may use environment variables) and directly
9503 return NULL from CONF_get_section.
9506 *) Fix potential buffer overrun for EBCDIC.
9509 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9510 keyUsage if basicConstraints absent for a CA.
9513 *) Make SMIME_write_PKCS7() write mail header values with a format that
9514 is more generally accepted (no spaces before the semicolon), since
9515 some programs can't parse those values properly otherwise. Also make
9516 sure BIO's that break lines after each write do not create invalid
9520 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9521 macros previously used would not encode an empty SEQUENCE OF
9522 and break the signature.
9524 [This change does not apply to 0.9.7.]
9526 *) Zero the premaster secret after deriving the master secret in
9530 *) Add some EVP_add_digest_alias registrations (as found in
9531 OpenSSL_add_all_digests()) to SSL_library_init()
9532 aka OpenSSL_add_ssl_algorithms(). This provides improved
9533 compatibility with peers using X.509 certificates
9534 with unconventional AlgorithmIdentifier OIDs.
9537 *) Fix for Irix with NO_ASM.
9538 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9540 *) ./config script fixes.
9541 [Ulf Moeller, Richard Levitte]
9543 *) Fix 'openssl passwd -1'.
9546 *) Change PKCS12_key_gen_asc() so it can cope with non null
9547 terminated strings whose length is passed in the passlen
9548 parameter, for example from PEM callbacks. This was done
9549 by adding an extra length parameter to asc2uni().
9550 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9552 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9553 call failed, free the DSA structure.
9556 *) Fix to uni2asc() to cope with zero length Unicode strings.
9557 These are present in some PKCS#12 files.
9560 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9561 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9562 when writing a 32767 byte record.
9563 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9565 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9566 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9568 (RSA objects have a reference count access to which is protected
9569 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9570 so they are meant to be shared between threads.)
9571 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9572 "Reddie, Steven" <Steven.Reddie@ca.com>]
9574 *) Fix a deadlock in CRYPTO_mem_leaks().
9577 *) Use better test patterns in bntest.
9580 *) rand_win.c fix for Borland C.
9583 *) BN_rshift bugfix for n == 0.
9586 *) Add a 'bctest' script that checks for some known 'bc' bugs
9587 so that 'make test' does not abort just because 'bc' is broken.
9590 *) Store verify_result within SSL_SESSION also for client side to
9591 avoid potential security hole. (Re-used sessions on the client side
9592 always resulted in verify_result==X509_V_OK, not using the original
9593 result of the server certificate verification.)
9596 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9597 SSL3_RT_APPLICATION_DATA, return 0.
9598 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9602 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9603 releases, have been re-implemented by renaming the previous
9604 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9605 and ssl3_read_internal, respectively, and adding 'peek' parameters
9606 to them. The new ssl[23]_{read,peek} functions are calls to
9607 ssl[23]_read_internal with the 'peek' flag set appropriately.
9608 A 'peek' parameter has also been added to ssl3_read_bytes, which
9609 does the actual work for ssl3_read_internal.
9612 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9613 the method-specific "init()" handler. Also clean up ex_data after
9614 calling the method-specific "finish()" handler. Previously, this was
9615 happening the other way round.
9618 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9619 The previous value, 12, was not always sufficient for BN_mod_exp().
9622 *) Make sure that shared libraries get the internal name engine with
9623 the full version number and not just 0. This should mark the
9624 shared libraries as not backward compatible. Of course, this should
9625 be changed again when we can guarantee backward binary compatibility.
9628 *) Fix typo in get_cert_by_subject() in by_dir.c
9629 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9631 *) Rework the system to generate shared libraries:
9633 - Make note of the expected extension for the shared libraries and
9634 if there is a need for symbolic links from for example libcrypto.so.0
9635 to libcrypto.so.0.9.7. There is extended info in Configure for
9638 - Make as few rebuilds of the shared libraries as possible.
9640 - Still avoid linking the OpenSSL programs with the shared libraries.
9642 - When installing, install the shared libraries separately from the
9646 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9648 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9649 and not in SSL_clear because the latter is also used by the
9650 accept/connect functions; previously, the settings made by
9651 SSL_set_read_ahead would be lost during the handshake.
9652 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9654 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9655 Previously, it would create entries for disabled algorithms no
9659 *) Added several new manual pages for SSL_* function.
9662 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9664 *) In ssl23_get_client_hello, generate an error message when faced
9665 with an initial SSL 3.0/TLS record that is too small to contain the
9666 first two bytes of the ClientHello message, i.e. client_version.
9667 (Note that this is a pathologic case that probably has never happened
9668 in real life.) The previous approach was to use the version number
9669 from the record header as a substitute; but our protocol choice
9670 should not depend on that one because it is not authenticated
9671 by the Finished messages.
9674 *) More robust randomness gathering functions for Windows.
9675 [Jeffrey Altman <jaltman@columbia.edu>]
9677 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9678 not set then we don't setup the error code for issuer check errors
9679 to avoid possibly overwriting other errors which the callback does
9680 handle. If an application does set the flag then we assume it knows
9681 what it is doing and can handle the new informational codes
9685 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9686 a general "ANY" type, as such it should be able to decode anything
9687 including tagged types. However it didn't check the class so it would
9688 wrongly interpret tagged types in the same way as their universal
9689 counterpart and unknown types were just rejected. Changed so that the
9690 tagged and unknown types are handled in the same way as a SEQUENCE:
9691 that is the encoding is stored intact. There is also a new type
9692 "V_ASN1_OTHER" which is used when the class is not universal, in this
9693 case we have no idea what the actual type is so we just lump them all
9697 *) On VMS, stdout may very well lead to a file that is written to
9698 in a record-oriented fashion. That means that every write() will
9699 write a separate record, which will be read separately by the
9700 programs trying to read from it. This can be very confusing.
9702 The solution is to put a BIO filter in the way that will buffer
9703 text until a linefeed is reached, and then write everything a
9704 line at a time, so every record written will be an actual line,
9705 not chunks of lines and not (usually doesn't happen, but I've
9706 seen it once) several lines in one record. BIO_f_linebuffer() is
9709 Currently, it's a VMS-only method, because that's where it has
9710 been tested well enough.
9713 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9714 it can return incorrect results.
9715 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9716 but it was in 0.9.6-beta[12].)
9719 *) Disable the check for content being present when verifying detached
9720 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9721 include zero length content when signing messages.
9724 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9725 BIO_ctrl (for BIO pairs).
9728 *) Add DSO method for VMS.
9731 *) Bug fix: Montgomery multiplication could produce results with the
9735 *) Add RPM specification openssl.spec and modify it to build three
9736 packages. The default package contains applications, application
9737 documentation and run-time libraries. The devel package contains
9738 include files, static libraries and function documentation. The
9739 doc package contains the contents of the doc directory. The original
9740 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9743 *) Add a large number of documentation files for many SSL routines.
9744 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9746 *) Add a configuration entry for Sony News 4.
9747 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9749 *) Don't set the two most significant bits to one when generating a
9750 random number < q in the DSA library.
9753 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9754 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9755 the underlying transport is blocking) if a handshake took place.
9756 (The default behaviour is needed by applications such as s_client
9757 and s_server that use select() to determine when to use SSL_read;
9758 but for applications that know in advance when to expect data, it
9759 just makes things more complicated.)
9762 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9766 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9767 work better on such systems.
9768 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9770 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9771 Update PKCS12_parse() so it copies the friendlyName and the
9772 keyid to the certificates aux info.
9775 *) Fix bug in PKCS7_verify() which caused an infinite loop
9776 if there was more than one signature.
9777 [Sven Uszpelkat <su@celocom.de>]
9779 *) Major change in util/mkdef.pl to include extra information
9780 about each symbol, as well as presenting variables as well
9781 as functions. This change means that there's n more need
9782 to rebuild the .num files when some algorithms are excluded.
9785 *) Allow the verify time to be set by an application,
9786 rather than always using the current time.
9789 *) Phase 2 verify code reorganisation. The certificate
9790 verify code now looks up an issuer certificate by a
9791 number of criteria: subject name, authority key id
9792 and key usage. It also verifies self signed certificates
9793 by the same criteria. The main comparison function is
9794 X509_check_issued() which performs these checks.
9796 Lot of changes were necessary in order to support this
9797 without completely rewriting the lookup code.
9799 Authority and subject key identifier are now cached.
9801 The LHASH 'certs' is X509_STORE has now been replaced
9802 by a STACK_OF(X509_OBJECT). This is mainly because an
9803 LHASH can't store or retrieve multiple objects with
9804 the same hash value.
9806 As a result various functions (which were all internal
9807 use only) have changed to handle the new X509_STORE
9808 structure. This will break anything that messed round
9809 with X509_STORE internally.
9811 The functions X509_STORE_add_cert() now checks for an
9812 exact match, rather than just subject name.
9814 The X509_STORE API doesn't directly support the retrieval
9815 of multiple certificates matching a given criteria, however
9816 this can be worked round by performing a lookup first
9817 (which will fill the cache with candidate certificates)
9818 and then examining the cache for matches. This is probably
9819 the best we can do without throwing out X509_LOOKUP
9820 entirely (maybe later...).
9822 The X509_VERIFY_CTX structure has been enhanced considerably.
9824 All certificate lookup operations now go via a get_issuer()
9825 callback. Although this currently uses an X509_STORE it
9826 can be replaced by custom lookups. This is a simple way
9827 to bypass the X509_STORE hackery necessary to make this
9828 work and makes it possible to use more efficient techniques
9829 in future. A very simple version which uses a simple
9830 STACK for its trusted certificate store is also provided
9831 using X509_STORE_CTX_trusted_stack().
9833 The verify_cb() and verify() callbacks now have equivalents
9834 in the X509_STORE_CTX structure.
9836 X509_STORE_CTX also has a 'flags' field which can be used
9837 to customise the verify behaviour.
9840 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9841 excludes S/MIME capabilities.
9844 *) When a certificate request is read in keep a copy of the
9845 original encoding of the signed data and use it when outputting
9846 again. Signatures then use the original encoding rather than
9847 a decoded, encoded version which may cause problems if the
9848 request is improperly encoded.
9851 *) For consistency with other BIO_puts implementations, call
9852 buffer_write(b, ...) directly in buffer_puts instead of calling
9855 In BIO_puts, increment b->num_write as in BIO_write.
9856 [Peter.Sylvester@EdelWeb.fr]
9858 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9859 BN_zero, we may not return a BIGNUM with an array consisting of
9863 *) Avoid calling abort() from within the library when problems are
9864 detected, except if preprocessor symbols have been defined
9865 (such as REF_CHECK, BN_DEBUG etc.).
9868 *) New openssl application 'rsautl'. This utility can be
9869 used for low level RSA operations. DER public key
9870 BIO/fp routines also added.
9873 *) New Configure entry and patches for compiling on QNX 4.
9874 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9876 *) A demo state-machine implementation was sponsored by
9877 Nuron (http://www.nuron.com/) and is now available in
9878 demos/state_machine.
9881 *) New options added to the 'dgst' utility for signature
9882 generation and verification.
9885 *) Unrecognized PKCS#7 content types are now handled via a
9886 catch all ASN1_TYPE structure. This allows unsupported
9887 types to be stored as a "blob" and an application can
9888 encode and decode it manually.
9891 *) Fix various signed/unsigned issues to make a_strex.c
9893 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9895 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9896 length if passed a buffer. ASN1_INTEGER_to_BN failed
9897 if passed a NULL BN and its argument was negative.
9898 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9900 *) Modification to PKCS#7 encoding routines to output definite
9901 length encoding. Since currently the whole structures are in
9902 memory there's not real point in using indefinite length
9903 constructed encoding. However if OpenSSL is compiled with
9904 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9907 *) Added BIO_vprintf() and BIO_vsnprintf().
9910 *) Added more prefixes to parse for in the strings written
9911 through a logging bio, to cover all the levels that are available
9912 through syslog. The prefixes are now:
9914 PANIC, EMERG, EMR => LOG_EMERG
9915 ALERT, ALR => LOG_ALERT
9916 CRIT, CRI => LOG_CRIT
9917 ERROR, ERR => LOG_ERR
9918 WARNING, WARN, WAR => LOG_WARNING
9919 NOTICE, NOTE, NOT => LOG_NOTICE
9920 INFO, INF => LOG_INFO
9921 DEBUG, DBG => LOG_DEBUG
9923 and as before, if none of those prefixes are present at the
9924 beginning of the string, LOG_ERR is chosen.
9926 On Win32, the LOG_* levels are mapped according to this:
9928 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9929 LOG_WARNING => EVENTLOG_WARNING_TYPE
9930 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9934 *) Made it possible to reconfigure with just the configuration
9935 argument "reconf" or "reconfigure". The command line arguments
9936 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9937 and are retrieved from there when reconfiguring.
9941 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9943 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9946 *) The obj_dat.pl script was messing up the sorting of object
9947 names. The reason was that it compared the quoted version
9948 of strings as a result "OCSP" > "OCSP Signing" because
9949 " > SPACE. Changed script to store unquoted versions of
9950 names and add quotes on output. It was also omitting some
9951 names from the lookup table if they were given a default
9952 value (that is if SN is missing it is given the same
9953 value as LN and vice versa), these are now added on the
9954 grounds that if an object has a name we should be able to
9955 look it up. Finally added warning output when duplicate
9956 short or long names are found.
9959 *) Changes needed for Tandem NSK.
9960 [Scott Uroff <scott@xypro.com>]
9962 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9963 RSA_padding_check_SSLv23(), special padding was never detected
9964 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9965 version rollback attacks was not effective.
9967 In s23_clnt.c, don't use special rollback-attack detection padding
9968 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9969 client; similarly, in s23_srvr.c, don't do the rollback check if
9970 SSL 2.0 is the only protocol enabled in the server.
9973 *) Make it possible to get hexdumps of unprintable data with 'openssl
9974 asn1parse'. By implication, the functions ASN1_parse_dump() and
9975 BIO_dump_indent() are added.
9978 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9979 these print out strings and name structures based on various
9980 flags including RFC2253 support and proper handling of
9981 multibyte characters. Added options to the 'x509' utility
9982 to allow the various flags to be set.
9985 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9986 Also change the functions X509_cmp_current_time() and
9987 X509_gmtime_adj() work with an ASN1_TIME structure,
9988 this will enable certificates using GeneralizedTime in validity
9989 dates to be checked.
9992 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9993 negative public key encodings) on by default,
9994 NO_NEG_PUBKEY_BUG can be set to disable it.
9997 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9998 content octets. An i2c_ASN1_OBJECT is unnecessary because
9999 the encoding can be trivially obtained from the structure.
10002 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10003 not read locks (CRYPTO_r_[un]lock).
10006 *) A first attempt at creating official support for shared
10007 libraries through configuration. I've kept it so the
10008 default is static libraries only, and the OpenSSL programs
10009 are always statically linked for now, but there are
10010 preparations for dynamic linking in place.
10011 This has been tested on Linux and Tru64.
10014 *) Randomness polling function for Win9x, as described in:
10015 Peter Gutmann, Software Generation of Practically Strong
10019 *) Fix so PRNG is seeded in req if using an already existing
10023 *) New options to smime application. -inform and -outform
10024 allow alternative formats for the S/MIME message including
10025 PEM and DER. The -content option allows the content to be
10026 specified separately. This should allow things like Netscape
10027 form signing output easier to verify.
10030 *) Fix the ASN1 encoding of tags using the 'long form'.
10033 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10034 STRING types. These convert content octets to and from the
10035 underlying type. The actual tag and length octets are
10036 already assumed to have been read in and checked. These
10037 are needed because all other string types have virtually
10038 identical handling apart from the tag. By having versions
10039 of the ASN1 functions that just operate on content octets
10040 IMPLICIT tagging can be handled properly. It also allows
10041 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10042 and ASN1_INTEGER are identical apart from the tag.
10045 *) Change the handling of OID objects as follows:
10047 - New object identifiers are inserted in objects.txt, following
10048 the syntax given in objects.README.
10049 - objects.pl is used to process obj_mac.num and create a new
10051 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10054 This is currently kind of a hack, and the perl code in objects.pl
10055 isn't very elegant, but it works as I intended. The simplest way
10056 to check that it worked correctly is to look in obj_dat.h and
10057 check the array nid_objs and make sure the objects haven't moved
10058 around (this is important!). Additions are OK, as well as
10059 consistent name changes.
10062 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10065 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10066 The given file adds to whatever has already been seeded into the
10067 random pool through the RANDFILE configuration file option or
10068 environment variable, or the default random state file.
10071 *) mkstack.pl now sorts each macro group into lexical order.
10072 Previously the output order depended on the order the files
10073 appeared in the directory, resulting in needless rewriting
10077 *) Patches to make OpenSSL compile under Win32 again. Mostly
10078 work arounds for the VC++ problem that it treats func() as
10079 func(void). Also stripped out the parts of mkdef.pl that
10080 added extra typesafe functions: these no longer exist.
10083 *) Reorganisation of the stack code. The macros are now all
10084 collected in safestack.h . Each macro is defined in terms of
10085 a "stack macro" of the form SKM_<name>(type, a, b). The
10086 DEBUG_SAFESTACK is now handled in terms of function casts,
10087 this has the advantage of retaining type safety without the
10088 use of additional functions. If DEBUG_SAFESTACK is not defined
10089 then the non typesafe macros are used instead. Also modified the
10090 mkstack.pl script to handle the new form. Needs testing to see
10091 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10092 the default if no major problems. Similar behaviour for ASN1_SET_OF
10093 and PKCS12_STACK_OF.
10096 *) When some versions of IIS use the 'NET' form of private key the
10097 key derivation algorithm is different. Normally MD5(password) is
10098 used as a 128 bit RC4 key. In the modified case
10099 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10100 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10101 as the old Netscape_RSA functions except they have an additional
10102 'sgckey' parameter which uses the modified algorithm. Also added
10103 an -sgckey command line option to the rsa utility. Thanks to
10104 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10105 algorithm to openssl-dev.
10108 *) The evp_local.h macros were using 'c.##kname' which resulted in
10109 invalid expansion on some systems (SCO 5.0.5 for example).
10110 Corrected to 'c.kname'.
10111 [Phillip Porch <root@theporch.com>]
10113 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10114 a STACK of email addresses from a certificate or request, these look
10115 in the subject name and the subject alternative name extensions and
10116 omit any duplicate addresses.
10119 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10120 This makes DSA verification about 2 % faster.
10123 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10124 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10125 plus overhead for 1024 bit moduli).
10126 This makes exponentiations about 0.5 % faster for 1024 bit
10127 exponents (as measured by "openssl speed rsa2048").
10130 *) Rename memory handling macros to avoid conflicts with other
10132 Malloc => OPENSSL_malloc
10133 Malloc_locked => OPENSSL_malloc_locked
10134 Realloc => OPENSSL_realloc
10135 Free => OPENSSL_free
10138 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10139 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10142 *) CygWin32 support.
10143 [John Jarvie <jjarvie@newsguy.com>]
10145 *) The type-safe stack code has been rejigged. It is now only compiled
10146 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10147 by default all type-specific stack functions are "#define"d back to
10148 standard stack functions. This results in more streamlined output
10149 but retains the type-safety checking possibilities of the original
10153 *) The STACK code has been cleaned up, and certain type declarations
10154 that didn't make a lot of sense have been brought in line. This has
10155 also involved a cleanup of sorts in safestack.h to more correctly
10156 map type-safe stack functions onto their plain stack counterparts.
10157 This work has also resulted in a variety of "const"ifications of
10158 lots of the code, especially "_cmp" operations which should normally
10159 be prototyped with "const" parameters anyway.
10162 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10163 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10164 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10165 where all of 'md' is used each time the PRNG is used, but 'state'
10166 is used only indexed by a cyclic counter. As entropy may not be
10167 well distributed from the beginning, 'md' is important as a
10168 chaining variable. However, the output function chains only half
10169 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10170 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10171 in all of 'state' being rewritten, with the new values depending
10172 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10175 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10176 the handshake is continued after ssl_verify_cert_chain();
10177 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10178 can lead to 'unexplainable' connection aborts later.
10179 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10181 *) Major EVP API cipher revision.
10182 Add hooks for extra EVP features. This allows various cipher
10183 parameters to be set in the EVP interface. Support added for variable
10184 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10185 setting of RC2 and RC5 parameters.
10187 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10190 Remove lots of duplicated code from the EVP library. For example *every*
10191 cipher init() function handles the 'iv' in the same way according to the
10192 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10193 for CFB and OFB modes they zero ctx->num.
10195 New functionality allows removal of S/MIME code RC2 hack.
10197 Most of the routines have the same form and so can be declared in terms
10200 By shifting this to the top level EVP_CipherInit() it can be removed from
10201 all individual ciphers. If the cipher wants to handle IVs or keys
10202 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10205 Change lots of functions like EVP_EncryptUpdate() to now return a
10206 value: although software versions of the algorithms cannot fail
10207 any installed hardware versions can.
10210 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10211 this option is set, tolerate broken clients that send the negotiated
10212 protocol version number instead of the requested protocol version
10216 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10217 i.e. non-zero for export ciphersuites, zero otherwise.
10218 Previous versions had this flag inverted, inconsistent with
10219 rsa_tmp_cb (..._TMP_RSA_CB).
10220 [Bodo Moeller; problem reported by Amit Chopra]
10222 *) Add missing DSA library text string. Work around for some IIS
10223 key files with invalid SEQUENCE encoding.
10226 *) Add a document (doc/standards.txt) that list all kinds of standards
10227 and so on that are implemented in OpenSSL.
10230 *) Enhance c_rehash script. Old version would mishandle certificates
10231 with the same subject name hash and wouldn't handle CRLs at all.
10232 Added -fingerprint option to crl utility, to support new c_rehash
10236 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10239 *) Fix for SSL server purpose checking. Server checking was
10240 rejecting certificates which had extended key usage present
10241 but no ssl client purpose.
10242 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10244 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10245 is a little unclear about how a blank password is handled.
10246 Since the password in encoded as a BMPString with terminating
10247 double NULL a zero length password would end up as just the
10248 double NULL. However no password at all is different and is
10249 handled differently in the PKCS#12 key generation code. NS
10250 treats a blank password as zero length. MSIE treats it as no
10251 password on export: but it will try both on import. We now do
10252 the same: PKCS12_parse() tries zero length and no password if
10253 the password is set to "" or NULL (NULL is now a valid password:
10254 it wasn't before) as does the pkcs12 application.
10257 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10258 perror when PEM_read_bio_X509_REQ fails, the error message must
10259 be obtained from the error queue.
10262 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10263 it in ERR_remove_state if appropriate, and change ERR_get_state
10264 accordingly to avoid race conditions (this is necessary because
10265 thread_hash is no longer constant once set).
10268 *) Bugfix for linux-elf makefile.one.
10271 *) RSA_get_default_method() will now cause a default
10272 RSA_METHOD to be chosen if one doesn't exist already.
10273 Previously this was only set during a call to RSA_new()
10274 or RSA_new_method(NULL) meaning it was possible for
10275 RSA_get_default_method() to return NULL.
10278 *) Added native name translation to the existing DSO code
10279 that will convert (if the flag to do so is set) filenames
10280 that are sufficiently small and have no path information
10281 into a canonical native form. Eg. "blah" converted to
10282 "libblah.so" or "blah.dll" etc.
10285 *) New function ERR_error_string_n(e, buf, len) which is like
10286 ERR_error_string(e, buf), but writes at most 'len' bytes
10287 including the 0 terminator. For ERR_error_string_n, 'buf'
10289 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10291 *) CONF library reworked to become more general. A new CONF
10292 configuration file reader "class" is implemented as well as a
10293 new functions (NCONF_*, for "New CONF") to handle it. The now
10294 old CONF_* functions are still there, but are reimplemented to
10295 work in terms of the new functions. Also, a set of functions
10296 to handle the internal storage of the configuration data is
10297 provided to make it easier to write new configuration file
10298 reader "classes" (I can definitely see something reading a
10299 configuration file in XML format, for example), called _CONF_*,
10300 or "the configuration storage API"...
10302 The new configuration file reading functions are:
10304 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10305 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10307 NCONF_default, NCONF_WIN32
10309 NCONF_dump_fp, NCONF_dump_bio
10311 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10312 NCONF_new creates a new CONF object. This works in the same way
10313 as other interfaces in OpenSSL, like the BIO interface.
10314 NCONF_dump_* dump the internal storage of the configuration file,
10315 which is useful for debugging. All other functions take the same
10316 arguments as the old CONF_* functions wth the exception of the
10317 first that must be a `CONF *' instead of a `LHASH *'.
10319 To make it easer to use the new classes with the old CONF_* functions,
10320 the function CONF_set_default_method is provided.
10323 *) Add '-tls1' option to 'openssl ciphers', which was already
10324 mentioned in the documentation but had not been implemented.
10325 (This option is not yet really useful because even the additional
10326 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10329 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10330 OpenSSL-based applications) load shared libraries and bind to
10331 them in a portable way.
10332 [Geoff Thorpe, with contributions from Richard Levitte]
10334 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10336 *) Make sure _lrotl and _lrotr are only used with MSVC.
10338 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10339 (the default implementation of RAND_status).
10341 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10342 to '-clrext' (= clear extensions), as intended and documented.
10343 [Bodo Moeller; inconsistency pointed out by Michael Attili
10344 <attili@amaxo.com>]
10346 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10347 was larger than the MD block size.
10348 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10350 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10351 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10352 using the passed key: if the passed key was a private key the result
10353 of X509_print(), for example, would be to print out all the private key
10357 *) des_quad_cksum() byte order bug fix.
10358 [Ulf Möller, using the problem description in krb4-0.9.7, where
10359 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10361 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10363 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10365 *) For easily testing in shell scripts whether some command
10366 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10367 returns with exit code 0 iff no command of the given name is available.
10368 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10369 the output goes to stdout and nothing is printed to stderr.
10370 Additional arguments are always ignored.
10372 Since for each cipher there is a command of the same name,
10373 the 'no-cipher' compilation switches can be tested this way.
10375 ('openssl no-XXX' is not able to detect pseudo-commands such
10376 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10379 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10382 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10383 is set; it will be thrown away anyway because each handshake creates
10385 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10386 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10387 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10388 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10391 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10392 'Q' and 'R' lose their special meanings (quit/renegotiate).
10393 This is part of what -quiet does; unlike -quiet, -ign_eof
10394 does not suppress any output.
10397 *) Add compatibility options to the purpose and trust code. The
10398 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10399 accepts a certificate or CA, this was the previous behaviour,
10400 with all the associated security issues.
10402 X509_TRUST_COMPAT is the old trust behaviour: only and
10403 automatically trust self signed roots in certificate store. A
10404 new trust setting X509_TRUST_DEFAULT is used to specify that
10405 a purpose has no associated trust setting and it should instead
10406 use the value in the default purpose.
10409 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10410 and fix a memory leak.
10413 *) In util/mkerr.pl (which implements 'make errors'), preserve
10414 reason strings from the previous version of the .c file, as
10415 the default to have only downcase letters (and digits) in
10416 automatically generated reasons codes is not always appropriate.
10419 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10420 using strerror. Previously, ERR_reason_error_string() returned
10421 library names as reason strings for SYSerr; but SYSerr is a special
10422 case where small numbers are errno values, not library numbers.
10425 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10426 converts DSA parameters into DH parameters. (When creating parameters,
10427 DSA_generate_parameters is used.)
10430 *) Include 'length' (recommended exponent length) in C code generated
10431 by 'openssl dhparam -C'.
10434 *) The second argument to set_label in perlasm was already being used
10435 so couldn't be used as a "file scope" flag. Moved to third argument
10439 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10440 instead of RAND_bytes for encryption IVs and salts.
10443 *) Include RAND_status() into RAND_METHOD instead of implementing
10444 it only for md_rand.c Otherwise replacing the PRNG by calling
10445 RAND_set_rand_method would be impossible.
10448 *) Don't let DSA_generate_key() enter an infinite loop if the random
10449 number generation fails.
10452 *) New 'rand' application for creating pseudo-random output.
10455 *) Added configuration support for Linux/IA64
10456 [Rolf Haberrecker <rolf@suse.de>]
10458 *) Assembler module support for Mingw32.
10461 *) Shared library support for HPUX (in shlib/).
10462 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10464 *) Shared library support for Solaris gcc.
10465 [Lutz Behnke <behnke@trustcenter.de>]
10467 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10469 *) PKCS7_encrypt() was adding text MIME headers twice because they
10470 were added manually and by SMIME_crlf_copy().
10473 *) In bntest.c don't call BN_rand with zero bits argument.
10474 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10476 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10477 case was implemented. This caused BN_div_recp() to fail occasionally.
10480 *) Add an optional second argument to the set_label() in the perl
10481 assembly language builder. If this argument exists and is set
10482 to 1 it signals that the assembler should use a symbol whose
10483 scope is the entire file, not just the current function. This
10484 is needed with MASM which uses the format label:: for this scope.
10485 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10487 *) Change the ASN1 types so they are typedefs by default. Before
10488 almost all types were #define'd to ASN1_STRING which was causing
10489 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10493 *) Change names of new functions to the new get1/get0 naming
10494 convention: After 'get1', the caller owns a reference count
10495 and has to call ..._free; 'get0' returns a pointer to some
10496 data structure without incrementing reference counters.
10497 (Some of the existing 'get' functions increment a reference
10498 counter, some don't.)
10499 Similarly, 'set1' and 'add1' functions increase reference
10500 counters or duplicate objects.
10503 *) Allow for the possibility of temp RSA key generation failure:
10504 the code used to assume it always worked and crashed on failure.
10507 *) Fix potential buffer overrun problem in BIO_printf().
10508 [Ulf Möller, using public domain code by Patrick Powell; problem
10509 pointed out by David Sacerdote <das33@cornell.edu>]
10511 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10512 RAND_egd() and RAND_status(). In the command line application,
10513 the EGD socket can be specified like a seed file using RANDFILE
10517 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10518 Some CAs (e.g. Verisign) distribute certificates in this form.
10521 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10522 list to exclude them. This means that no special compilation option
10523 is needed to use anonymous DH: it just needs to be included in the
10527 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10528 EVP_MD_type. The old functionality is available in a new macro called
10529 EVP_MD_md(). Change code that uses it and update docs.
10532 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10533 where the 'void *' argument is replaced by a function pointer argument.
10534 Previously 'void *' was abused to point to functions, which works on
10535 many platforms, but is not correct. As these functions are usually
10536 called by macros defined in OpenSSL header files, most source code
10537 should work without changes.
10540 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10541 sections with information on -D... compiler switches used for
10542 compiling the library so that applications can see them. To enable
10543 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10544 must be defined. E.g.,
10545 #define OPENSSL_ALGORITHM_DEFINES
10546 #include <openssl/opensslconf.h>
10547 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10548 [Richard Levitte, Ulf and Bodo Möller]
10550 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10554 *) Change the 'other' type in certificate aux info to a STACK_OF
10555 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10556 the required ASN1 format: arbitrary types determined by an OID.
10559 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10560 argument to 'req'. This is not because the function is newer or
10561 better than others it just uses the work 'NEW' in the certificate
10562 request header lines. Some software needs this.
10565 *) Reorganise password command line arguments: now passwords can be
10566 obtained from various sources. Delete the PEM_cb function and make
10567 it the default behaviour: i.e. if the callback is NULL and the
10568 usrdata argument is not NULL interpret it as a null terminated pass
10569 phrase. If usrdata and the callback are NULL then the pass phrase
10570 is prompted for as usual.
10573 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10574 the support is automatically enabled. The resulting binaries will
10575 autodetect the card and use it if present.
10576 [Ben Laurie and Compaq Inc.]
10578 *) Work around for Netscape hang bug. This sends certificate request
10579 and server done in one record. Since this is perfectly legal in the
10580 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10581 the bugs/SSLv3 entry for more info.
10584 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10587 *) Add -rand argument to smime and pkcs12 applications and read/write
10591 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10594 *) Add command line password options to the remaining applications.
10597 *) Bug fix for BN_div_recp() for numerators with an even number of
10601 *) More tests in bntest.c, and changed test_bn output.
10604 *) ./config recognizes MacOS X now.
10607 *) Bug fix for BN_div() when the first words of num and divisor are
10608 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10611 *) Add support for various broken PKCS#8 formats, and command line
10612 options to produce them.
10615 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10616 get temporary BIGNUMs from a BN_CTX.
10619 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10623 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10624 include a #define from the old name to the new. The original intent
10625 was that statically linked binaries could for example just call
10626 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10627 link with digests. This never worked because SSLeay_add_all_digests()
10628 and SSLeay_add_all_ciphers() were in the same source file so calling
10629 one would link with the other. They are now in separate source files.
10632 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10635 *) Use a less unusual form of the Miller-Rabin primality test (it used
10636 a binary algorithm for exponentiation integrated into the Miller-Rabin
10637 loop, our standard modexp algorithms are faster).
10640 *) Support for the EBCDIC character set completed.
10641 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10643 *) Source code cleanups: use const where appropriate, eliminate casts,
10644 use void * instead of char * in lhash.
10647 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10648 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10649 this the server could overwrite ephemeral keys that the client
10653 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10654 using 50 iterations of the Rabin-Miller test.
10656 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10657 iterations of the Rabin-Miller test as required by the appendix
10658 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10659 As BN_is_prime_fasttest includes trial division, DSA parameter
10660 generation becomes much faster.
10662 This implies a change for the callback functions in DSA_is_prime
10663 and DSA_generate_parameters: The callback function is called once
10664 for each positive witness in the Rabin-Miller test, not just
10665 occasionally in the inner loop; and the parameters to the
10666 callback function now provide an iteration count for the outer
10667 loop rather than for the current invocation of the inner loop.
10668 DSA_generate_parameters additionally can call the callback
10669 function with an 'iteration count' of -1, meaning that a
10670 candidate has passed the trial division test (when q is generated
10671 from an application-provided seed, trial division is skipped).
10674 *) New function BN_is_prime_fasttest that optionally does trial
10675 division before starting the Rabin-Miller test and has
10676 an additional BN_CTX * argument (whereas BN_is_prime always
10677 has to allocate at least one BN_CTX).
10678 'callback(1, -1, cb_arg)' is called when a number has passed the
10679 trial division stage.
10682 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10686 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10689 *) New function BN_pseudo_rand().
10692 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10693 bignum version of BN_from_montgomery() with the working code from
10694 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10698 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10699 made it impossible to use the same SSL_SESSION data structure in
10700 SSL2 clients in multiple threads.
10703 *) The return value of RAND_load_file() no longer counts bytes obtained
10704 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10705 to seed the PRNG (previously an explicit byte count was required).
10706 [Ulf Möller, Bodo Möller]
10708 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10709 used (char *) instead of (void *) and had casts all over the place.
10712 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10715 *) Retain source code compatibility for BN_prime_checks macro:
10716 BN_is_prime(..., BN_prime_checks, ...) now uses
10717 BN_prime_checks_for_size to determine the appropriate number of
10718 Rabin-Miller iterations.
10721 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10722 DH_CHECK_P_NOT_SAFE_PRIME.
10723 (Check if this is true? OpenPGP calls them "strong".)
10726 *) Merge the functionality of "dh" and "gendh" programs into a new program
10727 "dhparam". The old programs are retained for now but will handle DH keys
10728 (instead of parameters) in future.
10731 *) Make the ciphers, s_server and s_client programs check the return values
10732 when a new cipher list is set.
10735 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10736 ciphers. Before when the 56bit ciphers were enabled the sorting was
10739 The syntax for the cipher sorting has been extended to support sorting by
10740 cipher-strength (using the strength_bits hard coded in the tables).
10741 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10743 Fix a bug in the cipher-command parser: when supplying a cipher command
10744 string with an "undefined" symbol (neither command nor alphanumeric
10745 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10746 an error is flagged.
10748 Due to the strength-sorting extension, the code of the
10749 ssl_create_cipher_list() function was completely rearranged. I hope that
10750 the readability was also increased :-)
10751 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10753 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10754 for the first serial number and places 2 in the serial number file. This
10755 avoids problems when the root CA is created with serial number zero and
10756 the first user certificate has the same issuer name and serial number
10760 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10761 the new code. Add documentation for this stuff.
10764 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10765 X509_*() to X509at_*() on the grounds that they don't handle X509
10766 structures and behave in an analogous way to the X509v3 functions:
10767 they shouldn't be called directly but wrapper functions should be used
10770 So we also now have some wrapper functions that call the X509at functions
10771 when passed certificate requests. (TO DO: similar things can be done with
10772 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10773 things. Some of these need some d2i or i2d and print functionality
10774 because they handle more complex structures.)
10777 *) Add missing #ifndefs that caused missing symbols when building libssl
10778 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10779 NO_RSA in ssl/s2*.c.
10780 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10782 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10783 has a return value which indicates the quality of the random data
10784 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10785 error queue. New function RAND_pseudo_bytes() generates output that is
10786 guaranteed to be unique but not unpredictable. RAND_add is like
10787 RAND_seed, but takes an extra argument for an entropy estimate
10788 (RAND_seed always assumes full entropy).
10791 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10792 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10793 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10794 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10795 false-positive rate of at most 2^-80 for random input.
10798 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10801 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10802 in the 0.9.5 release), this returns the chain
10803 from an X509_CTX structure with a dup of the stack and all
10804 the X509 reference counts upped: so the stack will exist
10805 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10808 Also make SSL_SESSION_print() print out the verify return
10812 *) Add manpage for the pkcs12 command. Also change the default
10813 behaviour so MAC iteration counts are used unless the new
10814 -nomaciter option is used. This improves file security and
10815 only older versions of MSIE (4.0 for example) need it.
10818 *) Honor the no-xxx Configure options when creating .DEF files.
10821 *) Add PKCS#10 attributes to field table: challengePassword,
10822 unstructuredName and unstructuredAddress. These are taken from
10823 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10824 international characters are used.
10826 More changes to X509_ATTRIBUTE code: allow the setting of types
10827 based on strings. Remove the 'loc' parameter when adding
10828 attributes because these will be a SET OF encoding which is sorted
10832 *) Initial changes to the 'req' utility to allow request generation
10833 automation. This will allow an application to just generate a template
10834 file containing all the field values and have req construct the
10837 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10838 used all over the place including certificate requests and PKCS#7
10839 structures. They are currently handled manually where necessary with
10840 some primitive wrappers for PKCS#7. The new functions behave in a
10841 manner analogous to the X509 extension functions: they allow
10842 attributes to be looked up by NID and added.
10844 Later something similar to the X509V3 code would be desirable to
10845 automatically handle the encoding, decoding and printing of the
10846 more complex types. The string types like challengePassword can
10847 be handled by the string table functions.
10849 Also modified the multi byte string table handling. Now there is
10850 a 'global mask' which masks out certain types. The table itself
10851 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10852 is useful when for example there is only one permissible type
10853 (as in countryName) and using the mask might result in no valid
10857 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10858 SSL_get_peer_finished to allow applications to obtain the latest
10859 Finished messages sent to the peer or expected from the peer,
10860 respectively. (SSL_get_peer_finished is usually the Finished message
10861 actually received from the peer, otherwise the protocol will be aborted.)
10863 As the Finished message are message digests of the complete handshake
10864 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10865 be used for external authentication procedures when the authentication
10866 provided by SSL/TLS is not desired or is not enough.
10869 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10870 the host supports BWX extension and if Compaq C is present on the
10871 $PATH. Just exploiting of the BWX extension results in 20-30%
10872 performance kick for some algorithms, e.g. DES and RC4 to mention
10873 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10877 *) Add support for MS "fast SGC". This is arguably a violation of the
10878 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10879 weak crypto and after checking the certificate is SGC a second one
10880 with strong crypto. MS SGC stops the first handshake after receiving
10881 the server certificate message and sends a second client hello. Since
10882 a server will typically do all the time consuming operations before
10883 expecting any further messages from the client (server key exchange
10884 is the most expensive) there is little difference between the two.
10886 To get OpenSSL to support MS SGC we have to permit a second client
10887 hello message after we have sent server done. In addition we have to
10888 reset the MAC if we do get this second client hello.
10891 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10892 if a DER encoded private key is RSA or DSA traditional format. Changed
10893 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10894 format DER encoded private key. Newer code should use PKCS#8 format which
10895 has the key type encoded in the ASN1 structure. Added DER private key
10896 support to pkcs8 application.
10899 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10900 ciphersuites has been selected (as required by the SSL 3/TLS 1
10901 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10902 is set, we interpret this as a request to violate the specification
10903 (the worst that can happen is a handshake failure, and 'correct'
10904 behaviour would result in a handshake failure anyway).
10907 *) In SSL_CTX_add_session, take into account that there might be multiple
10908 SSL_SESSION structures with the same session ID (e.g. when two threads
10909 concurrently obtain them from an external cache).
10910 The internal cache can handle only one SSL_SESSION with a given ID,
10911 so if there's a conflict, we now throw out the old one to achieve
10915 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10916 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10917 some routines that use cipher OIDs: some ciphers do not have OIDs
10918 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10922 *) Simplify the trust setting structure and code. Now we just have
10923 two sequences of OIDs for trusted and rejected settings. These will
10924 typically have values the same as the extended key usage extension
10925 and any application specific purposes.
10927 The trust checking code now has a default behaviour: it will just
10928 check for an object with the same NID as the passed id. Functions can
10929 be provided to override either the default behaviour or the behaviour
10930 for a given id. SSL client, server and email already have functions
10931 in place for compatibility: they check the NID and also return "trusted"
10932 if the certificate is self signed.
10935 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10936 traditional format into an EVP_PKEY structure.
10939 *) Add a password callback function PEM_cb() which either prompts for
10940 a password if usr_data is NULL or otherwise assumes it is a null
10941 terminated password. Allow passwords to be passed on command line
10942 environment or config files in a few more utilities.
10945 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10946 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10947 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10948 Update documentation.
10951 *) Support for ASN1 "NULL" type. This could be handled before by using
10952 ASN1_TYPE but there wasn't any function that would try to read a NULL
10953 and produce an error if it couldn't. For compatibility we also have
10954 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10955 don't allocate anything because they don't need to.
10958 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10960 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10962 *) Rebuild of the memory allocation routines used by OpenSSL code and
10963 possibly others as well. The purpose is to make an interface that
10964 provide hooks so anyone can build a separate set of allocation and
10965 deallocation routines to be used by OpenSSL, for example memory
10966 pool implementations, or something else, which was previously hard
10967 since Malloc(), Realloc() and Free() were defined as macros having
10968 the values malloc, realloc and free, respectively (except for Win32
10969 compilations). The same is provided for memory debugging code.
10970 OpenSSL already comes with functionality to find memory leaks, but
10971 this gives people a chance to debug other memory problems.
10973 With these changes, a new set of functions and macros have appeared:
10975 CRYPTO_set_mem_debug_functions() [F]
10976 CRYPTO_get_mem_debug_functions() [F]
10977 CRYPTO_dbg_set_options() [F]
10978 CRYPTO_dbg_get_options() [F]
10979 CRYPTO_malloc_debug_init() [M]
10981 The memory debug functions are NULL by default, unless the library
10982 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10983 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10984 gives the standard debugging functions that come with OpenSSL) or
10985 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10986 provided by the library user) must be used. When the standard
10987 debugging functions are used, CRYPTO_dbg_set_options can be used to
10988 request additional information:
10989 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10990 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10992 Also, things like CRYPTO_set_mem_functions will always give the
10993 expected result (the new set of functions is used for allocation
10994 and deallocation) at all times, regardless of platform and compiler
10997 To finish it up, some functions that were never use in any other
10998 way than through macros have a new API and new semantic:
11000 CRYPTO_dbg_malloc()
11001 CRYPTO_dbg_realloc()
11004 All macros of value have retained their old syntax.
11005 [Richard Levitte and Bodo Moeller]
11007 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11008 ordering of SMIMECapabilities wasn't in "strength order" and there
11009 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11013 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11014 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11015 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11017 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11018 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11019 functionality to handle multipart/signed properly) and a utility
11020 called 'smime' to call all this stuff. This is based on code I
11021 originally wrote for Celo who have kindly allowed it to be
11022 included in OpenSSL.
11025 *) Add variants des_set_key_checked and des_set_key_unchecked of
11026 des_set_key (aka des_key_sched). Global variable des_check_key
11027 decides which of these is called by des_set_key; this way
11028 des_check_key behaves as it always did, but applications and
11029 the library itself, which was buggy for des_check_key == 1,
11030 have a cleaner way to pick the version they need.
11033 *) New function PKCS12_newpass() which changes the password of a
11037 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11038 dynamic mix. In both cases the ids can be used as an index into the
11039 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11040 functions so they accept a list of the field values and the
11041 application doesn't need to directly manipulate the X509_TRUST
11045 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11049 *) Modify the way the V3 extension code looks up extensions. This now
11050 works in a similar way to the object code: we have some "standard"
11051 extensions in a static table which is searched with OBJ_bsearch()
11052 and the application can add dynamic ones if needed. The file
11053 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11054 updated whenever a new extension is added to the core code and kept
11055 in ext_nid order. There is a simple program 'tabtest.c' which checks
11056 this. New extensions are not added too often so this file can readily
11057 be maintained manually.
11059 There are two big advantages in doing things this way. The extensions
11060 can be looked up immediately and no longer need to be "added" using
11061 X509V3_add_standard_extensions(): this function now does nothing.
11062 [Side note: I get *lots* of email saying the extension code doesn't
11063 work because people forget to call this function]
11064 Also no dynamic allocation is done unless new extensions are added:
11065 so if we don't add custom extensions there is no need to call
11066 X509V3_EXT_cleanup().
11069 *) Modify enc utility's salting as follows: make salting the default. Add a
11070 magic header, so unsalted files fail gracefully instead of just decrypting
11071 to garbage. This is because not salting is a big security hole, so people
11072 should be discouraged from doing it.
11075 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11076 digest to be passed on the command line but it only used this
11077 parameter when signing a certificate. Modified so all relevant
11078 operations are affected by the digest parameter including the
11079 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11080 DSA key was used because it didn't fix the digest.
11083 *) Initial certificate chain verify code. Currently tests the untrusted
11084 certificates for consistency with the verify purpose (which is set
11085 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11087 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11088 this is because it will reject chains with invalid extensions whereas
11089 every previous version of OpenSSL and SSLeay made no checks at all.
11091 Trust code: checks the root CA for the relevant trust settings. Trust
11092 settings have an initial value consistent with the verify purpose: e.g.
11093 if the verify purpose is for SSL client use it expects the CA to be
11094 trusted for SSL client use. However the default value can be changed to
11095 permit custom trust settings: one example of this would be to only trust
11096 certificates from a specific "secure" set of CAs.
11098 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11099 which should be used for version portability: especially since the
11100 verify structure is likely to change more often now.
11102 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11103 to set them. If not set then assume SSL clients will verify SSL servers
11106 Two new options to the verify program: -untrusted allows a set of
11107 untrusted certificates to be passed in and -purpose which sets the
11108 intended purpose of the certificate. If a purpose is set then the
11109 new chain verify code is used to check extension consistency.
11112 *) Support for the authority information access extension.
11115 *) Modify RSA and DSA PEM read routines to transparently handle
11116 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11117 public keys in a format compatible with certificate
11118 SubjectPublicKeyInfo structures. Unfortunately there were already
11119 functions called *_PublicKey_* which used various odd formats so
11120 these are retained for compatibility: however the DSA variants were
11121 never in a public release so they have been deleted. Changed dsa/rsa
11122 utilities to handle the new format: note no releases ever handled public
11123 keys so we should be OK.
11125 The primary motivation for this change is to avoid the same fiasco
11126 that dogs private keys: there are several incompatible private key
11127 formats some of which are standard and some OpenSSL specific and
11128 require various evil hacks to allow partial transparent handling and
11129 even then it doesn't work with DER formats. Given the option anything
11130 other than PKCS#8 should be dumped: but the other formats have to
11131 stay in the name of compatibility.
11133 With public keys and the benefit of hindsight one standard format
11134 is used which works with EVP_PKEY, RSA or DSA structures: though
11135 it clearly returns an error if you try to read the wrong kind of key.
11137 Added a -pubkey option to the 'x509' utility to output the public key.
11138 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11139 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11140 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11141 that do the same as the EVP_PKEY_assign_*() except they up the
11142 reference count of the added key (they don't "swallow" the
11146 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11147 CRLs would fail if the file contained no certificates or no CRLs:
11148 added a new function to read in both types and return the number
11149 read: this means that if none are read it will be an error. The
11150 DER versions of the certificate and CRL reader would always fail
11151 because it isn't possible to mix certificates and CRLs in DER format
11152 without choking one or the other routine. Changed this to just read
11153 a certificate: this is the best we can do. Also modified the code
11154 in apps/verify.c to take notice of return codes: it was previously
11155 attempting to read in certificates from NULL pointers and ignoring
11156 any errors: this is one reason why the cert and CRL reader seemed
11157 to work. It doesn't check return codes from the default certificate
11158 routines: these may well fail if the certificates aren't installed.
11161 *) Code to support otherName option in GeneralName.
11164 *) First update to verify code. Change the verify utility
11165 so it warns if it is passed a self signed certificate:
11166 for consistency with the normal behaviour. X509_verify
11167 has been modified to it will now verify a self signed
11168 certificate if *exactly* the same certificate appears
11169 in the store: it was previously impossible to trust a
11170 single self signed certificate. This means that:
11171 openssl verify ss.pem
11172 now gives a warning about a self signed certificate but
11173 openssl verify -CAfile ss.pem ss.pem
11177 *) For servers, store verify_result in SSL_SESSION data structure
11178 (and add it to external session representation).
11179 This is needed when client certificate verifications fails,
11180 but an application-provided verification callback (set by
11181 SSL_CTX_set_cert_verify_callback) allows accepting the session
11182 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11183 but returns 1): When the session is reused, we have to set
11184 ssl->verify_result to the appropriate error code to avoid
11186 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11188 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11189 case in PKCS7_dataInit() where the signed PKCS7 structure
11190 didn't contain any existing data because it was being created.
11191 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11193 *) Add a salt to the key derivation routines in enc.c. This
11194 forms the first 8 bytes of the encrypted file. Also add a
11195 -S option to allow a salt to be input on the command line.
11198 *) New function X509_cmp(). Oddly enough there wasn't a function
11199 to compare two certificates. We do this by working out the SHA1
11200 hash and comparing that. X509_cmp() will be needed by the trust
11204 *) SSL_get1_session() is like SSL_get_session(), but increments
11205 the reference count in the SSL_SESSION returned.
11206 [Geoff Thorpe <geoff@eu.c2.net>]
11208 *) Fix for 'req': it was adding a null to request attributes.
11209 Also change the X509_LOOKUP and X509_INFO code to handle
11210 certificate auxiliary information.
11213 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11217 *) Add the possibility to add extra information to the memory leak
11218 detecting output, to form tracebacks, showing from where each
11219 allocation was originated: CRYPTO_push_info("constant string") adds
11220 the string plus current file name and line number to a per-thread
11221 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11222 is like calling CYRPTO_pop_info() until the stack is empty.
11223 Also updated memory leak detection code to be multi-thread-safe.
11226 *) Add options -text and -noout to pkcs7 utility and delete the
11227 encryption options which never did anything. Update docs.
11230 *) Add options to some of the utilities to allow the pass phrase
11231 to be included on either the command line (not recommended on
11232 OSes like Unix) or read from the environment. Update the
11233 manpages and fix a few bugs.
11236 *) Add a few manpages for some of the openssl commands.
11239 *) Fix the -revoke option in ca. It was freeing up memory twice,
11240 leaking and not finding already revoked certificates.
11243 *) Extensive changes to support certificate auxiliary information.
11244 This involves the use of X509_CERT_AUX structure and X509_AUX
11245 functions. An X509_AUX function such as PEM_read_X509_AUX()
11246 can still read in a certificate file in the usual way but it
11247 will also read in any additional "auxiliary information". By
11248 doing things this way a fair degree of compatibility can be
11249 retained: existing certificates can have this information added
11250 using the new 'x509' options.
11252 Current auxiliary information includes an "alias" and some trust
11253 settings. The trust settings will ultimately be used in enhanced
11254 certificate chain verification routines: currently a certificate
11255 can only be trusted if it is self signed and then it is trusted
11259 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11260 The problem was that one of the replacement routines had not been working
11261 since SSLeay releases. For now the offending routine has been replaced
11262 with non-optimised assembler. Even so, this now gives around 95%
11263 performance improvement for 1024 bit RSA signs.
11266 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11267 handling. Most clients have the effective key size in bits equal to
11268 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11269 A few however don't do this and instead use the size of the decrypted key
11270 to determine the RC2 key length and the AlgorithmIdentifier to determine
11271 the effective key length. In this case the effective key length can still
11272 be 40 bits but the key length can be 168 bits for example. This is fixed
11273 by manually forcing an RC2 key into the EVP_PKEY structure because the
11274 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11275 the key length and effective key length are equal.
11278 *) Add a bunch of functions that should simplify the creation of
11279 X509_NAME structures. Now you should be able to do:
11280 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11281 and have it automatically work out the correct field type and fill in
11282 the structures. The more adventurous can try:
11283 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11284 and it will (hopefully) work out the correct multibyte encoding.
11287 *) Change the 'req' utility to use the new field handling and multibyte
11288 copy routines. Before the DN field creation was handled in an ad hoc
11289 way in req, ca, and x509 which was rather broken and didn't support
11290 BMPStrings or UTF8Strings. Since some software doesn't implement
11291 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11292 using the dirstring_type option. See the new comment in the default
11293 openssl.cnf for more info.
11296 *) Make crypto/rand/md_rand.c more robust:
11297 - Assure unique random numbers after fork().
11298 - Make sure that concurrent threads access the global counter and
11299 md serializably so that we never lose entropy in them
11300 or use exactly the same state in multiple threads.
11301 Access to the large state is not always serializable because
11302 the additional locking could be a performance killer, and
11303 md should be large enough anyway.
11306 *) New file apps/app_rand.c with commonly needed functionality
11307 for handling the random seed file.
11309 Use the random seed file in some applications that previously did not:
11311 dsaparam -genkey (which also ignored its '-rand' option),
11314 x509 (when signing).
11315 Except on systems with /dev/urandom, it is crucial to have a random
11316 seed file at least for key creation, DSA signing, and for DH exchanges;
11317 for RSA signatures we could do without one.
11319 gendh and gendsa (unlike genrsa) used to read only the first byte
11320 of each file listed in the '-rand' option. The function as previously
11321 found in genrsa is now in app_rand.c and is used by all programs
11322 that support '-rand'.
11325 *) In RAND_write_file, use mode 0600 for creating files;
11326 don't just chmod when it may be too late.
11329 *) Report an error from X509_STORE_load_locations
11330 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11333 *) New function ASN1_mbstring_copy() this copies a string in either
11334 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11335 into an ASN1_STRING type. A mask of permissible types is passed
11336 and it chooses the "minimal" type to use or an error if not type
11340 *) Add function equivalents to the various macros in asn1.h. The old
11341 macros are retained with an M_ prefix. Code inside the library can
11342 use the M_ macros. External code (including the openssl utility)
11343 should *NOT* in order to be "shared library friendly".
11346 *) Add various functions that can check a certificate's extensions
11347 to see if it usable for various purposes such as SSL client,
11348 server or S/MIME and CAs of these types. This is currently
11349 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11350 verification. Also added a -purpose flag to x509 utility to
11351 print out all the purposes.
11354 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11358 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11359 for, obtain and decode and extension and obtain its critical flag.
11360 This allows all the necessary extension code to be handled in a
11361 single function call.
11364 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11365 platforms. See crypto/rc4/rc4_enc.c for further details.
11368 *) New -noout option to asn1parse. This causes no output to be produced
11369 its main use is when combined with -strparse and -out to extract data
11370 from a file (which may not be in ASN.1 format).
11373 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11374 when producing the local key id.
11375 [Richard Levitte <levitte@stacken.kth.se>]
11377 *) New option -dhparam in s_server. This allows a DH parameter file to be
11378 stated explicitly. If it is not stated then it tries the first server
11379 certificate file. The previous behaviour hard coded the filename
11383 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11384 a public key to be input or output. For example:
11385 openssl rsa -in key.pem -pubout -out pubkey.pem
11386 Also added necessary DSA public key functions to handle this.
11389 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11390 in the message. This was handled by allowing
11391 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11392 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11394 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11395 to the end of the strings whereas this didn't. This would cause problems
11396 if strings read with d2i_ASN1_bytes() were later modified.
11397 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11399 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11400 data and it contains EOF it will end up returning an error. This is
11401 caused by input 46 bytes long. The cause is due to the way base64
11402 BIOs find the start of base64 encoded data. They do this by trying a
11403 trial decode on each line until they find one that works. When they
11404 do a flag is set and it starts again knowing it can pass all the
11405 data directly through the decoder. Unfortunately it doesn't reset
11406 the context it uses. This means that if EOF is reached an attempt
11407 is made to pass two EOFs through the context and this causes the
11408 resulting error. This can also cause other problems as well. As is
11409 usual with these problems it takes *ages* to find and the fix is
11410 trivial: move one line.
11411 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11413 *) Ugly workaround to get s_client and s_server working under Windows. The
11414 old code wouldn't work because it needed to select() on sockets and the
11415 tty (for keypresses and to see if data could be written). Win32 only
11416 supports select() on sockets so we select() with a 1s timeout on the
11417 sockets and then see if any characters are waiting to be read, if none
11418 are present then we retry, we also assume we can always write data to
11419 the tty. This isn't nice because the code then blocks until we've
11420 received a complete line of data and it is effectively polling the
11421 keyboard at 1s intervals: however it's quite a bit better than not
11422 working at all :-) A dedicated Windows application might handle this
11423 with an event loop for example.
11426 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11427 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11428 will be called when RSA_sign() and RSA_verify() are used. This is useful
11429 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11430 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11431 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11432 This necessitated the support of an extra signature type NID_md5_sha1
11433 for SSL signatures and modifications to the SSL library to use it instead
11434 of calling RSA_public_decrypt() and RSA_private_encrypt().
11437 *) Add new -verify -CAfile and -CApath options to the crl program, these
11438 will lookup a CRL issuers certificate and verify the signature in a
11439 similar way to the verify program. Tidy up the crl program so it
11440 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11441 less strict. It will now permit CRL extensions even if it is not
11442 a V2 CRL: this will allow it to tolerate some broken CRLs.
11445 *) Initialize all non-automatic variables each time one of the openssl
11446 sub-programs is started (this is necessary as they may be started
11447 multiple times from the "OpenSSL>" prompt).
11448 [Lennart Bang, Bodo Moeller]
11450 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11451 removing all other RSA functionality (this is what NO_RSA does). This
11452 is so (for example) those in the US can disable those operations covered
11453 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11457 *) Non-copying interface to BIO pairs.
11458 (still largely untested)
11461 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11462 ASCII string. This was handled independently in various places before.
11465 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11466 UTF8 strings a character at a time.
11469 *) Use client_version from client hello to select the protocol
11470 (s23_srvr.c) and for RSA client key exchange verification
11471 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11474 *) Add various utility functions to handle SPKACs, these were previously
11475 handled by poking round in the structure internals. Added new function
11476 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11477 print, verify and generate SPKACs. Based on an original idea from
11478 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11481 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11484 *) Allow the config file extension section to be overwritten on the
11485 command line. Based on an original idea from Massimiliano Pala
11486 <madwolf@comune.modena.it>. The new option is called -extensions
11487 and can be applied to ca, req and x509. Also -reqexts to override
11488 the request extensions in req and -crlexts to override the crl extensions
11492 *) Add new feature to the SPKAC handling in ca. Now you can include
11493 the same field multiple times by preceding it by "XXXX." for example:
11496 this is the same syntax as used in the req config file.
11499 *) Allow certificate extensions to be added to certificate requests. These
11500 are specified in a 'req_extensions' option of the req section of the
11501 config file. They can be printed out with the -text option to req but
11502 are otherwise ignored at present.
11505 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11506 data read consists of only the final block it would not decrypted because
11507 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11508 A misplaced 'break' also meant the decrypted final block might not be
11509 copied until the next read.
11512 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11513 a few extra parameters to the DH structure: these will be useful if
11514 for example we want the value of 'q' or implement X9.42 DH.
11517 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11518 provides hooks that allow the default DSA functions or functions on a
11519 "per key" basis to be replaced. This allows hardware acceleration and
11520 hardware key storage to be handled without major modification to the
11521 library. Also added low level modexp hooks and CRYPTO_EX structure and
11522 associated functions.
11525 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11526 as "read only": it can't be written to and the buffer it points to will
11527 not be freed. Reading from a read only BIO is much more efficient than
11528 a normal memory BIO. This was added because there are several times when
11529 an area of memory needs to be read from a BIO. The previous method was
11530 to create a memory BIO and write the data to it, this results in two
11531 copies of the data and an O(n^2) reading algorithm. There is a new
11532 function BIO_new_mem_buf() which creates a read only memory BIO from
11533 an area of memory. Also modified the PKCS#7 routines to use read only
11537 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11538 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11539 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11540 but a retry condition occurred while trying to read the rest.
11543 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11544 NID_pkcs7_encrypted by default: this was wrong since this should almost
11545 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11546 the encrypted data type: this is a more sensible place to put it and it
11547 allows the PKCS#12 code to be tidied up that duplicated this
11551 *) Changed obj_dat.pl script so it takes its input and output files on
11552 the command line. This should avoid shell escape redirection problems
11556 *) Initial support for certificate extension requests, these are included
11557 in things like Xenroll certificate requests. Included functions to allow
11558 extensions to be obtained and added.
11561 *) -crlf option to s_client and s_server for sending newlines as
11562 CRLF (as required by many protocols).
11565 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11567 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11568 [Ralf S. Engelschall]
11570 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11571 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11573 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11577 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11578 DH parameters/keys (q is lost during that conversion, but the resulting
11579 DH parameters contain its length).
11581 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11582 much faster than DH_generate_parameters (which creates parameters
11583 where p = 2*q + 1), and also the smaller q makes DH computations
11584 much more efficient (160-bit exponentiation instead of 1024-bit
11585 exponentiation); so this provides a convenient way to support DHE
11586 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11587 utter importance to use
11588 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11590 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11591 when such DH parameters are used, because otherwise small subgroup
11592 attacks may become possible!
11595 *) Avoid memory leak in i2d_DHparams.
11598 *) Allow the -k option to be used more than once in the enc program:
11599 this allows the same encrypted message to be read by multiple recipients.
11602 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11603 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11604 it will always use the numerical form of the OID, even if it has a short
11608 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11609 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11610 otherwise bn_mod_exp was called. In the case of hardware keys for example
11611 no private key components need be present and it might store extra data
11612 in the RSA structure, which cannot be accessed from bn_mod_exp.
11613 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11614 private key operations.
11617 *) Added support for SPARC Linux.
11620 *) pem_password_cb function type incompatibly changed from
11621 typedef int pem_password_cb(char *buf, int size, int rwflag);
11623 ....(char *buf, int size, int rwflag, void *userdata);
11624 so that applications can pass data to their callbacks:
11625 The PEM[_ASN1]_{read,write}... functions and macros now take an
11626 additional void * argument, which is just handed through whenever
11627 the password callback is called.
11628 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11630 New function SSL_CTX_set_default_passwd_cb_userdata.
11632 Compatibility note: As many C implementations push function arguments
11633 onto the stack in reverse order, the new library version is likely to
11634 interoperate with programs that have been compiled with the old
11635 pem_password_cb definition (PEM_whatever takes some data that
11636 happens to be on the stack as its last argument, and the callback
11637 just ignores this garbage); but there is no guarantee whatsoever that
11640 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11641 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11642 problems not only on Windows, but also on some Unix platforms.
11643 To avoid problematic command lines, these definitions are now in an
11644 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11645 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11648 *) MIPS III/IV assembler module is reimplemented.
11651 *) More DES library cleanups: remove references to srand/rand and
11652 delete an unused file.
11655 *) Add support for the free Netwide assembler (NASM) under Win32,
11656 since not many people have MASM (ml) and it can be hard to obtain.
11657 This is currently experimental but it seems to work OK and pass all
11658 the tests. Check out INSTALL.W32 for info.
11661 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11662 without temporary keys kept an extra copy of the server key,
11663 and connections with temporary keys did not free everything in case
11667 *) New function RSA_check_key and new openssl rsa option -check
11668 for verifying the consistency of RSA keys.
11669 [Ulf Moeller, Bodo Moeller]
11671 *) Various changes to make Win32 compile work:
11672 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11673 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11674 comparison" warnings.
11675 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11678 *) Add a debugging option to PKCS#5 v2 key generation function: when
11679 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11680 derived keys are printed to stderr.
11683 *) Copy the flags in ASN1_STRING_dup().
11684 [Roman E. Pavlov <pre@mo.msk.ru>]
11686 *) The x509 application mishandled signing requests containing DSA
11687 keys when the signing key was also DSA and the parameters didn't match.
11689 It was supposed to omit the parameters when they matched the signing key:
11690 the verifying software was then supposed to automatically use the CA's
11691 parameters if they were absent from the end user certificate.
11693 Omitting parameters is no longer recommended. The test was also
11694 the wrong way round! This was probably due to unusual behaviour in
11695 EVP_cmp_parameters() which returns 1 if the parameters match.
11696 This meant that parameters were omitted when they *didn't* match and
11697 the certificate was useless. Certificates signed with 'ca' didn't have
11699 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11701 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11702 The interface is as follows:
11703 Applications can use
11704 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11705 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11706 "off" is now the default.
11707 The library internally uses
11708 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11709 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11710 to disable memory-checking temporarily.
11712 Some inconsistent states that previously were possible (and were
11713 even the default) are now avoided.
11715 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11716 with each memory chunk allocated; this is occasionally more helpful
11717 than just having a counter.
11719 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11721 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11725 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11726 which largely parallels "options", but is for changing API behaviour,
11727 whereas "options" are about protocol behaviour.
11728 Initial "mode" flags are:
11730 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11731 a single record has been written.
11732 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11733 retries use the same buffer location.
11734 (But all of the contents must be
11738 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11741 *) Fix problems with no-hmac etc.
11742 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11744 *) New functions RSA_get_default_method(), RSA_set_method() and
11745 RSA_get_method(). These allows replacement of RSA_METHODs without having
11746 to mess around with the internals of an RSA structure.
11749 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11750 Also really enable memory leak checks in openssl.c and in some
11752 [Chad C. Mulligan, Bodo Moeller]
11754 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11755 up the length of negative integers. This has now been simplified to just
11756 store the length when it is first determined and use it later, rather
11757 than trying to keep track of where data is copied and updating it to
11759 [Steve Henson, reported by Brien Wheeler
11760 <bwheeler@authentica-security.com>]
11762 *) Add a new function PKCS7_signatureVerify. This allows the verification
11763 of a PKCS#7 signature but with the signing certificate passed to the
11764 function itself. This contrasts with PKCS7_dataVerify which assumes the
11765 certificate is present in the PKCS#7 structure. This isn't always the
11766 case: certificates can be omitted from a PKCS#7 structure and be
11767 distributed by "out of band" means (such as a certificate database).
11770 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11771 function prototypes in pem.h, also change util/mkdef.pl to add the
11772 necessary function names.
11775 *) mk1mf.pl (used by Windows builds) did not properly read the
11776 options set by Configure in the top level Makefile, and Configure
11777 was not even able to write more than one option correctly.
11778 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11781 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11782 file to be loaded from a BIO or FILE pointer. The BIO version will
11783 for example allow memory BIOs to contain config info.
11786 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11787 Whoever hopes to achieve shared-library compatibility across versions
11788 must use this, not the compile-time macro.
11789 (Exercise 0.9.4: Which is the minimum library version required by
11791 Note: All this applies only to multi-threaded programs, others don't
11795 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11796 through a BIO pair triggered the default case, i.e.
11797 SSLerr(...,SSL_R_UNKNOWN_STATE).
11800 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11801 can use the SSL library even if none of the specific BIOs is
11805 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11806 for the encoded length.
11807 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11809 *) Add initial documentation of the X509V3 functions.
11812 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11813 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11814 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11815 secure PKCS#8 private key format with a high iteration count.
11818 *) Fix determination of Perl interpreter: A perl or perl5
11819 _directory_ in $PATH was also accepted as the interpreter.
11820 [Ralf S. Engelschall]
11822 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11823 wrong with it but it was very old and did things like calling
11824 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11825 unusual formatting.
11828 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11829 to use the new extension code.
11832 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11833 with macros. This should make it easier to change their form, add extra
11834 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11838 *) Add to configuration table a new entry that can specify an alternative
11839 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11840 according to Mark Crispin <MRC@Panda.COM>.
11844 *) DES CBC did not update the IV. Weird.
11847 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11848 Changing the behaviour of the former might break existing programs --
11849 where IV updating is needed, des_ncbc_encrypt can be used.
11852 *) When bntest is run from "make test" it drives bc to check its
11853 calculations, as well as internally checking them. If an internal check
11854 fails, it needs to cause bc to give a non-zero result or make test carries
11855 on without noticing the failure. Fixed.
11858 *) DES library cleanups.
11861 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11862 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11863 ciphers. NOTE: although the key derivation function has been verified
11864 against some published test vectors it has not been extensively tested
11865 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11869 *) Instead of "mkdir -p", which is not fully portable, use new
11870 Perl script "util/mkdir-p.pl".
11873 *) Rewrite the way password based encryption (PBE) is handled. It used to
11874 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11875 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11876 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11877 the 'parameter' field of the AlgorithmIdentifier is passed to the
11878 underlying key generation function so it must do its own ASN1 parsing.
11879 This has also changed the EVP_PBE_CipherInit() function which now has a
11880 'parameter' argument instead of literal salt and iteration count values
11881 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11884 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11885 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11886 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11887 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11888 value was just used as a "magic string" and not used directly its
11889 value doesn't matter.
11892 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11896 *) "linux-sparc64" configuration (ultrapenguin).
11897 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11898 "linux-sparc" configuration.
11899 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11901 *) config now generates no-xxx options for missing ciphers.
11904 *) Support the EBCDIC character set (work in progress).
11905 File ebcdic.c not yet included because it has a different license.
11906 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11908 *) Support BS2000/OSD-POSIX.
11909 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11911 *) Make callbacks for key generation use void * instead of char *.
11914 *) Make S/MIME samples compile (not yet tested).
11917 *) Additional typesafe stacks.
11920 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11924 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11926 *) New configuration variant "sco5-gcc".
11928 *) Updated some demos.
11929 [Sean O Riordain, Wade Scholine]
11931 *) Add missing BIO_free at exit of pkcs12 application.
11934 *) Fix memory leak in conf.c.
11937 *) Updates for Win32 to assembler version of MD5.
11940 *) Set #! path to perl in apps/der_chop to where we found it
11941 instead of using a fixed path.
11944 *) SHA library changes for irix64-mips4-cc.
11947 *) Improvements for VMS support.
11951 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11953 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11954 This also avoids the problems with SC4.2 and unpatched SC5.
11955 [Andy Polyakov <appro@fy.chalmers.se>]
11957 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11958 These are required because of the typesafe stack would otherwise break
11959 existing code. If old code used a structure member which used to be STACK
11960 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11961 sk_num or sk_value it would produce an error because the num, data members
11962 are not present in STACK_OF. Now it just produces a warning. sk_set
11963 replaces the old method of assigning a value to sk_value
11964 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11965 that does this will no longer work (and should use sk_set instead) but
11966 this could be regarded as a "questionable" behaviour anyway.
11969 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11970 correctly handle encrypted S/MIME data.
11973 *) Change type of various DES function arguments from des_cblock
11974 (which means, in function argument declarations, pointer to char)
11975 to des_cblock * (meaning pointer to array with 8 char elements),
11976 which allows the compiler to do more typechecking; it was like
11977 that back in SSLeay, but with lots of ugly casts.
11979 Introduce new type const_des_cblock.
11982 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11983 problems: find RecipientInfo structure that matches recipient certificate
11984 and initialise the ASN1 structures properly based on passed cipher.
11987 *) Belatedly make the BN tests actually check the results.
11990 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11991 to and from BNs: it was completely broken. New compilation option
11992 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11993 key elements as negative integers.
11996 *) Reorganize and speed up MD5.
11997 [Andy Polyakov <appro@fy.chalmers.se>]
12000 [Richard Levitte <richard@levitte.org>]
12002 *) New option -out to asn1parse to allow the parsed structure to be
12003 output to a file. This is most useful when combined with the -strparse
12004 option to examine the output of things like OCTET STRINGS.
12007 *) Make SSL library a little more fool-proof by not requiring any longer
12008 that SSL_set_{accept,connect}_state be called before
12009 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12010 in many applications because usually everything *appeared* to work as
12011 intended anyway -- now it really works as intended).
12014 *) Move openssl.cnf out of lib/.
12017 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12018 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12019 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12020 [Ralf S. Engelschall]
12022 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12023 handle PKCS#7 enveloped data properly.
12024 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12026 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12027 copying pointers. The cert_st handling is changed by this in
12028 various ways (and thus what used to be known as ctx->default_cert
12029 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12030 any longer when s->cert does not give us what we need).
12031 ssl_cert_instantiate becomes obsolete by this change.
12032 As soon as we've got the new code right (possibly it already is?),
12033 we have solved a couple of bugs of the earlier code where s->cert
12034 was used as if it could not have been shared with other SSL structures.
12036 Note that using the SSL API in certain dirty ways now will result
12037 in different behaviour than observed with earlier library versions:
12038 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12039 does not influence s as it used to.
12041 In order to clean up things more thoroughly, inside SSL_SESSION
12042 we don't use CERT any longer, but a new structure SESS_CERT
12043 that holds per-session data (if available); currently, this is
12044 the peer's certificate chain and, for clients, the server's certificate
12045 and temporary key. CERT holds only those values that can have
12046 meaningful defaults in an SSL_CTX.
12049 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12050 from the internal representation. Various PKCS#7 fixes: remove some
12051 evil casts and set the enc_dig_alg field properly based on the signing
12055 *) Allow PKCS#12 password to be set from the command line or the
12056 environment. Let 'ca' get its config file name from the environment
12057 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12061 *) Allow certificate policies extension to use an IA5STRING for the
12062 organization field. This is contrary to the PKIX definition but
12063 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12067 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12068 without disallowing inline assembler and the like for non-pedantic builds.
12071 *) Support Borland C++ builder.
12072 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12074 *) Support Mingw32.
12077 *) SHA-1 cleanups and performance enhancements.
12078 [Andy Polyakov <appro@fy.chalmers.se>]
12080 *) Sparc v8plus assembler for the bignum library.
12081 [Andy Polyakov <appro@fy.chalmers.se>]
12083 *) Accept any -xxx and +xxx compiler options in Configure.
12086 *) Update HPUX configuration.
12089 *) Add missing sk_<type>_unshift() function to safestack.h
12090 [Ralf S. Engelschall]
12092 *) New function SSL_CTX_use_certificate_chain_file that sets the
12093 "extra_cert"s in addition to the certificate. (This makes sense
12094 only for "PEM" format files, as chains as a whole are not
12098 *) Support verify_depth from the SSL API.
12099 x509_vfy.c had what can be considered an off-by-one-error:
12100 Its depth (which was not part of the external interface)
12101 was actually counting the number of certificates in a chain;
12102 now it really counts the depth.
12105 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12106 instead of X509err, which often resulted in confusing error
12107 messages since the error codes are not globally unique
12108 (e.g. an alleged error in ssl3_accept when a certificate
12109 didn't match the private key).
12111 *) New function SSL_CTX_set_session_id_context that allows to set a default
12112 value (so that you don't need SSL_set_session_id_context for each
12113 connection using the SSL_CTX).
12116 *) OAEP decoding bug fix.
12119 *) Support INSTALL_PREFIX for package builders, as proposed by
12123 *) New Configure options "threads" and "no-threads". For systems
12124 where the proper compiler options are known (currently Solaris
12125 and Linux), "threads" is the default.
12128 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12131 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12132 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12133 such as /usr/local/bin.
12136 *) "make linux-shared" to build shared libraries.
12137 [Niels Poppe <niels@netbox.org>]
12139 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12142 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12143 extension adding in x509 utility.
12146 *) Remove NOPROTO sections and error code comments.
12149 *) Partial rewrite of the DEF file generator to now parse the ANSI
12153 *) New Configure options --prefix=DIR and --openssldir=DIR.
12156 *) Complete rewrite of the error code script(s). It is all now handled
12157 by one script at the top level which handles error code gathering,
12158 header rewriting and C source file generation. It should be much better
12159 than the old method: it now uses a modified version of Ulf's parser to
12160 read the ANSI prototypes in all header files (thus the old K&R definitions
12161 aren't needed for error creation any more) and do a better job of
12162 translating function codes into names. The old 'ASN1 error code imbedded
12163 in a comment' is no longer necessary and it doesn't use .err files which
12164 have now been deleted. Also the error code call doesn't have to appear all
12165 on one line (which resulted in some large lines...).
12168 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12171 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12172 0 (which usually indicates a closed connection), but continue reading.
12175 *) Fix some race conditions.
12178 *) Add support for CRL distribution points extension. Add Certificate
12179 Policies and CRL distribution points documentation.
12182 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12185 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12186 8 of keying material. Merlin has also confirmed interop with this fix
12187 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12188 [Merlin Hughes <merlin@baltimore.ie>]
12190 *) Fix lots of warnings.
12191 [Richard Levitte <levitte@stacken.kth.se>]
12193 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12194 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12195 [Richard Levitte <levitte@stacken.kth.se>]
12197 *) Fix problems with sizeof(long) == 8.
12198 [Andy Polyakov <appro@fy.chalmers.se>]
12200 *) Change functions to ANSI C.
12203 *) Fix typos in error codes.
12204 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12206 *) Remove defunct assembler files from Configure.
12209 *) SPARC v8 assembler BIGNUM implementation.
12210 [Andy Polyakov <appro@fy.chalmers.se>]
12212 *) Support for Certificate Policies extension: both print and set.
12213 Various additions to support the r2i method this uses.
12216 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12217 return a const string when you are expecting an allocated buffer.
12220 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12221 types DirectoryString and DisplayText.
12224 *) Add code to allow r2i extensions to access the configuration database,
12225 add an LHASH database driver and add several ctx helper functions.
12228 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12229 fail when they extended the size of a BIGNUM.
12232 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12233 support typesafe stack.
12236 *) Fix typo in SSL_[gs]et_options().
12237 [Nils Frostberg <nils@medcom.se>]
12239 *) Delete various functions and files that belonged to the (now obsolete)
12240 old X509V3 handling code.
12243 *) New Configure option "rsaref".
12246 *) Don't auto-generate pem.h.
12249 *) Introduce type-safe ASN.1 SETs.
12252 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12253 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12255 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12256 that links with OpenSSL (well at least cause lots of warnings), but fear
12257 not: the conversion is trivial, and it eliminates loads of evil casts. A
12258 few STACKed things have been converted already. Feel free to convert more.
12259 In the fullness of time, I'll do away with the STACK type altogether.
12262 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12263 specified in <certfile> by updating the entry in the index.txt file.
12264 This way one no longer has to edit the index.txt file manually for
12265 revoking a certificate. The -revoke option does the gory details now.
12266 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12268 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12269 `-text' option at all and this way the `-noout -text' combination was
12270 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12271 [Ralf S. Engelschall]
12273 *) Make sure a corresponding plain text error message exists for the
12274 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12275 verify callback function determined that a certificate was revoked.
12276 [Ralf S. Engelschall]
12278 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12279 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12280 all available ciphers including rc5, which was forgotten until now.
12281 In order to let the testing shell script know which algorithms
12282 are available, a new (up to now undocumented) command
12283 "openssl list-cipher-commands" is used.
12286 *) Bugfix: s_client occasionally would sleep in select() when
12287 it should have checked SSL_pending() first.
12290 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12291 the raw DSA values prior to ASN.1 encoding.
12294 *) Tweaks to Configure
12295 [Niels Poppe <niels@netbox.org>]
12297 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12301 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12304 *) New config option to avoid instructions that are illegal on the 80386.
12305 The default code is faster, but requires at least a 486.
12308 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12309 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12310 same as SSL2_VERSION anyway.
12313 *) New "-showcerts" option for s_client.
12316 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12317 application. Various cleanups and fixes.
12320 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12321 modify error routines to work internally. Add error codes and PBE init
12322 to library startup routines.
12325 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12326 packing functions to asn1 and evp. Changed function names and error
12327 codes along the way.
12330 *) PKCS12 integration: and so it begins... First of several patches to
12331 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12332 objects to objects.h
12335 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12336 and display support for Thawte strong extranet extension.
12339 *) Add LinuxPPC support.
12340 [Jeff Dubrule <igor@pobox.org>]
12342 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12343 bn_div_words in alpha.s.
12344 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12346 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12347 OAEP isn't supported when OpenSSL is built with RSAref.
12348 [Ulf Moeller <ulf@fitug.de>]
12350 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12351 so they no longer are missing under -DNOPROTO.
12352 [Soren S. Jorvang <soren@t.dk>]
12355 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12357 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12358 doesn't work when the session is reused. Coming soon!
12361 *) Fix a security hole, that allows sessions to be reused in the wrong
12362 context thus bypassing client cert protection! All software that uses
12363 client certs and session caches in multiple contexts NEEDS PATCHING to
12364 allow session reuse! A fuller solution is in the works.
12365 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12367 *) Some more source tree cleanups (removed obsolete files
12368 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12369 permission on "config" script to be executable) and a fix for the INSTALL
12371 [Ulf Moeller <ulf@fitug.de>]
12373 *) Remove some legacy and erroneous uses of malloc, free instead of
12375 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12377 *) Make rsa_oaep_test return non-zero on error.
12378 [Ulf Moeller <ulf@fitug.de>]
12380 *) Add support for native Solaris shared libraries. Configure
12381 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12382 if someone would make that last step automatic.
12383 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12385 *) ctx_size was not built with the right compiler during "make links". Fixed.
12388 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12389 except NULL ciphers". This means the default cipher list will no longer
12390 enable NULL ciphers. They need to be specifically enabled e.g. with
12391 the string "DEFAULT:eNULL".
12394 *) Fix to RSA private encryption routines: if p < q then it would
12395 occasionally produce an invalid result. This will only happen with
12396 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12399 *) Be less restrictive and allow also `perl util/perlpath.pl
12400 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12401 because this way one can also use an interpreter named `perl5' (which is
12402 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12403 installed as `perl').
12404 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12406 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12407 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12409 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12410 advapi32.lib to Win32 build and change the pem test comparison
12411 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12412 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12413 and crypto/des/ede_cbcm_enc.c.
12416 *) DES quad checksum was broken on big-endian architectures. Fixed.
12419 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12420 Win32 test batch file so it (might) work again. The Win32 test batch file
12421 is horrible: I feel ill....
12424 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12425 in e_os.h. Audit of header files to check ANSI and non ANSI
12426 sections: 10 functions were absent from non ANSI section and not exported
12427 from Windows DLLs. Fixed up libeay.num for new functions.
12430 *) Make `openssl version' output lines consistent.
12431 [Ralf S. Engelschall]
12433 *) Fix Win32 symbol export lists for BIO functions: Added
12434 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12435 to ms/libeay{16,32}.def.
12436 [Ralf S. Engelschall]
12438 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12439 fine under Unix and passes some trivial tests I've now added. But the
12440 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12441 added to make sure no one expects that this stuff really works in the
12442 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12443 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12445 [Ralf S. Engelschall]
12447 *) Fix the generation of two part addresses in perl.
12448 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12450 *) Add config entry for Linux on MIPS.
12451 [John Tobey <jtobey@channel1.com>]
12453 *) Make links whenever Configure is run, unless we are on Windoze.
12456 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12457 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12461 *) Add a useful kludge to allow package maintainers to specify compiler and
12462 other platforms details on the command line without having to patch the
12463 Configure script everytime: One now can use ``perl Configure
12464 <id>:<details>'', i.e. platform ids are allowed to have details appended
12465 to them (separated by colons). This is treated as there would be a static
12466 pre-configured entry in Configure's %table under key <id> with value
12467 <details> and ``perl Configure <id>'' is called. So, when you want to
12468 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12469 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12470 now, which overrides the FreeBSD-elf entry on-the-fly.
12471 [Ralf S. Engelschall]
12473 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12476 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12477 on the `perl Configure ...' command line. This way one can compile
12478 OpenSSL libraries with Position Independent Code (PIC) which is needed
12479 for linking it into DSOs.
12480 [Ralf S. Engelschall]
12482 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12486 *) Cleaned up the LICENSE document: The official contact for any license
12487 questions now is the OpenSSL core team under openssl-core@openssl.org.
12488 And add a paragraph about the dual-license situation to make sure people
12489 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12490 to the OpenSSL toolkit.
12491 [Ralf S. Engelschall]
12493 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12494 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12495 Additionally cleaned up the `make links' target: Remove unnecessary
12496 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12497 to speed processing and no longer clutter the display with confusing
12498 stuff. Instead only the actually done links are displayed.
12499 [Ralf S. Engelschall]
12501 *) Permit null encryption ciphersuites, used for authentication only. It used
12502 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12503 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12507 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12508 signed attributes when verifying signatures (this would break them),
12509 the detached data encoding was wrong and public keys obtained using
12510 X509_get_pubkey() weren't freed.
12513 *) Add text documentation for the BUFFER functions. Also added a work around
12514 to a Win95 console bug. This was triggered by the password read stuff: the
12515 last character typed gets carried over to the next fread(). If you were
12516 generating a new cert request using 'req' for example then the last
12517 character of the passphrase would be CR which would then enter the first
12521 *) Added the new `Includes OpenSSL Cryptography Software' button as
12522 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12523 button and can be used by applications based on OpenSSL to show the
12524 relationship to the OpenSSL project.
12525 [Ralf S. Engelschall]
12527 *) Remove confusing variables in function signatures in files
12528 ssl/ssl_lib.c and ssl/ssl.h.
12529 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12531 *) Don't install bss_file.c under PREFIX/include/
12532 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12534 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12535 functions that return function pointers and has support for NT specific
12536 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12537 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12538 unsigned to signed types: this was killing the Win32 compile.
12541 *) Add new certificate file to stack functions,
12542 SSL_add_dir_cert_subjects_to_stack() and
12543 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12544 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12545 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12546 This means that Apache-SSL and similar packages don't have to mess around
12547 to add as many CAs as they want to the preferred list.
12550 *) Experiment with doxygen documentation. Currently only partially applied to
12552 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12553 openssl.doxy as the configuration file.
12556 *) Get rid of remaining C++-style comments which strict C compilers hate.
12557 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12559 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12560 compiled in by default: it has problems with large keys.
12563 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12564 DH private keys and/or callback functions which directly correspond to
12565 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12566 is needed for applications which have to configure certificates on a
12567 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12569 For the RSA certificate situation is makes no difference, but
12570 for the DSA certificate situation this fixes the "no shared cipher"
12571 problem where the OpenSSL cipher selection procedure failed because the
12572 temporary keys were not overtaken from the context and the API provided
12573 no way to reconfigure them.
12574 The new functions now let applications reconfigure the stuff and they
12575 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12576 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12577 non-public-API function ssl_cert_instantiate() is used as a helper
12578 function and also to reduce code redundancy inside ssl_rsa.c.
12579 [Ralf S. Engelschall]
12581 *) Move s_server -dcert and -dkey options out of the undocumented feature
12582 area because they are useful for the DSA situation and should be
12583 recognized by the users.
12584 [Ralf S. Engelschall]
12586 *) Fix the cipher decision scheme for export ciphers: the export bits are
12587 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12588 SSL_EXP_MASK. So, the original variable has to be used instead of the
12589 already masked variable.
12590 [Richard Levitte <levitte@stacken.kth.se>]
12592 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12593 [Richard Levitte <levitte@stacken.kth.se>]
12595 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12596 from `int' to `unsigned int' because it's a length and initialized by
12597 EVP_DigestFinal() which expects an `unsigned int *'.
12598 [Richard Levitte <levitte@stacken.kth.se>]
12600 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12601 script. Instead use the usual Shell->Perl transition trick.
12602 [Ralf S. Engelschall]
12604 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12605 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12606 -noout -modulus' as it's already the case for `openssl rsa -noout
12607 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12608 currently the public key is printed (a decision which was already done by
12609 `openssl dsa -modulus' in the past) which serves a similar purpose.
12610 Additionally the NO_RSA no longer completely removes the whole -modulus
12611 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12613 [Ralf S. Engelschall]
12615 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12616 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12617 [Arne Ansper <arne@ats.cyber.ee>]
12619 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12620 to be added. Now both 'req' and 'ca' can use new objects defined in the
12624 *) Add cool BIO that does syslog (or event log on NT).
12625 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12627 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12628 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12629 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12630 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12633 *) Add preliminary config info for new extension code.
12636 *) Make RSA_NO_PADDING really use no padding.
12637 [Ulf Moeller <ulf@fitug.de>]
12639 *) Generate errors when private/public key check is done.
12642 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12643 for some CRL extensions and new objects added.
12646 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12647 key usage extension and fuller support for authority key id.
12650 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12651 padding method for RSA, which is recommended for new applications in PKCS
12652 #1 v2.0 (RFC 2437, October 1998).
12653 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12654 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12655 against Bleichbacher's attack on RSA.
12656 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12659 *) Updates to the new SSL compression code
12660 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12662 *) Fix so that the version number in the master secret, when passed
12663 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12664 (because the server will not accept higher), that the version number
12665 is 0x03,0x01, not 0x03,0x00
12666 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12668 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12669 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12670 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12673 *) Support for RAW extensions where an arbitrary extension can be
12674 created by including its DER encoding. See apps/openssl.cnf for
12678 *) Make sure latest Perl versions don't interpret some generated C array
12679 code as Perl array code in the crypto/err/err_genc.pl script.
12680 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12682 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12683 not many people have the assembler. Various Win32 compilation fixes and
12684 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12685 build instructions.
12688 *) Modify configure script 'Configure' to automatically create crypto/date.h
12689 file under Win32 and also build pem.h from pem.org. New script
12690 util/mkfiles.pl to create the MINFO file on environments that can't do a
12691 'make files': perl util/mkfiles.pl >MINFO should work.
12694 *) Major rework of DES function declarations, in the pursuit of correctness
12695 and purity. As a result, many evil casts evaporated, and some weirdness,
12696 too. You may find this causes warnings in your code. Zapping your evil
12697 casts will probably fix them. Mostly.
12700 *) Fix for a typo in asn1.h. Bug fix to object creation script
12701 obj_dat.pl. It considered a zero in an object definition to mean
12702 "end of object": none of the objects in objects.h have any zeros
12703 so it wasn't spotted.
12704 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12706 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12707 Masking (CBCM). In the absence of test vectors, the best I have been able
12708 to do is check that the decrypt undoes the encrypt, so far. Send me test
12709 vectors if you have them.
12712 *) Correct calculation of key length for export ciphers (too much space was
12713 allocated for null ciphers). This has not been tested!
12716 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12717 message is now correct (it understands "crypto" and "ssl" on its
12718 command line). There is also now an "update" option. This will update
12719 the util/ssleay.num and util/libeay.num files with any new functions.
12721 perl util/mkdef.pl crypto ssl update
12722 it will update them.
12725 *) Overhauled the Perl interface (perl/*):
12726 - ported BN stuff to OpenSSL's different BN library
12727 - made the perl/ source tree CVS-aware
12728 - renamed the package from SSLeay to OpenSSL (the files still contain
12729 their history because I've copied them in the repository)
12730 - removed obsolete files (the test scripts will be replaced
12731 by better Test::Harness variants in the future)
12732 [Ralf S. Engelschall]
12734 *) First cut for a very conservative source tree cleanup:
12735 1. merge various obsolete readme texts into doc/ssleay.txt
12736 where we collect the old documents and readme texts.
12737 2. remove the first part of files where I'm already sure that we no
12738 longer need them because of three reasons: either they are just temporary
12739 files which were left by Eric or they are preserved original files where
12740 I've verified that the diff is also available in the CVS via "cvs diff
12741 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12742 the crypto/md/ stuff).
12743 [Ralf S. Engelschall]
12745 *) More extension code. Incomplete support for subject and issuer alt
12746 name, issuer and authority key id. Change the i2v function parameters
12747 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12748 what that's for :-) Fix to ASN1 macro which messed up
12749 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12752 *) Preliminary support for ENUMERATED type. This is largely copied from the
12756 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12757 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12759 *) Make sure `make rehash' target really finds the `openssl' program.
12760 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12762 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12763 like to hear about it if this slows down other processors.
12766 *) Add CygWin32 platform information to Configure script.
12767 [Alan Batie <batie@aahz.jf.intel.com>]
12769 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12770 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12772 *) New program nseq to manipulate netscape certificate sequences
12775 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12779 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12780 but the BN code had some problems that would cause failures when
12781 doing certificate verification and some other functions.
12782 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12784 *) Add ASN1 and PEM code to support netscape certificate sequences.
12787 *) Add ASN1 and PEM code to support netscape certificate sequences.
12790 *) Add several PKIX and private extended key usage OIDs.
12793 *) Modify the 'ca' program to handle the new extension code. Modify
12794 openssl.cnf for new extension format, add comments.
12797 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12798 and add a sample to openssl.cnf so req -x509 now adds appropriate
12802 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12803 error code, add initial support to X509_print() and x509 application.
12806 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12807 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12808 stuff is currently isolated and isn't even compiled yet.
12811 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12812 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12813 Removed the versions check from X509 routines when loading extensions:
12814 this allows certain broken certificates that don't set the version
12815 properly to be processed.
12818 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12819 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12820 can still be regenerated with "make depend".
12823 *) Spelling mistake in C version of CAST-128.
12824 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12826 *) Changes to the error generation code. The perl script err-code.pl
12827 now reads in the old error codes and retains the old numbers, only
12828 adding new ones if necessary. It also only changes the .err files if new
12829 codes are added. The makefiles have been modified to only insert errors
12830 when needed (to avoid needlessly modifying header files). This is done
12831 by only inserting errors if the .err file is newer than the auto generated
12832 C file. To rebuild all the error codes from scratch (the old behaviour)
12833 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12834 or delete all the .err files.
12837 *) CAST-128 was incorrectly implemented for short keys. The C version has
12838 been fixed, but is untested. The assembler versions are also fixed, but
12839 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12840 to regenerate it if needed.
12841 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12842 Hagino <itojun@kame.net>]
12844 *) File was opened incorrectly in randfile.c.
12845 [Ulf Möller <ulf@fitug.de>]
12847 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12848 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12849 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12850 al: it's just almost always a UTCTime. Note this patch adds new error
12851 codes so do a "make errors" if there are problems.
12854 *) Correct Linux 1 recognition in config.
12855 [Ulf Möller <ulf@fitug.de>]
12857 *) Remove pointless MD5 hash when using DSA keys in ca.
12858 [Anonymous <nobody@replay.com>]
12860 *) Generate an error if given an empty string as a cert directory. Also
12861 generate an error if handed NULL (previously returned 0 to indicate an
12862 error, but didn't set one).
12863 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12865 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12868 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12869 parameters. This was causing a warning which killed off the Win32 compile.
12872 *) Remove C++ style comments from crypto/bn/bn_local.h.
12873 [Neil Costigan <neil.costigan@celocom.com>]
12875 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12876 based on a text string, looking up short and long names and finally
12877 "dot" format. The "dot" format stuff didn't work. Added new function
12878 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12879 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12880 OID is not part of the table.
12883 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12884 X509_LOOKUP_by_alias().
12887 *) Sort openssl functions by name.
12890 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12891 encryption from sample DSA keys (in case anyone is interested the password
12895 *) Make _all_ *_free functions accept a NULL pointer.
12896 [Frans Heymans <fheymans@isaserver.be>]
12898 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12900 [Anonymous <nobody@replay.com>]
12902 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12903 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12905 *) Don't blow it for numeric -newkey arguments to apps/req.
12906 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12908 *) Temp key "for export" tests were wrong in s3_srvr.c.
12909 [Anonymous <nobody@replay.com>]
12911 *) Add prototype for temp key callback functions
12912 SSL_CTX_set_tmp_{rsa,dh}_callback().
12915 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12916 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12919 *) X509_name_add_entry() freed the wrong thing after an error.
12920 [Arne Ansper <arne@ats.cyber.ee>]
12922 *) rsa_eay.c would attempt to free a NULL context.
12923 [Arne Ansper <arne@ats.cyber.ee>]
12925 *) BIO_s_socket() had a broken should_retry() on Windoze.
12926 [Arne Ansper <arne@ats.cyber.ee>]
12928 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12929 [Arne Ansper <arne@ats.cyber.ee>]
12931 *) Make sure the already existing X509_STORE->depth variable is initialized
12932 in X509_STORE_new(), but document the fact that this variable is still
12933 unused in the certificate verification process.
12934 [Ralf S. Engelschall]
12936 *) Fix the various library and apps files to free up pkeys obtained from
12937 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12940 *) Fix reference counting in X509_PUBKEY_get(). This makes
12941 demos/maurice/example2.c work, amongst others, probably.
12942 [Steve Henson and Ben Laurie]
12944 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12945 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12946 are no longer created. This way we have a single and consistent command
12947 line interface `openssl <command>', similar to `cvs <command>'.
12948 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12950 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12951 BIT STRING wrapper always have zero unused bits.
12954 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12957 *) Make the top-level INSTALL documentation easier to understand.
12960 *) Makefiles updated to exit if an error occurs in a sub-directory
12961 make (including if user presses ^C) [Paul Sutton]
12963 *) Make Montgomery context stuff explicit in RSA data structure.
12966 *) Fix build order of pem and err to allow for generated pem.h.
12969 *) Fix renumbering bug in X509_NAME_delete_entry().
12972 *) Enhanced the err-ins.pl script so it makes the error library number
12973 global and can add a library name. This is needed for external ASN1 and
12974 other error libraries.
12977 *) Fixed sk_insert which never worked properly.
12980 *) Fix ASN1 macros so they can handle indefinite length constructed
12981 EXPLICIT tags. Some non standard certificates use these: they can now
12985 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12986 into a single doc/ssleay.txt bundle. This way the information is still
12987 preserved but no longer messes up this directory. Now it's new room for
12988 the new set of documentation files.
12989 [Ralf S. Engelschall]
12991 *) SETs were incorrectly DER encoded. This was a major pain, because they
12992 shared code with SEQUENCEs, which aren't coded the same. This means that
12993 almost everything to do with SETs or SEQUENCEs has either changed name or
12994 number of arguments.
12995 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12997 *) Fix test data to work with the above.
13000 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13001 was already fixed by Eric for 0.9.1 it seems.
13002 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13004 *) Autodetect FreeBSD3.
13007 *) Fix various bugs in Configure. This affects the following platforms:
13011 unixware-2.0-pentium
13015 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13016 before they are needed.
13019 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13023 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13025 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13026 changed SSLeay to OpenSSL in version strings.
13027 [Ralf S. Engelschall]
13029 *) Some fixups to the top-level documents.
13032 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13033 because the symlink to include/ was missing.
13034 [Ralf S. Engelschall]
13036 *) Incorporated the popular no-RSA/DSA-only patches
13037 which allow to compile a RSA-free SSLeay.
13038 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13040 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13041 when "ssleay" is still not found.
13042 [Ralf S. Engelschall]
13044 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13045 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13047 *) Updated the README file.
13048 [Ralf S. Engelschall]
13050 *) Added various .cvsignore files in the CVS repository subdirs
13051 to make a "cvs update" really silent.
13052 [Ralf S. Engelschall]
13054 *) Recompiled the error-definition header files and added
13055 missing symbols to the Win32 linker tables.
13056 [Ralf S. Engelschall]
13058 *) Cleaned up the top-level documents;
13059 o new files: CHANGES and LICENSE
13060 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13061 o merged COPYRIGHT into LICENSE
13062 o removed obsolete TODO file
13063 o renamed MICROSOFT to INSTALL.W32
13064 [Ralf S. Engelschall]
13066 *) Removed dummy files from the 0.9.1b source tree:
13067 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13068 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13069 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13070 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13071 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13072 [Ralf S. Engelschall]
13074 *) Added various platform portability fixes.
13077 *) The Genesis of the OpenSSL rpject:
13078 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13079 Young and Tim J. Hudson created while they were working for C2Net until
13081 [The OpenSSL Project]
13084 Changes between 0.9.0b and 0.9.1b [not released]
13086 *) Updated a few CA certificates under certs/
13089 *) Changed some BIGNUM api stuff.
13092 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13093 DGUX x86, Linux Alpha, etc.
13096 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13097 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13101 *) Add -strparse option to asn1pars program which parses nested
13103 [Dr Stephen Henson <shenson@bigfoot.com>]
13105 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13108 *) DSA fix for "ca" program.
13111 *) Added "-genkey" option to "dsaparam" program.
13114 *) Added RIPE MD160 (rmd160) message digest.
13117 *) Added -a (all) option to "ssleay version" command.
13120 *) Added PLATFORM define which is the id given to Configure.
13123 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13126 *) Extended the ASN.1 parser routines.
13129 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13132 *) Added a BN_CTX to the BN library.
13135 *) Fixed the weak key values in DES library
13138 *) Changed API in EVP library for cipher aliases.
13141 *) Added support for RC2/64bit cipher.
13144 *) Converted the lhash library to the crypto/mem.c functions.
13147 *) Added more recognized ASN.1 object ids.
13150 *) Added more RSA padding checks for SSL/TLS.
13153 *) Added BIO proxy/filter functionality.
13156 *) Added extra_certs to SSL_CTX which can be used
13157 send extra CA certificates to the client in the CA cert chain sending
13158 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13161 *) Now Fortezza is denied in the authentication phase because
13162 this is key exchange mechanism is not supported by SSLeay at all.
13165 *) Additional PKCS1 checks.
13168 *) Support the string "TLSv1" for all TLS v1 ciphers.
13171 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13172 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13175 *) Fixed a few memory leaks.
13178 *) Fixed various code and comment typos.
13181 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13182 bytes sent in the client random.
13183 [Edward Bishop <ebishop@spyglass.com>]