5 Changes between 1.1.0e and 1.1.0f [xx XXX xxxx]
9 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
10 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
11 which is the minimum version we support.
14 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
16 *) Encrypt-Then-Mac renegotiation crash
18 During a renegotiation handshake if the Encrypt-Then-Mac extension is
19 negotiated where it was not in the original handshake (or vice-versa) then
20 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
21 and servers are affected.
23 This issue was reported to OpenSSL by Joe Orton (Red Hat).
27 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
29 *) Truncated packet could crash via OOB read
31 If one side of an SSL/TLS path is running on a 32-bit host and a specific
32 cipher is being used, then a truncated packet can cause that host to
33 perform an out-of-bounds read, usually resulting in a crash.
35 This issue was reported to OpenSSL by Robert Święcki of Google.
39 *) Bad (EC)DHE parameters cause a client crash
41 If a malicious server supplies bad parameters for a DHE or ECDHE key
42 exchange then this can result in the client attempting to dereference a
43 NULL pointer leading to a client crash. This could be exploited in a Denial
46 This issue was reported to OpenSSL by Guido Vranken.
50 *) BN_mod_exp may produce incorrect results on x86_64
52 There is a carry propagating bug in the x86_64 Montgomery squaring
53 procedure. No EC algorithms are affected. Analysis suggests that attacks
54 against RSA and DSA as a result of this defect would be very difficult to
55 perform and are not believed likely. Attacks against DH are considered just
56 feasible (although very difficult) because most of the work necessary to
57 deduce information about a private key may be performed offline. The amount
58 of resources required for such an attack would be very significant and
59 likely only accessible to a limited number of attackers. An attacker would
60 additionally need online access to an unpatched system using the target
61 private key in a scenario with persistent DH parameters and a private
62 key that is shared between multiple clients. For example this can occur by
63 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
64 similar to CVE-2015-3193 but must be treated as a separate problem.
66 This issue was reported to OpenSSL by the OSS-Fuzz project.
70 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
72 *) ChaCha20/Poly1305 heap-buffer-overflow
74 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
75 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
76 crash. This issue is not considered to be exploitable beyond a DoS.
78 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
82 *) CMS Null dereference
84 Applications parsing invalid CMS structures can crash with a NULL pointer
85 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
86 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
87 structure callback if an attempt is made to free certain invalid encodings.
88 Only CHOICE structures using a callback which do not handle NULL value are
91 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
95 *) Montgomery multiplication may produce incorrect results
97 There is a carry propagating bug in the Broadwell-specific Montgomery
98 multiplication procedure that handles input lengths divisible by, but
99 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
100 and DH private keys are impossible. This is because the subroutine in
101 question is not used in operations with the private key itself and an input
102 of the attacker's direct choice. Otherwise the bug can manifest itself as
103 transient authentication and key negotiation failures or reproducible
104 erroneous outcome of public-key operations with specially crafted input.
105 Among EC algorithms only Brainpool P-512 curves are affected and one
106 presumably can attack ECDH key negotiation. Impact was not analyzed in
107 detail, because pre-requisites for attack are considered unlikely. Namely
108 multiple clients have to choose the curve in question and the server has to
109 share the private key among them, neither of which is default behaviour.
110 Even then only clients that chose the curve will be affected.
112 This issue was publicly reported as transient failures and was not
113 initially recognized as a security issue. Thanks to Richard Morgan for
114 providing reproducible case.
118 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
119 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
120 prevent issues where no progress is being made and the peer continually
121 sends unrecognised record types, using up resources processing them.
124 *) Removed automatic addition of RPATH in shared libraries and executables,
125 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
128 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
130 *) Fix Use After Free for large message sizes
132 The patch applied to address CVE-2016-6307 resulted in an issue where if a
133 message larger than approx 16k is received then the underlying buffer to
134 store the incoming message is reallocated and moved. Unfortunately a
135 dangling pointer to the old location is left which results in an attempt to
136 write to the previously freed location. This is likely to result in a
137 crash, however it could potentially lead to execution of arbitrary code.
139 This issue only affects OpenSSL 1.1.0a.
141 This issue was reported to OpenSSL by Robert Święcki.
145 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
147 *) OCSP Status Request extension unbounded memory growth
149 A malicious client can send an excessively large OCSP Status Request
150 extension. If that client continually requests renegotiation, sending a
151 large OCSP Status Request extension each time, then there will be unbounded
152 memory growth on the server. This will eventually lead to a Denial Of
153 Service attack through memory exhaustion. Servers with a default
154 configuration are vulnerable even if they do not support OCSP. Builds using
155 the "no-ocsp" build time option are not affected.
157 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
161 *) SSL_peek() hang on empty record
163 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
164 sends an empty record. This could be exploited by a malicious peer in a
165 Denial Of Service attack.
167 This issue was reported to OpenSSL by Alex Gaynor.
171 *) Excessive allocation of memory in tls_get_message_header() and
172 dtls1_preprocess_fragment()
174 A (D)TLS message includes 3 bytes for its length in the header for the
175 message. This would allow for messages up to 16Mb in length. Messages of
176 this length are excessive and OpenSSL includes a check to ensure that a
177 peer is sending reasonably sized messages in order to avoid too much memory
178 being consumed to service a connection. A flaw in the logic of version
179 1.1.0 means that memory for the message is allocated too early, prior to
180 the excessive message length check. Due to way memory is allocated in
181 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
182 to service a connection. This could lead to a Denial of Service through
183 memory exhaustion. However, the excessive message length check still takes
184 place, and this would cause the connection to immediately fail. Assuming
185 that the application calls SSL_free() on the failed conneciton in a timely
186 manner then the 21Mb of allocated memory will then be immediately freed
187 again. Therefore the excessive memory allocation will be transitory in
188 nature. This then means that there is only a security impact if:
190 1) The application does not call SSL_free() in a timely manner in the event
191 that the connection fails
193 2) The application is working in a constrained environment where there is
194 very little free memory
196 3) The attacker initiates multiple connection attempts such that there are
197 multiple connections in a state where memory has been allocated for the
198 connection; SSL_free() has not yet been called; and there is insufficient
199 memory to service the multiple requests.
201 Except in the instance of (1) above any Denial Of Service is likely to be
202 transitory because as soon as the connection fails the memory is
203 subsequently freed again in the SSL_free() call. However there is an
204 increased risk during this period of application crashes due to the lack of
205 memory - which would then mean a more serious Denial of Service.
207 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
208 (CVE-2016-6307 and CVE-2016-6308)
211 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
212 had to be removed. Primary reason is that vendor assembler can't
213 assemble our modules with -KPIC flag. As result it, assembly
214 support, was not even available as option. But its lack means
215 lack of side-channel resistant code, which is incompatible with
216 security by todays standards. Fortunately gcc is readily available
217 prepackaged option, which we firmly point at...
220 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
222 *) Windows command-line tool supports UTF-8 opt-in option for arguments
223 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
224 (to any value) allows Windows user to access PKCS#12 file generated
225 with Windows CryptoAPI and protected with non-ASCII password, as well
226 as files generated under UTF-8 locale on Linux also protected with
230 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
231 have been disabled by default and removed from DEFAULT, just like RC4.
232 See the RC4 item below to re-enable both.
235 *) The method for finding the storage location for the Windows RAND seed file
236 has changed. First we check %RANDFILE%. If that is not set then we check
237 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
238 all else fails we fall back to C:\.
241 *) The EVP_EncryptUpdate() function has had its return type changed from void
242 to int. A return of 0 indicates and error while a return of 1 indicates
246 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
247 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
248 off the constant time implementation for RSA, DSA and DH have been made
249 no-ops and deprecated.
252 *) Windows RAND implementation was simplified to only get entropy by
253 calling CryptGenRandom(). Various other RAND-related tickets
255 [Joseph Wylie Yandle, Rich Salz]
257 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
258 and OPENSSL_LH_, respectively. The old names are available
259 with API compatibility. They new names are now completely documented.
262 *) Unify TYPE_up_ref(obj) methods signature.
263 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
264 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
265 int (instead of void) like all others TYPE_up_ref() methods.
266 So now these methods also check the return value of CRYPTO_atomic_add(),
267 and the validity of object reference counter.
268 [fdasilvayy@gmail.com]
270 *) With Windows Visual Studio builds, the .pdb files are installed
271 alongside the installed libraries and executables. For a static
272 library installation, ossl_static.pdb is the associate compiler
273 generated .pdb file to be used when linking programs.
276 *) Remove openssl.spec. Packaging files belong with the packagers.
279 *) Automatic Darwin/OSX configuration has had a refresh, it will now
280 recognise x86_64 architectures automatically. You can still decide
281 to build for a different bitness with the environment variable
282 KERNEL_BITS (can be 32 or 64), for example:
284 KERNEL_BITS=32 ./config
288 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
289 256 bit AES and HMAC with SHA256.
292 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
295 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
298 *) To enable users to have their own config files and build file templates,
299 Configure looks in the directory indicated by the environment variable
300 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
301 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
302 name and is used as is.
305 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
306 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
307 X509_CERT_FILE_CTX was removed.
310 *) "shared" builds are now the default. To create only static libraries use
311 the "no-shared" Configure option.
314 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
315 All of these option have not worked for some while and are fundamental
319 *) Make various cleanup routines no-ops and mark them as deprecated. Most
320 global cleanup functions are no longer required because they are handled
321 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
322 Explicitly de-initing can cause problems (e.g. where a library that uses
323 OpenSSL de-inits, but an application is still using it). The affected
324 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
325 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
326 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
330 *) --strict-warnings no longer enables runtime debugging options
331 such as REF_DEBUG. Instead, debug options are automatically
332 enabled with '--debug' builds.
333 [Andy Polyakov, Emilia Käsper]
335 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
336 have been moved out of the public header files. New functions for managing
337 these have been added.
340 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
341 objects have been moved out of the public header files. New
342 functions for managing these have been added.
345 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
346 have been moved out of the public header files. New functions for managing
347 these have been added.
350 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
351 moved out of the public header files. New functions for managing these
355 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
358 *) Removed the mk1mf build scripts.
361 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
362 it is always safe to #include a header now.
365 *) Removed the aged BC-32 config and all its supporting scripts
368 *) Removed support for Ultrix, Netware, and OS/2.
371 *) Add support for HKDF.
374 *) Add support for blake2b and blake2s
377 *) Added support for "pipelining". Ciphers that have the
378 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
379 encryptions/decryptions simultaneously. There are currently no built-in
380 ciphers with this property but the expectation is that engines will be able
381 to offer it to significantly improve throughput. Support has been extended
382 into libssl so that multiple records for a single connection can be
383 processed in one go (for >=TLS 1.1).
386 *) Added the AFALG engine. This is an async capable engine which is able to
387 offload work to the Linux kernel. In this initial version it only supports
388 AES128-CBC. The kernel must be version 4.1.0 or greater.
391 *) OpenSSL now uses a new threading API. It is no longer necessary to
392 set locking callbacks to use OpenSSL in a multi-threaded environment. There
393 are two supported threading models: pthreads and windows threads. It is
394 also possible to configure OpenSSL at compile time for "no-threads". The
395 old threading API should no longer be used. The functions have been
396 replaced with "no-op" compatibility macros.
397 [Alessandro Ghedini, Matt Caswell]
399 *) Modify behavior of ALPN to invoke callback after SNI/servername
400 callback, such that updates to the SSL_CTX affect ALPN.
403 *) Add SSL_CIPHER queries for authentication and key-exchange.
406 *) Changes to the DEFAULT cipherlist:
407 - Prefer (EC)DHE handshakes over plain RSA.
408 - Prefer AEAD ciphers over legacy ciphers.
409 - Prefer ECDSA over RSA when both certificates are available.
410 - Prefer TLSv1.2 ciphers/PRF.
411 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
415 *) Change the ECC default curve list to be this, in order: x25519,
416 secp256r1, secp521r1, secp384r1.
419 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
420 disabled by default. They can be re-enabled using the
421 enable-weak-ssl-ciphers option to Configure.
424 *) If the server has ALPN configured, but supports no protocols that the
425 client advertises, send a fatal "no_application_protocol" alert.
426 This behaviour is SHALL in RFC 7301, though it isn't universally
427 implemented by other servers.
430 *) Add X25519 support.
431 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
432 for public and private key encoding using the format documented in
433 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
434 key generation and key derivation.
436 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
440 *) Deprecate SRP_VBASE_get_by_user.
441 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
442 In order to fix an unavoidable memory leak (CVE-2016-0798),
443 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
444 seed, even if the seed is configured.
446 Users should use SRP_VBASE_get1_by_user instead. Note that in
447 SRP_VBASE_get1_by_user, caller must free the returned value. Note
448 also that even though configuring the SRP seed attempts to hide
449 invalid usernames by continuing the handshake with fake
450 credentials, this behaviour is not constant time and no strong
451 guarantees are made that the handshake is indistinguishable from
452 that of a valid user.
455 *) Configuration change; it's now possible to build dynamic engines
456 without having to build shared libraries and vice versa. This
457 only applies to the engines in engines/, those in crypto/engine/
458 will always be built into libcrypto (i.e. "static").
460 Building dynamic engines is enabled by default; to disable, use
461 the configuration option "disable-dynamic-engine".
463 The only requirements for building dynamic engines are the
464 presence of the DSO module and building with position independent
465 code, so they will also automatically be disabled if configuring
466 with "disable-dso" or "disable-pic".
468 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
469 are also taken away from openssl/opensslconf.h, as they are
473 *) Configuration change; if there is a known flag to compile
474 position independent code, it will always be applied on the
475 libcrypto and libssl object files, and never on the application
476 object files. This means other libraries that use routines from
477 libcrypto / libssl can be made into shared libraries regardless
478 of how OpenSSL was configured.
480 If this isn't desirable, the configuration options "disable-pic"
481 or "no-pic" can be used to disable the use of PIC. This will
482 also disable building shared libraries and dynamic engines.
485 *) Removed JPAKE code. It was experimental and has no wide use.
488 *) The INSTALL_PREFIX Makefile variable has been renamed to
489 DESTDIR. That makes for less confusion on what this variable
490 is for. Also, the configuration option --install_prefix is
494 *) Heartbeat for TLS has been removed and is disabled by default
495 for DTLS; configure with enable-heartbeats. Code that uses the
496 old #define's might need to be updated.
497 [Emilia Käsper, Rich Salz]
499 *) Rename REF_CHECK to REF_DEBUG.
502 *) New "unified" build system
504 The "unified" build system is aimed to be a common system for all
505 platforms we support. With it comes new support for VMS.
507 This system builds supports building in a different directory tree
508 than the source tree. It produces one Makefile (for unix family
509 or lookalikes), or one descrip.mms (for VMS).
511 The source of information to make the Makefile / descrip.mms is
512 small files called 'build.info', holding the necessary
513 information for each directory with source to compile, and a
514 template in Configurations, like unix-Makefile.tmpl or
517 With this change, the library names were also renamed on Windows
518 and on VMS. They now have names that are closer to the standard
519 on Unix, and include the major version number, and in certain
520 cases, the architecture they are built for. See "Notes on shared
521 libraries" in INSTALL.
523 We rely heavily on the perl module Text::Template.
526 *) Added support for auto-initialisation and de-initialisation of the library.
527 OpenSSL no longer requires explicit init or deinit routines to be called,
528 except in certain circumstances. See the OPENSSL_init_crypto() and
529 OPENSSL_init_ssl() man pages for further information.
532 *) The arguments to the DTLSv1_listen function have changed. Specifically the
533 "peer" argument is now expected to be a BIO_ADDR object.
535 *) Rewrite of BIO networking library. The BIO library lacked consistent
536 support of IPv6, and adding it required some more extensive
537 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
538 which hold all types of addresses and chains of address information.
539 It also introduces a new API, with functions like BIO_socket,
540 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
541 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
542 have been adapted accordingly.
545 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
549 *) CRIME protection: disable compression by default, even if OpenSSL is
550 compiled with zlib enabled. Applications can still enable compression
551 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
552 using the SSL_CONF library to configure compression.
555 *) The signature of the session callback configured with
556 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
557 was explicitly marked as 'const unsigned char*' instead of
561 *) Always DPURIFY. Remove the use of uninitialized memory in the
562 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
565 *) Removed many obsolete configuration items, including
566 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
567 MD2_CHAR, MD2_INT, MD2_LONG
569 IDEA_SHORT, IDEA_LONG
570 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
571 [Rich Salz, with advice from Andy Polyakov]
573 *) Many BN internals have been moved to an internal header file.
574 [Rich Salz with help from Andy Polyakov]
576 *) Configuration and writing out the results from it has changed.
577 Files such as Makefile include/openssl/opensslconf.h and are now
578 produced through general templates, such as Makefile.in and
579 crypto/opensslconf.h.in and some help from the perl module
582 Also, the center of configuration information is no longer
583 Makefile. Instead, Configure produces a perl module in
584 configdata.pm which holds most of the config data (in the hash
585 table %config), the target data that comes from the target
586 configuration in one of the Configurations/*.conf files (in
590 *) To clarify their intended purposes, the Configure options
591 --prefix and --openssldir change their semantics, and become more
592 straightforward and less interdependent.
594 --prefix shall be used exclusively to give the location INSTALLTOP
595 where programs, scripts, libraries, include files and manuals are
596 going to be installed. The default is now /usr/local.
598 --openssldir shall be used exclusively to give the default
599 location OPENSSLDIR where certificates, private keys, CRLs are
600 managed. This is also where the default openssl.cnf gets
602 If the directory given with this option is a relative path, the
603 values of both the --prefix value and the --openssldir value will
604 be combined to become OPENSSLDIR.
605 The default for --openssldir is INSTALLTOP/ssl.
607 Anyone who uses --openssldir to specify where OpenSSL is to be
608 installed MUST change to use --prefix instead.
611 *) The GOST engine was out of date and therefore it has been removed. An up
612 to date GOST engine is now being maintained in an external repository.
613 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
614 support for GOST ciphersuites (these are only activated if a GOST engine
618 *) EGD is no longer supported by default; use enable-egd when
620 [Ben Kaduk and Rich Salz]
622 *) The distribution now has Makefile.in files, which are used to
623 create Makefile's when Configure is run. *Configure must be run
624 before trying to build now.*
627 *) The return value for SSL_CIPHER_description() for error conditions
631 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
633 Obtaining and performing DNSSEC validation of TLSA records is
634 the application's responsibility. The application provides
635 the TLSA records of its choice to OpenSSL, and these are then
636 used to authenticate the peer.
638 The TLSA records need not even come from DNS. They can, for
639 example, be used to implement local end-entity certificate or
640 trust-anchor "pinning", where the "pin" data takes the form
641 of TLSA records, which can augment or replace verification
642 based on the usual WebPKI public certification authorities.
645 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
646 continues to support deprecated interfaces in default builds.
647 However, applications are strongly advised to compile their
648 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
649 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
650 or the 1.1.0 releases.
652 In environments in which all applications have been ported to
653 not use any deprecated interfaces OpenSSL's Configure script
654 should be used with the --api=1.1.0 option to entirely remove
655 support for the deprecated features from the library and
656 unconditionally disable them in the installed headers.
657 Essentially the same effect can be achieved with the "no-deprecated"
658 argument to Configure, except that this will always restrict
659 the build to just the latest API, rather than a fixed API
662 As applications are ported to future revisions of the API,
663 they should update their compile-time OPENSSL_API_COMPAT define
664 accordingly, but in most cases should be able to continue to
665 compile with later releases.
667 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
668 0x10000000L and 0x00908000L, respectively. However those
669 versions did not support the OPENSSL_API_COMPAT feature, and
670 so applications are not typically tested for explicit support
671 of just the undeprecated features of either release.
674 *) Add support for setting the minimum and maximum supported protocol.
675 It can bet set via the SSL_set_min_proto_version() and
676 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
677 MaxProtcol. It's recommended to use the new APIs to disable
678 protocols instead of disabling individual protocols using
679 SSL_set_options() or SSL_CONF's Protocol. This change also
680 removes support for disabling TLS 1.2 in the OpenSSL TLS
681 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
684 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
687 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
688 and integrates ECDSA and ECDH functionality into EC. Implementations can
689 now redirect key generation and no longer need to convert to or from
692 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
693 include the ec.h header file instead.
696 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
697 ciphers who are no longer supported and drops support the ephemeral RSA key
698 exchange. The LOW ciphers currently doesn't have any ciphers in it.
701 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
702 opaque. For HMAC_CTX, the following constructors and destructors
705 HMAC_CTX *HMAC_CTX_new(void);
706 void HMAC_CTX_free(HMAC_CTX *ctx);
708 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
709 destroy such methods has been added. See EVP_MD_meth_new(3) and
710 EVP_CIPHER_meth_new(3) for documentation.
713 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
714 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
715 EVP_MD_CTX_reset() should be called instead to reinitialise
716 an already created structure.
717 2) For consistency with the majority of our object creators and
718 destructors, EVP_MD_CTX_(create|destroy) were renamed to
719 EVP_MD_CTX_(new|free). The old names are retained as macros
720 for deprecated builds.
723 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
724 cryptographic operations to be performed asynchronously as long as an
725 asynchronous capable engine is used. See the ASYNC_start_job() man page for
726 further details. Libssl has also had this capability integrated with the
727 introduction of the new mode SSL_MODE_ASYNC and associated error
728 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
729 pages. This work was developed in partnership with Intel Corp.
732 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
733 always enabled now. If you want to disable the support you should
734 exclude it using the list of supported ciphers. This also means that the
735 "-no_ecdhe" option has been removed from s_server.
738 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
739 SSL_{CTX_}set1_curves() which can set a list.
742 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
743 curve you want to support using SSL_{CTX_}set1_curves().
746 *) State machine rewrite. The state machine code has been significantly
747 refactored in order to remove much duplication of code and solve issues
748 with the old code (see ssl/statem/README for further details). This change
749 does have some associated API changes. Notably the SSL_state() function
750 has been removed and replaced by SSL_get_state which now returns an
751 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
752 altogether. The previous handshake states defined in ssl.h and ssl3.h have
756 *) All instances of the string "ssleay" in the public API were replaced
757 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
758 Some error codes related to internal RSA_eay API's were renamed.
761 *) The demo files in crypto/threads were moved to demo/threads.
764 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
766 [Matt Caswell, Rich Salz]
768 *) New ASN.1 embed macro.
770 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
771 structure is not allocated: it is part of the parent. That is instead of
779 This reduces memory fragmentation and make it impossible to accidentally
780 set a mandatory field to NULL.
782 This currently only works for some fields specifically a SEQUENCE, CHOICE,
783 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
784 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
788 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
791 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
792 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
793 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
794 DES and RC4 ciphersuites.
797 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
798 This changes the decoding behaviour for some invalid messages,
799 though the change is mostly in the more lenient direction, and
800 legacy behaviour is preserved as much as possible.
803 *) Fix no-stdio build.
804 [ David Woodhouse <David.Woodhouse@intel.com> and also
805 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
807 *) New testing framework
808 The testing framework has been largely rewritten and is now using
809 perl and the perl modules Test::Harness and an extended variant of
810 Test::More called OpenSSL::Test to do its work. All test scripts in
811 test/ have been rewritten into test recipes, and all direct calls to
812 executables in test/Makefile have become individual recipes using the
813 simplified testing OpenSSL::Test::Simple.
815 For documentation on our testing modules, do:
817 perldoc test/testlib/OpenSSL/Test/Simple.pm
818 perldoc test/testlib/OpenSSL/Test.pm
822 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
823 are used; the latter aborts on memory leaks (usually checked on exit).
824 Some undocumented "set malloc, etc., hooks" functions were removed
825 and others were changed. All are now documented.
828 *) In DSA_generate_parameters_ex, if the provided seed is too short,
830 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
832 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
833 from RFC4279, RFC4785, RFC5487, RFC5489.
835 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
836 original RSA_PSK patch.
839 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
840 era flag was never set throughout the codebase (only read). Also removed
841 SSL3_FLAGS_POP_BUFFER which was only used if
842 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
845 *) Changed the default name options in the "ca", "crl", "req" and "x509"
846 to be "oneline" instead of "compat".
849 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
850 not aware of clients that still exhibit this bug, and the workaround
851 hasn't been working properly for a while.
854 *) The return type of BIO_number_read() and BIO_number_written() as well as
855 the corresponding num_read and num_write members in the BIO structure has
856 changed from unsigned long to uint64_t. On platforms where an unsigned
857 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
861 *) Given the pervasive nature of TLS extensions it is inadvisable to run
862 OpenSSL without support for them. It also means that maintaining
863 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
864 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
867 *) Removed support for the two export grade static DH ciphersuites
868 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
869 were newly added (along with a number of other static DH ciphersuites) to
870 1.0.2. However the two export ones have *never* worked since they were
871 introduced. It seems strange in any case to be adding new export
872 ciphersuites, and given "logjam" it also does not seem correct to fix them.
875 *) Version negotiation has been rewritten. In particular SSLv23_method(),
876 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
877 and turned into macros which simply call the new preferred function names
878 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
879 should use the new names instead. Also as part of this change the ssl23.h
880 header file has been removed.
883 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
884 code and the associated standard is no longer considered fit-for-purpose.
887 *) RT2547 was closed. When generating a private key, try to make the
888 output file readable only by the owner. This behavior change might
889 be noticeable when interacting with other software.
891 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
895 *) Added HTTP GET support to the ocsp command.
898 *) Changed default digest for the dgst and enc commands from MD5 to
902 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
905 *) Added support for TLS extended master secret from
906 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
907 initial patch which was a great help during development.
910 *) All libssl internal structures have been removed from the public header
911 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
912 now redundant). Users should not attempt to access internal structures
913 directly. Instead they should use the provided API functions.
916 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
917 Access to deprecated functions can be re-enabled by running config with
918 "enable-deprecated". In addition applications wishing to use deprecated
919 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
920 will, by default, disable some transitive includes that previously existed
921 in the header files (e.g. ec.h will no longer, by default, include bn.h)
924 *) Added support for OCB mode. OpenSSL has been granted a patent license
925 compatible with the OpenSSL license for use of OCB. Details are available
926 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
927 for OCB can be removed by calling config with no-ocb.
930 *) SSLv2 support has been removed. It still supports receiving a SSLv2
931 compatible client hello.
934 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
935 done while fixing the error code for the key-too-small case.
936 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
938 *) CA.sh has been removmed; use CA.pl instead.
941 *) Removed old DES API.
944 *) Remove various unsupported platforms:
950 Sinix/ReliantUNIX RM400
955 16-bit platforms such as WIN16
958 *) Clean up OPENSSL_NO_xxx #define's
959 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
960 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
961 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
962 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
963 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
964 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
965 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
966 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
967 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
968 Remove MS_STATIC; it's a relic from platforms <32 bits.
971 *) Cleaned up dead code
972 Remove all but one '#ifdef undef' which is to be looked at.
975 *) Clean up calling of xxx_free routines.
976 Just like free(), fix most of the xxx_free routines to accept
977 NULL. Remove the non-null checks from callers. Save much code.
980 *) Add secure heap for storage of private keys (when possible).
981 Add BIO_s_secmem(), CBIGNUM, etc.
982 Contributed by Akamai Technologies under our Corporate CLA.
985 *) Experimental support for a new, fast, unbiased prime candidate generator,
986 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
987 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
989 *) New output format NSS in the sess_id command line tool. This allows
990 exporting the session id and the master key in NSS keylog format.
991 [Martin Kaiser <martin@kaiser.cx>]
993 *) Harmonize version and its documentation. -f flag is used to display
995 [mancha <mancha1@zoho.com>]
997 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
998 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
999 [mancha <mancha1@zoho.com>]
1001 *) Fix some double frees. These are not thought to be exploitable.
1002 [mancha <mancha1@zoho.com>]
1004 *) A missing bounds check in the handling of the TLS heartbeat extension
1005 can be used to reveal up to 64k of memory to a connected client or
1008 Thanks for Neel Mehta of Google Security for discovering this bug and to
1009 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1010 preparing the fix (CVE-2014-0160)
1011 [Adam Langley, Bodo Moeller]
1013 *) Fix for the attack described in the paper "Recovering OpenSSL
1014 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1015 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1016 http://eprint.iacr.org/2014/140
1018 Thanks to Yuval Yarom and Naomi Benger for discovering this
1019 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1020 [Yuval Yarom and Naomi Benger]
1022 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1023 this fixes a limitation in previous versions of OpenSSL.
1026 *) Experimental encrypt-then-mac support.
1028 Experimental support for encrypt then mac from
1029 draft-gutmann-tls-encrypt-then-mac-02.txt
1031 To enable it set the appropriate extension number (0x42 for the test
1032 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1034 For non-compliant peers (i.e. just about everything) this should have no
1037 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1041 *) Add EVP support for key wrapping algorithms, to avoid problems with
1042 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1043 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1044 algorithms and include tests cases.
1047 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1051 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1052 MGF1 digest and OAEP label.
1055 *) Make openssl verify return errors.
1056 [Chris Palmer <palmer@google.com> and Ben Laurie]
1058 *) New function ASN1_TIME_diff to calculate the difference between two
1059 ASN1_TIME structures or one structure and the current time.
1062 *) Update fips_test_suite to support multiple command line options. New
1063 test to induce all self test errors in sequence and check expected
1067 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1068 sign or verify all in one operation.
1071 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1072 test programs and fips_test_suite. Includes functionality to parse
1073 the minimal script output of fipsalgest.pl directly.
1076 *) Add authorisation parameter to FIPS_module_mode_set().
1079 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1082 *) Use separate DRBG fields for internal and external flags. New function
1083 FIPS_drbg_health_check() to perform on demand health checking. Add
1084 generation tests to fips_test_suite with reduced health check interval to
1085 demonstrate periodic health checking. Add "nodh" option to
1086 fips_test_suite to skip very slow DH test.
1089 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1093 *) More extensive health check for DRBG checking many more failure modes.
1094 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1095 combination: call this in fips_test_suite.
1098 *) Add support for canonical generation of DSA parameter 'g'. See
1101 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1102 POST to handle HMAC cases.
1105 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1106 to return numerical and string versions of the FIPS module number.
1109 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1110 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1111 outside the validated module in the FIPS capable OpenSSL.
1114 *) Minor change to DRBG entropy callback semantics. In some cases
1115 there is no multiple of the block length between min_len and
1116 max_len. Allow the callback to return more than max_len bytes
1117 of entropy but discard any extra: it is the callback's responsibility
1118 to ensure that the extra data discarded does not impact the
1119 requested amount of entropy.
1122 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1123 information in FIPS186-3, SP800-57 and SP800-131A.
1126 *) CCM support via EVP. Interface is very similar to GCM case except we
1127 must supply all data in one chunk (i.e. no update, final) and the
1128 message length must be supplied if AAD is used. Add algorithm test
1132 *) Initial version of POST overhaul. Add POST callback to allow the status
1133 of POST to be monitored and/or failures induced. Modify fips_test_suite
1134 to use callback. Always run all selftests even if one fails.
1137 *) XTS support including algorithm test driver in the fips_gcmtest program.
1138 Note: this does increase the maximum key length from 32 to 64 bytes but
1139 there should be no binary compatibility issues as existing applications
1140 will never use XTS mode.
1143 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1144 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1145 performs algorithm blocking for unapproved PRNG types. Also do not
1146 set PRNG type in FIPS_mode_set(): leave this to the application.
1147 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1148 the standard OpenSSL PRNG: set additional data to a date time vector.
1151 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1152 This shouldn't present any incompatibility problems because applications
1153 shouldn't be using these directly and any that are will need to rethink
1154 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1157 *) Extensive self tests and health checking required by SP800-90 DRBG.
1158 Remove strength parameter from FIPS_drbg_instantiate and always
1159 instantiate at maximum supported strength.
1162 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1165 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1168 *) New function DH_compute_key_padded() to compute a DH key and pad with
1169 leading zeroes if needed: this complies with SP800-56A et al.
1172 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1173 anything, incomplete, subject to change and largely untested at present.
1176 *) Modify fipscanisteronly build option to only build the necessary object
1177 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1180 *) Add experimental option FIPSSYMS to give all symbols in
1181 fipscanister.o and FIPS or fips prefix. This will avoid
1182 conflicts with future versions of OpenSSL. Add perl script
1183 util/fipsas.pl to preprocess assembly language source files
1184 and rename any affected symbols.
1187 *) Add selftest checks and algorithm block of non-fips algorithms in
1188 FIPS mode. Remove DES2 from selftests.
1191 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1192 return internal method without any ENGINE dependencies. Add new
1193 tiny fips sign and verify functions.
1196 *) New build option no-ec2m to disable characteristic 2 code.
1199 *) New build option "fipscanisteronly". This only builds fipscanister.o
1200 and (currently) associated fips utilities. Uses the file Makefile.fips
1201 instead of Makefile.org as the prototype.
1204 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1205 Update fips_gcmtest to use IV generator.
1208 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1209 setting output buffer to NULL. The *Final function must be
1210 called although it will not retrieve any additional data. The tag
1211 can be set or retrieved with a ctrl. The IV length is by default 12
1212 bytes (96 bits) but can be set to an alternative value. If the IV
1213 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1217 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1218 underlying do_cipher function handles all cipher semantics itself
1219 including padding and finalisation. This is useful if (for example)
1220 an ENGINE cipher handles block padding itself. The behaviour of
1221 do_cipher is subtly changed if this flag is set: the return value
1222 is the number of characters written to the output buffer (zero is
1223 no longer an error code) or a negative error code. Also if the
1224 input buffer is NULL and length 0 finalisation should be performed.
1227 *) If a candidate issuer certificate is already part of the constructed
1228 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1231 *) Improve forward-security support: add functions
1233 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1234 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1236 for use by SSL/TLS servers; the callback function will be called whenever a
1237 new session is created, and gets to decide whether the session may be
1238 cached to make it resumable (return 0) or not (return 1). (As by the
1239 SSL/TLS protocol specifications, the session_id sent by the server will be
1240 empty to indicate that the session is not resumable; also, the server will
1241 not generate RFC 4507 (RFC 5077) session tickets.)
1243 A simple reasonable callback implementation is to return is_forward_secure.
1244 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1245 by the SSL/TLS server library, indicating whether it can provide forward
1247 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1249 *) New -verify_name option in command line utilities to set verification
1253 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1254 Add CMAC pkey methods.
1257 *) Experimental renegotiation in s_server -www mode. If the client
1258 browses /reneg connection is renegotiated. If /renegcert it is
1259 renegotiated requesting a certificate.
1262 *) Add an "external" session cache for debugging purposes to s_server. This
1263 should help trace issues which normally are only apparent in deployed
1264 multi-process servers.
1267 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1268 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1269 BIO_set_cipher() and some obscure PEM functions were changed so they
1270 can now return an error. The RAND changes required a change to the
1271 RAND_METHOD structure.
1274 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1275 a gcc attribute to warn if the result of a function is ignored. This
1276 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1277 whose return value is often ignored.
1280 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1281 These allow SCTs (signed certificate timestamps) to be requested and
1282 validated when establishing a connection.
1283 [Rob Percival <robpercival@google.com>]
1285 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1287 *) Prevent padding oracle in AES-NI CBC MAC check
1289 A MITM attacker can use a padding oracle attack to decrypt traffic
1290 when the connection uses an AES CBC cipher and the server support
1293 This issue was introduced as part of the fix for Lucky 13 padding
1294 attack (CVE-2013-0169). The padding check was rewritten to be in
1295 constant time by making sure that always the same bytes are read and
1296 compared against either the MAC or padding bytes. But it no longer
1297 checked that there was enough data to have both the MAC and padding
1300 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1304 *) Fix EVP_EncodeUpdate overflow
1306 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1307 Base64 encoding of binary data. If an attacker is able to supply very large
1308 amounts of input data then a length check can overflow resulting in a heap
1311 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1312 the PEM_write_bio* family of functions. These are mainly used within the
1313 OpenSSL command line applications, so any application which processes data
1314 from an untrusted source and outputs it as a PEM file should be considered
1315 vulnerable to this issue. User applications that call these APIs directly
1316 with large amounts of untrusted data may also be vulnerable.
1318 This issue was reported by Guido Vranken.
1322 *) Fix EVP_EncryptUpdate overflow
1324 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1325 is able to supply very large amounts of input data after a previous call to
1326 EVP_EncryptUpdate() with a partial block then a length check can overflow
1327 resulting in a heap corruption. Following an analysis of all OpenSSL
1328 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1329 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1330 the first called function after an EVP_EncryptInit(), and therefore that
1331 specific call must be safe. The second form is where the length passed to
1332 EVP_EncryptUpdate() can be seen from the code to be some small value and
1333 therefore there is no possibility of an overflow. Since all instances are
1334 one of these two forms, it is believed that there can be no overflows in
1335 internal code due to this problem. It should be noted that
1336 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1337 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1338 of these calls have also been analysed too and it is believed there are no
1339 instances in internal usage where an overflow could occur.
1341 This issue was reported by Guido Vranken.
1345 *) Prevent ASN.1 BIO excessive memory allocation
1347 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1348 a short invalid encoding can cause allocation of large amounts of memory
1349 potentially consuming excessive resources or exhausting memory.
1351 Any application parsing untrusted data through d2i BIO functions is
1352 affected. The memory based functions such as d2i_X509() are *not* affected.
1353 Since the memory based functions are used by the TLS library, TLS
1354 applications are not affected.
1356 This issue was reported by Brian Carpenter.
1362 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1363 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1364 in arbitrary stack data being returned in the buffer.
1366 This issue was reported by Guido Vranken.
1370 *) Modify behavior of ALPN to invoke callback after SNI/servername
1371 callback, such that updates to the SSL_CTX affect ALPN.
1374 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1378 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1379 methods are enabled and ssl2 is disabled the methods return NULL.
1382 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1384 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1385 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1386 provide any "EXPORT" or "LOW" strength ciphers.
1389 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1390 is by default disabled at build-time. Builds that are not configured with
1391 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1392 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1393 will need to explicitly call either of:
1395 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1397 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1399 as appropriate. Even if either of those is used, or the application
1400 explicitly uses the version-specific SSLv2_method() or its client and
1401 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1402 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1403 ciphers, and SSLv2 56-bit DES are no longer available.
1407 *) Fix a double-free in DSA code
1409 A double free bug was discovered when OpenSSL parses malformed DSA private
1410 keys and could lead to a DoS attack or memory corruption for applications
1411 that receive DSA private keys from untrusted sources. This scenario is
1414 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1419 *) Disable SRP fake user seed to address a server memory leak.
1421 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1423 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1424 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1425 was changed to ignore the "fake user" SRP seed, even if the seed
1428 Users should use SRP_VBASE_get1_by_user instead. Note that in
1429 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1430 also that even though configuring the SRP seed attempts to hide
1431 invalid usernames by continuing the handshake with fake
1432 credentials, this behaviour is not constant time and no strong
1433 guarantees are made that the handshake is indistinguishable from
1434 that of a valid user.
1438 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1440 In the BN_hex2bn function the number of hex digits is calculated using an
1441 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1442 large values of |i| this can result in |bn_expand| not allocating any
1443 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1444 field as NULL leading to a subsequent NULL ptr deref. For very large values
1445 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1446 In this case memory is allocated to the internal BIGNUM data field, but it
1447 is insufficiently sized leading to heap corruption. A similar issue exists
1448 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1449 is ever called by user applications with very large untrusted hex/dec data.
1450 This is anticipated to be a rare occurrence.
1452 All OpenSSL internal usage of these functions use data that is not expected
1453 to be untrusted, e.g. config file data or application command line
1454 arguments. If user developed applications generate config file data based
1455 on untrusted data then it is possible that this could also lead to security
1456 consequences. This is also anticipated to be rare.
1458 This issue was reported to OpenSSL by Guido Vranken.
1462 *) Fix memory issues in BIO_*printf functions
1464 The internal |fmtstr| function used in processing a "%s" format string in
1465 the BIO_*printf functions could overflow while calculating the length of a
1466 string and cause an OOB read when printing very long strings.
1468 Additionally the internal |doapr_outch| function can attempt to write to an
1469 OOB memory location (at an offset from the NULL pointer) in the event of a
1470 memory allocation failure. In 1.0.2 and below this could be caused where
1471 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1472 could be in processing a very long "%s" format string. Memory leaks can
1475 The first issue may mask the second issue dependent on compiler behaviour.
1476 These problems could enable attacks where large amounts of untrusted data
1477 is passed to the BIO_*printf functions. If applications use these functions
1478 in this way then they could be vulnerable. OpenSSL itself uses these
1479 functions when printing out human-readable dumps of ASN.1 data. Therefore
1480 applications that print this data could be vulnerable if the data is from
1481 untrusted sources. OpenSSL command line applications could also be
1482 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1483 as command line arguments.
1485 Libssl is not considered directly vulnerable. Additionally certificates etc
1486 received via remote connections via libssl are also unlikely to be able to
1487 trigger these issues because of message size limits enforced within libssl.
1489 This issue was reported to OpenSSL Guido Vranken.
1493 *) Side channel attack on modular exponentiation
1495 A side-channel attack was found which makes use of cache-bank conflicts on
1496 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1497 of RSA keys. The ability to exploit this issue is limited as it relies on
1498 an attacker who has control of code in a thread running on the same
1499 hyper-threaded core as the victim thread which is performing decryptions.
1501 This issue was reported to OpenSSL by Yuval Yarom, The University of
1502 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1503 Nadia Heninger, University of Pennsylvania with more information at
1504 http://cachebleed.info.
1508 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1509 if no keysize is specified with default_bits. This fixes an
1510 omission in an earlier change that changed all RSA/DSA key generation
1511 apps to use 2048 bits by default.
1514 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1515 *) DH small subgroups
1517 Historically OpenSSL only ever generated DH parameters based on "safe"
1518 primes. More recently (in version 1.0.2) support was provided for
1519 generating X9.42 style parameter files such as those required for RFC 5114
1520 support. The primes used in such files may not be "safe". Where an
1521 application is using DH configured with parameters based on primes that are
1522 not "safe" then an attacker could use this fact to find a peer's private
1523 DH exponent. This attack requires that the attacker complete multiple
1524 handshakes in which the peer uses the same private DH exponent. For example
1525 this could be used to discover a TLS server's private DH exponent if it's
1526 reusing the private DH exponent or it's using a static DH ciphersuite.
1528 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1529 TLS. It is not on by default. If the option is not set then the server
1530 reuses the same private DH exponent for the life of the server process and
1531 would be vulnerable to this attack. It is believed that many popular
1532 applications do set this option and would therefore not be at risk.
1534 The fix for this issue adds an additional check where a "q" parameter is
1535 available (as is the case in X9.42 based parameters). This detects the
1536 only known attack, and is the only possible defense for static DH
1537 ciphersuites. This could have some performance impact.
1539 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1540 default and cannot be disabled. This could have some performance impact.
1542 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1546 *) SSLv2 doesn't block disabled ciphers
1548 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1549 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1550 been disabled, provided that the SSLv2 protocol was not also disabled via
1553 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1554 and Sebastian Schinzel.
1558 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1560 *) BN_mod_exp may produce incorrect results on x86_64
1562 There is a carry propagating bug in the x86_64 Montgomery squaring
1563 procedure. No EC algorithms are affected. Analysis suggests that attacks
1564 against RSA and DSA as a result of this defect would be very difficult to
1565 perform and are not believed likely. Attacks against DH are considered just
1566 feasible (although very difficult) because most of the work necessary to
1567 deduce information about a private key may be performed offline. The amount
1568 of resources required for such an attack would be very significant and
1569 likely only accessible to a limited number of attackers. An attacker would
1570 additionally need online access to an unpatched system using the target
1571 private key in a scenario with persistent DH parameters and a private
1572 key that is shared between multiple clients. For example this can occur by
1573 default in OpenSSL DHE based SSL/TLS ciphersuites.
1575 This issue was reported to OpenSSL by Hanno Böck.
1579 *) Certificate verify crash with missing PSS parameter
1581 The signature verification routines will crash with a NULL pointer
1582 dereference if presented with an ASN.1 signature using the RSA PSS
1583 algorithm and absent mask generation function parameter. Since these
1584 routines are used to verify certificate signature algorithms this can be
1585 used to crash any certificate verification operation and exploited in a
1586 DoS attack. Any application which performs certificate verification is
1587 vulnerable including OpenSSL clients and servers which enable client
1590 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1594 *) X509_ATTRIBUTE memory leak
1596 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1597 memory. This structure is used by the PKCS#7 and CMS routines so any
1598 application which reads PKCS#7 or CMS data from untrusted sources is
1599 affected. SSL/TLS is not affected.
1601 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1606 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1607 This changes the decoding behaviour for some invalid messages,
1608 though the change is mostly in the more lenient direction, and
1609 legacy behaviour is preserved as much as possible.
1612 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1614 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1616 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1618 *) Alternate chains certificate forgery
1620 During certificate verification, OpenSSL will attempt to find an
1621 alternative certificate chain if the first attempt to build such a chain
1622 fails. An error in the implementation of this logic can mean that an
1623 attacker could cause certain checks on untrusted certificates to be
1624 bypassed, such as the CA flag, enabling them to use a valid leaf
1625 certificate to act as a CA and "issue" an invalid certificate.
1627 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1631 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1633 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1634 incompatibility in the handling of HMAC. The previous ABI has now been
1638 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1640 *) Malformed ECParameters causes infinite loop
1642 When processing an ECParameters structure OpenSSL enters an infinite loop
1643 if the curve specified is over a specially malformed binary polynomial
1646 This can be used to perform denial of service against any
1647 system which processes public keys, certificate requests or
1648 certificates. This includes TLS clients and TLS servers with
1649 client authentication enabled.
1651 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1655 *) Exploitable out-of-bounds read in X509_cmp_time
1657 X509_cmp_time does not properly check the length of the ASN1_TIME
1658 string and can read a few bytes out of bounds. In addition,
1659 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1662 An attacker can use this to craft malformed certificates and CRLs of
1663 various sizes and potentially cause a segmentation fault, resulting in
1664 a DoS on applications that verify certificates or CRLs. TLS clients
1665 that verify CRLs are affected. TLS clients and servers with client
1666 authentication enabled may be affected if they use custom verification
1669 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1670 independently by Hanno Böck.
1674 *) PKCS7 crash with missing EnvelopedContent
1676 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1677 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1678 with missing content and trigger a NULL pointer dereference on parsing.
1680 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1681 structures from untrusted sources are affected. OpenSSL clients and
1682 servers are not affected.
1684 This issue was reported to OpenSSL by Michal Zalewski (Google).
1688 *) CMS verify infinite loop with unknown hash function
1690 When verifying a signedData message the CMS code can enter an infinite loop
1691 if presented with an unknown hash function OID. This can be used to perform
1692 denial of service against any system which verifies signedData messages using
1694 This issue was reported to OpenSSL by Johannes Bauer.
1698 *) Race condition handling NewSessionTicket
1700 If a NewSessionTicket is received by a multi-threaded client when attempting to
1701 reuse a previous ticket then a race condition can occur potentially leading to
1702 a double free of the ticket data.
1706 *) Only support 256-bit or stronger elliptic curves with the
1707 'ecdh_auto' setting (server) or by default (client). Of supported
1708 curves, prefer P-256 (both).
1711 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1713 *) ClientHello sigalgs DoS fix
1715 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1716 invalid signature algorithms extension a NULL pointer dereference will
1717 occur. This can be exploited in a DoS attack against the server.
1719 This issue was was reported to OpenSSL by David Ramos of Stanford
1722 [Stephen Henson and Matt Caswell]
1724 *) Multiblock corrupted pointer fix
1726 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1727 feature only applies on 64 bit x86 architecture platforms that support AES
1728 NI instructions. A defect in the implementation of "multiblock" can cause
1729 OpenSSL's internal write buffer to become incorrectly set to NULL when
1730 using non-blocking IO. Typically, when the user application is using a
1731 socket BIO for writing, this will only result in a failed connection.
1732 However if some other BIO is used then it is likely that a segmentation
1733 fault will be triggered, thus enabling a potential DoS attack.
1735 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1739 *) Segmentation fault in DTLSv1_listen fix
1741 The DTLSv1_listen function is intended to be stateless and processes the
1742 initial ClientHello from many peers. It is common for user code to loop
1743 over the call to DTLSv1_listen until a valid ClientHello is received with
1744 an associated cookie. A defect in the implementation of DTLSv1_listen means
1745 that state is preserved in the SSL object from one invocation to the next
1746 that can lead to a segmentation fault. Errors processing the initial
1747 ClientHello can trigger this scenario. An example of such an error could be
1748 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1751 This issue was reported to OpenSSL by Per Allansson.
1755 *) Segmentation fault in ASN1_TYPE_cmp fix
1757 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1758 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1759 certificate signature algorithm consistency this can be used to crash any
1760 certificate verification operation and exploited in a DoS attack. Any
1761 application which performs certificate verification is vulnerable including
1762 OpenSSL clients and servers which enable client authentication.
1766 *) Segmentation fault for invalid PSS parameters fix
1768 The signature verification routines will crash with a NULL pointer
1769 dereference if presented with an ASN.1 signature using the RSA PSS
1770 algorithm and invalid parameters. Since these routines are used to verify
1771 certificate signature algorithms this can be used to crash any
1772 certificate verification operation and exploited in a DoS attack. Any
1773 application which performs certificate verification is vulnerable including
1774 OpenSSL clients and servers which enable client authentication.
1776 This issue was was reported to OpenSSL by Brian Carpenter.
1780 *) ASN.1 structure reuse memory corruption fix
1782 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1783 memory corruption via an invalid write. Such reuse is and has been
1784 strongly discouraged and is believed to be rare.
1786 Applications that parse structures containing CHOICE or ANY DEFINED BY
1787 components may be affected. Certificate parsing (d2i_X509 and related
1788 functions) are however not affected. OpenSSL clients and servers are
1793 *) PKCS7 NULL pointer dereferences fix
1795 The PKCS#7 parsing code does not handle missing outer ContentInfo
1796 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1797 missing content and trigger a NULL pointer dereference on parsing.
1799 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1800 otherwise parse PKCS#7 structures from untrusted sources are
1801 affected. OpenSSL clients and servers are not affected.
1803 This issue was reported to OpenSSL by Michal Zalewski (Google).
1807 *) DoS via reachable assert in SSLv2 servers fix
1809 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1810 servers that both support SSLv2 and enable export cipher suites by sending
1811 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1813 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1814 (OpenSSL development team).
1818 *) Empty CKE with client auth and DHE fix
1820 If client auth is used then a server can seg fault in the event of a DHE
1821 ciphersuite being selected and a zero length ClientKeyExchange message
1822 being sent by the client. This could be exploited in a DoS attack.
1826 *) Handshake with unseeded PRNG fix
1828 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1829 with an unseeded PRNG. The conditions are:
1830 - The client is on a platform where the PRNG has not been seeded
1831 automatically, and the user has not seeded manually
1832 - A protocol specific client method version has been used (i.e. not
1833 SSL_client_methodv23)
1834 - A ciphersuite is used that does not require additional random data from
1835 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1837 If the handshake succeeds then the client random that has been used will
1838 have been generated from a PRNG with insufficient entropy and therefore the
1839 output may be predictable.
1841 For example using the following command with an unseeded openssl will
1842 succeed on an unpatched platform:
1844 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1848 *) Use After Free following d2i_ECPrivatekey error fix
1850 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1851 could cause a use after free condition. This, in turn, could cause a double
1852 free in several private key parsing functions (such as d2i_PrivateKey
1853 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1854 for applications that receive EC private keys from untrusted
1855 sources. This scenario is considered rare.
1857 This issue was discovered by the BoringSSL project and fixed in their
1862 *) X509_to_X509_REQ NULL pointer deref fix
1864 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1865 the certificate key is invalid. This function is rarely used in practice.
1867 This issue was discovered by Brian Carpenter.
1871 *) Removed the export ciphers from the DEFAULT ciphers
1874 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1876 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1877 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1878 So far those who have to target multiple platforms would compromise
1879 and argue that binary targeting say ARMv5 would still execute on
1880 ARMv8. "Universal" build resolves this compromise by providing
1881 near-optimal performance even on newer platforms.
1884 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1885 (other platforms pending).
1886 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1888 *) Add support for the SignedCertificateTimestampList certificate and
1889 OCSP response extensions from RFC6962.
1892 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1893 for corner cases. (Certain input points at infinity could lead to
1894 bogus results, with non-infinity inputs mapped to infinity too.)
1897 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1898 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1899 common cases are optimized and there still is room for further
1900 improvements. Vector Permutation AES for Altivec is also added.
1903 *) Add support for little-endian ppc64 Linux target.
1904 [Marcelo Cerri (IBM)]
1906 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1907 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1908 are optimized and there still is room for further improvements.
1909 Both 32- and 64-bit modes are supported.
1910 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1912 *) Improved ARMv7 NEON support.
1915 *) Support for SPARC Architecture 2011 crypto extensions, first
1916 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1917 SHA256/512, MD5, GHASH and modular exponentiation.
1918 [Andy Polyakov, David Miller]
1920 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1922 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1924 *) Support for new and upcoming Intel processors, including AVX2,
1925 BMI and SHA ISA extensions. This includes additional "stitched"
1926 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1929 This work was sponsored by Intel Corp.
1932 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1933 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1934 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1937 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1938 this fixes a limitation in previous versions of OpenSSL.
1941 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1942 MGF1 digest and OAEP label.
1945 *) Add EVP support for key wrapping algorithms, to avoid problems with
1946 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1947 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1948 algorithms and include tests cases.
1951 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1953 [Douglas E. Engert, Steve Henson]
1955 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1956 difference in days and seconds between two tm or ASN1_TIME structures.
1959 *) Add -rev test option to s_server to just reverse order of characters
1960 received by client and send back to server. Also prints an abbreviated
1961 summary of the connection parameters.
1964 *) New option -brief for s_client and s_server to print out a brief summary
1965 of connection parameters.
1968 *) Add callbacks for arbitrary TLS extensions.
1969 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1971 *) New option -crl_download in several openssl utilities to download CRLs
1972 from CRLDP extension in certificates.
1975 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1978 *) New function X509_CRL_diff to generate a delta CRL from the difference
1979 of two full CRLs. Add support to "crl" utility.
1982 *) New functions to set lookup_crls function and to retrieve
1983 X509_STORE from X509_STORE_CTX.
1986 *) Print out deprecated issuer and subject unique ID fields in
1990 *) Extend OCSP I/O functions so they can be used for simple general purpose
1991 HTTP as well as OCSP. New wrapper function which can be used to download
1992 CRLs using the OCSP API.
1995 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1998 *) SSL_CONF* functions. These provide a common framework for application
1999 configuration using configuration files or command lines.
2002 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2003 message callback and prints the results. Needs compile time option
2004 "enable-ssl-trace". New options to s_client and s_server to enable
2008 *) New ctrl and macro to retrieve supported points extensions.
2009 Print out extension in s_server and s_client.
2012 *) New functions to retrieve certificate signature and signature
2016 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2020 *) New Suite B modes for TLS code. These use and enforce the requirements
2021 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2022 only use Suite B curves. The Suite B modes can be set by using the
2023 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2026 *) New chain verification flags for Suite B levels of security. Check
2027 algorithms are acceptable when flags are set in X509_verify_cert.
2030 *) Make tls1_check_chain return a set of flags indicating checks passed
2031 by a certificate chain. Add additional tests to handle client
2032 certificates: checks for matching certificate type and issuer name
2036 *) If an attempt is made to use a signature algorithm not in the peer
2037 preference list abort the handshake. If client has no suitable
2038 signature algorithms in response to a certificate request do not
2039 use the certificate.
2042 *) If server EC tmp key is not in client preference list abort handshake.
2045 *) Add support for certificate stores in CERT structure. This makes it
2046 possible to have different stores per SSL structure or one store in
2047 the parent SSL_CTX. Include distinct stores for certificate chain
2048 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2049 to build and store a certificate chain in CERT structure: returning
2050 an error if the chain cannot be built: this will allow applications
2051 to test if a chain is correctly configured.
2053 Note: if the CERT based stores are not set then the parent SSL_CTX
2054 store is used to retain compatibility with existing behaviour.
2058 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2059 mask based on the current session, check mask when sending client
2060 hello and checking the requested ciphersuite.
2063 *) New ctrls to retrieve and set certificate types in a certificate
2064 request message. Print out received values in s_client. If certificate
2065 types is not set with custom values set sensible values based on
2066 supported signature algorithms.
2069 *) Support for distinct client and server supported signature algorithms.
2072 *) Add certificate callback. If set this is called whenever a certificate
2073 is required by client or server. An application can decide which
2074 certificate chain to present based on arbitrary criteria: for example
2075 supported signature algorithms. Add very simple example to s_server.
2076 This fixes many of the problems and restrictions of the existing client
2077 certificate callback: for example you can now clear an existing
2078 certificate and specify the whole chain.
2081 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2082 the certificate can be used for (if anything). Set valid_flags field
2083 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2084 to have similar checks in it.
2086 Add new "cert_flags" field to CERT structure and include a "strict mode".
2087 This enforces some TLS certificate requirements (such as only permitting
2088 certificate signature algorithms contained in the supported algorithms
2089 extension) which some implementations ignore: this option should be used
2090 with caution as it could cause interoperability issues.
2093 *) Update and tidy signature algorithm extension processing. Work out
2094 shared signature algorithms based on preferences and peer algorithms
2095 and print them out in s_client and s_server. Abort handshake if no
2096 shared signature algorithms.
2099 *) Add new functions to allow customised supported signature algorithms
2100 for SSL and SSL_CTX structures. Add options to s_client and s_server
2104 *) New function SSL_certs_clear() to delete all references to certificates
2105 from an SSL structure. Before this once a certificate had been added
2106 it couldn't be removed.
2109 *) Integrate hostname, email address and IP address checking with certificate
2110 verification. New verify options supporting checking in openssl utility.
2113 *) Fixes and wildcard matching support to hostname and email checking
2114 functions. Add manual page.
2115 [Florian Weimer (Red Hat Product Security Team)]
2117 *) New functions to check a hostname email or IP address against a
2118 certificate. Add options x509 utility to print results of checks against
2122 *) Fix OCSP checking.
2123 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2125 *) Initial experimental support for explicitly trusted non-root CAs.
2126 OpenSSL still tries to build a complete chain to a root but if an
2127 intermediate CA has a trust setting included that is used. The first
2128 setting is used: whether to trust (e.g., -addtrust option to the x509
2132 *) Add -trusted_first option which attempts to find certificates in the
2133 trusted store even if an untrusted chain is also supplied.
2136 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2137 platform support for Linux and Android.
2140 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2143 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2144 When in FIPS mode the approved implementations are used as normal,
2145 when not in FIPS mode the internal unapproved versions are used instead.
2146 This means that the FIPS capable OpenSSL isn't forced to use the
2147 (often lower performance) FIPS implementations outside FIPS mode.
2150 *) Transparently support X9.42 DH parameters when calling
2151 PEM_read_bio_DHparameters. This means existing applications can handle
2152 the new parameter format automatically.
2155 *) Initial experimental support for X9.42 DH parameter format: mainly
2156 to support use of 'q' parameter for RFC5114 parameters.
2159 *) Add DH parameters from RFC5114 including test data to dhtest.
2162 *) Support for automatic EC temporary key parameter selection. If enabled
2163 the most preferred EC parameters are automatically used instead of
2164 hardcoded fixed parameters. Now a server just has to call:
2165 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2166 support ECDH and use the most appropriate parameters.
2169 *) Enhance and tidy EC curve and point format TLS extension code. Use
2170 static structures instead of allocation if default values are used.
2171 New ctrls to set curves we wish to support and to retrieve shared curves.
2172 Print out shared curves in s_server. New options to s_server and s_client
2173 to set list of supported curves.
2176 *) New ctrls to retrieve supported signature algorithms and
2177 supported curve values as an array of NIDs. Extend openssl utility
2178 to print out received values.
2181 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2182 between NIDs and the more common NIST names such as "P-256". Enhance
2183 ecparam utility and ECC method to recognise the NIST names for curves.
2186 *) Enhance SSL/TLS certificate chain handling to support different
2187 chains for each certificate instead of one chain in the parent SSL_CTX.
2190 *) Support for fixed DH ciphersuite client authentication: where both
2191 server and client use DH certificates with common parameters.
2194 *) Support for fixed DH ciphersuites: those requiring DH server
2198 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2200 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2201 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2202 X509_CINF_get_signature were reverted post internal team review.
2204 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2206 *) Build fixes for the Windows and OpenVMS platforms
2207 [Matt Caswell and Richard Levitte]
2209 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2211 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2212 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2213 dereference. This could lead to a Denial Of Service attack. Thanks to
2214 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2218 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2219 dtls1_buffer_record function under certain conditions. In particular this
2220 could occur if an attacker sent repeated DTLS records with the same
2221 sequence number but for the next epoch. The memory leak could be exploited
2222 by an attacker in a Denial of Service attack through memory exhaustion.
2223 Thanks to Chris Mueller for reporting this issue.
2227 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2228 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2229 method would be set to NULL which could later result in a NULL pointer
2230 dereference. Thanks to Frank Schmirler for reporting this issue.
2234 *) Abort handshake if server key exchange message is omitted for ephemeral
2237 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2238 reporting this issue.
2242 *) Remove non-export ephemeral RSA code on client and server. This code
2243 violated the TLS standard by allowing the use of temporary RSA keys in
2244 non-export ciphersuites and could be used by a server to effectively
2245 downgrade the RSA key length used to a value smaller than the server
2246 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2247 INRIA or reporting this issue.
2251 *) Fixed issue where DH client certificates are accepted without verification.
2252 An OpenSSL server will accept a DH certificate for client authentication
2253 without the certificate verify message. This effectively allows a client to
2254 authenticate without the use of a private key. This only affects servers
2255 which trust a client certificate authority which issues certificates
2256 containing DH keys: these are extremely rare and hardly ever encountered.
2257 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2262 *) Ensure that the session ID context of an SSL is updated when its
2263 SSL_CTX is updated via SSL_set_SSL_CTX.
2265 The session ID context is typically set from the parent SSL_CTX,
2266 and can vary with the CTX.
2269 *) Fix various certificate fingerprint issues.
2271 By using non-DER or invalid encodings outside the signed portion of a
2272 certificate the fingerprint can be changed without breaking the signature.
2273 Although no details of the signed portion of the certificate can be changed
2274 this can cause problems with some applications: e.g. those using the
2275 certificate fingerprint for blacklists.
2277 1. Reject signatures with non zero unused bits.
2279 If the BIT STRING containing the signature has non zero unused bits reject
2280 the signature. All current signature algorithms require zero unused bits.
2282 2. Check certificate algorithm consistency.
2284 Check the AlgorithmIdentifier inside TBS matches the one in the
2285 certificate signature. NB: this will result in signature failure
2286 errors for some broken certificates.
2288 Thanks to Konrad Kraszewski from Google for reporting this issue.
2290 3. Check DSA/ECDSA signatures use DER.
2292 Re-encode DSA/ECDSA signatures and compare with the original received
2293 signature. Return an error if there is a mismatch.
2295 This will reject various cases including garbage after signature
2296 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2297 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2298 (negative or with leading zeroes).
2300 Further analysis was conducted and fixes were developed by Stephen Henson
2301 of the OpenSSL core team.
2306 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2307 results on some platforms, including x86_64. This bug occurs at random
2308 with a very low probability, and is not known to be exploitable in any
2309 way, though its exact impact is difficult to determine. Thanks to Pieter
2310 Wuille (Blockstream) who reported this issue and also suggested an initial
2311 fix. Further analysis was conducted by the OpenSSL development team and
2312 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2313 the OpenSSL core team.
2317 *) Do not resume sessions on the server if the negotiated protocol
2318 version does not match the session's version. Resuming with a different
2319 version, while not strictly forbidden by the RFC, is of questionable
2320 sanity and breaks all known clients.
2321 [David Benjamin, Emilia Käsper]
2323 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2324 early CCS messages during renegotiation. (Note that because
2325 renegotiation is encrypted, this early CCS was not exploitable.)
2328 *) Tighten client-side session ticket handling during renegotiation:
2329 ensure that the client only accepts a session ticket if the server sends
2330 the extension anew in the ServerHello. Previously, a TLS client would
2331 reuse the old extension state and thus accept a session ticket if one was
2332 announced in the initial ServerHello.
2334 Similarly, ensure that the client requires a session ticket if one
2335 was advertised in the ServerHello. Previously, a TLS client would
2336 ignore a missing NewSessionTicket message.
2339 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2341 *) SRTP Memory Leak.
2343 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2344 sends a carefully crafted handshake message, to cause OpenSSL to fail
2345 to free up to 64k of memory causing a memory leak. This could be
2346 exploited in a Denial Of Service attack. This issue affects OpenSSL
2347 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2348 whether SRTP is used or configured. Implementations of OpenSSL that
2349 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2351 The fix was developed by the OpenSSL team.
2355 *) Session Ticket Memory Leak.
2357 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2358 integrity of that ticket is first verified. In the event of a session
2359 ticket integrity check failing, OpenSSL will fail to free memory
2360 causing a memory leak. By sending a large number of invalid session
2361 tickets an attacker could exploit this issue in a Denial Of Service
2366 *) Build option no-ssl3 is incomplete.
2368 When OpenSSL is configured with "no-ssl3" as a build option, servers
2369 could accept and complete a SSL 3.0 handshake, and clients could be
2370 configured to send them.
2372 [Akamai and the OpenSSL team]
2374 *) Add support for TLS_FALLBACK_SCSV.
2375 Client applications doing fallback retries should call
2376 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2378 [Adam Langley, Bodo Moeller]
2380 *) Add additional DigestInfo checks.
2382 Re-encode DigestInto in DER and check against the original when
2383 verifying RSA signature: this will reject any improperly encoded
2384 DigestInfo structures.
2386 Note: this is a precautionary measure and no attacks are currently known.
2390 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2392 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2393 SRP code can be overrun an internal buffer. Add sanity check that
2394 g, A, B < N to SRP code.
2396 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2397 Group for discovering this issue.
2401 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2402 TLS 1.0 instead of higher protocol versions when the ClientHello message
2403 is badly fragmented. This allows a man-in-the-middle attacker to force a
2404 downgrade to TLS 1.0 even if both the server and the client support a
2405 higher protocol version, by modifying the client's TLS records.
2407 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2408 researching this issue.
2412 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2413 to a denial of service attack. A malicious server can crash the client
2414 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2415 ciphersuite and sending carefully crafted handshake messages.
2417 Thanks to Felix Gröbert (Google) for discovering and researching this
2422 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2423 to leak memory. This can be exploited through a Denial of Service attack.
2424 Thanks to Adam Langley for discovering and researching this issue.
2428 *) An attacker can force openssl to consume large amounts of memory whilst
2429 processing DTLS handshake messages. This can be exploited through a
2430 Denial of Service attack.
2431 Thanks to Adam Langley for discovering and researching this issue.
2435 *) An attacker can force an error condition which causes openssl to crash
2436 whilst processing DTLS packets due to memory being freed twice. This
2437 can be exploited through a Denial of Service attack.
2438 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2443 *) If a multithreaded client connects to a malicious server using a resumed
2444 session and the server sends an ec point format extension it could write
2445 up to 255 bytes to freed memory.
2447 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2452 *) A malicious server can crash an OpenSSL client with a null pointer
2453 dereference (read) by specifying an SRP ciphersuite even though it was not
2454 properly negotiated with the client. This can be exploited through a
2455 Denial of Service attack.
2457 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2458 discovering and researching this issue.
2462 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2463 X509_name_oneline, X509_name_print_ex et al. to leak some information
2464 from the stack. Applications may be affected if they echo pretty printing
2465 output to the attacker.
2467 Thanks to Ivan Fratric (Google) for discovering this issue.
2469 [Emilia Käsper, and Steve Henson]
2471 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2472 for corner cases. (Certain input points at infinity could lead to
2473 bogus results, with non-infinity inputs mapped to infinity too.)
2476 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2478 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2479 handshake can force the use of weak keying material in OpenSSL
2480 SSL/TLS clients and servers.
2482 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2483 researching this issue. (CVE-2014-0224)
2484 [KIKUCHI Masashi, Steve Henson]
2486 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2487 OpenSSL DTLS client the code can be made to recurse eventually crashing
2490 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2492 [Imre Rad, Steve Henson]
2494 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2495 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2496 client or server. This is potentially exploitable to run arbitrary
2497 code on a vulnerable client or server.
2499 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2500 [Jüri Aedla, Steve Henson]
2502 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2503 are subject to a denial of service attack.
2505 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2506 this issue. (CVE-2014-3470)
2507 [Felix Gröbert, Ivan Fratric, Steve Henson]
2509 *) Harmonize version and its documentation. -f flag is used to display
2511 [mancha <mancha1@zoho.com>]
2513 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2514 in i2d_ECPrivateKey.
2515 [mancha <mancha1@zoho.com>]
2517 *) Fix some double frees. These are not thought to be exploitable.
2518 [mancha <mancha1@zoho.com>]
2520 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2522 *) A missing bounds check in the handling of the TLS heartbeat extension
2523 can be used to reveal up to 64k of memory to a connected client or
2526 Thanks for Neel Mehta of Google Security for discovering this bug and to
2527 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2528 preparing the fix (CVE-2014-0160)
2529 [Adam Langley, Bodo Moeller]
2531 *) Fix for the attack described in the paper "Recovering OpenSSL
2532 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2533 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2534 http://eprint.iacr.org/2014/140
2536 Thanks to Yuval Yarom and Naomi Benger for discovering this
2537 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2538 [Yuval Yarom and Naomi Benger]
2540 *) TLS pad extension: draft-agl-tls-padding-03
2542 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2543 TLS client Hello record length value would otherwise be > 255 and
2544 less that 512 pad with a dummy extension containing zeroes so it
2545 is at least 512 bytes long.
2547 [Adam Langley, Steve Henson]
2549 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2551 *) Fix for TLS record tampering bug. A carefully crafted invalid
2552 handshake could crash OpenSSL with a NULL pointer exception.
2553 Thanks to Anton Johansson for reporting this issues.
2556 *) Keep original DTLS digest and encryption contexts in retransmission
2557 structures so we can use the previous session parameters if they need
2558 to be resent. (CVE-2013-6450)
2561 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2562 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2563 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2564 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2565 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2566 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2567 [Rob Stradling, Adam Langley]
2569 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2571 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2572 supporting platforms or when small records were transferred.
2573 [Andy Polyakov, Steve Henson]
2575 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2577 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2579 This addresses the flaw in CBC record processing discovered by
2580 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2581 at: http://www.isg.rhul.ac.uk/tls/
2583 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2584 Security Group at Royal Holloway, University of London
2585 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2586 Emilia Käsper for the initial patch.
2588 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2590 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2591 ciphersuites which can be exploited in a denial of service attack.
2592 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2593 and detecting this bug and to Wolfgang Ettlinger
2594 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2598 *) Return an error when checking OCSP signatures when key is NULL.
2599 This fixes a DoS attack. (CVE-2013-0166)
2602 *) Make openssl verify return errors.
2603 [Chris Palmer <palmer@google.com> and Ben Laurie]
2605 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2606 the right response is stapled. Also change SSL_get_certificate()
2607 so it returns the certificate actually sent.
2608 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2609 [Rob Stradling <rob.stradling@comodo.com>]
2611 *) Fix possible deadlock when decoding public keys.
2614 *) Don't use TLS 1.0 record version number in initial client hello
2618 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2620 *) Sanity check record length before skipping explicit IV in TLS
2621 1.2, 1.1 and DTLS to fix DoS attack.
2623 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2624 fuzzing as a service testing platform.
2628 *) Initialise tkeylen properly when encrypting CMS messages.
2629 Thanks to Solar Designer of Openwall for reporting this issue.
2632 *) In FIPS mode don't try to use composite ciphers as they are not
2636 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2638 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2639 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2640 mean any application compiled against OpenSSL 1.0.0 headers setting
2641 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2642 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2643 0x10000000L Any application which was previously compiled against
2644 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2645 will need to be recompiled as a result. Letting be results in
2646 inability to disable specifically TLS 1.1 and in client context,
2647 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2650 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2651 disable just protocol X, but all protocols above X *if* there are
2652 protocols *below* X still enabled. In more practical terms it means
2653 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2654 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2655 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2659 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2661 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2662 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2663 in CRYPTO_realloc_clean.
2665 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2666 issue and to Adam Langley <agl@chromium.org> for fixing it.
2668 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2670 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2673 *) Workarounds for some broken servers that "hang" if a client hello
2674 record length exceeds 255 bytes.
2676 1. Do not use record version number > TLS 1.0 in initial client
2677 hello: some (but not all) hanging servers will now work.
2678 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2679 the number of ciphers sent in the client hello. This should be
2680 set to an even number, such as 50, for example by passing:
2681 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2682 Most broken servers should now work.
2683 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2684 TLS 1.2 client support entirely.
2687 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2690 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2692 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2693 STRING form instead of a DigestInfo.
2696 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2697 and the RSA_sign/RSA_verify functions. This was made more apparent when
2698 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2699 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2700 the correct format in RSA_verify so both forms transparently work.
2703 *) Some servers which support TLS 1.0 can choke if we initially indicate
2704 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2705 encrypted premaster secret. As a workaround use the maximum permitted
2706 client version in client hello, this should keep such servers happy
2707 and still work with previous versions of OpenSSL.
2710 *) Add support for TLS/DTLS heartbeats.
2711 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2713 *) Add support for SCTP.
2714 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2716 *) Improved PRNG seeding for VOS.
2717 [Paul Green <Paul.Green@stratus.com>]
2719 *) Extensive assembler packs updates, most notably:
2721 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2722 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2723 - x86_64: bit-sliced AES implementation;
2724 - ARM: NEON support, contemporary platforms optimizations;
2725 - s390x: z196 support;
2726 - *: GHASH and GF(2^m) multiplication implementations;
2730 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2731 (removal of unnecessary code)
2732 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2734 *) Add TLS key material exporter from RFC 5705.
2737 *) Add DTLS-SRTP negotiation from RFC 5764.
2740 *) Add Next Protocol Negotiation,
2741 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2742 disabled with a no-npn flag to config or Configure. Code donated
2744 [Adam Langley <agl@google.com> and Ben Laurie]
2746 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2747 NIST-P256, NIST-P521, with constant-time single point multiplication on
2748 typical inputs. Compiler support for the nonstandard type __uint128_t is
2749 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2750 Code made available under Apache License version 2.0.
2752 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2753 line to include this in your build of OpenSSL, and run "make depend" (or
2754 "make update"). This enables the following EC_METHODs:
2756 EC_GFp_nistp224_method()
2757 EC_GFp_nistp256_method()
2758 EC_GFp_nistp521_method()
2760 EC_GROUP_new_by_curve_name() will automatically use these (while
2761 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2763 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2765 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2766 all platforms. Move ssize_t definition from e_os.h to the public
2767 header file e_os2.h as it now appears in public header file cms.h
2770 *) New -sigopt option to the ca, req and x509 utilities. Additional
2771 signature parameters can be passed using this option and in
2775 *) Add RSA PSS signing function. This will generate and set the
2776 appropriate AlgorithmIdentifiers for PSS based on those in the
2777 corresponding EVP_MD_CTX structure. No application support yet.
2780 *) Support for companion algorithm specific ASN1 signing routines.
2781 New function ASN1_item_sign_ctx() signs a pre-initialised
2782 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2783 the appropriate parameters.
2786 *) Add new algorithm specific ASN1 verification initialisation function
2787 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2788 handling will be the same no matter what EVP_PKEY_METHOD is used.
2789 Add a PSS handler to support verification of PSS signatures: checked
2790 against a number of sample certificates.
2793 *) Add signature printing for PSS. Add PSS OIDs.
2794 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2796 *) Add algorithm specific signature printing. An individual ASN1 method
2797 can now print out signatures instead of the standard hex dump.
2799 More complex signatures (e.g. PSS) can print out more meaningful
2800 information. Include DSA version that prints out the signature
2804 *) Password based recipient info support for CMS library: implementing
2808 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2809 neatly separates the code into cipher and PBE sections and is required
2810 for some algorithms that split PBES2 into separate pieces (such as
2811 password based CMS).
2814 *) Session-handling fixes:
2815 - Fix handling of connections that are resuming with a session ID,
2816 but also support Session Tickets.
2817 - Fix a bug that suppressed issuing of a new ticket if the client
2818 presented a ticket with an expired session.
2819 - Try to set the ticket lifetime hint to something reasonable.
2820 - Make tickets shorter by excluding irrelevant information.
2821 - On the client side, don't ignore renewed tickets.
2822 [Adam Langley, Bodo Moeller (Google)]
2824 *) Fix PSK session representation.
2827 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2829 This work was sponsored by Intel.
2832 *) Add GCM support to TLS library. Some custom code is needed to split
2833 the IV between the fixed (from PRF) and explicit (from TLS record)
2834 portions. This adds all GCM ciphersuites supported by RFC5288 and
2835 RFC5289. Generalise some AES* cipherstrings to include GCM and
2836 add a special AESGCM string for GCM only.
2839 *) Expand range of ctrls for AES GCM. Permit setting invocation
2840 field on decrypt and retrieval of invocation field only on encrypt.
2843 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2844 As required by RFC5289 these ciphersuites cannot be used if for
2845 versions of TLS earlier than 1.2.
2848 *) For FIPS capable OpenSSL interpret a NULL default public key method
2849 as unset and return the appropriate default but do *not* set the default.
2850 This means we can return the appropriate method in applications that
2851 switch between FIPS and non-FIPS modes.
2854 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2855 ENGINE is used then we cannot handle that in the FIPS module so we
2856 keep original code iff non-FIPS operations are allowed.
2859 *) Add -attime option to openssl utilities.
2860 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2862 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2865 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2866 FIPS EC methods unconditionally for now.
2869 *) New build option no-ec2m to disable characteristic 2 code.
2872 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2873 all cases can be covered as some introduce binary incompatibilities.
2876 *) Redirect RSA operations to FIPS module including keygen,
2877 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2880 *) Add similar low level API blocking to ciphers.
2883 *) Low level digest APIs are not approved in FIPS mode: any attempt
2884 to use these will cause a fatal error. Applications that *really* want
2885 to use them can use the private_* version instead.
2888 *) Redirect cipher operations to FIPS module for FIPS builds.
2891 *) Redirect digest operations to FIPS module for FIPS builds.
2894 *) Update build system to add "fips" flag which will link in fipscanister.o
2895 for static and shared library builds embedding a signature if needed.
2898 *) Output TLS supported curves in preference order instead of numerical
2899 order. This is currently hardcoded for the highest order curves first.
2900 This should be configurable so applications can judge speed vs strength.
2903 *) Add TLS v1.2 server support for client authentication.
2906 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2910 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2911 FIPS modules versions.
2914 *) Add TLS v1.2 client side support for client authentication. Keep cache
2915 of handshake records longer as we don't know the hash algorithm to use
2916 until after the certificate request message is received.
2919 *) Initial TLS v1.2 client support. Add a default signature algorithms
2920 extension including all the algorithms we support. Parse new signature
2921 format in client key exchange. Relax some ECC signing restrictions for
2922 TLS v1.2 as indicated in RFC5246.
2925 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2926 to new signature format when needed using client digest preference.
2927 All server ciphersuites should now work correctly in TLS v1.2. No client
2928 support yet and no support for client certificates.
2931 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2932 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2933 ciphersuites. At present only RSA key exchange ciphersuites work with
2934 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2935 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2936 and version checking.
2939 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2940 with this defined it will not be affected by any changes to ssl internal
2941 structures. Add several utility functions to allow openssl application
2942 to work with OPENSSL_NO_SSL_INTERN defined.
2946 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2948 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2951 *) Permit abbreviated handshakes when renegotiating using the function
2952 SSL_renegotiate_abbreviated().
2953 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2955 *) Add call to ENGINE_register_all_complete() to
2956 ENGINE_load_builtin_engines(), so some implementations get used
2957 automatically instead of needing explicit application support.
2960 *) Add support for TLS key exporter as described in RFC5705.
2961 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2963 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2964 a few changes are required:
2966 Add SSL_OP_NO_TLSv1_1 flag.
2967 Add TLSv1_1 methods.
2968 Update version checking logic to handle version 1.1.
2969 Add explicit IV handling (ported from DTLS code).
2970 Add command line options to s_client/s_server.
2973 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2975 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2976 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2977 content decryption and always return the same error. Note: this attack
2978 needs on average 2^20 messages so it only affects automated senders. The
2979 old behaviour can be re-enabled in the CMS code by setting the
2980 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2981 an MMA defence is not necessary.
2982 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2983 this issue. (CVE-2012-0884)
2986 *) Fix CVE-2011-4619: make sure we really are receiving a
2987 client hello before rejecting multiple SGC restarts. Thanks to
2988 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2991 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2993 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2994 Thanks to Antonio Martin, Enterprise Secure Access Research and
2995 Development, Cisco Systems, Inc. for discovering this bug and
2996 preparing a fix. (CVE-2012-0050)
2999 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3001 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3002 of the Vaudenay padding oracle attack on CBC mode encryption
3003 which enables an efficient plaintext recovery attack against
3004 the OpenSSL implementation of DTLS. Their attack exploits timing
3005 differences arising during decryption processing. A research
3006 paper describing this attack can be found at:
3007 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3008 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3009 Security Group at Royal Holloway, University of London
3010 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3011 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3012 for preparing the fix. (CVE-2011-4108)
3013 [Robin Seggelmann, Michael Tuexen]
3015 *) Clear bytes used for block padding of SSL 3.0 records.
3017 [Adam Langley (Google)]
3019 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3020 Kadianakis <desnacked@gmail.com> for discovering this issue and
3021 Adam Langley for preparing the fix. (CVE-2011-4619)
3022 [Adam Langley (Google)]
3024 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3025 [Andrey Kulikov <amdeich@gmail.com>]
3027 *) Prevent malformed RFC3779 data triggering an assertion failure.
3028 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3029 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3030 [Rob Austein <sra@hactrn.net>]
3032 *) Improved PRNG seeding for VOS.
3033 [Paul Green <Paul.Green@stratus.com>]
3035 *) Fix ssl_ciph.c set-up race.
3036 [Adam Langley (Google)]
3038 *) Fix spurious failures in ecdsatest.c.
3039 [Emilia Käsper (Google)]
3041 *) Fix the BIO_f_buffer() implementation (which was mixing different
3042 interpretations of the '..._len' fields).
3043 [Adam Langley (Google)]
3045 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3046 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3047 threads won't reuse the same blinding coefficients.
3049 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3050 lock to call BN_BLINDING_invert_ex, and avoids one use of
3051 BN_BLINDING_update for each BN_BLINDING structure (previously,
3052 the last update always remained unused).
3053 [Emilia Käsper (Google)]
3055 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3056 [Bob Buckholz (Google)]
3058 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3060 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3061 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3062 [Kaspar Brand <ossl@velox.ch>]
3064 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3065 for multi-threaded use of ECDH. (CVE-2011-3210)
3066 [Adam Langley (Google)]
3068 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3071 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3072 signature public key algorithm by using OID xref utilities instead.
3073 Before this you could only use some ECC ciphersuites with SHA1 only.
3076 *) Add protection against ECDSA timing attacks as mentioned in the paper
3077 by Billy Bob Brumley and Nicola Tuveri, see:
3079 http://eprint.iacr.org/2011/232.pdf
3081 [Billy Bob Brumley and Nicola Tuveri]
3083 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3085 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3086 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3088 *) Fix bug in string printing code: if *any* escaping is enabled we must
3089 escape the escape character (backslash) or the resulting string is
3093 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3095 *) Disable code workaround for ancient and obsolete Netscape browsers
3096 and servers: an attacker can use it in a ciphersuite downgrade attack.
3097 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3100 *) Fixed J-PAKE implementation error, originally discovered by
3101 Sebastien Martini, further info and confirmation from Stefan
3102 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3105 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3107 *) Fix extension code to avoid race conditions which can result in a buffer
3108 overrun vulnerability: resumed sessions must not be modified as they can
3109 be shared by multiple threads. CVE-2010-3864
3112 *) Fix WIN32 build system to correctly link an ENGINE directory into
3116 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3118 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3120 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3122 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3124 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3125 context. The operation can be customised via the ctrl mechanism in
3126 case ENGINEs want to include additional functionality.
3129 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3132 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3133 output hashes compatible with older versions of OpenSSL.
3134 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3136 *) Fix compression algorithm handling: if resuming a session use the
3137 compression algorithm of the resumed session instead of determining
3138 it from client hello again. Don't allow server to change algorithm.
3141 *) Add load_crls() function to apps tidying load_certs() too. Add option
3142 to verify utility to allow additional CRLs to be included.
3145 *) Update OCSP request code to permit adding custom headers to the request:
3146 some responders need this.
3149 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3151 [Julia Lawall <julia@diku.dk>]
3153 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3154 needlessly dereferenced structures, used obsolete functions and
3155 didn't handle all updated verify codes correctly.
3158 *) Disable MD2 in the default configuration.
3161 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3162 indicate the initial BIO being pushed or popped. This makes it possible
3163 to determine whether the BIO is the one explicitly called or as a result
3164 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3165 it handles reference counts correctly and doesn't zero out the I/O bio
3166 when it is not being explicitly popped. WARNING: applications which
3167 included workarounds for the old buggy behaviour will need to be modified
3168 or they could free up already freed BIOs.
3171 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3172 renaming to all platforms (within the 0.9.8 branch, this was
3173 done conditionally on Netware platforms to avoid a name clash).
3174 [Guenter <lists@gknw.net>]
3176 *) Add ECDHE and PSK support to DTLS.
3177 [Michael Tuexen <tuexen@fh-muenster.de>]
3179 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3183 *) Add "missing" function EVP_MD_flags() (without this the only way to
3184 retrieve a digest flags is by accessing the structure directly. Update
3185 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3186 or cipher is registered as in the "from" argument. Print out all
3187 registered digests in the dgst usage message instead of manually
3188 attempting to work them out.
3191 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3192 this allows the use of compression and extensions. Change default cipher
3193 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3194 by default unless an application cipher string requests it.
3197 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3198 key ids to find matching certificates and keys but some PKCS#12 files
3199 don't follow the (somewhat unwritten) rules and this strategy fails.
3200 Now just gather all certificates together and the first private key
3201 then look for the first certificate that matches the key.
3204 *) Support use of registered digest and cipher names for dgst and cipher
3205 commands instead of having to add each one as a special case. So now
3212 openssl dgst -sha256 foo
3214 and this works for ENGINE based algorithms too.
3218 *) Update Gost ENGINE to support parameter files.
3219 [Victor B. Wagner <vitus@cryptocom.ru>]
3221 *) Support GeneralizedTime in ca utility.
3222 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3224 *) Enhance the hash format used for certificate directory links. The new
3225 form uses the canonical encoding (meaning equivalent names will work
3226 even if they aren't identical) and uses SHA1 instead of MD5. This form
3227 is incompatible with the older format and as a result c_rehash should
3228 be used to rebuild symbolic links.
3231 *) Make PKCS#8 the default write format for private keys, replacing the
3232 traditional format. This form is standardised, more secure and doesn't
3233 include an implicit MD5 dependency.
3236 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3237 committed to OpenSSL should pass this lot as a minimum.
3240 *) Add session ticket override functionality for use by EAP-FAST.
3241 [Jouni Malinen <j@w1.fi>]
3243 *) Modify HMAC functions to return a value. Since these can be implemented
3244 in an ENGINE errors can occur.
3247 *) Type-checked OBJ_bsearch_ex.
3250 *) Type-checked OBJ_bsearch. Also some constification necessitated
3251 by type-checking. Still to come: TXT_DB, bsearch(?),
3252 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3256 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3257 seconds to a tm structure directly, instead of going through OS
3258 specific date routines. This avoids any issues with OS routines such
3259 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3260 and X509_time_adj_ex() to cover the extended range. The existing
3261 X509_time_adj() is still usable and will no longer have any date issues.
3264 *) Delta CRL support. New use deltas option which will attempt to locate
3265 and search any appropriate delta CRLs available.
3267 This work was sponsored by Google.
3270 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3271 code and add additional score elements. Validate alternate CRL paths
3272 as part of the CRL checking and indicate a new error "CRL path validation
3273 error" in this case. Applications wanting additional details can use
3274 the verify callback and check the new "parent" field. If this is not
3275 NULL CRL path validation is taking place. Existing applications won't
3276 see this because it requires extended CRL support which is off by
3279 This work was sponsored by Google.
3282 *) Support for freshest CRL extension.
3284 This work was sponsored by Google.
3287 *) Initial indirect CRL support. Currently only supported in the CRLs
3288 passed directly and not via lookup. Process certificate issuer
3289 CRL entry extension and lookup CRL entries by bother issuer name
3290 and serial number. Check and process CRL issuer entry in IDP extension.
3292 This work was sponsored by Google.
3295 *) Add support for distinct certificate and CRL paths. The CRL issuer
3296 certificate is validated separately in this case. Only enabled if
3297 an extended CRL support flag is set: this flag will enable additional
3298 CRL functionality in future.
3300 This work was sponsored by Google.
3303 *) Add support for policy mappings extension.
3305 This work was sponsored by Google.
3308 *) Fixes to pathlength constraint, self issued certificate handling,
3309 policy processing to align with RFC3280 and PKITS tests.
3311 This work was sponsored by Google.
3314 *) Support for name constraints certificate extension. DN, email, DNS
3315 and URI types are currently supported.
3317 This work was sponsored by Google.
3320 *) To cater for systems that provide a pointer-based thread ID rather
3321 than numeric, deprecate the current numeric thread ID mechanism and
3322 replace it with a structure and associated callback type. This
3323 mechanism allows a numeric "hash" to be extracted from a thread ID in
3324 either case, and on platforms where pointers are larger than 'long',
3325 mixing is done to help ensure the numeric 'hash' is usable even if it
3326 can't be guaranteed unique. The default mechanism is to use "&errno"
3327 as a pointer-based thread ID to distinguish between threads.
3329 Applications that want to provide their own thread IDs should now use
3330 CRYPTO_THREADID_set_callback() to register a callback that will call
3331 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3333 Note that ERR_remove_state() is now deprecated, because it is tied
3334 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3335 to free the current thread's error state should be replaced by
3336 ERR_remove_thread_state(NULL).
3338 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3339 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3340 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3341 application was previously providing a numeric thread callback that
3342 was inappropriate for distinguishing threads, then uniqueness might
3343 have been obtained with &errno that happened immediately in the
3344 intermediate development versions of OpenSSL; this is no longer the
3345 case, the numeric thread callback will now override the automatic use
3347 [Geoff Thorpe, with help from Bodo Moeller]
3349 *) Initial support for different CRL issuing certificates. This covers a
3350 simple case where the self issued certificates in the chain exist and
3351 the real CRL issuer is higher in the existing chain.
3353 This work was sponsored by Google.
3356 *) Removed effectively defunct crypto/store from the build.
3359 *) Revamp of STACK to provide stronger type-checking. Still to come:
3360 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3361 ASN1_STRING, CONF_VALUE.
3364 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3365 RAM on SSL connections. This option can save about 34k per idle SSL.
3368 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3369 STACK, TXT_DB, bsearch, qsort.
3372 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3373 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3374 support for data, signedData, compressedData, digestedData and
3375 encryptedData, envelopedData types included. Scripts to check against
3376 RFC4134 examples draft and interop and consistency checks of many
3377 content types and variants.
3380 *) Add options to enc utility to support use of zlib compression BIO.
3383 *) Extend mk1mf to support importing of options and assembly language
3384 files from Configure script, currently only included in VC-WIN32.
3385 The assembly language rules can now optionally generate the source
3386 files from the associated perl scripts.
3389 *) Implement remaining functionality needed to support GOST ciphersuites.
3390 Interop testing has been performed using CryptoPro implementations.
3391 [Victor B. Wagner <vitus@cryptocom.ru>]
3393 *) s390x assembler pack.
3396 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3400 *) Implement Opaque PRF Input TLS extension as specified in
3401 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3402 official specification yet and no extension type assignment by
3403 IANA exists, this extension (for now) will have to be explicitly
3404 enabled when building OpenSSL by providing the extension number
3405 to use. For example, specify an option
3407 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3409 to the "config" or "Configure" script to enable the extension,
3410 assuming extension number 0x9527 (which is a completely arbitrary
3411 and unofficial assignment based on the MD5 hash of the Internet
3412 Draft). Note that by doing so, you potentially lose
3413 interoperability with other TLS implementations since these might
3414 be using the same extension number for other purposes.
3416 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3417 opaque PRF input value to use in the handshake. This will create
3418 an interal copy of the length-'len' string at 'src', and will
3419 return non-zero for success.
3421 To get more control and flexibility, provide a callback function
3424 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3425 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3429 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3432 Callback function 'cb' will be called in handshakes, and is
3433 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3434 Argument 'arg' is for application purposes (the value as given to
3435 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3436 be provided to the callback function). The callback function
3437 has to return non-zero to report success: usually 1 to use opaque
3438 PRF input just if possible, or 2 to enforce use of the opaque PRF
3439 input. In the latter case, the library will abort the handshake
3440 if opaque PRF input is not successfully negotiated.
3442 Arguments 'peerinput' and 'len' given to the callback function
3443 will always be NULL and 0 in the case of a client. A server will
3444 see the client's opaque PRF input through these variables if
3445 available (NULL and 0 otherwise). Note that if the server
3446 provides an opaque PRF input, the length must be the same as the
3447 length of the client's opaque PRF input.
3449 Note that the callback function will only be called when creating
3450 a new session (session resumption can resume whatever was
3451 previously negotiated), and will not be called in SSL 2.0
3452 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3453 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3454 for applications that need to enforce opaque PRF input.
3458 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3461 [Victor B. Wagner <vitus@cryptocom.ru>]
3463 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3464 RFC4507bis. The encrypted ticket format is an encrypted encoded
3465 SSL_SESSION structure, that way new session features are automatically
3468 If a client application caches session in an SSL_SESSION structure
3469 support is transparent because tickets are now stored in the encoded
3472 The SSL_CTX structure automatically generates keys for ticket
3473 protection in servers so again support should be possible
3474 with no application modification.
3476 If a client or server wishes to disable RFC4507 support then the option
3477 SSL_OP_NO_TICKET can be set.
3479 Add a TLS extension debugging callback to allow the contents of any client
3480 or server extensions to be examined.
3482 This work was sponsored by Google.
3485 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3486 OpenSSL should now compile cleanly on gcc 4.2
3487 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3489 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3490 support including streaming MAC support: this is required for GOST
3491 ciphersuite support.
3492 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3494 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3495 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3496 to output in BER and PEM format.
3499 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3500 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3501 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3502 ENGINE support for HMAC keys which are unextractable. New -mac and
3503 -macopt options to dgst utility.
3506 *) New option -sigopt to dgst utility. Update dgst to use
3507 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3508 alternative signing parameters such as X9.31 or PSS in the dgst
3512 *) Change ssl_cipher_apply_rule(), the internal function that does
3513 the work each time a ciphersuite string requests enabling
3514 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3515 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3516 the order of disabled ciphersuites such that those ciphersuites
3517 that most recently went from enabled to disabled not only stay
3518 in order with respect to each other, but also have higher priority
3519 than other disabled ciphersuites the next time ciphersuites are
3522 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3523 the same ciphersuites as with "HIGH" alone, but in a specific
3524 order where the PSK ciphersuites come first (since they are the
3525 most recently disabled ciphersuites when "HIGH" is parsed).
3527 Also, change ssl_create_cipher_list() (using this new
3528 funcionality) such that between otherwise identical
3529 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3533 *) Change ssl_create_cipher_list() so that it automatically
3534 arranges the ciphersuites in reasonable order before starting
3535 to process the rule string. Thus, the definition for "DEFAULT"
3536 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3537 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3538 This makes it much easier to arrive at a reasonable default order
3539 in applications for which anonymous ciphers are OK (meaning
3540 that you can't actually use DEFAULT).
3541 [Bodo Moeller; suggested by Victor Duchovni]
3543 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3544 processing) into multiple integers instead of setting
3545 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3546 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3547 (These masks as well as the individual bit definitions are hidden
3548 away into the non-exported interface ssl/ssl_locl.h, so this
3549 change to the definition of the SSL_CIPHER structure shouldn't
3550 affect applications.) This give us more bits for each of these
3551 categories, so there is no longer a need to coagulate AES128 and
3552 AES256 into a single algorithm bit, and to coagulate Camellia128
3553 and Camellia256 into a single algorithm bit, which has led to all
3556 Thus, among other things, the kludge introduced in 0.9.7m and
3557 0.9.8e for masking out AES256 independently of AES128 or masking
3558 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3560 With the change, we also introduce new ciphersuite aliases that
3561 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3565 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3566 Use the leftmost N bytes of the signature input if the input is
3567 larger than the prime q (with N being the size in bytes of q).
3570 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3571 it yet and it is largely untested.
3574 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3577 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3578 some compilers (gcc 4.2 and later) reject their use. Safestack is
3579 reimplemented. Update ASN1 to avoid use of legacy functions.
3582 *) Win32/64 targets are linked with Winsock2.
3585 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3586 to external functions. This can be used to increase CRL handling
3587 efficiency especially when CRLs are very large by (for example) storing
3588 the CRL revoked certificates in a database.
3591 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3592 new CRLs added to a directory can be used. New command line option
3593 -verify_return_error to s_client and s_server. This causes real errors
3594 to be returned by the verify callback instead of carrying on no matter
3595 what. This reflects the way a "real world" verify callback would behave.
3598 *) GOST engine, supporting several GOST algorithms and public key formats.
3599 Kindly donated by Cryptocom.
3602 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3603 partitioned by DP are handled but no indirect CRL or reason partitioning
3604 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3605 selected via a scoring technique which handles IDP and AKID in CRLs.
3608 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3609 will ultimately be used for all verify operations: this will remove the
3610 X509_STORE dependency on certificate verification and allow alternative
3611 lookup methods. X509_STORE based implementations of these two callbacks.
3614 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3615 Modify get_crl() to find a valid (unexpired) CRL if possible.
3618 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3619 this would be called X509_CRL_cmp() but that name is already used by
3620 a function that just compares CRL issuer names. Cache several CRL
3621 extensions in X509_CRL structure and cache CRLDP in X509.
3624 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3625 this maps equivalent X509_NAME structures into a consistent structure.
3626 Name comparison can then be performed rapidly using memcmp().
3629 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3633 *) Allow digests to supply their own micalg string for S/MIME type using
3634 the ctrl EVP_MD_CTRL_MICALG.
3637 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3638 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3639 ctrl. It can then customise the structure before and/or after signing
3643 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3644 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3645 to free up any added signature OIDs.
3648 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3649 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3650 digest and cipher tables. New options added to openssl utility:
3651 list-message-digest-algorithms and list-cipher-algorithms.
3654 *) Change the array representation of binary polynomials: the list
3655 of degrees of non-zero coefficients is now terminated with -1.
3656 Previously it was terminated with 0, which was also part of the
3657 value; thus, the array representation was not applicable to
3658 polynomials where t^0 has coefficient zero. This change makes
3659 the array representation useful in a more general context.
3662 *) Various modifications and fixes to SSL/TLS cipher string
3663 handling. For ECC, the code now distinguishes between fixed ECDH
3664 with RSA certificates on the one hand and with ECDSA certificates
3665 on the other hand, since these are separate ciphersuites. The
3666 unused code for Fortezza ciphersuites has been removed.
3668 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3669 (not "ECDHE"). For consistency with the code for DH
3670 certificates, use of ECDH certificates is now considered ECDH
3671 authentication, not RSA or ECDSA authentication (the latter is
3672 merely the CA's signing algorithm and not actively used in the
3675 The temporary ciphersuite alias "ECCdraft" is no longer
3676 available, and ECC ciphersuites are no longer excluded from "ALL"
3677 and "DEFAULT". The following aliases now exist for RFC 4492
3678 ciphersuites, most of these by analogy with the DH case:
3680 kECDHr - ECDH cert, signed with RSA
3681 kECDHe - ECDH cert, signed with ECDSA
3682 kECDH - ECDH cert (signed with either RSA or ECDSA)
3683 kEECDH - ephemeral ECDH
3684 ECDH - ECDH cert or ephemeral ECDH
3690 AECDH - anonymous ECDH
3691 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3695 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3696 Use correct micalg parameters depending on digest(s) in signed message.
3699 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3700 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3703 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3704 an engine to register a method. Add ENGINE lookups for methods and
3705 functional reference processing.
3708 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3709 EVP_{Sign,Verify}* which allow an application to customise the signature
3713 *) New -resign option to smime utility. This adds one or more signers
3714 to an existing PKCS#7 signedData structure. Also -md option to use an
3715 alternative message digest algorithm for signing.
3718 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3719 create PKCS7 structures containing multiple signers. Update smime
3720 application to support multiple signers.
3723 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3727 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3728 Reorganize PBE internals to lookup from a static table using NIDs,
3729 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3730 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3731 PRF which will be automatically used with PBES2.
3734 *) Replace the algorithm specific calls to generate keys in "req" with the
3738 *) Update PKCS#7 enveloped data routines to use new API. This is now
3739 supported by any public key method supporting the encrypt operation. A
3740 ctrl is added to allow the public key algorithm to examine or modify
3741 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3745 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3746 a default digest type to use. In most cases this will be SHA1 but some
3747 algorithms (such as GOST) need to specify an alternative digest. The
3748 return value indicates how strong the preference is 1 means optional and
3749 2 is mandatory (that is it is the only supported type). Modify
3750 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3751 use the default md. Update openssl utilities to use the default digest
3752 type for signing if it is not explicitly indicated.
3755 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3756 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3757 signing method from the key type. This effectively removes the link
3758 between digests and public key types.
3761 *) Add an OID cross reference table and utility functions. Its purpose is to
3762 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3763 rsaEncryption. This will allow some of the algorithm specific hackery
3764 needed to use the correct OID to be removed.
3767 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3768 structures for PKCS7_sign(). They are now set up by the relevant public
3772 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3775 *) Add support for key derivation (agreement) in the API, DH method and
3779 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3780 public and private key formats. As a side effect these add additional
3781 command line functionality not previously available: DSA signatures can be
3782 generated and verified using pkeyutl and DH key support and generation in
3787 [Oliver Tappe <zooey@hirschkaefer.de>]
3789 *) New make target "install_html_docs" installs HTML renditions of the
3791 [Oliver Tappe <zooey@hirschkaefer.de>]
3793 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3794 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3795 support key and parameter generation and add initial key generation
3796 functionality for RSA.
3799 *) Add functions for main EVP_PKEY_method operations. The undocumented
3800 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3801 EVP_PKEY_{encrypt,decrypt}_old.
3804 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3805 key API, doesn't do much yet.
3808 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3809 public key algorithms. New option to openssl utility:
3810 "list-public-key-algorithms" to print out info.
3813 *) Implement the Supported Elliptic Curves Extension for
3814 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3817 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3818 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3821 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3822 utilities such as rsa, dsa, dsaparam etc except they process any key
3826 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3827 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3828 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3832 *) Initial support for pluggable public key ASN1.
3833 De-spaghettify the public key ASN1 handling. Move public and private
3834 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3835 algorithm specific handling to a single module within the relevant
3836 algorithm directory. Add functions to allow (near) opaque processing
3837 of public and private key structures.
3840 *) Implement the Supported Point Formats Extension for
3841 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3844 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3845 for the psk identity [hint] and the psk callback functions to the
3846 SSL_SESSION, SSL and SSL_CTX structure.
3849 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3853 SSL_CTX_use_psk_identity_hint
3854 SSL_get_psk_identity_hint
3855 SSL_get_psk_identity
3856 SSL_use_psk_identity_hint
3858 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3860 *) Add RFC 3161 compliant time stamp request creation, response generation
3861 and response verification functionality.
3862 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3864 *) Add initial support for TLS extensions, specifically for the server_name
3865 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3866 have new members for a host name. The SSL data structure has an
3867 additional member SSL_CTX *initial_ctx so that new sessions can be
3868 stored in that context to allow for session resumption, even after the
3869 SSL has been switched to a new SSL_CTX in reaction to a client's
3870 server_name extension.
3872 New functions (subject to change):
3874 SSL_get_servername()
3875 SSL_get_servername_type()
3878 New CTRL codes and macros (subject to change):
3880 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3881 - SSL_CTX_set_tlsext_servername_callback()
3882 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3883 - SSL_CTX_set_tlsext_servername_arg()
3884 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3886 openssl s_client has a new '-servername ...' option.
3888 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3889 '-key2 ...', '-servername_fatal' (subject to change). This allows
3890 testing the HostName extension for a specific single host name ('-cert'
3891 and '-key' remain fallbacks for handshakes without HostName
3892 negotiation). If the unrecognized_name alert has to be sent, this by
3893 default is a warning; it becomes fatal with the '-servername_fatal'
3896 [Peter Sylvester, Remy Allais, Christophe Renou]
3898 *) Whirlpool hash implementation is added.
3901 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3902 bn(64,32). Because of instruction set limitations it doesn't have
3903 any negative impact on performance. This was done mostly in order
3904 to make it possible to share assembler modules, such as bn_mul_mont
3905 implementations, between 32- and 64-bit builds without hassle.
3908 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3909 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3913 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3914 dedicated Montgomery multiplication procedure, is introduced.
3915 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3916 "64-bit" performance on certain 32-bit targets.
3919 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3920 in SSL structures. New SSL ctrl to set maximum send fragment size.
3921 Save memory by seeting the I/O buffer sizes dynamically instead of
3922 using the maximum available value.
3925 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3926 in addition to the text details.
3929 *) Very, very preliminary EXPERIMENTAL support for printing of general
3930 ASN1 structures. This currently produces rather ugly output and doesn't
3931 handle several customised structures at all.
3934 *) Integrated support for PVK file format and some related formats such
3935 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3936 these in the 'rsa' and 'dsa' utilities.
3939 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3942 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3943 place for the (very old) "NETSCAPE" format certificates which are now
3944 handled using new ASN1 code equivalents.
3947 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3948 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3949 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3952 *) Modify CRL distribution points extension code to print out previously
3953 unsupported fields. Enhance extension setting code to allow setting of
3957 *) Add print and set support for Issuing Distribution Point CRL extension.
3960 *) Change 'Configure' script to enable Camellia by default.
3963 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3965 *) When rejecting SSL/TLS records due to an incorrect version number, never
3966 update s->server with a new major version number. As of
3967 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3968 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3969 the previous behavior could result in a read attempt at NULL when
3970 receiving specific incorrect SSL/TLS records once record payload
3971 protection is active. (CVE-2010-0740)
3972 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3974 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3975 could be crashed if the relevant tables were not present (e.g. chrooted).
3976 [Tomas Hoger <thoger@redhat.com>]
3978 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3980 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3981 [Martin Olsson, Neel Mehta]
3983 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3984 accommodate for stack sorting, always a write lock!).
3987 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3988 excessive delays in the RAND_poll(): over a minute. As a workaround
3989 include a time check in the inner Heap32Next loop too.
3992 *) The code that handled flushing of data in SSL/TLS originally used the
3993 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3994 the problem outlined in PR#1949. The fix suggested there however can
3995 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3996 of Apache). So instead simplify the code to flush unconditionally.
3997 This should be fine since flushing with no data to flush is a no op.
4000 *) Handle TLS versions 2.0 and later properly and correctly use the
4001 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4002 off ancient servers have a habit of sticking around for a while...
4005 *) Modify compression code so it frees up structures without using the
4006 ex_data callbacks. This works around a problem where some applications
4007 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4008 restarting) then use compression (e.g. SSL with compression) later.
4009 This results in significant per-connection memory leaks and
4010 has caused some security issues including CVE-2008-1678 and
4014 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4015 change when encrypting or decrypting.
4018 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4019 connect and renegotiate with servers which do not support RI.
4020 Until RI is more widely deployed this option is enabled by default.
4023 *) Add "missing" ssl ctrls to clear options and mode.
4026 *) If client attempts to renegotiate and doesn't support RI respond with
4027 a no_renegotiation alert as required by RFC5746. Some renegotiating
4028 TLS clients will continue a connection gracefully when they receive
4029 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4030 waiting for a server hello which it will never receive. Now we treat a
4031 received no_renegotiation alert as a fatal error. This is because
4032 applications requesting a renegotiation might well expect it to succeed
4033 and would have no code in place to handle the server denying it so the
4034 only safe thing to do is to terminate the connection.
4037 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4038 peer supports secure renegotiation and 0 otherwise. Print out peer
4039 renegotiation support in s_client/s_server.
4042 *) Replace the highly broken and deprecated SPKAC certification method with
4043 the updated NID creation version. This should correctly handle UTF8.
4046 *) Implement RFC5746. Re-enable renegotiation but require the extension
4047 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4048 turns out to be a bad idea. It has been replaced by
4049 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4050 SSL_CTX_set_options(). This is really not recommended unless you
4051 know what you are doing.
4052 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4054 *) Fixes to stateless session resumption handling. Use initial_ctx when
4055 issuing and attempting to decrypt tickets in case it has changed during
4056 servername handling. Use a non-zero length session ID when attempting
4057 stateless session resumption: this makes it possible to determine if
4058 a resumption has occurred immediately after receiving server hello
4059 (several places in OpenSSL subtly assume this) instead of later in
4063 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4064 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4065 fixes for a few places where the return code is not checked
4067 [Julia Lawall <julia@diku.dk>]
4069 *) Add --strict-warnings option to Configure script to include devteam
4070 warnings in other configurations.
4073 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4074 makes it possible to install openssl libraries in locations which
4075 have names other than "lib", for example "/usr/lib64" which some
4077 [Steve Henson, based on patch from Jeremy Utley]
4079 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4080 X690 8.9.12 and can produce some misleading textual output of OIDs.
4081 [Steve Henson, reported by Dan Kaminsky]
4083 *) Delete MD2 from algorithm tables. This follows the recommendation in
4084 several standards that it is not used in new applications due to
4085 several cryptographic weaknesses. For binary compatibility reasons
4086 the MD2 API is still compiled in by default.
4089 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4093 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4094 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4096 [Guenter <lists@gknw.net>]
4098 *) Fix the server certificate chain building code to use X509_verify_cert(),
4099 it used to have an ad-hoc builder which was unable to cope with anything
4100 other than a simple chain.
4101 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4103 *) Don't check self signed certificate signatures in X509_verify_cert()
4104 by default (a flag can override this): it just wastes time without
4105 adding any security. As a useful side effect self signed root CAs
4106 with non-FIPS digests are now usable in FIPS mode.
4109 *) In dtls1_process_out_of_seq_message() the check if the current message
4110 is already buffered was missing. For every new message was memory
4111 allocated, allowing an attacker to perform an denial of service attack
4112 with sending out of seq handshake messages until there is no memory
4113 left. Additionally every future messege was buffered, even if the
4114 sequence number made no sense and would be part of another handshake.
4115 So only messages with sequence numbers less than 10 in advance will be
4116 buffered. (CVE-2009-1378)
4117 [Robin Seggelmann, discovered by Daniel Mentz]
4119 *) Records are buffered if they arrive with a future epoch to be
4120 processed after finishing the corresponding handshake. There is
4121 currently no limitation to this buffer allowing an attacker to perform
4122 a DOS attack with sending records with future epochs until there is no
4123 memory left. This patch adds the pqueue_size() function to determine
4124 the size of a buffer and limits the record buffer to 100 entries.
4126 [Robin Seggelmann, discovered by Daniel Mentz]
4128 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4129 parent structure is freed. (CVE-2009-1379)
4132 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4133 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4136 [Ilya O. <vrghost@gmail.com>]
4138 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4140 *) Disable renegotiation completely - this fixes a severe security
4141 problem (CVE-2009-3555) at the cost of breaking all
4142 renegotiation. Renegotiation can be re-enabled by setting
4143 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4144 run-time. This is really not recommended unless you know what
4148 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4150 *) Don't set val to NULL when freeing up structures, it is freed up by
4151 underlying code. If sizeof(void *) > sizeof(long) this can result in
4152 zeroing past the valid field. (CVE-2009-0789)
4153 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4155 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4156 checked correctly. This would allow some invalid signed attributes to
4157 appear to verify correctly. (CVE-2009-0591)
4158 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4160 *) Reject UniversalString and BMPString types with invalid lengths. This
4161 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4162 a legal length. (CVE-2009-0590)
4165 *) Set S/MIME signing as the default purpose rather than setting it
4166 unconditionally. This allows applications to override it at the store
4170 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4171 to handle some structures.
4174 *) Improve efficiency of mem_gets: don't search whole buffer each time
4176 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4178 *) New -hex option for openssl rand.
4181 *) Print out UTF8String and NumericString when parsing ASN1.
4184 *) Support NumericString type for name components.
4187 *) Allow CC in the environment to override the automatically chosen
4188 compiler. Note that nothing is done to ensure flags work with the
4192 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4194 *) Properly check EVP_VerifyFinal() and similar return values
4196 [Ben Laurie, Bodo Moeller, Google Security Team]
4198 *) Enable TLS extensions by default.
4201 *) Allow the CHIL engine to be loaded, whether the application is
4202 multithreaded or not. (This does not release the developer from the
4203 obligation to set up the dynamic locking callbacks.)
4204 [Sander Temme <sander@temme.net>]
4206 *) Use correct exit code if there is an error in dgst command.
4207 [Steve Henson; problem pointed out by Roland Dirlewanger]
4209 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4210 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4213 *) Add experimental JPAKE support, including demo authentication in
4214 s_client and s_server.
4217 *) Set the comparison function in v3_addr_canonize().
4218 [Rob Austein <sra@hactrn.net>]
4220 *) Add support for XMPP STARTTLS in s_client.
4221 [Philip Paeps <philip@freebsd.org>]
4223 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4224 to ensure that even with this option, only ciphersuites in the
4225 server's preference list will be accepted. (Note that the option
4226 applies only when resuming a session, so the earlier behavior was
4227 just about the algorithm choice for symmetric cryptography.)
4230 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4232 *) Fix NULL pointer dereference if a DTLS server received
4233 ChangeCipherSpec as first record (CVE-2009-1386).
4236 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4237 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4240 *) The fix in 0.9.8c that supposedly got rid of unsafe
4241 double-checked locking was incomplete for RSA blinding,
4242 addressing just one layer of what turns out to have been
4243 doubly unsafe triple-checked locking.
4245 So now fix this for real by retiring the MONT_HELPER macro
4246 in crypto/rsa/rsa_eay.c.
4248 [Bodo Moeller; problem pointed out by Marius Schilder]
4250 *) Various precautionary measures:
4252 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4254 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4255 (NB: This would require knowledge of the secret session ticket key
4256 to exploit, in which case you'd be SOL either way.)
4258 - Change bn_nist.c so that it will properly handle input BIGNUMs
4259 outside the expected range.
4261 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4264 [Neel Mehta, Bodo Moeller]
4266 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4267 the load fails. Useful for distros.
4268 [Ben Laurie and the FreeBSD team]
4270 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4273 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4276 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4278 This work was sponsored by Logica.
4281 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4282 keystores. Support for SSL/TLS client authentication too.
4283 Not compiled unless enable-capieng specified to Configure.
4285 This work was sponsored by Logica.
4288 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4289 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4290 attribute creation routines such as certificate requests and PKCS#12
4294 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4296 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4297 handshake which could lead to a cilent crash as found using the
4298 Codenomicon TLS test suite (CVE-2008-1672)
4299 [Steve Henson, Mark Cox]
4301 *) Fix double free in TLS server name extensions which could lead to
4302 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4305 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4307 Clear the error queue to ensure that error entries left from
4308 older function calls do not interfere with the correct operation.
4309 [Lutz Jaenicke, Erik de Castro Lopo]
4311 *) Remove root CA certificates of commercial CAs:
4313 The OpenSSL project does not recommend any specific CA and does not
4314 have any policy with respect to including or excluding any CA.
4315 Therefore it does not make any sense to ship an arbitrary selection
4316 of root CA certificates with the OpenSSL software.
4319 *) RSA OAEP patches to fix two separate invalid memory reads.
4320 The first one involves inputs when 'lzero' is greater than
4321 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4322 before the beginning of from). The second one involves inputs where
4323 the 'db' section contains nothing but zeroes (there is a one-byte
4324 invalid read after the end of 'db').
4325 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4327 *) Partial backport from 0.9.9-dev:
4329 Introduce bn_mul_mont (dedicated Montgomery multiplication
4330 procedure) as a candidate for BIGNUM assembler implementation.
4331 While 0.9.9-dev uses assembler for various architectures, only
4332 x86_64 is available by default here in the 0.9.8 branch, and
4333 32-bit x86 is available through a compile-time setting.
4335 To try the 32-bit x86 assembler implementation, use Configure
4336 option "enable-montasm" (which exists only for this backport).
4338 As "enable-montasm" for 32-bit x86 disclaims code stability
4339 anyway, in this constellation we activate additional code
4340 backported from 0.9.9-dev for further performance improvements,
4341 namely BN_from_montgomery_word. (To enable this otherwise,
4342 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4344 [Andy Polyakov (backport partially by Bodo Moeller)]
4346 *) Add TLS session ticket callback. This allows an application to set
4347 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4348 values. This is useful for key rollover for example where several key
4349 sets may exist with different names.
4352 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4353 This was broken until now in 0.9.8 releases, such that the only way
4354 a registered ENGINE could be used (assuming it initialises
4355 successfully on the host) was to explicitly set it as the default
4356 for the relevant algorithms. This is in contradiction with 0.9.7
4357 behaviour and the documentation. With this fix, when an ENGINE is
4358 registered into a given algorithm's table of implementations, the
4359 'uptodate' flag is reset so that auto-discovery will be used next
4360 time a new context for that algorithm attempts to select an
4362 [Ian Lister (tweaked by Geoff Thorpe)]
4364 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4365 implementation in the following ways:
4367 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4370 Lack of BER streaming support means one pass streaming processing is
4371 only supported if data is detached: setting the streaming flag is
4372 ignored for embedded content.
4374 CMS support is disabled by default and must be explicitly enabled
4375 with the enable-cms configuration option.
4378 *) Update the GMP engine glue to do direct copies between BIGNUM and
4379 mpz_t when openssl and GMP use the same limb size. Otherwise the
4380 existing "conversion via a text string export" trick is still used.
4381 [Paul Sheer <paulsheer@gmail.com>]
4383 *) Zlib compression BIO. This is a filter BIO which compressed and
4384 uncompresses any data passed through it.
4387 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4388 RFC3394 compatible AES key wrapping.
4391 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4392 sets string data without copying. X509_ALGOR_set0() and
4393 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4394 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4395 from an X509_ATTRIBUTE structure optionally checking it occurs only
4396 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4400 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4401 to get the expected BN_FLG_CONSTTIME behavior.
4402 [Bodo Moeller (Google)]
4406 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4407 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4408 - added some more tests to do_tests.pl
4409 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4410 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4411 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4412 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4413 - various changes to netware.pl to enable gcc-cross builds on Win32
4415 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4416 - various changes to fix missing prototype warnings
4417 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4418 - added AES, WHIRLPOOL and CPUID assembler code to build files
4419 - added missing AES assembler make rules to mk1mf.pl
4420 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4421 [Guenter Knauf <eflash@gmx.net>]
4423 *) Implement certificate status request TLS extension defined in RFC3546.
4424 A client can set the appropriate parameters and receive the encoded
4425 OCSP response via a callback. A server can query the supplied parameters
4426 and set the encoded OCSP response in the callback. Add simplified examples
4427 to s_client and s_server.
4430 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4432 *) Fix various bugs:
4433 + Binary incompatibility of ssl_ctx_st structure
4434 + DTLS interoperation with non-compliant servers
4435 + Don't call get_session_cb() without proposed session
4436 + Fix ia64 assembler code
4437 [Andy Polyakov, Steve Henson]
4439 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4441 *) DTLS Handshake overhaul. There were longstanding issues with
4442 OpenSSL DTLS implementation, which were making it impossible for
4443 RFC 4347 compliant client to communicate with OpenSSL server.
4444 Unfortunately just fixing these incompatibilities would "cut off"
4445 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4446 server keeps tolerating non RFC compliant syntax. The opposite is
4447 not true, 0.9.8f client can not communicate with earlier server.
4448 This update even addresses CVE-2007-4995.
4451 *) Changes to avoid need for function casts in OpenSSL: some compilers
4452 (gcc 4.2 and later) reject their use.
4453 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4456 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4457 RFC4507bis. The encrypted ticket format is an encrypted encoded
4458 SSL_SESSION structure, that way new session features are automatically
4461 If a client application caches session in an SSL_SESSION structure
4462 support is transparent because tickets are now stored in the encoded
4465 The SSL_CTX structure automatically generates keys for ticket
4466 protection in servers so again support should be possible
4467 with no application modification.
4469 If a client or server wishes to disable RFC4507 support then the option
4470 SSL_OP_NO_TICKET can be set.
4472 Add a TLS extension debugging callback to allow the contents of any client
4473 or server extensions to be examined.
4475 This work was sponsored by Google.
4478 *) Add initial support for TLS extensions, specifically for the server_name
4479 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4480 have new members for a host name. The SSL data structure has an
4481 additional member SSL_CTX *initial_ctx so that new sessions can be
4482 stored in that context to allow for session resumption, even after the
4483 SSL has been switched to a new SSL_CTX in reaction to a client's
4484 server_name extension.
4486 New functions (subject to change):
4488 SSL_get_servername()
4489 SSL_get_servername_type()
4492 New CTRL codes and macros (subject to change):
4494 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4495 - SSL_CTX_set_tlsext_servername_callback()
4496 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4497 - SSL_CTX_set_tlsext_servername_arg()
4498 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4500 openssl s_client has a new '-servername ...' option.
4502 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4503 '-key2 ...', '-servername_fatal' (subject to change). This allows
4504 testing the HostName extension for a specific single host name ('-cert'
4505 and '-key' remain fallbacks for handshakes without HostName
4506 negotiation). If the unrecognized_name alert has to be sent, this by
4507 default is a warning; it becomes fatal with the '-servername_fatal'
4510 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4512 *) Add AES and SSE2 assembly language support to VC++ build.
4515 *) Mitigate attack on final subtraction in Montgomery reduction.
4518 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4519 (which previously caused an internal error).
4522 *) Squeeze another 10% out of IGE mode when in != out.
4525 *) AES IGE mode speedup.
4526 [Dean Gaudet (Google)]
4528 *) Add the Korean symmetric 128-bit cipher SEED (see
4529 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4530 add SEED ciphersuites from RFC 4162:
4532 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4533 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4534 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4535 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4537 To minimize changes between patchlevels in the OpenSSL 0.9.8
4538 series, SEED remains excluded from compilation unless OpenSSL
4539 is configured with 'enable-seed'.
4540 [KISA, Bodo Moeller]
4542 *) Mitigate branch prediction attacks, which can be practical if a
4543 single processor is shared, allowing a spy process to extract
4544 information. For detailed background information, see
4545 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4546 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4547 and Necessary Software Countermeasures"). The core of the change
4548 are new versions BN_div_no_branch() and
4549 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4550 respectively, which are slower, but avoid the security-relevant
4551 conditional branches. These are automatically called by BN_div()
4552 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4553 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4554 remove a conditional branch.
4556 BN_FLG_CONSTTIME is the new name for the previous
4557 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4558 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4559 in the exponent causes BN_mod_exp_mont() to use the alternative
4560 implementation in BN_mod_exp_mont_consttime().) The old name
4561 remains as a deprecated alias.
4563 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4564 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4565 constant-time implementations for more than just exponentiation.
4566 Here too the old name is kept as a deprecated alias.
4568 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4569 the BN_BLINDING structure gets an independent copy of the
4570 modulus. This means that the previous "BIGNUM *m" argument to
4571 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4572 essentially becomes "const BIGNUM *m", although we can't actually
4573 change this in the header file before 0.9.9. It allows
4574 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4575 enable BN_FLG_CONSTTIME.
4577 [Matthew D Wood (Intel Corp)]
4579 *) In the SSL/TLS server implementation, be strict about session ID
4580 context matching (which matters if an application uses a single
4581 external cache for different purposes). Previously,
4582 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4583 set. This did ensure strict client verification, but meant that,
4584 with applications using a single external cache for quite
4585 different requirements, clients could circumvent ciphersuite
4586 restrictions for a given session ID context by starting a session
4587 in a different context.
4590 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4591 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4592 authentication-only ciphersuites.
4595 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4596 not complete and could lead to a possible single byte overflow
4597 (CVE-2007-5135) [Ben Laurie]
4599 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4601 *) Since AES128 and AES256 (and similarly Camellia128 and
4602 Camellia256) share a single mask bit in the logic of
4603 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4604 kludge to work properly if AES128 is available and AES256 isn't
4605 (or if Camellia128 is available and Camellia256 isn't).
4608 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4609 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4610 When a point or a seed is encoded in a BIT STRING, we need to
4611 prevent the removal of trailing zero bits to get the proper DER
4612 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4613 of a NamedBitList, for which trailing 0 bits need to be removed.)
4616 *) Have SSL/TLS server implementation tolerate "mismatched" record
4617 protocol version while receiving ClientHello even if the
4618 ClientHello is fragmented. (The server can't insist on the
4619 particular protocol version it has chosen before the ServerHello
4620 message has informed the client about his choice.)
4623 *) Add RFC 3779 support.
4624 [Rob Austein for ARIN, Ben Laurie]
4626 *) Load error codes if they are not already present instead of using a
4627 static variable. This allows them to be cleanly unloaded and reloaded.
4628 Improve header file function name parsing.
4631 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4632 or CAPABILITY handshake as required by RFCs.
4635 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4637 *) Introduce limits to prevent malicious keys being able to
4638 cause a denial of service. (CVE-2006-2940)
4639 [Steve Henson, Bodo Moeller]
4641 *) Fix ASN.1 parsing of certain invalid structures that can result
4642 in a denial of service. (CVE-2006-2937) [Steve Henson]
4644 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4645 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4647 *) Fix SSL client code which could crash if connecting to a
4648 malicious SSLv2 server. (CVE-2006-4343)
4649 [Tavis Ormandy and Will Drewry, Google Security Team]
4651 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4652 match only those. Before that, "AES256-SHA" would be interpreted
4653 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4654 the same strength classification in 0.9.7h) as we currently only
4655 have a single AES bit in the ciphersuite description bitmap.
4656 That change, however, also applied to ciphersuite strings such as
4657 "RC4-MD5" that intentionally matched multiple ciphersuites --
4658 namely, SSL 2.0 ciphersuites in addition to the more common ones
4659 from SSL 3.0/TLS 1.0.
4661 So we change the selection algorithm again: Naming an explicit
4662 ciphersuite selects this one ciphersuite, and any other similar
4663 ciphersuite (same bitmap) from *other* protocol versions.
4664 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4665 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4667 Since SSL 2.0 does not have any ciphersuites for which the
4668 128/256 bit distinction would be relevant, this works for now.
4669 The proper fix will be to use different bits for AES128 and
4670 AES256, which would have avoided the problems from the beginning;
4671 however, bits are scarce, so we can only do this in a new release
4672 (not just a patchlevel) when we can change the SSL_CIPHER
4673 definition to split the single 'unsigned long mask' bitmap into
4674 multiple values to extend the available space.
4678 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4680 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4681 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4683 *) Add AES IGE and biIGE modes.
4686 *) Change the Unix randomness entropy gathering to use poll() when
4687 possible instead of select(), since the latter has some
4688 undesirable limitations.
4689 [Darryl Miles via Richard Levitte and Bodo Moeller]
4691 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4692 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4693 cannot be implicitly activated as part of, e.g., the "AES" alias.
4694 However, please upgrade to OpenSSL 0.9.9[-dev] for
4695 non-experimental use of the ECC ciphersuites to get TLS extension
4696 support, which is required for curve and point format negotiation
4697 to avoid potential handshake problems.
4700 *) Disable rogue ciphersuites:
4702 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4703 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4704 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4706 The latter two were purportedly from
4707 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4710 Also deactivate the remaining ciphersuites from
4711 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4712 unofficial, and the ID has long expired.
4715 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4716 dual-core machines) and other potential thread-safety issues.
4719 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4720 versions), which is now available for royalty-free use
4721 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4722 Also, add Camellia TLS ciphersuites from RFC 4132.
4724 To minimize changes between patchlevels in the OpenSSL 0.9.8
4725 series, Camellia remains excluded from compilation unless OpenSSL
4726 is configured with 'enable-camellia'.
4729 *) Disable the padding bug check when compression is in use. The padding
4730 bug check assumes the first packet is of even length, this is not
4731 necessarily true if compresssion is enabled and can result in false
4732 positives causing handshake failure. The actual bug test is ancient
4733 code so it is hoped that implementations will either have fixed it by
4734 now or any which still have the bug do not support compression.
4737 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4739 *) When applying a cipher rule check to see if string match is an explicit
4740 cipher suite and only match that one cipher suite if it is.
4743 *) Link in manifests for VC++ if needed.
4744 [Austin Ziegler <halostatue@gmail.com>]
4746 *) Update support for ECC-based TLS ciphersuites according to
4747 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4748 TLS extensions, which are supported starting with the 0.9.9
4749 branch, not in the OpenSSL 0.9.8 branch).
4752 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4753 opaque EVP_CIPHER_CTX handling.
4756 *) Fixes and enhancements to zlib compression code. We now only use
4757 "zlib1.dll" and use the default __cdecl calling convention on Win32
4758 to conform with the standards mentioned here:
4759 http://www.zlib.net/DLL_FAQ.txt
4760 Static zlib linking now works on Windows and the new --with-zlib-include
4761 --with-zlib-lib options to Configure can be used to supply the location
4762 of the headers and library. Gracefully handle case where zlib library
4766 *) Several fixes and enhancements to the OID generation code. The old code
4767 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4768 handle numbers larger than ULONG_MAX, truncated printing and had a
4769 non standard OBJ_obj2txt() behaviour.
4772 *) Add support for building of engines under engine/ as shared libraries
4773 under VC++ build system.
4776 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4777 Hopefully, we will not see any false combination of paths any more.
4780 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4782 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4783 (part of SSL_OP_ALL). This option used to disable the
4784 countermeasure against man-in-the-middle protocol-version
4785 rollback in the SSL 2.0 server implementation, which is a bad
4786 idea. (CVE-2005-2969)
4788 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4789 for Information Security, National Institute of Advanced Industrial
4790 Science and Technology [AIST], Japan)]
4792 *) Add two function to clear and return the verify parameter flags.
4795 *) Keep cipherlists sorted in the source instead of sorting them at
4796 runtime, thus removing the need for a lock.
4799 *) Avoid some small subgroup attacks in Diffie-Hellman.
4800 [Nick Mathewson and Ben Laurie]
4802 *) Add functions for well-known primes.
4805 *) Extended Windows CE support.
4806 [Satoshi Nakamura and Andy Polyakov]
4808 *) Initialize SSL_METHOD structures at compile time instead of during
4809 runtime, thus removing the need for a lock.
4812 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4813 attempting to decrypt each encrypted key in turn. Add support to
4817 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4819 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4822 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4825 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4826 key into the same file any more.
4829 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4832 *) Add -utf8 command line and config file option to 'ca'.
4833 [Stefan <stf@udoma.org]
4835 *) Removed the macro des_crypt(), as it seems to conflict with some
4836 libraries. Use DES_crypt().
4839 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4840 involves renaming the source and generated shared-libs for
4841 both. The engines will accept the corrected or legacy ids
4842 ('ncipher' and '4758_cca' respectively) when binding. NB,
4843 this only applies when building 'shared'.
4844 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4846 *) Add attribute functions to EVP_PKEY structure. Modify
4847 PKCS12_create() to recognize a CSP name attribute and
4848 use it. Make -CSP option work again in pkcs12 utility.
4851 *) Add new functionality to the bn blinding code:
4852 - automatic re-creation of the BN_BLINDING parameters after
4853 a fixed number of uses (currently 32)
4854 - add new function for parameter creation
4855 - introduce flags to control the update behaviour of the
4856 BN_BLINDING parameters
4857 - hide BN_BLINDING structure
4858 Add a second BN_BLINDING slot to the RSA structure to improve
4859 performance when a single RSA object is shared among several
4863 *) Add support for DTLS.
4864 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4866 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4867 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4870 *) Remove buggy and incomplete DH cert support from
4871 ssl/ssl_rsa.c and ssl/s3_both.c
4874 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4875 the apps/openssl applications.
4878 *) Compile clean with "-Wall -Wmissing-prototypes
4879 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4880 DEBUG_SAFESTACK must also be set.
4883 *) Change ./Configure so that certain algorithms can be disabled by default.
4884 The new counterpiece to "no-xxx" is "enable-xxx".
4886 The patented RC5 and MDC2 algorithms will now be disabled unless
4887 "enable-rc5" and "enable-mdc2", respectively, are specified.
4889 (IDEA remains enabled despite being patented. This is because IDEA
4890 is frequently required for interoperability, and there is no license
4891 fee for non-commercial use. As before, "no-idea" can be used to
4892 avoid this algorithm.)
4896 *) Add processing of proxy certificates (see RFC 3820). This work was
4897 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4898 EGEE (Enabling Grids for E-science in Europe).
4901 *) RC4 performance overhaul on modern architectures/implementations, such
4902 as Intel P4, IA-64 and AMD64.
4905 *) New utility extract-section.pl. This can be used specify an alternative
4906 section number in a pod file instead of having to treat each file as
4907 a separate case in Makefile. This can be done by adding two lines to the
4910 =for comment openssl_section:XXX
4912 The blank line is mandatory.
4916 *) New arguments -certform, -keyform and -pass for s_client and s_server
4917 to allow alternative format key and certificate files and passphrase
4921 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4922 update associated structures and add various utility functions.
4924 Add new policy related verify parameters, include policy checking in
4925 standard verify code. Enhance 'smime' application with extra parameters
4926 to support policy checking and print out.
4929 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4930 Nehemiah processors. These extensions support AES encryption in hardware
4931 as well as RNG (though RNG support is currently disabled).
4932 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4934 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4937 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4938 [Andy Polyakov and a number of other people]
4940 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4941 implementation contributed by IBM.
4942 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4944 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4945 exponent rather than 'unsigned long'. There is a corresponding change to
4946 the new 'rsa_keygen' element of the RSA_METHOD structure.
4947 [Jelte Jansen, Geoff Thorpe]
4949 *) Functionality for creating the initial serial number file is now
4950 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4952 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4953 number file to 1, which is bound to cause problems. To avoid
4954 the problems while respecting compatibility between different 0.9.7
4955 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4956 CA.pl for serial number initialization. With the new release 0.9.8,
4957 we can fix the problem directly in the 'ca' utility.)
4960 *) Reduced header interdepencies by declaring more opaque objects in
4961 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4962 give fewer recursive includes, which could break lazy source code - so
4963 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4964 developers should define this symbol when building and using openssl to
4965 ensure they track the recommended behaviour, interfaces, [etc], but
4966 backwards-compatible behaviour prevails when this isn't defined.
4969 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4972 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4973 This will generate a random key of the appropriate length based on the
4974 cipher context. The EVP_CIPHER can provide its own random key generation
4975 routine to support keys of a specific form. This is used in the des and
4976 3des routines to generate a key of the correct parity. Update S/MIME
4977 code to use new functions and hence generate correct parity DES keys.
4978 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4979 valid (weak or incorrect parity).
4982 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4983 as looking them up. This is useful when the verified structure may contain
4984 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4985 present unless the new PKCS7_NO_CRL flag is asserted.
4988 *) Extend ASN1 oid configuration module. It now additionally accepts the
4991 shortName = some long name, 1.2.3.4
4994 *) Reimplemented the BN_CTX implementation. There is now no more static
4995 limitation on the number of variables it can handle nor the depth of the
4996 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4997 information can now expand as required, and rather than having a single
4998 static array of bignums, BN_CTX now uses a linked-list of such arrays
4999 allowing it to expand on demand whilst maintaining the usefulness of
5000 BN_CTX's "bundling".
5003 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5004 to allow all RSA operations to function using a single BN_CTX.
5007 *) Preliminary support for certificate policy evaluation and checking. This
5008 is initially intended to pass the tests outlined in "Conformance Testing
5009 of Relying Party Client Certificate Path Processing Logic" v1.07.
5012 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5013 remained unused and not that useful. A variety of other little bignum
5014 tweaks and fixes have also been made continuing on from the audit (see
5018 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5019 associated ASN1, EVP and SSL functions and old ASN1 macros.
5022 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5023 and this should never fail. So the return value from the use of
5024 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5025 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5028 *) BN_CTX_get() should return zero-valued bignums, providing the same
5029 initialised value as BN_new().
5030 [Geoff Thorpe, suggested by Ulf Möller]
5032 *) Support for inhibitAnyPolicy certificate extension.
5035 *) An audit of the BIGNUM code is underway, for which debugging code is
5036 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5037 is considered valid when processing BIGNUMs, and causes execution to
5038 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5039 further steps are taken to deliberately pollute unused data in BIGNUM
5040 structures to try and expose faulty code further on. For now, openssl will
5041 (in its default mode of operation) continue to tolerate the inconsistent
5042 forms that it has tolerated in the past, but authors and packagers should
5043 consider trying openssl and their own applications when compiled with
5044 these debugging symbols defined. It will help highlight potential bugs in
5045 their own code, and will improve the test coverage for OpenSSL itself. At
5046 some point, these tighter rules will become openssl's default to improve
5047 maintainability, though the assert()s and other overheads will remain only
5048 in debugging configurations. See bn.h for more details.
5049 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5051 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5052 that can only be obtained through BN_CTX_new() (which implicitly
5053 initialises it). The presence of this function only made it possible
5054 to overwrite an existing structure (and cause memory leaks).
5057 *) Because of the callback-based approach for implementing LHASH as a
5058 template type, lh_insert() adds opaque objects to hash-tables and
5059 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5060 to clean up those corresponding objects before destroying the hash table
5061 (and losing the object pointers). So some over-zealous constifications in
5062 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5063 objects as "const" and the lh_doall[_arg] callback wrappers are not
5064 prototyped to have "const" restrictions on the object pointers they are
5065 given (and so aren't required to cast them away any more).
5068 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5069 (speed) prefers to use its own implementation. The two implementations
5070 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5071 its object type properly exposed (MS_TM) instead of casting to/from "char
5072 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5073 aren't necessarily the greatest nomenclatures - but this is what was used
5074 internally to the implementation so I've used that for now.
5077 *) Ensure that deprecated functions do not get compiled when
5078 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5079 the self-tests were still using deprecated key-generation functions so
5080 these have been updated also.
5083 *) Reorganise PKCS#7 code to separate the digest location functionality
5084 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5085 New function PKCS7_set_digest() to set the digest type for PKCS#7
5086 digestedData type. Add additional code to correctly generate the
5087 digestedData type and add support for this type in PKCS7 initialization
5091 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5092 structure of type "other".
5095 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5096 sure the loop does correctly stop and breaking ("division by zero")
5097 modulus operations are not performed. The (pre-generated) prime
5098 table crypto/bn/bn_prime.h was already correct, but it could not be
5099 re-generated on some platforms because of the "division by zero"
5100 situation in the script.
5101 [Ralf S. Engelschall]
5103 *) Update support for ECC-based TLS ciphersuites according to
5104 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5105 SHA-1 now is only used for "small" curves (where the
5106 representation of a field element takes up to 24 bytes); for
5107 larger curves, the field element resulting from ECDH is directly
5108 used as premaster secret.
5109 [Douglas Stebila (Sun Microsystems Laboratories)]
5111 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5112 curve secp160r1 to the tests.
5113 [Douglas Stebila (Sun Microsystems Laboratories)]
5115 *) Add the possibility to load symbols globally with DSO.
5116 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5118 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5119 control of the error stack.
5122 *) Add support for STORE in ENGINE.
5125 *) Add the STORE type. The intention is to provide a common interface
5126 to certificate and key stores, be they simple file-based stores, or
5127 HSM-type store, or LDAP stores, or...
5128 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5131 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5132 pass a list of arguments to any function as well as provide a way
5133 for a function to pass data back to the caller.
5136 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5137 works like BUF_strdup() but can be used to duplicate a portion of
5138 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5142 *) Add the function sk_find_ex() which works like sk_find(), but will
5143 return an index to an element even if an exact match couldn't be
5144 found. The index is guaranteed to point at the element where the
5145 searched-for key would be inserted to preserve sorting order.
5148 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5149 takes an extra flags argument for optional functionality. Currently,
5150 the following flags are defined:
5152 OBJ_BSEARCH_VALUE_ON_NOMATCH
5153 This one gets OBJ_bsearch_ex() to return a pointer to the first
5154 element where the comparing function returns a negative or zero
5157 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5158 This one gets OBJ_bsearch_ex() to return a pointer to the first
5159 element where the comparing function returns zero. This is useful
5160 if there are more than one element where the comparing function
5164 *) Make it possible to create self-signed certificates with 'openssl ca'
5165 in such a way that the self-signed certificate becomes part of the
5166 CA database and uses the same mechanisms for serial number generation
5167 as all other certificate signing. The new flag '-selfsign' enables
5168 this functionality. Adapt CA.sh and CA.pl.in.
5171 *) Add functionality to check the public key of a certificate request
5172 against a given private. This is useful to check that a certificate
5173 request can be signed by that key (self-signing).
5176 *) Make it possible to have multiple active certificates with the same
5177 subject in the CA index file. This is done only if the keyword
5178 'unique_subject' is set to 'no' in the main CA section (default
5179 if 'CA_default') of the configuration file. The value is saved
5180 with the database itself in a separate index attribute file,
5181 named like the index file with '.attr' appended to the name.
5184 *) Generate muti valued AVAs using '+' notation in config files for
5188 *) Support for nameConstraints certificate extension.
5191 *) Support for policyConstraints certificate extension.
5194 *) Support for policyMappings certificate extension.
5197 *) Make sure the default DSA_METHOD implementation only uses its
5198 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5199 and change its own handlers to be NULL so as to remove unnecessary
5200 indirection. This lets alternative implementations fallback to the
5201 default implementation more easily.
5204 *) Support for directoryName in GeneralName related extensions
5208 *) Make it possible to link applications using Makefile.shared.
5209 Make that possible even when linking against static libraries!
5212 *) Support for single pass processing for S/MIME signing. This now
5213 means that S/MIME signing can be done from a pipe, in addition
5214 cleartext signing (multipart/signed type) is effectively streaming
5215 and the signed data does not need to be all held in memory.
5217 This is done with a new flag PKCS7_STREAM. When this flag is set
5218 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5219 is done after the data is output (and digests calculated) in
5220 SMIME_write_PKCS7().
5223 *) Add full support for -rpath/-R, both in shared libraries and
5224 applications, at least on the platforms where it's known how
5228 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5229 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5230 will now compute a table of multiples of the generator that
5231 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5232 faster (notably in the case of a single point multiplication,
5233 scalar * generator).
5234 [Nils Larsch, Bodo Moeller]
5236 *) IPv6 support for certificate extensions. The various extensions
5237 which use the IP:a.b.c.d can now take IPv6 addresses using the
5238 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5242 *) Added an ENGINE that implements RSA by performing private key
5243 exponentiations with the GMP library. The conversions to and from
5244 GMP's mpz_t format aren't optimised nor are any montgomery forms
5245 cached, and on x86 it appears OpenSSL's own performance has caught up.
5246 However there are likely to be other architectures where GMP could
5247 provide a boost. This ENGINE is not built in by default, but it can be
5248 specified at Configure time and should be accompanied by the necessary
5249 linker additions, eg;
5250 ./config -DOPENSSL_USE_GMP -lgmp
5253 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5254 testing availability of engines with "-t" - the old behaviour is
5255 produced by increasing the feature's verbosity with "-tt".
5258 *) ECDSA routines: under certain error conditions uninitialized BN objects
5259 could be freed. Solution: make sure initialization is performed early
5260 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5264 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5265 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5266 software implementations. For DSA and DH, parameter generation can
5267 also be overridden by providing the appropriate method callbacks.
5270 *) Change the "progress" mechanism used in key-generation and
5271 primality testing to functions that take a new BN_GENCB pointer in
5272 place of callback/argument pairs. The new API functions have "_ex"
5273 postfixes and the older functions are reimplemented as wrappers for
5274 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5275 declarations of the old functions to help (graceful) attempts to
5276 migrate to the new functions. Also, the new key-generation API
5277 functions operate on a caller-supplied key-structure and return
5278 success/failure rather than returning a key or NULL - this is to
5279 help make "keygen" another member function of RSA_METHOD etc.
5281 Example for using the new callback interface:
5283 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5287 BN_GENCB_set(&my_cb, my_callback, my_arg);
5289 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5290 /* For the meaning of a, b in calls to my_callback(), see the
5291 * documentation of the function that calls the callback.
5292 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5293 * my_callback should return 1 if it wants BN_is_prime_ex()
5294 * to continue, or 0 to stop.
5299 *) Change the ZLIB compression method to be stateful, and make it
5300 available to TLS with the number defined in
5301 draft-ietf-tls-compression-04.txt.
5304 *) Add the ASN.1 structures and functions for CertificatePair, which
5305 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5307 CertificatePair ::= SEQUENCE {
5308 forward [0] Certificate OPTIONAL,
5309 reverse [1] Certificate OPTIONAL,
5310 -- at least one of the pair shall be present -- }
5312 Also implement the PEM functions to read and write certificate
5313 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5315 This needed to be defined, mostly for the sake of the LDAP
5316 attribute crossCertificatePair, but may prove useful elsewhere as
5320 *) Make it possible to inhibit symlinking of shared libraries in
5321 Makefile.shared, for Cygwin's sake.
5324 *) Extend the BIGNUM API by creating a function
5325 void BN_set_negative(BIGNUM *a, int neg);
5326 and a macro that behave like
5327 int BN_is_negative(const BIGNUM *a);
5329 to avoid the need to access 'a->neg' directly in applications.
5332 *) Implement fast modular reduction for pseudo-Mersenne primes
5333 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5334 EC_GROUP_new_curve_GFp() will now automatically use this
5336 [Nils Larsch <nla@trustcenter.de>]
5338 *) Add new lock type (CRYPTO_LOCK_BN).
5341 *) Change the ENGINE framework to automatically load engines
5342 dynamically from specific directories unless they could be
5343 found to already be built in or loaded. Move all the
5344 current engines except for the cryptodev one to a new
5346 The engines in engines/ are built as shared libraries if
5347 the "shared" options was given to ./Configure or ./config.
5348 Otherwise, they are inserted in libcrypto.a.
5349 /usr/local/ssl/engines is the default directory for dynamic
5350 engines, but that can be overridden at configure time through
5351 the usual use of --prefix and/or --openssldir, and at run
5352 time with the environment variable OPENSSL_ENGINES.
5353 [Geoff Thorpe and Richard Levitte]
5355 *) Add Makefile.shared, a helper makefile to build shared
5356 libraries. Adapt Makefile.org.
5359 *) Add version info to Win32 DLLs.
5360 [Peter 'Luna' Runestig" <peter@runestig.com>]
5362 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5363 can be added using this API to created arbitrary PKCS#12
5364 files while avoiding the low level API.
5366 New options to PKCS12_create(), key or cert can be NULL and
5367 will then be omitted from the output file. The encryption
5368 algorithm NIDs can be set to -1 for no encryption, the mac
5369 iteration count can be set to 0 to omit the mac.
5371 Enhance pkcs12 utility by making the -nokeys and -nocerts
5372 options work when creating a PKCS#12 file. New option -nomac
5373 to omit the mac, NONE can be set for an encryption algorithm.
5374 New code is modified to use the enhanced PKCS12_create()
5375 instead of the low level API.
5378 *) Extend ASN1 encoder to support indefinite length constructed
5379 encoding. This can output sequences tags and octet strings in
5380 this form. Modify pk7_asn1.c to support indefinite length
5381 encoding. This is experimental and needs additional code to
5382 be useful, such as an ASN1 bio and some enhanced streaming
5385 Extend template encode functionality so that tagging is passed
5386 down to the template encoder.
5389 *) Let 'openssl req' fail if an argument to '-newkey' is not
5390 recognized instead of using RSA as a default.
5393 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5394 As these are not official, they are not included in "ALL";
5395 the "ECCdraft" ciphersuite group alias can be used to select them.
5396 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5398 *) Add ECDH engine support.
5399 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5401 *) Add ECDH in new directory crypto/ecdh/.
5402 [Douglas Stebila (Sun Microsystems Laboratories)]
5404 *) Let BN_rand_range() abort with an error after 100 iterations
5405 without success (which indicates a broken PRNG).
5408 *) Change BN_mod_sqrt() so that it verifies that the input value
5409 is really the square of the return value. (Previously,
5410 BN_mod_sqrt would show GIGO behaviour.)
5413 *) Add named elliptic curves over binary fields from X9.62, SECG,
5414 and WAP/WTLS; add OIDs that were still missing.
5416 [Sheueling Chang Shantz and Douglas Stebila
5417 (Sun Microsystems Laboratories)]
5419 *) Extend the EC library for elliptic curves over binary fields
5420 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5423 EC_GF2m_simple_method
5427 EC_GROUP_new_curve_GF2m
5428 EC_GROUP_set_curve_GF2m
5429 EC_GROUP_get_curve_GF2m
5430 EC_POINT_set_affine_coordinates_GF2m
5431 EC_POINT_get_affine_coordinates_GF2m
5432 EC_POINT_set_compressed_coordinates_GF2m
5434 Point compression for binary fields is disabled by default for
5435 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5438 As binary polynomials are represented as BIGNUMs, various members
5439 of the EC_GROUP and EC_POINT data structures can be shared
5440 between the implementations for prime fields and binary fields;
5441 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5442 are essentially identical to their ..._GFp counterparts.
5443 (For simplicity, the '..._GFp' prefix has been dropped from
5444 various internal method names.)
5446 An internal 'field_div' method (similar to 'field_mul' and
5447 'field_sqr') has been added; this is used only for binary fields.
5449 [Sheueling Chang Shantz and Douglas Stebila
5450 (Sun Microsystems Laboratories)]
5452 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5453 through methods ('mul', 'precompute_mult').
5455 The generic implementations (now internally called 'ec_wNAF_mul'
5456 and 'ec_wNAF_precomputed_mult') remain the default if these
5457 methods are undefined.
5459 [Sheueling Chang Shantz and Douglas Stebila
5460 (Sun Microsystems Laboratories)]
5462 *) New function EC_GROUP_get_degree, which is defined through
5463 EC_METHOD. For curves over prime fields, this returns the bit
5464 length of the modulus.
5466 [Sheueling Chang Shantz and Douglas Stebila
5467 (Sun Microsystems Laboratories)]
5469 *) New functions EC_GROUP_dup, EC_POINT_dup.
5470 (These simply call ..._new and ..._copy).
5472 [Sheueling Chang Shantz and Douglas Stebila
5473 (Sun Microsystems Laboratories)]
5475 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5476 Polynomials are represented as BIGNUMs (where the sign bit is not
5477 used) in the following functions [macros]:
5480 BN_GF2m_sub [= BN_GF2m_add]
5481 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5482 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5483 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5485 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5486 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5487 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5488 BN_GF2m_cmp [= BN_ucmp]
5490 (Note that only the 'mod' functions are actually for fields GF(2^m).
5491 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5493 For some functions, an the irreducible polynomial defining a
5494 field can be given as an 'unsigned int[]' with strictly
5495 decreasing elements giving the indices of those bits that are set;
5496 i.e., p[] represents the polynomial
5497 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5499 p[0] > p[1] > ... > p[k] = 0.
5500 This applies to the following functions:
5505 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5506 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5508 BN_GF2m_mod_sqrt_arr
5509 BN_GF2m_mod_solve_quad_arr
5513 Conversion can be performed by the following functions:
5518 bntest.c has additional tests for binary polynomial arithmetic.
5520 Two implementations for BN_GF2m_mod_div() are available.
5521 The default algorithm simply uses BN_GF2m_mod_inv() and
5522 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5523 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5524 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5526 [Sheueling Chang Shantz and Douglas Stebila
5527 (Sun Microsystems Laboratories)]
5529 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5530 functionality is disabled at compile-time.
5531 [Douglas Stebila <douglas.stebila@sun.com>]
5533 *) Change default behaviour of 'openssl asn1parse' so that more
5534 information is visible when viewing, e.g., a certificate:
5536 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5537 mode the content of non-printable OCTET STRINGs is output in a
5538 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5539 avoid the appearance of a printable string.
5540 [Nils Larsch <nla@trustcenter.de>]
5542 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5544 EC_GROUP_set_asn1_flag()
5545 EC_GROUP_get_asn1_flag()
5546 EC_GROUP_set_point_conversion_form()
5547 EC_GROUP_get_point_conversion_form()
5548 These control ASN1 encoding details:
5549 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5550 has been set to OPENSSL_EC_NAMED_CURVE.
5551 - Points are encoded in uncompressed form by default; options for
5552 asn1_for are as for point2oct, namely
5553 POINT_CONVERSION_COMPRESSED
5554 POINT_CONVERSION_UNCOMPRESSED
5555 POINT_CONVERSION_HYBRID
5557 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5560 EC_GROUP_get0_seed()
5561 EC_GROUP_get_seed_len()
5562 This is used only for ASN1 purposes (so far).
5563 [Nils Larsch <nla@trustcenter.de>]
5565 *) Add 'field_type' member to EC_METHOD, which holds the NID
5566 of the appropriate field type OID. The new function
5567 EC_METHOD_get_field_type() returns this value.
5568 [Nils Larsch <nla@trustcenter.de>]
5573 EC_POINT_point2hex()
5574 EC_POINT_hex2point()
5575 providing useful interfaces to EC_POINT_point2oct() and
5576 EC_POINT_oct2point().
5577 [Nils Larsch <nla@trustcenter.de>]
5579 *) Change internals of the EC library so that the functions
5580 EC_GROUP_set_generator()
5581 EC_GROUP_get_generator()
5582 EC_GROUP_get_order()
5583 EC_GROUP_get_cofactor()
5584 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5585 to methods, which would lead to unnecessary code duplication when
5586 adding different types of curves.
5587 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5589 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5590 arithmetic, and such that modified wNAFs are generated
5591 (which avoid length expansion in many cases).
5594 *) Add a function EC_GROUP_check_discriminant() (defined via
5595 EC_METHOD) that verifies that the curve discriminant is non-zero.
5597 Add a function EC_GROUP_check() that makes some sanity tests
5598 on a EC_GROUP, its generator and order. This includes
5599 EC_GROUP_check_discriminant().
5600 [Nils Larsch <nla@trustcenter.de>]
5602 *) Add ECDSA in new directory crypto/ecdsa/.
5604 Add applications 'openssl ecparam' and 'openssl ecdsa'
5605 (these are based on 'openssl dsaparam' and 'openssl dsa').
5607 ECDSA support is also included in various other files across the
5608 library. Most notably,
5609 - 'openssl req' now has a '-newkey ecdsa:file' option;
5610 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5611 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5612 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5613 them suitable for ECDSA where domain parameters must be
5614 extracted before the specific public key;
5615 - ECDSA engine support has been added.
5616 [Nils Larsch <nla@trustcenter.de>]
5618 *) Include some named elliptic curves, and add OIDs from X9.62,
5619 SECG, and WAP/WTLS. Each curve can be obtained from the new
5621 EC_GROUP_new_by_curve_name(),
5622 and the list of available named curves can be obtained with
5623 EC_get_builtin_curves().
5624 Also add a 'curve_name' member to EC_GROUP objects, which can be
5626 EC_GROUP_set_curve_name()
5627 EC_GROUP_get_curve_name()
5628 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5630 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5631 was actually never needed) and in BN_mul(). The removal in BN_mul()
5632 required a small change in bn_mul_part_recursive() and the addition
5633 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5634 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5635 bn_sub_words() and bn_add_words() except they take arrays with
5639 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5641 *) Cleanse PEM buffers before freeing them since they may contain
5643 [Benjamin Bennett <ben@psc.edu>]
5645 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5646 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5647 authentication-only ciphersuites.
5650 *) Since AES128 and AES256 share a single mask bit in the logic of
5651 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5652 kludge to work properly if AES128 is available and AES256 isn't.
5655 *) Expand security boundary to match 1.1.1 module.
5658 *) Remove redundant features: hash file source, editing of test vectors
5659 modify fipsld to use external fips_premain.c signature.
5662 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5663 run algorithm test programs.
5666 *) Make algorithm test programs more tolerant of whitespace.
5669 *) Have SSL/TLS server implementation tolerate "mismatched" record
5670 protocol version while receiving ClientHello even if the
5671 ClientHello is fragmented. (The server can't insist on the
5672 particular protocol version it has chosen before the ServerHello
5673 message has informed the client about his choice.)
5676 *) Load error codes if they are not already present instead of using a
5677 static variable. This allows them to be cleanly unloaded and reloaded.
5680 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5682 *) Introduce limits to prevent malicious keys being able to
5683 cause a denial of service. (CVE-2006-2940)
5684 [Steve Henson, Bodo Moeller]
5686 *) Fix ASN.1 parsing of certain invalid structures that can result
5687 in a denial of service. (CVE-2006-2937) [Steve Henson]
5689 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5690 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5692 *) Fix SSL client code which could crash if connecting to a
5693 malicious SSLv2 server. (CVE-2006-4343)
5694 [Tavis Ormandy and Will Drewry, Google Security Team]
5696 *) Change ciphersuite string processing so that an explicit
5697 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5698 will no longer include "AES128-SHA"), and any other similar
5699 ciphersuite (same bitmap) from *other* protocol versions (so that
5700 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5701 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5702 changes from 0.9.8b and 0.9.8d.
5705 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5707 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5708 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5710 *) Change the Unix randomness entropy gathering to use poll() when
5711 possible instead of select(), since the latter has some
5712 undesirable limitations.
5713 [Darryl Miles via Richard Levitte and Bodo Moeller]
5715 *) Disable rogue ciphersuites:
5717 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5718 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5719 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5721 The latter two were purportedly from
5722 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5725 Also deactive the remaining ciphersuites from
5726 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5727 unofficial, and the ID has long expired.
5730 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5731 dual-core machines) and other potential thread-safety issues.
5734 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5736 *) Adapt fipsld and the build system to link against the validated FIPS
5737 module in FIPS mode.
5740 *) Fixes for VC++ 2005 build under Windows.
5743 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5744 from a Windows bash shell such as MSYS. It is autodetected from the
5745 "config" script when run from a VC++ environment. Modify standard VC++
5746 build to use fipscanister.o from the GNU make build.
5749 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5751 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5752 The value now differs depending on if you build for FIPS or not.
5753 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5754 safely run with a non-FIPSed libcrypto, as it may crash because of
5755 the difference induced by this change.
5758 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5760 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5761 (part of SSL_OP_ALL). This option used to disable the
5762 countermeasure against man-in-the-middle protocol-version
5763 rollback in the SSL 2.0 server implementation, which is a bad
5764 idea. (CVE-2005-2969)
5766 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5767 for Information Security, National Institute of Advanced Industrial
5768 Science and Technology [AIST], Japan)]
5770 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5771 mainly for FIPS compliance and not fully integrated at this stage.
5774 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5775 the exponentiation using a fixed-length exponent. (Otherwise,
5776 the information leaked through timing could expose the secret key
5777 after many signatures; cf. Bleichenbacher's attack on DSA with
5781 *) Make a new fixed-window mod_exp implementation the default for
5782 RSA, DSA, and DH private-key operations so that the sequence of
5783 squares and multiplies and the memory access pattern are
5784 independent of the particular secret key. This will mitigate
5785 cache-timing and potential related attacks.
5787 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5788 and this is automatically used by BN_mod_exp_mont() if the new flag
5789 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5790 will use this BN flag for private exponents unless the flag
5791 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5792 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5794 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5796 *) Change the client implementation for SSLv23_method() and
5797 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5798 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5799 (Previously, the SSL 2.0 backwards compatible Client Hello
5800 message format would be used even with SSL_OP_NO_SSLv2.)
5803 *) Add support for smime-type MIME parameter in S/MIME messages which some
5807 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5808 a threadsafe manner. Modify rsa code to use new function and add calls
5809 to dsa and dh code (which had race conditions before).
5812 *) Include the fixed error library code in the C error file definitions
5813 instead of fixing them up at runtime. This keeps the error code
5814 structures constant.
5817 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5819 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5822 *) Fixes for newer kerberos headers. NB: the casts are needed because
5823 the 'length' field is signed on one version and unsigned on another
5824 with no (?) obvious way to tell the difference, without these VC++
5825 complains. Also the "definition" of FAR (blank) is no longer included
5826 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5827 some needed definitions.
5830 *) Undo Cygwin change.
5833 *) Added support for proxy certificates according to RFC 3820.
5834 Because they may be a security thread to unaware applications,
5835 they must be explicitly allowed in run-time. See
5836 docs/HOWTO/proxy_certificates.txt for further information.
5839 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5841 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5842 server and client random values. Previously
5843 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5844 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5846 This change has negligible security impact because:
5848 1. Server and client random values still have 24 bytes of pseudo random
5851 2. Server and client random values are sent in the clear in the initial
5854 3. The master secret is derived using the premaster secret (48 bytes in
5855 size for static RSA ciphersuites) as well as client server and random
5858 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5861 [Stephen Henson, reported by UK NISCC]
5863 *) Use Windows randomness collection on Cygwin.
5866 *) Fix hang in EGD/PRNGD query when communication socket is closed
5867 prematurely by EGD/PRNGD.
5868 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5870 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5873 *) Back-port of selected performance improvements from development
5874 branch, as well as improved support for PowerPC platforms.
5877 *) Add lots of checks for memory allocation failure, error codes to indicate
5878 failure and freeing up memory if a failure occurs.
5879 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5881 *) Add new -passin argument to dgst.
5884 *) Perform some character comparisons of different types in X509_NAME_cmp:
5885 this is needed for some certificates that re-encode DNs into UTF8Strings
5886 (in violation of RFC3280) and can't or won't issue name rollover
5890 *) Make an explicit check during certificate validation to see that
5891 the CA setting in each certificate on the chain is correct. As a
5892 side effect always do the following basic checks on extensions,
5893 not just when there's an associated purpose to the check:
5895 - if there is an unhandled critical extension (unless the user
5896 has chosen to ignore this fault)
5897 - if the path length has been exceeded (if one is set at all)
5898 - that certain extensions fit the associated purpose (if one has
5902 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5904 *) Avoid a race condition when CRLs are checked in a multi threaded
5905 environment. This would happen due to the reordering of the revoked
5906 entries during signature checking and serial number lookup. Now the
5907 encoding is cached and the serial number sort performed under a lock.
5908 Add new STACK function sk_is_sorted().
5911 *) Add Delta CRL to the extension code.
5914 *) Various fixes to s3_pkt.c so alerts are sent properly.
5915 [David Holmes <d.holmes@f5.com>]
5917 *) Reduce the chances of duplicate issuer name and serial numbers (in
5918 violation of RFC3280) using the OpenSSL certificate creation utilities.
5919 This is done by creating a random 64 bit value for the initial serial
5920 number when a serial number file is created or when a self signed
5921 certificate is created using 'openssl req -x509'. The initial serial
5922 number file is created using 'openssl x509 -next_serial' in CA.pl
5923 rather than being initialized to 1.
5926 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5928 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5929 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5930 [Joe Orton, Steve Henson]
5932 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5934 [Joe Orton, Steve Henson]
5936 *) Make it possible to have multiple active certificates with the same
5937 subject in the CA index file. This is done only if the keyword
5938 'unique_subject' is set to 'no' in the main CA section (default
5939 if 'CA_default') of the configuration file. The value is saved
5940 with the database itself in a separate index attribute file,
5941 named like the index file with '.attr' appended to the name.
5944 *) X509 verify fixes. Disable broken certificate workarounds when
5945 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5946 keyUsage extension present. Don't accept CRLs with unhandled critical
5947 extensions: since verify currently doesn't process CRL extensions this
5948 rejects a CRL with *any* critical extensions. Add new verify error codes
5952 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5953 A clarification of RFC2560 will require the use of OCTET STRINGs and
5954 some implementations cannot handle the current raw format. Since OpenSSL
5955 copies and compares OCSP nonces as opaque blobs without any attempt at
5956 parsing them this should not create any compatibility issues.
5959 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5960 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5961 this HMAC (and other) operations are several times slower than OpenSSL
5965 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5966 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5968 *) Use the correct content when signing type "other".
5971 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5973 *) Fix various bugs revealed by running the NISCC test suite:
5975 Stop out of bounds reads in the ASN1 code when presented with
5976 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5978 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5980 If verify callback ignores invalid public key errors don't try to check
5981 certificate signature with the NULL public key.
5985 *) New -ignore_err option in ocsp application to stop the server
5986 exiting on the first error in a request.
5989 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5990 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5994 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5995 extra data after the compression methods not only for TLS 1.0
5996 but also for SSL 3.0 (as required by the specification).
5997 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5999 *) Change X509_certificate_type() to mark the key as exported/exportable
6000 when it's 512 *bits* long, not 512 bytes.
6003 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6004 blocks during encryption.
6007 *) Various fixes to base64 BIO and non blocking I/O. On write
6008 flushes were not handled properly if the BIO retried. On read
6009 data was not being buffered properly and had various logic bugs.
6010 This also affects blocking I/O when the data being decoded is a
6014 *) Various S/MIME bugfixes and compatibility changes:
6015 output correct application/pkcs7 MIME type if
6016 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6017 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6018 of files as .eml work). Correctly handle very long lines in MIME
6022 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6024 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6025 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6026 a protocol version number mismatch like a decryption error
6027 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6030 *) Turn on RSA blinding by default in the default implementation
6031 to avoid a timing attack. Applications that don't want it can call
6032 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6033 They would be ill-advised to do so in most cases.
6034 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6036 *) Change RSA blinding code so that it works when the PRNG is not
6037 seeded (in this case, the secret RSA exponent is abused as
6038 an unpredictable seed -- if it is not unpredictable, there
6039 is no point in blinding anyway). Make RSA blinding thread-safe
6040 by remembering the creator's thread ID in rsa->blinding and
6041 having all other threads use local one-time blinding factors
6042 (this requires more computation than sharing rsa->blinding, but
6043 avoids excessive locking; and if an RSA object is not shared
6044 between threads, blinding will still be very fast).
6047 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6048 ENGINE as defaults for all supported algorithms irrespective of
6049 the 'flags' parameter. 'flags' is now honoured, so applications
6050 should make sure they are passing it correctly.
6053 *) Target "mingw" now allows native Windows code to be generated in
6054 the Cygwin environment as well as with the MinGW compiler.
6057 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6059 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6060 via timing by performing a MAC computation even if incorrect
6061 block cipher padding has been found. This is a countermeasure
6062 against active attacks where the attacker has to distinguish
6063 between bad padding and a MAC verification error. (CVE-2003-0078)
6065 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6066 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6067 Martin Vuagnoux (EPFL, Ilion)]
6069 *) Make the no-err option work as intended. The intention with no-err
6070 is not to have the whole error stack handling routines removed from
6071 libcrypto, it's only intended to remove all the function name and
6072 reason texts, thereby removing some of the footprint that may not
6073 be interesting if those errors aren't displayed anyway.
6075 NOTE: it's still possible for any application or module to have it's
6076 own set of error texts inserted. The routines are there, just not
6077 used by default when no-err is given.
6080 *) Add support for FreeBSD on IA64.
6081 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6083 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6084 Kerberos function mit_des_cbc_cksum(). Before this change,
6085 the value returned by DES_cbc_cksum() was like the one from
6086 mit_des_cbc_cksum(), except the bytes were swapped.
6087 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6089 *) Allow an application to disable the automatic SSL chain building.
6090 Before this a rather primitive chain build was always performed in
6091 ssl3_output_cert_chain(): an application had no way to send the
6092 correct chain if the automatic operation produced an incorrect result.
6094 Now the chain builder is disabled if either:
6096 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6098 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6100 The reasoning behind this is that an application would not want the
6101 auto chain building to take place if extra chain certificates are
6102 present and it might also want a means of sending no additional
6103 certificates (for example the chain has two certificates and the
6107 *) Add the possibility to build without the ENGINE framework.
6108 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6110 *) Under Win32 gmtime() can return NULL: check return value in
6111 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6114 *) DSA routines: under certain error conditions uninitialized BN objects
6115 could be freed. Solution: make sure initialization is performed early
6116 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6117 Nils Larsch <nla@trustcenter.de> via PR#459)
6120 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6121 checked on reconnect on the client side, therefore session resumption
6122 could still fail with a "ssl session id is different" error. This
6123 behaviour is masked when SSL_OP_ALL is used due to
6124 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6125 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6126 followup to PR #377.
6129 *) IA-32 assembler support enhancements: unified ELF targets, support
6130 for SCO/Caldera platforms, fix for Cygwin shared build.
6133 *) Add support for FreeBSD on sparc64. As a consequence, support for
6134 FreeBSD on non-x86 processors is separate from x86 processors on
6135 the config script, much like the NetBSD support.
6136 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6138 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6140 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6143 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6144 code (06) was taken as the first octet of the session ID and the last
6145 octet was ignored consequently. As a result SSLv2 client side session
6146 caching could not have worked due to the session ID mismatch between
6148 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6152 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6153 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6157 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6158 seems that in spite of existing for more than a year, many application
6159 author have done nothing to provide the necessary callbacks, which
6160 means that this particular engine will not work properly anywhere.
6161 This is a very unfortunate situation which forces us, in the name
6162 of usability, to give the hw_ncipher.c a static lock, which is part
6164 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6165 appear in 0.9.8 or later. We EXPECT application authors to have
6166 dealt properly with this when 0.9.8 is released (unless we actually
6167 make such changes in the libcrypto locking code that changes will
6168 have to be made anyway).
6171 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6172 octets have been read, EOF or an error occurs. Without this change
6173 some truncated ASN1 structures will not produce an error.
6176 *) Disable Heimdal support, since it hasn't been fully implemented.
6177 Still give the possibility to force the use of Heimdal, but with
6178 warnings and a request that patches get sent to openssl-dev.
6181 *) Add the VC-CE target, introduce the WINCE sysname, and add
6182 INSTALL.WCE and appropriate conditionals to make it build.
6183 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6185 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6186 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6187 edit numbers of the version.
6188 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6190 *) Introduce safe string copy and catenation functions
6191 (BUF_strlcpy() and BUF_strlcat()).
6192 [Ben Laurie (CHATS) and Richard Levitte]
6194 *) Avoid using fixed-size buffers for one-line DNs.
6195 [Ben Laurie (CHATS)]
6197 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6198 resizing buffers containing secrets, and use where appropriate.
6199 [Ben Laurie (CHATS)]
6201 *) Avoid using fixed size buffers for configuration file location.
6202 [Ben Laurie (CHATS)]
6204 *) Avoid filename truncation for various CA files.
6205 [Ben Laurie (CHATS)]
6207 *) Use sizeof in preference to magic numbers.
6208 [Ben Laurie (CHATS)]
6210 *) Avoid filename truncation in cert requests.
6211 [Ben Laurie (CHATS)]
6213 *) Add assertions to check for (supposedly impossible) buffer
6215 [Ben Laurie (CHATS)]
6217 *) Don't cache truncated DNS entries in the local cache (this could
6218 potentially lead to a spoofing attack).
6219 [Ben Laurie (CHATS)]
6221 *) Fix various buffers to be large enough for hex/decimal
6222 representations in a platform independent manner.
6223 [Ben Laurie (CHATS)]
6225 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6226 resizing buffers containing secrets, and use where appropriate.
6227 [Ben Laurie (CHATS)]
6229 *) Add BIO_indent() to avoid much slightly worrying code to do
6231 [Ben Laurie (CHATS)]
6233 *) Convert sprintf()/BIO_puts() to BIO_printf().
6234 [Ben Laurie (CHATS)]
6236 *) buffer_gets() could terminate with the buffer only half
6238 [Ben Laurie (CHATS)]
6240 *) Add assertions to prevent user-supplied crypto functions from
6241 overflowing internal buffers by having large block sizes, etc.
6242 [Ben Laurie (CHATS)]
6244 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6246 [Ben Laurie (CHATS)]
6248 *) Eliminate unused copy of key in RC4.
6249 [Ben Laurie (CHATS)]
6251 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6252 [Ben Laurie (CHATS)]
6254 *) Fix off-by-one error in EGD path.
6255 [Ben Laurie (CHATS)]
6257 *) If RANDFILE path is too long, ignore instead of truncating.
6258 [Ben Laurie (CHATS)]
6260 *) Eliminate unused and incorrectly sized X.509 structure
6262 [Ben Laurie (CHATS)]
6264 *) Eliminate unused and dangerous function knumber().
6265 [Ben Laurie (CHATS)]
6267 *) Eliminate unused and dangerous structure, KSSL_ERR.
6268 [Ben Laurie (CHATS)]
6270 *) Protect against overlong session ID context length in an encoded
6271 session object. Since these are local, this does not appear to be
6273 [Ben Laurie (CHATS)]
6275 *) Change from security patch (see 0.9.6e below) that did not affect
6276 the 0.9.6 release series:
6278 Remote buffer overflow in SSL3 protocol - an attacker could
6279 supply an oversized master key in Kerberos-enabled versions.
6281 [Ben Laurie (CHATS)]
6283 *) Change the SSL kerb5 codes to match RFC 2712.
6286 *) Make -nameopt work fully for req and add -reqopt switch.
6287 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6289 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6290 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6292 *) Make sure tests can be performed even if the corresponding algorithms
6293 have been removed entirely. This was also the last step to make
6294 OpenSSL compilable with DJGPP under all reasonable conditions.
6295 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6297 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6298 to allow version independent disabling of normally unselected ciphers,
6299 which may be activated as a side-effect of selecting a single cipher.
6301 (E.g., cipher list string "RSA" enables ciphersuites that are left
6302 out of "ALL" because they do not provide symmetric encryption.
6303 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6304 [Lutz Jaenicke, Bodo Moeller]
6306 *) Add appropriate support for separate platform-dependent build
6307 directories. The recommended way to make a platform-dependent
6308 build directory is the following (tested on Linux), maybe with
6311 # Place yourself outside of the OpenSSL source tree. In
6312 # this example, the environment variable OPENSSL_SOURCE
6313 # is assumed to contain the absolute OpenSSL source directory.
6314 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6315 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6316 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6317 mkdir -p `dirname $F`
6318 ln -s $OPENSSL_SOURCE/$F $F
6321 To be absolutely sure not to disturb the source tree, a "make clean"
6322 is a good thing. If it isn't successful, don't worry about it,
6323 it probably means the source directory is very clean.
6326 *) Make sure any ENGINE control commands make local copies of string
6327 pointers passed to them whenever necessary. Otherwise it is possible
6328 the caller may have overwritten (or deallocated) the original string
6329 data when a later ENGINE operation tries to use the stored values.
6330 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6332 *) Improve diagnostics in file reading and command-line digests.
6333 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6335 *) Add AES modes CFB and OFB to the object database. Correct an
6336 error in AES-CFB decryption.
6339 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6340 allows existing EVP_CIPHER_CTX structures to be reused after
6341 calling EVP_*Final(). This behaviour is used by encryption
6342 BIOs and some applications. This has the side effect that
6343 applications must explicitly clean up cipher contexts with
6344 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6347 *) Check the values of dna and dnb in bn_mul_recursive before calling
6348 bn_mul_comba (a non zero value means the a or b arrays do not contain
6349 n2 elements) and fallback to bn_mul_normal if either is not zero.
6352 *) Fix escaping of non-ASCII characters when using the -subj option
6353 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6356 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6357 form for "surname", serialNumber has no short form.
6358 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6359 therefore remove "mail" short name for "internet 7".
6360 The OID for unique identifiers in X509 certificates is
6361 x500UniqueIdentifier, not uniqueIdentifier.
6362 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6365 *) Add an "init" command to the ENGINE config module and auto initialize
6366 ENGINEs. Without any "init" command the ENGINE will be initialized
6367 after all ctrl commands have been executed on it. If init=1 the
6368 ENGINE is initailized at that point (ctrls before that point are run
6369 on the uninitialized ENGINE and after on the initialized one). If
6370 init=0 then the ENGINE will not be iniatialized at all.
6373 *) Fix the 'app_verify_callback' interface so that the user-defined
6374 argument is actually passed to the callback: In the
6375 SSL_CTX_set_cert_verify_callback() prototype, the callback
6376 declaration has been changed from
6379 int (*cb)(X509_STORE_CTX *,void *);
6380 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6381 i=s->ctx->app_verify_callback(&ctx)
6382 has been changed into
6383 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6385 To update applications using SSL_CTX_set_cert_verify_callback(),
6386 a dummy argument can be added to their callback functions.
6387 [D. K. Smetters <smetters@parc.xerox.com>]
6389 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6390 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6392 *) Add and OPENSSL_LOAD_CONF define which will cause
6393 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6394 This allows older applications to transparently support certain
6395 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6396 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6397 load the config file and OPENSSL_add_all_algorithms_conf() which will
6398 always load it have also been added.
6401 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6402 Adjust NIDs and EVP layer.
6403 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6405 *) Config modules support in openssl utility.
6407 Most commands now load modules from the config file,
6408 though in a few (such as version) this isn't done
6409 because it couldn't be used for anything.
6411 In the case of ca and req the config file used is
6412 the same as the utility itself: that is the -config
6413 command line option can be used to specify an
6417 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6418 use "openssl_conf" if filename is NULL use default openssl config file.
6421 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6422 config section name. Add a new flag to tolerate a missing config file
6423 and move code to CONF_modules_load_file().
6426 *) Support for crypto accelerator cards from Accelerated Encryption
6427 Processing, www.aep.ie. (Use engine 'aep')
6428 The support was copied from 0.9.6c [engine] and adapted/corrected
6429 to work with the new engine framework.
6430 [AEP Inc. and Richard Levitte]
6432 *) Support for SureWare crypto accelerator cards from Baltimore
6433 Technologies. (Use engine 'sureware')
6434 The support was copied from 0.9.6c [engine] and adapted
6435 to work with the new engine framework.
6438 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6439 make the newer ENGINE framework commands for the CHIL engine work.
6440 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6442 *) Make it possible to produce shared libraries on ReliantUNIX.
6443 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6445 *) Add the configuration target debug-linux-ppro.
6446 Make 'openssl rsa' use the general key loading routines
6447 implemented in apps.c, and make those routines able to
6448 handle the key format FORMAT_NETSCAPE and the variant
6450 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6452 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6453 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6455 *) Add -keyform to rsautl, and document -engine.
6456 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6458 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6459 BIO_R_NO_SUCH_FILE error code rather than the generic
6460 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6463 *) Add new functions
6465 ERR_peek_last_error_line
6466 ERR_peek_last_error_line_data.
6467 These are similar to
6470 ERR_peek_error_line_data,
6471 but report on the latest error recorded rather than the first one
6472 still in the error queue.
6473 [Ben Laurie, Bodo Moeller]
6475 *) default_algorithms option in ENGINE config module. This allows things
6477 default_algorithms = ALL
6478 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6481 *) Preliminary ENGINE config module.
6484 *) New experimental application configuration code.
6487 *) Change the AES code to follow the same name structure as all other
6488 symmetric ciphers, and behave the same way. Move everything to
6489 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6490 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6492 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6493 [Ben Laurie and Theo de Raadt]
6495 *) Add option to output public keys in req command.
6496 [Massimiliano Pala madwolf@openca.org]
6498 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6499 (up to about 10% better than before for P-192 and P-224).
6502 *) New functions/macros
6504 SSL_CTX_set_msg_callback(ctx, cb)
6505 SSL_CTX_set_msg_callback_arg(ctx, arg)
6506 SSL_set_msg_callback(ssl, cb)
6507 SSL_set_msg_callback_arg(ssl, arg)
6509 to request calling a callback function
6511 void cb(int write_p, int version, int content_type,
6512 const void *buf, size_t len, SSL *ssl, void *arg)
6514 whenever a protocol message has been completely received
6515 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6516 protocol version according to which the SSL library interprets
6517 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6518 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6519 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6520 specification (change_cipher_spec(20), alert(21), handshake(22)).
6521 'buf' and 'len' point to the actual message, 'ssl' to the
6522 SSL object, and 'arg' is the application-defined value set by
6523 SSL[_CTX]_set_msg_callback_arg().
6525 'openssl s_client' and 'openssl s_server' have new '-msg' options
6526 to enable a callback that displays all protocol messages.
6529 *) Change the shared library support so shared libraries are built as
6530 soon as the corresponding static library is finished, and thereby get
6531 openssl and the test programs linked against the shared library.
6532 This still only happens when the keyword "shard" has been given to
6533 the configuration scripts.
6535 NOTE: shared library support is still an experimental thing, and
6536 backward binary compatibility is still not guaranteed.
6537 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6539 *) Add support for Subject Information Access extension.
6540 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6542 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6543 additional bytes when new memory had to be allocated, not just
6544 when reusing an existing buffer.
6547 *) New command line and configuration option 'utf8' for the req command.
6548 This allows field values to be specified as UTF8 strings.
6551 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6552 runs for the former and machine-readable output for the latter.
6555 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6556 of the e-mail address in the DN (i.e., it will go into a certificate
6557 extension only). The new configuration file option 'email_in_dn = no'
6558 has the same effect.
6559 [Massimiliano Pala madwolf@openca.org]
6561 *) Change all functions with names starting with des_ to be starting
6562 with DES_ instead. Add wrappers that are compatible with libdes,
6563 but are named _ossl_old_des_*. Finally, add macros that map the
6564 des_* symbols to the corresponding _ossl_old_des_* if libdes
6565 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6566 desired, the des_* symbols will be mapped to DES_*, with one
6569 Since we provide two compatibility mappings, the user needs to
6570 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6571 compatibility is desired. The default (i.e., when that macro
6572 isn't defined) is OpenSSL 0.9.6c compatibility.
6574 There are also macros that enable and disable the support of old
6575 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6576 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6577 are defined, the default will apply: to support the old des routines.
6579 In either case, one must include openssl/des.h to get the correct
6580 definitions. Do not try to just include openssl/des_old.h, that
6583 NOTE: This is a major break of an old API into a new one. Software
6584 authors are encouraged to switch to the DES_ style functions. Some
6585 time in the future, des_old.h and the libdes compatibility functions
6586 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6587 default), and then completely removed.
6590 *) Test for certificates which contain unsupported critical extensions.
6591 If such a certificate is found during a verify operation it is
6592 rejected by default: this behaviour can be overridden by either
6593 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6594 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6595 X509_supported_extension() has also been added which returns 1 if a
6596 particular extension is supported.
6599 *) Modify the behaviour of EVP cipher functions in similar way to digests
6600 to retain compatibility with existing code.
6603 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6604 compatibility with existing code. In particular the 'ctx' parameter does
6605 not have to be to be initialized before the call to EVP_DigestInit() and
6606 it is tidied up after a call to EVP_DigestFinal(). New function
6607 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6608 EVP_MD_CTX_copy() changed to not require the destination to be
6609 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6610 requires the destination to be valid.
6612 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6613 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6616 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6617 so that complete 'Handshake' protocol structures are kept in memory
6618 instead of overwriting 'msg_type' and 'length' with 'body' data.
6621 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6622 [Massimo Santin via Richard Levitte]
6624 *) Major restructuring to the underlying ENGINE code. This includes
6625 reduction of linker bloat, separation of pure "ENGINE" manipulation
6626 (initialisation, etc) from functionality dealing with implementations
6627 of specific crypto iterfaces. This change also introduces integrated
6628 support for symmetric ciphers and digest implementations - so ENGINEs
6629 can now accelerate these by providing EVP_CIPHER and EVP_MD
6630 implementations of their own. This is detailed in crypto/engine/README
6631 as it couldn't be adequately described here. However, there are a few
6632 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6633 were changed in the original introduction of ENGINE code have now
6634 reverted back - the hooking from this code to ENGINE is now a good
6635 deal more passive and at run-time, operations deal directly with
6636 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6637 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6638 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6639 they were not being used by the framework as there is no concept of a
6640 BIGNUM_METHOD and they could not be generalised to the new
6641 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6642 ENGINE_cpy() has been removed as it cannot be consistently defined in
6646 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6649 *) Change mkdef.pl to sort symbols that get the same entry number,
6650 and make sure the automatically generated functions ERR_load_*
6651 become part of libeay.num as well.
6654 *) New function SSL_renegotiate_pending(). This returns true once
6655 renegotiation has been requested (either SSL_renegotiate() call
6656 or HelloRequest/ClientHello received from the peer) and becomes
6657 false once a handshake has been completed.
6658 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6659 sends a HelloRequest, but does not ensure that a handshake takes
6660 place. SSL_renegotiate_pending() is useful for checking if the
6661 client has followed the request.)
6664 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6665 By default, clients may request session resumption even during
6666 renegotiation (if session ID contexts permit); with this option,
6667 session resumption is possible only in the first handshake.
6669 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6670 more bits available for options that should not be part of
6671 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6674 *) Add some demos for certificate and certificate request creation.
6677 *) Make maximum certificate chain size accepted from the peer application
6678 settable (SSL*_get/set_max_cert_list()), as proposed by
6679 "Douglas E. Engert" <deengert@anl.gov>.
6682 *) Add support for shared libraries for Unixware-7
6683 (Boyd Lynn Gerber <gerberb@zenez.com>).
6686 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6687 be done prior to destruction. Use this to unload error strings from
6688 ENGINEs that load their own error strings. NB: This adds two new API
6689 functions to "get" and "set" this destroy handler in an ENGINE.
6692 *) Alter all existing ENGINE implementations (except "openssl" and
6693 "openbsd") to dynamically instantiate their own error strings. This
6694 makes them more flexible to be built both as statically-linked ENGINEs
6695 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6696 Also, add stub code to each that makes building them as self-contained
6697 shared-libraries easier (see README.ENGINE).
6700 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6701 implementations into applications that are completely implemented in
6702 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6703 commands that can be used to configure what shared-library to load and
6704 to control aspects of the way it is handled. Also, made an update to
6705 the README.ENGINE file that brings its information up-to-date and
6706 provides some information and instructions on the "dynamic" ENGINE
6707 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6710 *) Make it possible to unload ranges of ERR strings with a new
6711 "ERR_unload_strings" function.
6714 *) Add a copy() function to EVP_MD.
6717 *) Make EVP_MD routines take a context pointer instead of just the
6718 md_data void pointer.
6721 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6722 that the digest can only process a single chunk of data
6723 (typically because it is provided by a piece of
6724 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6725 is only going to provide a single chunk of data, and hence the
6726 framework needn't accumulate the data for oneshot drivers.
6729 *) As with "ERR", make it possible to replace the underlying "ex_data"
6730 functions. This change also alters the storage and management of global
6731 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6732 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6733 index counters. The API functions that use this state have been changed
6734 to take a "class_index" rather than pointers to the class's local STACK
6735 and counter, and there is now an API function to dynamically create new
6736 classes. This centralisation allows us to (a) plug a lot of the
6737 thread-safety problems that existed, and (b) makes it possible to clean
6738 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6739 such data would previously have always leaked in application code and
6740 workarounds were in place to make the memory debugging turn a blind eye
6741 to it. Application code that doesn't use this new function will still
6742 leak as before, but their memory debugging output will announce it now
6743 rather than letting it slide.
6745 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6746 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6747 has a return value to indicate success or failure.
6750 *) Make it possible to replace the underlying "ERR" functions such that the
6751 global state (2 LHASH tables and 2 locks) is only used by the "default"
6752 implementation. This change also adds two functions to "get" and "set"
6753 the implementation prior to it being automatically set the first time
6754 any other ERR function takes place. Ie. an application can call "get",
6755 pass the return value to a module it has just loaded, and that module
6756 can call its own "set" function using that value. This means the
6757 module's "ERR" operations will use (and modify) the error state in the
6758 application and not in its own statically linked copy of OpenSSL code.
6761 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6762 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6763 the operation, and provides a more encapsulated way for external code
6764 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6765 to use these functions rather than manually incrementing the counts.
6767 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6770 *) Add EVP test program.
6773 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6776 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6777 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6778 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6779 These allow a CRL to be built without having to access X509_CRL fields
6780 directly. Modify 'ca' application to use new functions.
6783 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6784 bug workarounds. Rollback attack detection is a security feature.
6785 The problem will only arise on OpenSSL servers when TLSv1 is not
6786 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6787 Software authors not wanting to support TLSv1 will have special reasons
6788 for their choice and can explicitly enable this option.
6789 [Bodo Moeller, Lutz Jaenicke]
6791 *) Rationalise EVP so it can be extended: don't include a union of
6792 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6793 (similar to those existing for EVP_CIPHER_CTX).
6798 EVP_MD_CTX_init(&md); /* new function call */
6799 EVP_DigestInit(&md, EVP_sha1());
6800 EVP_DigestUpdate(&md, in, len);
6801 EVP_DigestFinal(&md, out, NULL);
6802 EVP_MD_CTX_cleanup(&md); /* new function call */
6806 *) Make DES key schedule conform to the usual scheme, as well as
6807 correcting its structure. This means that calls to DES functions
6808 now have to pass a pointer to a des_key_schedule instead of a
6809 plain des_key_schedule (which was actually always a pointer
6812 des_key_schedule ks;
6814 des_set_key_checked(..., &ks);
6815 des_ncbc_encrypt(..., &ks, ...);
6817 (Note that a later change renames 'des_...' into 'DES_...'.)
6820 *) Initial reduction of linker bloat: the use of some functions, such as
6821 PEM causes large amounts of unused functions to be linked in due to
6822 poor organisation. For example pem_all.c contains every PEM function
6823 which has a knock on effect of linking in large amounts of (unused)
6824 ASN1 code. Grouping together similar functions and splitting unrelated
6825 functions prevents this.
6828 *) Cleanup of EVP macros.
6831 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6832 correct _ecb suffix.
6835 *) Add initial OCSP responder support to ocsp application. The
6836 revocation information is handled using the text based index
6837 use by the ca application. The responder can either handle
6838 requests generated internally, supplied in files (for example
6839 via a CGI script) or using an internal minimal server.
6842 *) Add configuration choices to get zlib compression for TLS.
6845 *) Changes to Kerberos SSL for RFC 2712 compliance:
6846 1. Implemented real KerberosWrapper, instead of just using
6847 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6848 2. Implemented optional authenticator field of KerberosWrapper.
6850 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6851 and authenticator structs; see crypto/krb5/.
6853 Generalized Kerberos calls to support multiple Kerberos libraries.
6854 [Vern Staats <staatsvr@asc.hpc.mil>,
6855 Jeffrey Altman <jaltman@columbia.edu>
6856 via Richard Levitte]
6858 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6859 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6860 values for each of the key sizes rather than having just
6861 parameters (and 'speed' generating keys each time).
6864 *) Speed up EVP routines.
6867 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6868 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6869 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6870 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6872 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6873 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6874 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6877 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6879 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6882 *) Added the OS2-EMX target.
6883 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6885 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6886 to support NCONF routines in extension code. New function CONF_set_nconf()
6887 to allow functions which take an NCONF to also handle the old LHASH
6888 structure: this means that the old CONF compatible routines can be
6889 retained (in particular wrt extensions) without having to duplicate the
6890 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6893 *) Enhance the general user interface with mechanisms for inner control
6894 and with possibilities to have yes/no kind of prompts.
6897 *) Change all calls to low level digest routines in the library and
6898 applications to use EVP. Add missing calls to HMAC_cleanup() and
6899 don't assume HMAC_CTX can be copied using memcpy().
6900 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6902 *) Add the possibility to control engines through control names but with
6903 arbitrary arguments instead of just a string.
6904 Change the key loaders to take a UI_METHOD instead of a callback
6905 function pointer. NOTE: this breaks binary compatibility with earlier
6906 versions of OpenSSL [engine].
6907 Adapt the nCipher code for these new conditions and add a card insertion
6911 *) Enhance the general user interface with mechanisms to better support
6912 dialog box interfaces, application-defined prompts, the possibility
6913 to use defaults (for example default passwords from somewhere else)
6914 and interrupts/cancellations.
6917 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6918 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6921 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6922 tidy up some unnecessarily weird code in 'sk_new()').
6923 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6925 *) Change the key loading routines for ENGINEs to use the same kind
6926 callback (pem_password_cb) as all other routines that need this
6930 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6931 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6932 than this minimum value is recommended.
6935 *) New random seeder for OpenVMS, using the system process statistics
6936 that are easily reachable.
6939 *) Windows apparently can't transparently handle global
6940 variables defined in DLLs. Initialisations such as:
6942 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6944 won't compile. This is used by the any applications that need to
6945 declare their own ASN1 modules. This was fixed by adding the option
6946 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6947 needed for static libraries under Win32.
6950 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6951 setting of purpose and trust fields. New X509_STORE trust and
6952 purpose functions and tidy up setting in other SSL functions.
6955 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6956 structure. These are inherited by X509_STORE_CTX when it is
6957 initialised. This allows various defaults to be set in the
6958 X509_STORE structure (such as flags for CRL checking and custom
6959 purpose or trust settings) for functions which only use X509_STORE_CTX
6960 internally such as S/MIME.
6962 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6963 trust settings if they are not set in X509_STORE. This allows X509_STORE
6964 purposes and trust (in S/MIME for example) to override any set by default.
6966 Add command line options for CRL checking to smime, s_client and s_server
6970 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6971 are set then the CRL is looked up in the X509_STORE structure and
6972 its validity and signature checked, then if the certificate is found
6973 in the CRL the verify fails with a revoked error.
6975 Various new CRL related callbacks added to X509_STORE_CTX structure.
6977 Command line options added to 'verify' application to support this.
6979 This needs some additional work, such as being able to handle multiple
6980 CRLs with different times, extension based lookup (rather than just
6981 by subject name) and ultimately more complete V2 CRL extension
6985 *) Add a general user interface API (crypto/ui/). This is designed
6986 to replace things like des_read_password and friends (backward
6987 compatibility functions using this new API are provided).
6988 The purpose is to remove prompting functions from the DES code
6989 section as well as provide for prompting through dialog boxes in
6990 a window system and the like.
6993 *) Add "ex_data" support to ENGINE so implementations can add state at a
6994 per-structure level rather than having to store it globally.
6997 *) Make it possible for ENGINE structures to be copied when retrieved by
6998 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6999 This causes the "original" ENGINE structure to act like a template,
7000 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7001 operational state can be localised to each ENGINE structure, despite the
7002 fact they all share the same "methods". New ENGINE structures returned in
7003 this case have no functional references and the return value is the single
7004 structural reference. This matches the single structural reference returned
7005 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7009 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7010 needs to match any other type at all we need to manually clear the
7014 *) Changes to the "openssl engine" utility to include;
7015 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7016 about an ENGINE's available control commands.
7017 - executing control commands from command line arguments using the
7018 '-pre' and '-post' switches. '-post' is only used if '-t' is
7019 specified and the ENGINE is successfully initialised. The syntax for
7020 the individual commands are colon-separated, for example;
7021 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7024 *) New dynamic control command support for ENGINEs. ENGINEs can now
7025 declare their own commands (numbers), names (strings), descriptions,
7026 and input types for run-time discovery by calling applications. A
7027 subset of these commands are implicitly classed as "executable"
7028 depending on their input type, and only these can be invoked through
7029 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7030 can be based on user input, config files, etc). The distinction is
7031 that "executable" commands cannot return anything other than a boolean
7032 result and can only support numeric or string input, whereas some
7033 discoverable commands may only be for direct use through
7034 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7035 pointers, or other custom uses. The "executable" commands are to
7036 support parameterisations of ENGINE behaviour that can be
7037 unambiguously defined by ENGINEs and used consistently across any
7038 OpenSSL-based application. Commands have been added to all the
7039 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7040 control over shared-library paths without source code alterations.
7043 *) Changed all ENGINE implementations to dynamically allocate their
7044 ENGINEs rather than declaring them statically. Apart from this being
7045 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7046 this also allows the implementations to compile without using the
7047 internal engine_int.h header.
7050 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7051 'const' value. Any code that should be able to modify a RAND_METHOD
7052 should already have non-const pointers to it (ie. they should only
7053 modify their own ones).
7056 *) Made a variety of little tweaks to the ENGINE code.
7057 - "atalla" and "ubsec" string definitions were moved from header files
7058 to C code. "nuron" string definitions were placed in variables
7059 rather than hard-coded - allowing parameterisation of these values
7060 later on via ctrl() commands.
7061 - Removed unused "#if 0"'d code.
7062 - Fixed engine list iteration code so it uses ENGINE_free() to release
7063 structural references.
7064 - Constified the RAND_METHOD element of ENGINE structures.
7065 - Constified various get/set functions as appropriate and added
7066 missing functions (including a catch-all ENGINE_cpy that duplicates
7067 all ENGINE values onto a new ENGINE except reference counts/state).
7068 - Removed NULL parameter checks in get/set functions. Setting a method
7069 or function to NULL is a way of cancelling out a previously set
7070 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7071 and doesn't justify the extra error symbols and code.
7072 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7073 flags from engine_int.h to engine.h.
7074 - Changed prototypes for ENGINE handler functions (init(), finish(),
7075 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7078 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7079 to the algorithm using long division. The binary algorithm can be
7080 used only if the modulus is odd. On 32-bit systems, it is faster
7081 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7082 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7083 up to 450 bits. In 64-bit environments, the binary algorithm
7084 appears to be advantageous for much longer moduli; here we use it
7085 for moduli up to 2048 bits.
7088 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7089 could not support the combine flag in choice fields.
7092 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7093 extensions from a certificate request to the certificate.
7096 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7097 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7098 file: this allows the display of the certificate about to be
7099 signed to be customised, to allow certain fields to be included
7100 or excluded and extension details. The old system didn't display
7101 multicharacter strings properly, omitted fields not in the policy
7102 and couldn't display additional details such as extensions.
7105 *) Function EC_POINTs_mul for multiple scalar multiplication
7106 of an arbitrary number of elliptic curve points
7107 \sum scalars[i]*points[i],
7108 optionally including the generator defined for the EC_GROUP:
7109 scalar*generator + \sum scalars[i]*points[i].
7111 EC_POINT_mul is a simple wrapper function for the typical case
7112 that the point list has just one item (besides the optional
7116 *) First EC_METHODs for curves over GF(p):
7118 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7119 operations and provides various method functions that can also
7120 operate with faster implementations of modular arithmetic.
7122 EC_GFp_mont_method() reuses most functions that are part of
7123 EC_GFp_simple_method, but uses Montgomery arithmetic.
7125 [Bodo Moeller; point addition and point doubling
7126 implementation directly derived from source code provided by
7127 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7129 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7130 crypto/ec/ec_lib.c):
7132 Curves are EC_GROUP objects (with an optional group generator)
7133 based on EC_METHODs that are built into the library.
7135 Points are EC_POINT objects based on EC_GROUP objects.
7137 Most of the framework would be able to handle curves over arbitrary
7138 finite fields, but as there are no obvious types for fields other
7139 than GF(p), some functions are limited to that for now.
7142 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7143 that the file contains a complete HTTP response.
7146 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7147 change the def and num file printf format specifier from "%-40sXXX"
7148 to "%-39s XXX". The latter will always guarantee a space after the
7149 field while the former will cause them to run together if the field
7150 is 40 of more characters long.
7153 *) Constify the cipher and digest 'method' functions and structures
7154 and modify related functions to take constant EVP_MD and EVP_CIPHER
7158 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7159 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7162 *) Modify EVP_Digest*() routines so they now return values. Although the
7163 internal software routines can never fail additional hardware versions
7167 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7169 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7170 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7173 ERR_R_NESTED_ASN1_ERROR
7175 ERR_R_MISSING_ASN1_EOS
7176 were 4 .. 9, conflicting with
7177 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7179 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7180 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7182 Add new error code 'ERR_R_INTERNAL_ERROR'.
7185 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7189 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7190 sets the subject name for a new request or supersedes the
7191 subject name in a given request. Formats that can be parsed are
7192 'CN=Some Name, OU=myOU, C=IT'
7194 'CN=Some Name/OU=myOU/C=IT'.
7196 Add options '-batch' and '-verbose' to 'openssl req'.
7197 [Massimiliano Pala <madwolf@hackmasters.net>]
7199 *) Introduce the possibility to access global variables through
7200 functions on platform were that's the best way to handle exporting
7201 global variables in shared libraries. To enable this functionality,
7202 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7203 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7204 is normally done by Configure or something similar).
7206 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7207 in the source file (foo.c) like this:
7209 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7210 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7212 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7213 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7215 OPENSSL_DECLARE_GLOBAL(int,foo);
7216 #define foo OPENSSL_GLOBAL_REF(foo)
7217 OPENSSL_DECLARE_GLOBAL(double,bar);
7218 #define bar OPENSSL_GLOBAL_REF(bar)
7220 The #defines are very important, and therefore so is including the
7221 header file everywhere where the defined globals are used.
7223 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7224 of ASN.1 items, but that structure is a bit different.
7226 The largest change is in util/mkdef.pl which has been enhanced with
7227 better and easier to understand logic to choose which symbols should
7228 go into the Windows .def files as well as a number of fixes and code
7229 cleanup (among others, algorithm keywords are now sorted
7230 lexicographically to avoid constant rewrites).
7233 *) In BN_div() keep a copy of the sign of 'num' before writing the
7234 result to 'rm' because if rm==num the value will be overwritten
7235 and produce the wrong result if 'num' is negative: this caused
7236 problems with BN_mod() and BN_nnmod().
7239 *) Function OCSP_request_verify(). This checks the signature on an
7240 OCSP request and verifies the signer certificate. The signer
7241 certificate is just checked for a generic purpose and OCSP request
7245 *) Add OCSP_check_validity() function to check the validity of OCSP
7246 responses. OCSP responses are prepared in real time and may only
7247 be a few seconds old. Simply checking that the current time lies
7248 between thisUpdate and nextUpdate max reject otherwise valid responses
7249 caused by either OCSP responder or client clock inaccuracy. Instead
7250 we allow thisUpdate and nextUpdate to fall within a certain period of
7251 the current time. The age of the response can also optionally be
7252 checked. Two new options -validity_period and -status_age added to
7256 *) If signature or public key algorithm is unrecognized print out its
7257 OID rather that just UNKNOWN.
7260 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7261 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7262 ID to be generated from the issuer certificate alone which can then be
7263 passed to OCSP_id_issuer_cmp().
7266 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7267 ASN1 modules to export functions returning ASN1_ITEM pointers
7268 instead of the ASN1_ITEM structures themselves. This adds several
7269 new macros which allow the underlying ASN1 function/structure to
7270 be accessed transparently. As a result code should not use ASN1_ITEM
7271 references directly (such as &X509_it) but instead use the relevant
7272 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7273 use of the new ASN1 code on platforms where exporting structures
7274 is problematical (for example in shared libraries) but exporting
7275 functions returning pointers to structures is not.
7278 *) Add support for overriding the generation of SSL/TLS session IDs.
7279 These callbacks can be registered either in an SSL_CTX or per SSL.
7280 The purpose of this is to allow applications to control, if they wish,
7281 the arbitrary values chosen for use as session IDs, particularly as it
7282 can be useful for session caching in multiple-server environments. A
7283 command-line switch for testing this (and any client code that wishes
7284 to use such a feature) has been added to "s_server".
7285 [Geoff Thorpe, Lutz Jaenicke]
7287 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7288 of the form '#if defined(...) || defined(...) || ...' and
7289 '#if !defined(...) && !defined(...) && ...'. This also avoids
7290 the growing number of special cases it was previously handling.
7293 *) Make all configuration macros available for application by making
7294 sure they are available in opensslconf.h, by giving them names starting
7295 with "OPENSSL_" to avoid conflicts with other packages and by making
7296 sure e_os2.h will cover all platform-specific cases together with
7298 Additionally, it is now possible to define configuration/platform-
7299 specific names (called "system identities"). In the C code, these
7300 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7301 macro with the name beginning with "OPENSSL_SYS_", which is determined
7302 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7306 *) New option -set_serial to 'req' and 'x509' this allows the serial
7307 number to use to be specified on the command line. Previously self
7308 signed certificates were hard coded with serial number 0 and the
7309 CA options of 'x509' had to use a serial number in a file which was
7313 *) New options to 'ca' utility to support V2 CRL entry extensions.
7314 Currently CRL reason, invalidity date and hold instruction are
7315 supported. Add new CRL extensions to V3 code and some new objects.
7318 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7319 disable standard block padding (aka PKCS#5 padding) in the EVP
7320 API, which was previously mandatory. This means that the data is
7321 not padded in any way and so the total length much be a multiple
7322 of the block size, otherwise an error occurs.
7325 *) Initial (incomplete) OCSP SSL support.
7328 *) New function OCSP_parse_url(). This splits up a URL into its host,
7329 port and path components: primarily to parse OCSP URLs. New -url
7330 option to ocsp utility.
7333 *) New nonce behavior. The return value of OCSP_check_nonce() now
7334 reflects the various checks performed. Applications can decide
7335 whether to tolerate certain situations such as an absent nonce
7336 in a response when one was present in a request: the ocsp application
7337 just prints out a warning. New function OCSP_add1_basic_nonce()
7338 this is to allow responders to include a nonce in a response even if
7339 the request is nonce-less.
7342 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7343 skipped when using openssl x509 multiple times on a single input file,
7344 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7347 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7348 set string type: to handle setting ASN1_TIME structures. Fix ca
7349 utility to correctly initialize revocation date of CRLs.
7352 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7353 the clients preferred ciphersuites and rather use its own preferences.
7354 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7355 Internet Explorer by ensuring unchanged hash method during stepup.
7356 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7359 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7360 to aes and add a new 'exist' option to print out symbols that don't
7364 *) Additional options to ocsp utility to allow flags to be set and
7365 additional certificates supplied.
7368 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7369 OCSP client a number of certificate to only verify the response
7373 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7374 handle the new API. Currently only ECB, CBC modes supported. Add new
7377 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7378 Encryption Standard (AES) Ciphersuites for Transport Layer
7379 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7380 not enabled by default and were not part of the "ALL" ciphersuite
7381 alias because they were not yet official; they could be
7382 explicitly requested by specifying the "AESdraft" ciphersuite
7383 group alias. In the final release of OpenSSL 0.9.7, the group
7384 alias is called "AES" and is part of "ALL".)
7385 [Ben Laurie, Steve Henson, Bodo Moeller]
7387 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7388 request to response.
7391 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7392 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7393 extract information from a certificate request. OCSP_response_create()
7394 creates a response and optionally adds a basic response structure.
7395 OCSP_basic_add1_status() adds a complete single response to a basic
7396 response and returns the OCSP_SINGLERESP structure just added (to allow
7397 extensions to be included for example). OCSP_basic_add1_cert() adds a
7398 certificate to a basic response and OCSP_basic_sign() signs a basic
7399 response with various flags. New helper functions ASN1_TIME_check()
7400 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7401 (converts ASN1_TIME to GeneralizedTime).
7404 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7405 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7406 structure from a certificate. X509_pubkey_digest() digests the public_key
7407 contents: this is used in various key identifiers.
7410 *) Make sk_sort() tolerate a NULL argument.
7411 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7413 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7414 passed by the function are trusted implicitly. If any of them signed the
7415 response then it is assumed to be valid and is not verified.
7418 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7419 to data. This was previously part of the PKCS7 ASN1 code. This
7420 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7421 [Steve Henson, reported by Kenneth R. Robinette
7422 <support@securenetterm.com>]
7424 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7425 routines: without these tracing memory leaks is very painful.
7426 Fix leaks in PKCS12 and PKCS7 routines.
7429 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7430 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7431 effectively meant GeneralizedTime would never be used. Now it
7432 is initialised to -1 but X509_time_adj() now has to check the value
7433 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7434 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7435 [Steve Henson, reported by Kenneth R. Robinette
7436 <support@securenetterm.com>]
7438 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7439 result in a zero length in the ASN1_INTEGER structure which was
7440 not consistent with the structure when d2i_ASN1_INTEGER() was used
7441 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7442 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7443 where it did not print out a minus for negative ASN1_INTEGER.
7446 *) Add summary printout to ocsp utility. The various functions which
7447 convert status values to strings have been renamed to:
7448 OCSP_response_status_str(), OCSP_cert_status_str() and
7449 OCSP_crl_reason_str() and are no longer static. New options
7450 to verify nonce values and to disable verification. OCSP response
7451 printout format cleaned up.
7454 *) Add additional OCSP certificate checks. These are those specified
7455 in RFC2560. This consists of two separate checks: the CA of the
7456 certificate being checked must either be the OCSP signer certificate
7457 or the issuer of the OCSP signer certificate. In the latter case the
7458 OCSP signer certificate must contain the OCSP signing extended key
7459 usage. This check is performed by attempting to match the OCSP
7460 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7461 in the OCSP_CERTID structures of the response.
7464 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7465 and related routines. This uses the standard OpenSSL certificate
7466 verify routines to perform initial checks (just CA validity) and
7467 to obtain the certificate chain. Then additional checks will be
7468 performed on the chain. Currently the root CA is checked to see
7469 if it is explicitly trusted for OCSP signing. This is used to set
7470 a root CA as a global signing root: that is any certificate that
7471 chains to that CA is an acceptable OCSP signing certificate.
7474 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7475 extensions from a separate configuration file.
7476 As when reading extensions from the main configuration file,
7477 the '-extensions ...' option may be used for specifying the
7479 [Massimiliano Pala <madwolf@comune.modena.it>]
7481 *) New OCSP utility. Allows OCSP requests to be generated or
7482 read. The request can be sent to a responder and the output
7483 parsed, outputed or printed in text form. Not complete yet:
7484 still needs to check the OCSP response validity.
7487 *) New subcommands for 'openssl ca':
7488 'openssl ca -status <serial>' prints the status of the cert with
7489 the given serial number (according to the index file).
7490 'openssl ca -updatedb' updates the expiry status of certificates
7492 [Massimiliano Pala <madwolf@comune.modena.it>]
7494 *) New '-newreq-nodes' command option to CA.pl. This is like
7495 '-newreq', but calls 'openssl req' with the '-nodes' option
7496 so that the resulting key is not encrypted.
7497 [Damien Miller <djm@mindrot.org>]
7499 *) New configuration for the GNU Hurd.
7500 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7502 *) Initial code to implement OCSP basic response verify. This
7503 is currently incomplete. Currently just finds the signer's
7504 certificate and verifies the signature on the response.
7507 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7508 value of OPENSSLDIR. This is available via the new '-d' option
7509 to 'openssl version', and is also included in 'openssl version -a'.
7512 *) Allowing defining memory allocation callbacks that will be given
7513 file name and line number information in additional arguments
7514 (a const char* and an int). The basic functionality remains, as
7515 well as the original possibility to just replace malloc(),
7516 realloc() and free() by functions that do not know about these
7517 additional arguments. To register and find out the current
7518 settings for extended allocation functions, the following
7519 functions are provided:
7521 CRYPTO_set_mem_ex_functions
7522 CRYPTO_set_locked_mem_ex_functions
7523 CRYPTO_get_mem_ex_functions
7524 CRYPTO_get_locked_mem_ex_functions
7526 These work the same way as CRYPTO_set_mem_functions and friends.
7527 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7528 extended allocation function is enabled.
7529 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7530 a conventional allocation function is enabled.
7531 [Richard Levitte, Bodo Moeller]
7533 *) Finish off removing the remaining LHASH function pointer casts.
7534 There should no longer be any prototype-casting required when using
7535 the LHASH abstraction, and any casts that remain are "bugs". See
7536 the callback types and macros at the head of lhash.h for details
7537 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7540 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7541 If /dev/[u]random devices are not available or do not return enough
7542 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7544 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7545 /etc/entropy will be queried once each in this sequence, querying stops
7546 when enough entropy was collected without querying more sockets.
7549 *) Change the Unix RAND_poll() variant to be able to poll several
7550 random devices, as specified by DEVRANDOM, until a sufficient amount
7551 of data has been collected. We spend at most 10 ms on each file
7552 (select timeout) and read in non-blocking mode. DEVRANDOM now
7553 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7554 (previously it was just the string "/dev/urandom"), so on typical
7555 platforms the 10 ms delay will never occur.
7556 Also separate out the Unix variant to its own file, rand_unix.c.
7557 For VMS, there's a currently-empty rand_vms.c.
7560 *) Move OCSP client related routines to ocsp_cl.c. These
7561 provide utility functions which an application needing
7562 to issue a request to an OCSP responder and analyse the
7563 response will typically need: as opposed to those which an
7564 OCSP responder itself would need which will be added later.
7566 OCSP_request_sign() signs an OCSP request with an API similar
7567 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7568 response. OCSP_response_get1_basic() extracts basic response
7569 from response. OCSP_resp_find_status(): finds and extracts status
7570 information from an OCSP_CERTID structure (which will be created
7571 when the request structure is built). These are built from lower
7572 level functions which work on OCSP_SINGLERESP structures but
7573 won't normally be used unless the application wishes to examine
7574 extensions in the OCSP response for example.
7576 Replace nonce routines with a pair of functions.
7577 OCSP_request_add1_nonce() adds a nonce value and optionally
7578 generates a random value. OCSP_check_nonce() checks the
7579 validity of the nonce in an OCSP response.
7582 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7583 This doesn't copy the supplied OCSP_CERTID and avoids the
7584 need to free up the newly created id. Change return type
7585 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7586 This can then be used to add extensions to the request.
7587 Deleted OCSP_request_new(), since most of its functionality
7588 is now in OCSP_REQUEST_new() (and the case insensitive name
7589 clash) apart from the ability to set the request name which
7590 will be added elsewhere.
7593 *) Update OCSP API. Remove obsolete extensions argument from
7594 various functions. Extensions are now handled using the new
7595 OCSP extension code. New simple OCSP HTTP function which
7596 can be used to send requests and parse the response.
7599 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7600 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7601 uses the special reorder version of SET OF to sort the attributes
7602 and reorder them to match the encoded order. This resolves a long
7603 standing problem: a verify on a PKCS7 structure just after signing
7604 it used to fail because the attribute order did not match the
7605 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7606 it uses the received order. This is necessary to tolerate some broken
7607 software that does not order SET OF. This is handled by encoding
7608 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7609 to produce the required SET OF.
7612 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7613 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7614 files to get correct declarations of the ASN.1 item variables.
7617 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7618 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7619 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7620 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7621 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7622 ASN1_ITEM and no wrapper functions.
7625 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7626 replace the old function pointer based I/O routines. Change most of
7627 the *_d2i_bio() and *_d2i_fp() functions to use these.
7630 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7631 lines, recognice more "algorithms" that can be deselected, and make
7632 it complain about algorithm deselection that isn't recognised.
7635 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7636 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7637 to use new functions. Add NO_ASN1_OLD which can be set to remove
7638 some old style ASN1 functions: this can be used to determine if old
7639 code will still work when these eventually go away.
7642 *) New extension functions for OCSP structures, these follow the
7643 same conventions as certificates and CRLs.
7646 *) New function X509V3_add1_i2d(). This automatically encodes and
7647 adds an extension. Its behaviour can be customised with various
7648 flags to append, replace or delete. Various wrappers added for
7649 certificates and CRLs.
7652 *) Fix to avoid calling the underlying ASN1 print routine when
7653 an extension cannot be parsed. Correct a typo in the
7654 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7657 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7658 entries for variables.
7661 *) Add functionality to apps/openssl.c for detecting locking
7662 problems: As the program is single-threaded, all we have
7663 to do is register a locking callback using an array for
7664 storing which locks are currently held by the program.
7667 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7668 SSL_get_ex_data_X509_STORE_idx(), which is used in
7669 ssl_verify_cert_chain() and thus can be called at any time
7670 during TLS/SSL handshakes so that thread-safety is essential.
7671 Unfortunately, the ex_data design is not at all suited
7672 for multi-threaded use, so it probably should be abolished.
7675 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7676 [Broadcom, tweaked and integrated by Geoff Thorpe]
7678 *) Move common extension printing code to new function
7679 X509V3_print_extensions(). Reorganise OCSP print routines and
7680 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7683 *) New function X509_signature_print() to remove duplication in some
7687 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7688 set (this was treated exactly the same as SET OF previously). This
7689 is used to reorder the STACK representing the structure to match the
7690 encoding. This will be used to get round a problem where a PKCS7
7691 structure which was signed could not be verified because the STACK
7692 order did not reflect the encoded order.
7695 *) Reimplement the OCSP ASN1 module using the new code.
7698 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7699 for its ASN1 operations. The old style function pointers still exist
7700 for now but they will eventually go away.
7703 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7704 completely replaces the old ASN1 functionality with a table driven
7705 encoder and decoder which interprets an ASN1_ITEM structure describing
7706 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7707 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7708 has also been converted to the new form.
7711 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7712 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7713 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7714 for negative moduli.
7717 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7718 of not touching the result's sign bit.
7721 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7725 *) Changed the LHASH code to use prototypes for callbacks, and created
7726 macros to declare and implement thin (optionally static) functions
7727 that provide type-safety and avoid function pointer casting for the
7728 type-specific callbacks.
7731 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7733 [Veers Staats <staatsvr@asc.hpc.mil>,
7734 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7736 *) Reformat the FAQ so the different questions and answers can be divided
7737 in sections depending on the subject.
7740 *) Have the zlib compression code load ZLIB.DLL dynamically under
7744 *) New function BN_mod_sqrt for computing square roots modulo a prime
7745 (using the probabilistic Tonelli-Shanks algorithm unless
7746 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7747 be handled deterministically).
7748 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7750 *) Make BN_mod_inverse faster by explicitly handling small quotients
7751 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7752 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7755 *) New function BN_kronecker.
7758 *) Fix BN_gcd so that it works on negative inputs; the result is
7759 positive unless both parameters are zero.
7760 Previously something reasonably close to an infinite loop was
7761 possible because numbers could be growing instead of shrinking
7762 in the implementation of Euclid's algorithm.
7765 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7766 sign of the number in question.
7768 Fix BN_is_word(a,w) to work correctly for w == 0.
7770 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7771 because its test if the absolute value of 'a' equals 'w'.
7772 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7773 it exists mostly for use in the implementations of BN_is_zero(),
7774 BN_is_one(), and BN_is_word().
7777 *) New function BN_swap.
7780 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7781 the exponentiation functions are more likely to produce reasonable
7782 results on negative inputs.
7785 *) Change BN_mod_mul so that the result is always non-negative.
7786 Previously, it could be negative if one of the factors was negative;
7787 I don't think anyone really wanted that behaviour.
7790 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7791 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7792 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7793 and add new functions:
7802 BN_mod_lshift1_quick
7806 These functions always generate non-negative results.
7808 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7809 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7811 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7812 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7813 be reduced modulo m.
7814 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7817 The following entry accidentally appeared in the CHANGES file
7818 distributed with OpenSSL 0.9.7. The modifications described in
7819 it do *not* apply to OpenSSL 0.9.7.
7821 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7822 was actually never needed) and in BN_mul(). The removal in BN_mul()
7823 required a small change in bn_mul_part_recursive() and the addition
7824 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7825 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7826 bn_sub_words() and bn_add_words() except they take arrays with
7831 *) In 'openssl passwd', verify passwords read from the terminal
7832 unless the '-salt' option is used (which usually means that
7833 verification would just waste user's time since the resulting
7834 hash is going to be compared with some given password hash)
7835 or the new '-noverify' option is used.
7837 This is an incompatible change, but it does not affect
7838 non-interactive use of 'openssl passwd' (passwords on the command
7839 line, '-stdin' option, '-in ...' option) and thus should not
7843 *) Remove all references to RSAref, since there's no more need for it.
7846 *) Make DSO load along a path given through an environment variable
7847 (SHLIB_PATH) with shl_load().
7850 *) Constify the ENGINE code as a result of BIGNUM constification.
7851 Also constify the RSA code and most things related to it. In a
7852 few places, most notable in the depth of the ASN.1 code, ugly
7853 casts back to non-const were required (to be solved at a later
7857 *) Make it so the openssl application has all engines loaded by default.
7860 *) Constify the BIGNUM routines a little more.
7863 *) Add the following functions:
7865 ENGINE_load_cswift()
7867 ENGINE_load_atalla()
7869 ENGINE_load_builtin_engines()
7871 That way, an application can itself choose if external engines that
7872 are built-in in OpenSSL shall ever be used or not. The benefit is
7873 that applications won't have to be linked with libdl or other dso
7874 libraries unless it's really needed.
7876 Changed 'openssl engine' to load all engines on demand.
7877 Changed the engine header files to avoid the duplication of some
7878 declarations (they differed!).
7881 *) 'openssl engine' can now list capabilities.
7884 *) Better error reporting in 'openssl engine'.
7887 *) Never call load_dh_param(NULL) in s_server.
7890 *) Add engine application. It can currently list engines by name and
7891 identity, and test if they are actually available.
7894 *) Improve RPM specification file by forcing symbolic linking and making
7895 sure the installed documentation is also owned by root.root.
7896 [Damien Miller <djm@mindrot.org>]
7898 *) Give the OpenSSL applications more possibilities to make use of
7899 keys (public as well as private) handled by engines.
7902 *) Add OCSP code that comes from CertCo.
7905 *) Add VMS support for the Rijndael code.
7908 *) Added untested support for Nuron crypto accelerator.
7911 *) Add support for external cryptographic devices. This code was
7912 previously distributed separately as the "engine" branch.
7913 [Geoff Thorpe, Richard Levitte]
7915 *) Rework the filename-translation in the DSO code. It is now possible to
7916 have far greater control over how a "name" is turned into a filename
7917 depending on the operating environment and any oddities about the
7918 different shared library filenames on each system.
7921 *) Support threads on FreeBSD-elf in Configure.
7924 *) Fix for SHA1 assembly problem with MASM: it produces
7925 warnings about corrupt line number information when assembling
7926 with debugging information. This is caused by the overlapping
7928 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7931 NCONF_get_number() has no error checking at all. As a replacement,
7932 NCONF_get_number_e() is defined (_e for "error checking") and is
7933 promoted strongly. The old NCONF_get_number is kept around for
7934 binary backward compatibility.
7935 Make it possible for methods to load from something other than a BIO,
7936 by providing a function pointer that is given a name instead of a BIO.
7937 For example, this could be used to load configuration data from an
7941 *) Fix for non blocking accept BIOs. Added new I/O special reason
7942 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7943 with non blocking I/O was not possible because no retry code was
7944 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7948 *) Added the beginnings of Rijndael support.
7951 *) Fix for bug in DirectoryString mask setting. Add support for
7952 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7953 to allow certificate printing to more controllable, additional
7954 'certopt' option to 'x509' to allow new printing options to be
7958 *) Clean old EAY MD5 hack from e_os.h.
7961 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7963 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7964 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7965 [Joe Orton, Steve Henson]
7967 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7969 *) Fix additional bug revealed by the NISCC test suite:
7971 Stop bug triggering large recursion when presented with
7972 certain ASN.1 tags (CVE-2003-0851)
7975 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7977 *) Fix various bugs revealed by running the NISCC test suite:
7979 Stop out of bounds reads in the ASN1 code when presented with
7980 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7982 If verify callback ignores invalid public key errors don't try to check
7983 certificate signature with the NULL public key.
7987 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7988 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7992 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7993 extra data after the compression methods not only for TLS 1.0
7994 but also for SSL 3.0 (as required by the specification).
7995 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7997 *) Change X509_certificate_type() to mark the key as exported/exportable
7998 when it's 512 *bits* long, not 512 bytes.
8001 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8003 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8004 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8005 a protocol version number mismatch like a decryption error
8006 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8009 *) Turn on RSA blinding by default in the default implementation
8010 to avoid a timing attack. Applications that don't want it can call
8011 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8012 They would be ill-advised to do so in most cases.
8013 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8015 *) Change RSA blinding code so that it works when the PRNG is not
8016 seeded (in this case, the secret RSA exponent is abused as
8017 an unpredictable seed -- if it is not unpredictable, there
8018 is no point in blinding anyway). Make RSA blinding thread-safe
8019 by remembering the creator's thread ID in rsa->blinding and
8020 having all other threads use local one-time blinding factors
8021 (this requires more computation than sharing rsa->blinding, but
8022 avoids excessive locking; and if an RSA object is not shared
8023 between threads, blinding will still be very fast).
8026 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8028 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8029 via timing by performing a MAC computation even if incorrrect
8030 block cipher padding has been found. This is a countermeasure
8031 against active attacks where the attacker has to distinguish
8032 between bad padding and a MAC verification error. (CVE-2003-0078)
8034 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8035 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8036 Martin Vuagnoux (EPFL, Ilion)]
8038 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8040 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8041 memory from it's contents. This is done with a counter that will
8042 place alternating values in each byte. This can be used to solve
8043 two issues: 1) the removal of calls to memset() by highly optimizing
8044 compilers, and 2) cleansing with other values than 0, since those can
8045 be read through on certain media, for example a swap space on disk.
8048 *) Bugfix: client side session caching did not work with external caching,
8049 because the session->cipher setting was not restored when reloading
8050 from the external cache. This problem was masked, when
8051 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8052 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8055 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8056 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8057 [Zeev Lieber <zeev-l@yahoo.com>]
8059 *) Undo an undocumented change introduced in 0.9.6e which caused
8060 repeated calls to OpenSSL_add_all_ciphers() and
8061 OpenSSL_add_all_digests() to be ignored, even after calling
8065 *) Change the default configuration reader to deal with last line not
8066 being properly terminated.
8069 *) Change X509_NAME_cmp() so it applies the special rules on handling
8070 DN values that are of type PrintableString, as well as RDNs of type
8071 emailAddress where the value has the type ia5String.
8072 [stefank@valicert.com via Richard Levitte]
8074 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8075 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8076 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8077 the bitwise-OR of the two for use by the majority of applications
8078 wanting this behaviour, and update the docs. The documented
8079 behaviour and actual behaviour were inconsistent and had been
8080 changing anyway, so this is more a bug-fix than a behavioural
8082 [Geoff Thorpe, diagnosed by Nadav Har'El]
8084 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8085 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8088 *) Fix initialization code race conditions in
8089 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8090 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8091 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8092 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8093 ssl2_get_cipher_by_char(),
8094 ssl3_get_cipher_by_char().
8095 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8097 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8098 the cached sessions are flushed, as the remove_cb() might use ex_data
8099 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8100 (see [openssl.org #212]).
8101 [Geoff Thorpe, Lutz Jaenicke]
8103 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8104 length, instead of the encoding length to d2i_ASN1_OBJECT.
8107 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8109 *) [In 0.9.6g-engine release:]
8110 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8111 [Lynn Gazis <lgazis@rainbow.com>]
8113 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8115 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8116 and get fix the header length calculation.
8117 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8118 Alon Kantor <alonk@checkpoint.com> (and others),
8121 *) Use proper error handling instead of 'assertions' in buffer
8122 overflow checks added in 0.9.6e. This prevents DoS (the
8123 assertions could call abort()).
8124 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8126 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8128 *) Add various sanity checks to asn1_get_length() to reject
8129 the ASN1 length bytes if they exceed sizeof(long), will appear
8130 negative or the content length exceeds the length of the
8132 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8134 *) Fix cipher selection routines: ciphers without encryption had no flags
8135 for the cipher strength set and where therefore not handled correctly
8136 by the selection routines (PR #130).
8139 *) Fix EVP_dsa_sha macro.
8143 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8144 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8145 that was added in OpenSSL 0.9.6d.
8147 As the countermeasure turned out to be incompatible with some
8148 broken SSL implementations, the new option is part of SSL_OP_ALL.
8149 SSL_OP_ALL is usually employed when compatibility with weird SSL
8150 implementations is desired (e.g. '-bugs' option to 's_client' and
8151 's_server'), so the new option is automatically set in many
8155 *) Changes in security patch:
8157 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8158 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8159 Air Force Materiel Command, USAF, under agreement number
8162 *) Add various sanity checks to asn1_get_length() to reject
8163 the ASN1 length bytes if they exceed sizeof(long), will appear
8164 negative or the content length exceeds the length of the
8165 supplied buffer. (CVE-2002-0659)
8166 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8168 *) Assertions for various potential buffer overflows, not known to
8170 [Ben Laurie (CHATS)]
8172 *) Various temporary buffers to hold ASCII versions of integers were
8173 too small for 64 bit platforms. (CVE-2002-0655)
8174 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8176 *) Remote buffer overflow in SSL3 protocol - an attacker could
8177 supply an oversized session ID to a client. (CVE-2002-0656)
8178 [Ben Laurie (CHATS)]
8180 *) Remote buffer overflow in SSL2 protocol - an attacker could
8181 supply an oversized client master key. (CVE-2002-0656)
8182 [Ben Laurie (CHATS)]
8184 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8186 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8187 encoded as NULL) with id-dsa-with-sha1.
8188 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8190 *) Check various X509_...() return values in apps/req.c.
8191 [Nils Larsch <nla@trustcenter.de>]
8193 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8194 an end-of-file condition would erroneously be flagged, when the CRLF
8195 was just at the end of a processed block. The bug was discovered when
8196 processing data through a buffering memory BIO handing the data to a
8197 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8198 <ptsekov@syntrex.com> and Nedelcho Stanev.
8201 *) Implement a countermeasure against a vulnerability recently found
8202 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8203 before application data chunks to avoid the use of known IVs
8204 with data potentially chosen by the attacker.
8207 *) Fix length checks in ssl3_get_client_hello().
8210 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8211 to prevent ssl3_read_internal() from incorrectly assuming that
8212 ssl3_read_bytes() found application data while handshake
8213 processing was enabled when in fact s->s3->in_read_app_data was
8214 merely automatically cleared during the initial handshake.
8215 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8217 *) Fix object definitions for Private and Enterprise: they were not
8218 recognized in their shortname (=lowercase) representation. Extend
8219 obj_dat.pl to issue an error when using undefined keywords instead
8220 of silently ignoring the problem (Svenning Sorensen
8221 <sss@sss.dnsalias.net>).
8224 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8225 generators, i.e. generators other than 2 and 5. (Previously, the
8226 code did not properly initialise the 'add' and 'rem' values to
8227 BN_generate_prime().)
8229 In the new general case, we do not insist that 'generator' is
8230 actually a primitive root: This requirement is rather pointless;
8231 a generator of the order-q subgroup is just as good, if not
8235 *) Map new X509 verification errors to alerts. Discovered and submitted by
8236 Tom Wu <tom@arcot.com>.
8239 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8240 returning non-zero before the data has been completely received
8241 when using non-blocking I/O.
8242 [Bodo Moeller; problem pointed out by John Hughes]
8244 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8245 [Ben Laurie, Lutz Jaenicke]
8247 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8248 Yoram Zahavi <YoramZ@gilian.com>).
8251 *) Add information about CygWin 1.3 and on, and preserve proper
8252 configuration for the versions before that.
8253 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8255 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8256 check whether we deal with a copy of a session and do not delete from
8257 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8258 <izhar@checkpoint.com>.
8261 *) Do not store session data into the internal session cache, if it
8262 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8263 flag is set). Proposed by Aslam <aslam@funk.com>.
8266 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8270 *) [In 0.9.6d-engine release:]
8271 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8272 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8274 *) Add the configuration target linux-s390x.
8275 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8277 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8278 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8279 variable as an indication that a ClientHello message has been
8280 received. As the flag value will be lost between multiple
8281 invocations of ssl3_accept when using non-blocking I/O, the
8282 function may not be aware that a handshake has actually taken
8283 place, thus preventing a new session from being added to the
8286 To avoid this problem, we now set s->new_session to 2 instead of
8287 using a local variable.
8288 [Lutz Jaenicke, Bodo Moeller]
8290 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8291 if the SSL_R_LENGTH_MISMATCH error is detected.
8292 [Geoff Thorpe, Bodo Moeller]
8294 *) New 'shared_ldflag' column in Configure platform table.
8297 *) Fix EVP_CIPHER_mode macro.
8298 ["Dan S. Camper" <dan@bti.net>]
8300 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8301 type, we must throw them away by setting rr->length to 0.
8302 [D P Chang <dpc@qualys.com>]
8304 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8306 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8307 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8308 worked incorrectly for those cases where range = 10..._2 and
8309 3*range is two bits longer than range.)
8312 *) Only add signing time to PKCS7 structures if it is not already
8316 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8317 OBJ_ld_ce should be OBJ_id_ce.
8318 Also some ip-pda OIDs in crypto/objects/objects.txt were
8319 incorrect (cf. RFC 3039).
8320 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8322 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8323 returns early because it has nothing to do.
8324 [Andy Schneider <andy.schneider@bjss.co.uk>]
8326 *) [In 0.9.6c-engine release:]
8327 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8328 [Andy Schneider <andy.schneider@bjss.co.uk>]
8330 *) [In 0.9.6c-engine release:]
8331 Add support for Cryptographic Appliance's keyserver technology.
8332 (Use engine 'keyclient')
8333 [Cryptographic Appliances and Geoff Thorpe]
8335 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8336 is called via tools/c89.sh because arguments have to be
8337 rearranged (all '-L' options must appear before the first object
8339 [Richard Shapiro <rshapiro@abinitio.com>]
8341 *) [In 0.9.6c-engine release:]
8342 Add support for Broadcom crypto accelerator cards, backported
8344 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8346 *) [In 0.9.6c-engine release:]
8347 Add support for SureWare crypto accelerator cards from
8348 Baltimore Technologies. (Use engine 'sureware')
8349 [Baltimore Technologies and Mark Cox]
8351 *) [In 0.9.6c-engine release:]
8352 Add support for crypto accelerator cards from Accelerated
8353 Encryption Processing, www.aep.ie. (Use engine 'aep')
8354 [AEP Inc. and Mark Cox]
8356 *) Add a configuration entry for gcc on UnixWare.
8357 [Gary Benson <gbenson@redhat.com>]
8359 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8360 messages are stored in a single piece (fixed-length part and
8361 variable-length part combined) and fix various bugs found on the way.
8364 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8365 instead. BIO_gethostbyname() does not know what timeouts are
8366 appropriate, so entries would stay in cache even when they have
8368 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8370 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8371 faced with a pathologically small ClientHello fragment that does
8372 not contain client_version: Instead of aborting with an error,
8373 simply choose the highest available protocol version (i.e.,
8374 TLS 1.0 unless it is disabled). In practice, ClientHello
8375 messages are never sent like this, but this change gives us
8376 strictly correct behaviour at least for TLS.
8379 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8380 never resets s->method to s->ctx->method when called from within
8381 one of the SSL handshake functions.
8382 [Bodo Moeller; problem pointed out by Niko Baric]
8384 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8385 (sent using the client's version number) if client_version is
8386 smaller than the protocol version in use. Also change
8387 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8388 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8389 the client will at least see that alert.
8392 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8396 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8397 client receives HelloRequest while in a handshake.
8398 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8400 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8401 should end in 'break', not 'goto end' which circumvents various
8402 cleanups done in state SSL_ST_OK. But session related stuff
8403 must be disabled for SSL_ST_OK in the case that we just sent a
8406 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8407 before just sending a HelloRequest.
8408 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8410 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8411 reveal whether illegal block cipher padding was found or a MAC
8412 verification error occurred. (Neither SSLerr() codes nor alerts
8413 are directly visible to potential attackers, but the information
8414 may leak via logfiles.)
8416 Similar changes are not required for the SSL 2.0 implementation
8417 because the number of padding bytes is sent in clear for SSL 2.0,
8418 and the extra bytes are just ignored. However ssl/s2_pkt.c
8419 failed to verify that the purported number of padding bytes is in
8423 *) Add OpenUNIX-8 support including shared libraries
8424 (Boyd Lynn Gerber <gerberb@zenez.com>).
8427 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8428 'wristwatch attack' using huge encoding parameters (cf.
8429 James H. Manger's CRYPTO 2001 paper). Note that the
8430 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8431 encoding parameters and hence was not vulnerable.
8434 *) BN_sqr() bug fix.
8435 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8437 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8438 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8439 followed by modular reduction.
8440 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8442 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8443 equivalent based on BN_pseudo_rand() instead of BN_rand().
8446 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8447 This function was broken, as the check for a new client hello message
8448 to handle SGC did not allow these large messages.
8449 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8452 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8455 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8456 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8459 *) Rework the configuration and shared library support for Tru64 Unix.
8460 The configuration part makes use of modern compiler features and
8461 still retains old compiler behavior for those that run older versions
8462 of the OS. The shared library support part includes a variant that
8463 uses the RPATH feature, and is available through the special
8464 configuration target "alpha-cc-rpath", which will never be selected
8466 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8468 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8469 with the same message size as in ssl3_get_certificate_request().
8470 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8471 messages might inadvertently be reject as too long.
8472 [Petr Lampa <lampa@fee.vutbr.cz>]
8474 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8477 *) Modified SSL library such that the verify_callback that has been set
8478 specificly for an SSL object with SSL_set_verify() is actually being
8479 used. Before the change, a verify_callback set with this function was
8480 ignored and the verify_callback() set in the SSL_CTX at the time of
8481 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8482 to allow the necessary settings.
8485 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8486 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8487 done automatically (in contradiction to the requirements of the C
8488 standard). This made problems when used from OpenSSH.
8491 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8492 dh->length and always used
8494 BN_rand_range(priv_key, dh->p).
8496 BN_rand_range() is not necessary for Diffie-Hellman, and this
8497 specific range makes Diffie-Hellman unnecessarily inefficient if
8498 dh->length (recommended exponent length) is much smaller than the
8499 length of dh->p. We could use BN_rand_range() if the order of
8500 the subgroup was stored in the DH structure, but we only have
8505 BN_rand(priv_key, l, ...)
8507 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8513 RSA_eay_public_encrypt
8514 RSA_eay_private_decrypt
8515 RSA_eay_private_encrypt (signing)
8516 RSA_eay_public_decrypt (signature verification)
8518 (default implementations for RSA_public_encrypt,
8519 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8520 always reject numbers >= n.
8523 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8524 to synchronize access to 'locking_thread'. This is necessary on
8525 systems where access to 'locking_thread' (an 'unsigned long'
8526 variable) is not atomic.
8529 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8530 *before* setting the 'crypto_lock_rand' flag. The previous code had
8531 a race condition if 0 is a valid thread ID.
8532 [Travis Vitek <vitek@roguewave.com>]
8534 *) Add support for shared libraries under Irix.
8535 [Albert Chin-A-Young <china@thewrittenword.com>]
8537 *) Add configuration option to build on Linux on both big-endian and
8539 [Ralf Baechle <ralf@uni-koblenz.de>]
8541 *) Add the possibility to create shared libraries on HP-UX.
8544 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8546 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8547 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8548 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8549 PRNG state recovery was possible based on the output of
8550 one PRNG request appropriately sized to gain knowledge on
8551 'md' followed by enough consecutive 1-byte PRNG requests
8552 to traverse all of 'state'.
8554 1. When updating 'md_local' (the current thread's copy of 'md')
8555 during PRNG output generation, hash all of the previous
8556 'md_local' value, not just the half used for PRNG output.
8558 2. Make the number of bytes from 'state' included into the hash
8559 independent from the number of PRNG bytes requested.
8561 The first measure alone would be sufficient to avoid
8562 Markku-Juhani's attack. (Actually it had never occurred
8563 to me that the half of 'md_local' used for chaining was the
8564 half from which PRNG output bytes were taken -- I had always
8565 assumed that the secret half would be used.) The second
8566 measure makes sure that additional data from 'state' is never
8567 mixed into 'md_local' in small portions; this heuristically
8568 further strengthens the PRNG.
8571 *) Fix crypto/bn/asm/mips3.s.
8574 *) When only the key is given to "enc", the IV is undefined. Print out
8575 an error message in this case.
8578 *) Handle special case when X509_NAME is empty in X509 printing routines.
8581 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8582 positive and less than q.
8585 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8586 used: it isn't thread safe and the add_lock_callback should handle
8588 [Paul Rose <Paul.Rose@bridge.com>]
8590 *) Verify that incoming data obeys the block size in
8591 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8595 [Ulf Möller, Bodo Möller]
8597 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8598 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8599 when fixing the server behaviour for backwards-compatible 'client
8600 hello' messages. (Note that the attack is impractical against
8601 SSL 3.0 and TLS 1.0 anyway because length and version checking
8602 means that the probability of guessing a valid ciphertext is
8603 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8606 Before 0.9.5, the countermeasure (hide the error by generating a
8607 random 'decryption result') did not work properly because
8608 ERR_clear_error() was missing, meaning that SSL_get_error() would
8609 detect the supposedly ignored error.
8611 Both problems are now fixed.
8614 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8615 (previously it was 1024).
8618 *) Fix for compatibility mode trust settings: ignore trust settings
8619 unless some valid trust or reject settings are present.
8622 *) Fix for blowfish EVP: its a variable length cipher.
8625 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8626 parameters in DSA public key structures and return an error in the
8627 DSA routines if parameters are absent.
8630 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8631 in the current directory if neither $RANDFILE nor $HOME was set.
8632 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8633 caused some confusion to Windows users who haven't defined $HOME.
8634 Thus RAND_file_name() is changed again: e_os.h can define a
8635 DEFAULT_HOME, which will be used if $HOME is not set.
8636 For Windows, we use "C:"; on other platforms, we still require
8637 environment variables.
8639 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8640 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8641 having multiple threads call RAND_poll() concurrently.
8644 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8645 combination of a flag and a thread ID variable.
8646 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8647 flag), *other* threads can enter ssleay_add_bytes without obeying
8648 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8649 that they do not hold after the first thread unsets add_do_not_lock).
8652 *) Change bctest again: '-x' expressions are not available in all
8656 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8658 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8659 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8661 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8662 the default extension for executables, if any. Also, make the perl
8663 scripts that use symlink() to test if it really exists and use "cp"
8664 if it doesn't. All this made OpenSSL compilable and installable in
8668 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8669 If SEQUENCE is length is indefinite just set c->slen to the total
8670 amount of data available.
8671 [Steve Henson, reported by shige@FreeBSD.org]
8672 [This change does not apply to 0.9.7.]
8674 *) Change bctest to avoid here-documents inside command substitution
8675 (workaround for FreeBSD /bin/sh bug).
8676 For compatibility with Ultrix, avoid shell functions (introduced
8677 in the bctest version that searches along $PATH).
8680 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8681 with des_encrypt() defined on some operating systems, like Solaris
8685 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8686 On the Importance of Eliminating Errors in Cryptographic
8687 Computations, J. Cryptology 14 (2001) 2, 101-119,
8688 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8691 *) MIPS assembler BIGNUM division bug fix.
8694 *) Disabled incorrect Alpha assembler code.
8697 *) Fix PKCS#7 decode routines so they correctly update the length
8698 after reading an EOC for the EXPLICIT tag.
8700 [This change does not apply to 0.9.7.]
8702 *) Fix bug in PKCS#12 key generation routines. This was triggered
8703 if a 3DES key was generated with a 0 initial byte. Include
8704 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8705 (but broken) behaviour.
8708 *) Enhance bctest to search for a working bc along $PATH and print
8710 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8712 *) Fix memory leaks in err.c: free err_data string if necessary;
8713 don't write to the wrong index in ERR_set_error_data.
8716 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8720 *) Replace rdtsc with _emit statements for VC++ version 5.
8721 [Jeremy Cooper <jeremy@baymoo.org>]
8723 *) Make it possible to reuse SSLv2 sessions.
8726 *) In copy_email() check for >= 0 as a return value for
8727 X509_NAME_get_index_by_NID() since 0 is a valid index.
8728 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8730 *) Avoid coredump with unsupported or invalid public keys by checking if
8731 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8732 PKCS7_verify() fails with non detached data.
8735 *) Don't use getenv in library functions when run as setuid/setgid.
8736 New function OPENSSL_issetugid().
8739 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8740 due to incorrect handling of multi-threading:
8742 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8744 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8746 3. Count how many times MemCheck_off() has been called so that
8747 nested use can be treated correctly. This also avoids
8748 inband-signalling in the previous code (which relied on the
8749 assumption that thread ID 0 is impossible).
8752 *) Add "-rand" option also to s_client and s_server.
8755 *) Fix CPU detection on Irix 6.x.
8756 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8757 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8759 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8762 [This change does not apply to 0.9.7.]
8764 *) Use the cached encoding of an X509_NAME structure rather than
8765 copying it. This is apparently the reason for the libsafe "errors"
8766 but the code is actually correct.
8769 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8770 Bleichenbacher's DSA attack.
8771 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8772 to be set and top=0 forces the highest bit to be set; top=-1 is new
8773 and leaves the highest bit random.
8774 [Ulf Moeller, Bodo Moeller]
8776 *) In the NCONF_...-based implementations for CONF_... queries
8777 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8778 a temporary CONF structure with the data component set to NULL
8779 (which gives segmentation faults in lh_retrieve).
8780 Instead, use NULL for the CONF pointer in CONF_get_string and
8781 CONF_get_number (which may use environment variables) and directly
8782 return NULL from CONF_get_section.
8785 *) Fix potential buffer overrun for EBCDIC.
8788 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8789 keyUsage if basicConstraints absent for a CA.
8792 *) Make SMIME_write_PKCS7() write mail header values with a format that
8793 is more generally accepted (no spaces before the semicolon), since
8794 some programs can't parse those values properly otherwise. Also make
8795 sure BIO's that break lines after each write do not create invalid
8799 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8800 macros previously used would not encode an empty SEQUENCE OF
8801 and break the signature.
8803 [This change does not apply to 0.9.7.]
8805 *) Zero the premaster secret after deriving the master secret in
8809 *) Add some EVP_add_digest_alias registrations (as found in
8810 OpenSSL_add_all_digests()) to SSL_library_init()
8811 aka OpenSSL_add_ssl_algorithms(). This provides improved
8812 compatibility with peers using X.509 certificates
8813 with unconventional AlgorithmIdentifier OIDs.
8816 *) Fix for Irix with NO_ASM.
8817 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8819 *) ./config script fixes.
8820 [Ulf Moeller, Richard Levitte]
8822 *) Fix 'openssl passwd -1'.
8825 *) Change PKCS12_key_gen_asc() so it can cope with non null
8826 terminated strings whose length is passed in the passlen
8827 parameter, for example from PEM callbacks. This was done
8828 by adding an extra length parameter to asc2uni().
8829 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8831 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8832 call failed, free the DSA structure.
8835 *) Fix to uni2asc() to cope with zero length Unicode strings.
8836 These are present in some PKCS#12 files.
8839 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8840 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8841 when writing a 32767 byte record.
8842 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8844 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8845 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8847 (RSA objects have a reference count access to which is protected
8848 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8849 so they are meant to be shared between threads.)
8850 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8851 "Reddie, Steven" <Steven.Reddie@ca.com>]
8853 *) Fix a deadlock in CRYPTO_mem_leaks().
8856 *) Use better test patterns in bntest.
8859 *) rand_win.c fix for Borland C.
8862 *) BN_rshift bugfix for n == 0.
8865 *) Add a 'bctest' script that checks for some known 'bc' bugs
8866 so that 'make test' does not abort just because 'bc' is broken.
8869 *) Store verify_result within SSL_SESSION also for client side to
8870 avoid potential security hole. (Re-used sessions on the client side
8871 always resulted in verify_result==X509_V_OK, not using the original
8872 result of the server certificate verification.)
8875 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8876 SSL3_RT_APPLICATION_DATA, return 0.
8877 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8881 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8882 releases, have been re-implemented by renaming the previous
8883 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8884 and ssl3_read_internal, respectively, and adding 'peek' parameters
8885 to them. The new ssl[23]_{read,peek} functions are calls to
8886 ssl[23]_read_internal with the 'peek' flag set appropriately.
8887 A 'peek' parameter has also been added to ssl3_read_bytes, which
8888 does the actual work for ssl3_read_internal.
8891 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8892 the method-specific "init()" handler. Also clean up ex_data after
8893 calling the method-specific "finish()" handler. Previously, this was
8894 happening the other way round.
8897 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8898 The previous value, 12, was not always sufficient for BN_mod_exp().
8901 *) Make sure that shared libraries get the internal name engine with
8902 the full version number and not just 0. This should mark the
8903 shared libraries as not backward compatible. Of course, this should
8904 be changed again when we can guarantee backward binary compatibility.
8907 *) Fix typo in get_cert_by_subject() in by_dir.c
8908 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8910 *) Rework the system to generate shared libraries:
8912 - Make note of the expected extension for the shared libraries and
8913 if there is a need for symbolic links from for example libcrypto.so.0
8914 to libcrypto.so.0.9.7. There is extended info in Configure for
8917 - Make as few rebuilds of the shared libraries as possible.
8919 - Still avoid linking the OpenSSL programs with the shared libraries.
8921 - When installing, install the shared libraries separately from the
8925 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8927 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8928 and not in SSL_clear because the latter is also used by the
8929 accept/connect functions; previously, the settings made by
8930 SSL_set_read_ahead would be lost during the handshake.
8931 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8933 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8934 Previously, it would create entries for disabled algorithms no
8938 *) Added several new manual pages for SSL_* function.
8941 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8943 *) In ssl23_get_client_hello, generate an error message when faced
8944 with an initial SSL 3.0/TLS record that is too small to contain the
8945 first two bytes of the ClientHello message, i.e. client_version.
8946 (Note that this is a pathologic case that probably has never happened
8947 in real life.) The previous approach was to use the version number
8948 from the record header as a substitute; but our protocol choice
8949 should not depend on that one because it is not authenticated
8950 by the Finished messages.
8953 *) More robust randomness gathering functions for Windows.
8954 [Jeffrey Altman <jaltman@columbia.edu>]
8956 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8957 not set then we don't setup the error code for issuer check errors
8958 to avoid possibly overwriting other errors which the callback does
8959 handle. If an application does set the flag then we assume it knows
8960 what it is doing and can handle the new informational codes
8964 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8965 a general "ANY" type, as such it should be able to decode anything
8966 including tagged types. However it didn't check the class so it would
8967 wrongly interpret tagged types in the same way as their universal
8968 counterpart and unknown types were just rejected. Changed so that the
8969 tagged and unknown types are handled in the same way as a SEQUENCE:
8970 that is the encoding is stored intact. There is also a new type
8971 "V_ASN1_OTHER" which is used when the class is not universal, in this
8972 case we have no idea what the actual type is so we just lump them all
8976 *) On VMS, stdout may very well lead to a file that is written to
8977 in a record-oriented fashion. That means that every write() will
8978 write a separate record, which will be read separately by the
8979 programs trying to read from it. This can be very confusing.
8981 The solution is to put a BIO filter in the way that will buffer
8982 text until a linefeed is reached, and then write everything a
8983 line at a time, so every record written will be an actual line,
8984 not chunks of lines and not (usually doesn't happen, but I've
8985 seen it once) several lines in one record. BIO_f_linebuffer() is
8988 Currently, it's a VMS-only method, because that's where it has
8989 been tested well enough.
8992 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8993 it can return incorrect results.
8994 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8995 but it was in 0.9.6-beta[12].)
8998 *) Disable the check for content being present when verifying detached
8999 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9000 include zero length content when signing messages.
9003 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9004 BIO_ctrl (for BIO pairs).
9007 *) Add DSO method for VMS.
9010 *) Bug fix: Montgomery multiplication could produce results with the
9014 *) Add RPM specification openssl.spec and modify it to build three
9015 packages. The default package contains applications, application
9016 documentation and run-time libraries. The devel package contains
9017 include files, static libraries and function documentation. The
9018 doc package contains the contents of the doc directory. The original
9019 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9022 *) Add a large number of documentation files for many SSL routines.
9023 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9025 *) Add a configuration entry for Sony News 4.
9026 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9028 *) Don't set the two most significant bits to one when generating a
9029 random number < q in the DSA library.
9032 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9033 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9034 the underlying transport is blocking) if a handshake took place.
9035 (The default behaviour is needed by applications such as s_client
9036 and s_server that use select() to determine when to use SSL_read;
9037 but for applications that know in advance when to expect data, it
9038 just makes things more complicated.)
9041 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9045 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9046 work better on such systems.
9047 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9049 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9050 Update PKCS12_parse() so it copies the friendlyName and the
9051 keyid to the certificates aux info.
9054 *) Fix bug in PKCS7_verify() which caused an infinite loop
9055 if there was more than one signature.
9056 [Sven Uszpelkat <su@celocom.de>]
9058 *) Major change in util/mkdef.pl to include extra information
9059 about each symbol, as well as presenting variables as well
9060 as functions. This change means that there's n more need
9061 to rebuild the .num files when some algorithms are excluded.
9064 *) Allow the verify time to be set by an application,
9065 rather than always using the current time.
9068 *) Phase 2 verify code reorganisation. The certificate
9069 verify code now looks up an issuer certificate by a
9070 number of criteria: subject name, authority key id
9071 and key usage. It also verifies self signed certificates
9072 by the same criteria. The main comparison function is
9073 X509_check_issued() which performs these checks.
9075 Lot of changes were necessary in order to support this
9076 without completely rewriting the lookup code.
9078 Authority and subject key identifier are now cached.
9080 The LHASH 'certs' is X509_STORE has now been replaced
9081 by a STACK_OF(X509_OBJECT). This is mainly because an
9082 LHASH can't store or retrieve multiple objects with
9083 the same hash value.
9085 As a result various functions (which were all internal
9086 use only) have changed to handle the new X509_STORE
9087 structure. This will break anything that messed round
9088 with X509_STORE internally.
9090 The functions X509_STORE_add_cert() now checks for an
9091 exact match, rather than just subject name.
9093 The X509_STORE API doesn't directly support the retrieval
9094 of multiple certificates matching a given criteria, however
9095 this can be worked round by performing a lookup first
9096 (which will fill the cache with candidate certificates)
9097 and then examining the cache for matches. This is probably
9098 the best we can do without throwing out X509_LOOKUP
9099 entirely (maybe later...).
9101 The X509_VERIFY_CTX structure has been enhanced considerably.
9103 All certificate lookup operations now go via a get_issuer()
9104 callback. Although this currently uses an X509_STORE it
9105 can be replaced by custom lookups. This is a simple way
9106 to bypass the X509_STORE hackery necessary to make this
9107 work and makes it possible to use more efficient techniques
9108 in future. A very simple version which uses a simple
9109 STACK for its trusted certificate store is also provided
9110 using X509_STORE_CTX_trusted_stack().
9112 The verify_cb() and verify() callbacks now have equivalents
9113 in the X509_STORE_CTX structure.
9115 X509_STORE_CTX also has a 'flags' field which can be used
9116 to customise the verify behaviour.
9119 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9120 excludes S/MIME capabilities.
9123 *) When a certificate request is read in keep a copy of the
9124 original encoding of the signed data and use it when outputting
9125 again. Signatures then use the original encoding rather than
9126 a decoded, encoded version which may cause problems if the
9127 request is improperly encoded.
9130 *) For consistency with other BIO_puts implementations, call
9131 buffer_write(b, ...) directly in buffer_puts instead of calling
9134 In BIO_puts, increment b->num_write as in BIO_write.
9135 [Peter.Sylvester@EdelWeb.fr]
9137 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9138 BN_zero, we may not return a BIGNUM with an array consisting of
9142 *) Avoid calling abort() from within the library when problems are
9143 detected, except if preprocessor symbols have been defined
9144 (such as REF_CHECK, BN_DEBUG etc.).
9147 *) New openssl application 'rsautl'. This utility can be
9148 used for low level RSA operations. DER public key
9149 BIO/fp routines also added.
9152 *) New Configure entry and patches for compiling on QNX 4.
9153 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9155 *) A demo state-machine implementation was sponsored by
9156 Nuron (http://www.nuron.com/) and is now available in
9157 demos/state_machine.
9160 *) New options added to the 'dgst' utility for signature
9161 generation and verification.
9164 *) Unrecognized PKCS#7 content types are now handled via a
9165 catch all ASN1_TYPE structure. This allows unsupported
9166 types to be stored as a "blob" and an application can
9167 encode and decode it manually.
9170 *) Fix various signed/unsigned issues to make a_strex.c
9172 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9174 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9175 length if passed a buffer. ASN1_INTEGER_to_BN failed
9176 if passed a NULL BN and its argument was negative.
9177 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9179 *) Modification to PKCS#7 encoding routines to output definite
9180 length encoding. Since currently the whole structures are in
9181 memory there's not real point in using indefinite length
9182 constructed encoding. However if OpenSSL is compiled with
9183 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9186 *) Added BIO_vprintf() and BIO_vsnprintf().
9189 *) Added more prefixes to parse for in the the strings written
9190 through a logging bio, to cover all the levels that are available
9191 through syslog. The prefixes are now:
9193 PANIC, EMERG, EMR => LOG_EMERG
9194 ALERT, ALR => LOG_ALERT
9195 CRIT, CRI => LOG_CRIT
9196 ERROR, ERR => LOG_ERR
9197 WARNING, WARN, WAR => LOG_WARNING
9198 NOTICE, NOTE, NOT => LOG_NOTICE
9199 INFO, INF => LOG_INFO
9200 DEBUG, DBG => LOG_DEBUG
9202 and as before, if none of those prefixes are present at the
9203 beginning of the string, LOG_ERR is chosen.
9205 On Win32, the LOG_* levels are mapped according to this:
9207 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9208 LOG_WARNING => EVENTLOG_WARNING_TYPE
9209 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9213 *) Made it possible to reconfigure with just the configuration
9214 argument "reconf" or "reconfigure". The command line arguments
9215 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9216 and are retrieved from there when reconfiguring.
9220 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9222 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9225 *) The obj_dat.pl script was messing up the sorting of object
9226 names. The reason was that it compared the quoted version
9227 of strings as a result "OCSP" > "OCSP Signing" because
9228 " > SPACE. Changed script to store unquoted versions of
9229 names and add quotes on output. It was also omitting some
9230 names from the lookup table if they were given a default
9231 value (that is if SN is missing it is given the same
9232 value as LN and vice versa), these are now added on the
9233 grounds that if an object has a name we should be able to
9234 look it up. Finally added warning output when duplicate
9235 short or long names are found.
9238 *) Changes needed for Tandem NSK.
9239 [Scott Uroff <scott@xypro.com>]
9241 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9242 RSA_padding_check_SSLv23(), special padding was never detected
9243 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9244 version rollback attacks was not effective.
9246 In s23_clnt.c, don't use special rollback-attack detection padding
9247 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9248 client; similarly, in s23_srvr.c, don't do the rollback check if
9249 SSL 2.0 is the only protocol enabled in the server.
9252 *) Make it possible to get hexdumps of unprintable data with 'openssl
9253 asn1parse'. By implication, the functions ASN1_parse_dump() and
9254 BIO_dump_indent() are added.
9257 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9258 these print out strings and name structures based on various
9259 flags including RFC2253 support and proper handling of
9260 multibyte characters. Added options to the 'x509' utility
9261 to allow the various flags to be set.
9264 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9265 Also change the functions X509_cmp_current_time() and
9266 X509_gmtime_adj() work with an ASN1_TIME structure,
9267 this will enable certificates using GeneralizedTime in validity
9268 dates to be checked.
9271 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9272 negative public key encodings) on by default,
9273 NO_NEG_PUBKEY_BUG can be set to disable it.
9276 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9277 content octets. An i2c_ASN1_OBJECT is unnecessary because
9278 the encoding can be trivially obtained from the structure.
9281 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9282 not read locks (CRYPTO_r_[un]lock).
9285 *) A first attempt at creating official support for shared
9286 libraries through configuration. I've kept it so the
9287 default is static libraries only, and the OpenSSL programs
9288 are always statically linked for now, but there are
9289 preparations for dynamic linking in place.
9290 This has been tested on Linux and Tru64.
9293 *) Randomness polling function for Win9x, as described in:
9294 Peter Gutmann, Software Generation of Practically Strong
9298 *) Fix so PRNG is seeded in req if using an already existing
9302 *) New options to smime application. -inform and -outform
9303 allow alternative formats for the S/MIME message including
9304 PEM and DER. The -content option allows the content to be
9305 specified separately. This should allow things like Netscape
9306 form signing output easier to verify.
9309 *) Fix the ASN1 encoding of tags using the 'long form'.
9312 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9313 STRING types. These convert content octets to and from the
9314 underlying type. The actual tag and length octets are
9315 already assumed to have been read in and checked. These
9316 are needed because all other string types have virtually
9317 identical handling apart from the tag. By having versions
9318 of the ASN1 functions that just operate on content octets
9319 IMPLICIT tagging can be handled properly. It also allows
9320 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9321 and ASN1_INTEGER are identical apart from the tag.
9324 *) Change the handling of OID objects as follows:
9326 - New object identifiers are inserted in objects.txt, following
9327 the syntax given in objects.README.
9328 - objects.pl is used to process obj_mac.num and create a new
9330 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9333 This is currently kind of a hack, and the perl code in objects.pl
9334 isn't very elegant, but it works as I intended. The simplest way
9335 to check that it worked correctly is to look in obj_dat.h and
9336 check the array nid_objs and make sure the objects haven't moved
9337 around (this is important!). Additions are OK, as well as
9338 consistent name changes.
9341 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9344 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9345 The given file adds to whatever has already been seeded into the
9346 random pool through the RANDFILE configuration file option or
9347 environment variable, or the default random state file.
9350 *) mkstack.pl now sorts each macro group into lexical order.
9351 Previously the output order depended on the order the files
9352 appeared in the directory, resulting in needless rewriting
9356 *) Patches to make OpenSSL compile under Win32 again. Mostly
9357 work arounds for the VC++ problem that it treats func() as
9358 func(void). Also stripped out the parts of mkdef.pl that
9359 added extra typesafe functions: these no longer exist.
9362 *) Reorganisation of the stack code. The macros are now all
9363 collected in safestack.h . Each macro is defined in terms of
9364 a "stack macro" of the form SKM_<name>(type, a, b). The
9365 DEBUG_SAFESTACK is now handled in terms of function casts,
9366 this has the advantage of retaining type safety without the
9367 use of additional functions. If DEBUG_SAFESTACK is not defined
9368 then the non typesafe macros are used instead. Also modified the
9369 mkstack.pl script to handle the new form. Needs testing to see
9370 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9371 the default if no major problems. Similar behaviour for ASN1_SET_OF
9372 and PKCS12_STACK_OF.
9375 *) When some versions of IIS use the 'NET' form of private key the
9376 key derivation algorithm is different. Normally MD5(password) is
9377 used as a 128 bit RC4 key. In the modified case
9378 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9379 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9380 as the old Netscape_RSA functions except they have an additional
9381 'sgckey' parameter which uses the modified algorithm. Also added
9382 an -sgckey command line option to the rsa utility. Thanks to
9383 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9384 algorithm to openssl-dev.
9387 *) The evp_local.h macros were using 'c.##kname' which resulted in
9388 invalid expansion on some systems (SCO 5.0.5 for example).
9389 Corrected to 'c.kname'.
9390 [Phillip Porch <root@theporch.com>]
9392 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9393 a STACK of email addresses from a certificate or request, these look
9394 in the subject name and the subject alternative name extensions and
9395 omit any duplicate addresses.
9398 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9399 This makes DSA verification about 2 % faster.
9402 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9403 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9404 plus overhead for 1024 bit moduli).
9405 This makes exponentiations about 0.5 % faster for 1024 bit
9406 exponents (as measured by "openssl speed rsa2048").
9409 *) Rename memory handling macros to avoid conflicts with other
9411 Malloc => OPENSSL_malloc
9412 Malloc_locked => OPENSSL_malloc_locked
9413 Realloc => OPENSSL_realloc
9414 Free => OPENSSL_free
9417 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9418 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9421 *) CygWin32 support.
9422 [John Jarvie <jjarvie@newsguy.com>]
9424 *) The type-safe stack code has been rejigged. It is now only compiled
9425 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9426 by default all type-specific stack functions are "#define"d back to
9427 standard stack functions. This results in more streamlined output
9428 but retains the type-safety checking possibilities of the original
9432 *) The STACK code has been cleaned up, and certain type declarations
9433 that didn't make a lot of sense have been brought in line. This has
9434 also involved a cleanup of sorts in safestack.h to more correctly
9435 map type-safe stack functions onto their plain stack counterparts.
9436 This work has also resulted in a variety of "const"ifications of
9437 lots of the code, especially "_cmp" operations which should normally
9438 be prototyped with "const" parameters anyway.
9441 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9442 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9443 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9444 where all of 'md' is used each time the PRNG is used, but 'state'
9445 is used only indexed by a cyclic counter. As entropy may not be
9446 well distributed from the beginning, 'md' is important as a
9447 chaining variable. However, the output function chains only half
9448 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9449 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9450 in all of 'state' being rewritten, with the new values depending
9451 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9454 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9455 the handshake is continued after ssl_verify_cert_chain();
9456 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9457 can lead to 'unexplainable' connection aborts later.
9458 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9460 *) Major EVP API cipher revision.
9461 Add hooks for extra EVP features. This allows various cipher
9462 parameters to be set in the EVP interface. Support added for variable
9463 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9464 setting of RC2 and RC5 parameters.
9466 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9469 Remove lots of duplicated code from the EVP library. For example *every*
9470 cipher init() function handles the 'iv' in the same way according to the
9471 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9472 for CFB and OFB modes they zero ctx->num.
9474 New functionality allows removal of S/MIME code RC2 hack.
9476 Most of the routines have the same form and so can be declared in terms
9479 By shifting this to the top level EVP_CipherInit() it can be removed from
9480 all individual ciphers. If the cipher wants to handle IVs or keys
9481 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9484 Change lots of functions like EVP_EncryptUpdate() to now return a
9485 value: although software versions of the algorithms cannot fail
9486 any installed hardware versions can.
9489 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9490 this option is set, tolerate broken clients that send the negotiated
9491 protocol version number instead of the requested protocol version
9495 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9496 i.e. non-zero for export ciphersuites, zero otherwise.
9497 Previous versions had this flag inverted, inconsistent with
9498 rsa_tmp_cb (..._TMP_RSA_CB).
9499 [Bodo Moeller; problem reported by Amit Chopra]
9501 *) Add missing DSA library text string. Work around for some IIS
9502 key files with invalid SEQUENCE encoding.
9505 *) Add a document (doc/standards.txt) that list all kinds of standards
9506 and so on that are implemented in OpenSSL.
9509 *) Enhance c_rehash script. Old version would mishandle certificates
9510 with the same subject name hash and wouldn't handle CRLs at all.
9511 Added -fingerprint option to crl utility, to support new c_rehash
9515 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9518 *) Fix for SSL server purpose checking. Server checking was
9519 rejecting certificates which had extended key usage present
9520 but no ssl client purpose.
9521 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9523 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9524 is a little unclear about how a blank password is handled.
9525 Since the password in encoded as a BMPString with terminating
9526 double NULL a zero length password would end up as just the
9527 double NULL. However no password at all is different and is
9528 handled differently in the PKCS#12 key generation code. NS
9529 treats a blank password as zero length. MSIE treats it as no
9530 password on export: but it will try both on import. We now do
9531 the same: PKCS12_parse() tries zero length and no password if
9532 the password is set to "" or NULL (NULL is now a valid password:
9533 it wasn't before) as does the pkcs12 application.
9536 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9537 perror when PEM_read_bio_X509_REQ fails, the error message must
9538 be obtained from the error queue.
9541 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9542 it in ERR_remove_state if appropriate, and change ERR_get_state
9543 accordingly to avoid race conditions (this is necessary because
9544 thread_hash is no longer constant once set).
9547 *) Bugfix for linux-elf makefile.one.
9550 *) RSA_get_default_method() will now cause a default
9551 RSA_METHOD to be chosen if one doesn't exist already.
9552 Previously this was only set during a call to RSA_new()
9553 or RSA_new_method(NULL) meaning it was possible for
9554 RSA_get_default_method() to return NULL.
9557 *) Added native name translation to the existing DSO code
9558 that will convert (if the flag to do so is set) filenames
9559 that are sufficiently small and have no path information
9560 into a canonical native form. Eg. "blah" converted to
9561 "libblah.so" or "blah.dll" etc.
9564 *) New function ERR_error_string_n(e, buf, len) which is like
9565 ERR_error_string(e, buf), but writes at most 'len' bytes
9566 including the 0 terminator. For ERR_error_string_n, 'buf'
9568 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9570 *) CONF library reworked to become more general. A new CONF
9571 configuration file reader "class" is implemented as well as a
9572 new functions (NCONF_*, for "New CONF") to handle it. The now
9573 old CONF_* functions are still there, but are reimplemented to
9574 work in terms of the new functions. Also, a set of functions
9575 to handle the internal storage of the configuration data is
9576 provided to make it easier to write new configuration file
9577 reader "classes" (I can definitely see something reading a
9578 configuration file in XML format, for example), called _CONF_*,
9579 or "the configuration storage API"...
9581 The new configuration file reading functions are:
9583 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9584 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9586 NCONF_default, NCONF_WIN32
9588 NCONF_dump_fp, NCONF_dump_bio
9590 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9591 NCONF_new creates a new CONF object. This works in the same way
9592 as other interfaces in OpenSSL, like the BIO interface.
9593 NCONF_dump_* dump the internal storage of the configuration file,
9594 which is useful for debugging. All other functions take the same
9595 arguments as the old CONF_* functions wth the exception of the
9596 first that must be a `CONF *' instead of a `LHASH *'.
9598 To make it easer to use the new classes with the old CONF_* functions,
9599 the function CONF_set_default_method is provided.
9602 *) Add '-tls1' option to 'openssl ciphers', which was already
9603 mentioned in the documentation but had not been implemented.
9604 (This option is not yet really useful because even the additional
9605 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9608 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9609 OpenSSL-based applications) load shared libraries and bind to
9610 them in a portable way.
9611 [Geoff Thorpe, with contributions from Richard Levitte]
9613 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9615 *) Make sure _lrotl and _lrotr are only used with MSVC.
9617 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9618 (the default implementation of RAND_status).
9620 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9621 to '-clrext' (= clear extensions), as intended and documented.
9622 [Bodo Moeller; inconsistency pointed out by Michael Attili
9625 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9626 was larger than the MD block size.
9627 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9629 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9630 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9631 using the passed key: if the passed key was a private key the result
9632 of X509_print(), for example, would be to print out all the private key
9636 *) des_quad_cksum() byte order bug fix.
9637 [Ulf Möller, using the problem description in krb4-0.9.7, where
9638 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9640 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9642 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9644 *) For easily testing in shell scripts whether some command
9645 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9646 returns with exit code 0 iff no command of the given name is available.
9647 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9648 the output goes to stdout and nothing is printed to stderr.
9649 Additional arguments are always ignored.
9651 Since for each cipher there is a command of the same name,
9652 the 'no-cipher' compilation switches can be tested this way.
9654 ('openssl no-XXX' is not able to detect pseudo-commands such
9655 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9658 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9661 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9662 is set; it will be thrown away anyway because each handshake creates
9664 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9665 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9666 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9667 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9670 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9671 'Q' and 'R' lose their special meanings (quit/renegotiate).
9672 This is part of what -quiet does; unlike -quiet, -ign_eof
9673 does not suppress any output.
9676 *) Add compatibility options to the purpose and trust code. The
9677 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9678 accepts a certificate or CA, this was the previous behaviour,
9679 with all the associated security issues.
9681 X509_TRUST_COMPAT is the old trust behaviour: only and
9682 automatically trust self signed roots in certificate store. A
9683 new trust setting X509_TRUST_DEFAULT is used to specify that
9684 a purpose has no associated trust setting and it should instead
9685 use the value in the default purpose.
9688 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9689 and fix a memory leak.
9692 *) In util/mkerr.pl (which implements 'make errors'), preserve
9693 reason strings from the previous version of the .c file, as
9694 the default to have only downcase letters (and digits) in
9695 automatically generated reasons codes is not always appropriate.
9698 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9699 using strerror. Previously, ERR_reason_error_string() returned
9700 library names as reason strings for SYSerr; but SYSerr is a special
9701 case where small numbers are errno values, not library numbers.
9704 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9705 converts DSA parameters into DH parameters. (When creating parameters,
9706 DSA_generate_parameters is used.)
9709 *) Include 'length' (recommended exponent length) in C code generated
9710 by 'openssl dhparam -C'.
9713 *) The second argument to set_label in perlasm was already being used
9714 so couldn't be used as a "file scope" flag. Moved to third argument
9718 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9719 instead of RAND_bytes for encryption IVs and salts.
9722 *) Include RAND_status() into RAND_METHOD instead of implementing
9723 it only for md_rand.c Otherwise replacing the PRNG by calling
9724 RAND_set_rand_method would be impossible.
9727 *) Don't let DSA_generate_key() enter an infinite loop if the random
9728 number generation fails.
9731 *) New 'rand' application for creating pseudo-random output.
9734 *) Added configuration support for Linux/IA64
9735 [Rolf Haberrecker <rolf@suse.de>]
9737 *) Assembler module support for Mingw32.
9740 *) Shared library support for HPUX (in shlib/).
9741 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9743 *) Shared library support for Solaris gcc.
9744 [Lutz Behnke <behnke@trustcenter.de>]
9746 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9748 *) PKCS7_encrypt() was adding text MIME headers twice because they
9749 were added manually and by SMIME_crlf_copy().
9752 *) In bntest.c don't call BN_rand with zero bits argument.
9753 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9755 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9756 case was implemented. This caused BN_div_recp() to fail occasionally.
9759 *) Add an optional second argument to the set_label() in the perl
9760 assembly language builder. If this argument exists and is set
9761 to 1 it signals that the assembler should use a symbol whose
9762 scope is the entire file, not just the current function. This
9763 is needed with MASM which uses the format label:: for this scope.
9764 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9766 *) Change the ASN1 types so they are typedefs by default. Before
9767 almost all types were #define'd to ASN1_STRING which was causing
9768 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9772 *) Change names of new functions to the new get1/get0 naming
9773 convention: After 'get1', the caller owns a reference count
9774 and has to call ..._free; 'get0' returns a pointer to some
9775 data structure without incrementing reference counters.
9776 (Some of the existing 'get' functions increment a reference
9777 counter, some don't.)
9778 Similarly, 'set1' and 'add1' functions increase reference
9779 counters or duplicate objects.
9782 *) Allow for the possibility of temp RSA key generation failure:
9783 the code used to assume it always worked and crashed on failure.
9786 *) Fix potential buffer overrun problem in BIO_printf().
9787 [Ulf Möller, using public domain code by Patrick Powell; problem
9788 pointed out by David Sacerdote <das33@cornell.edu>]
9790 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9791 RAND_egd() and RAND_status(). In the command line application,
9792 the EGD socket can be specified like a seed file using RANDFILE
9796 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9797 Some CAs (e.g. Verisign) distribute certificates in this form.
9800 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9801 list to exclude them. This means that no special compilation option
9802 is needed to use anonymous DH: it just needs to be included in the
9806 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9807 EVP_MD_type. The old functionality is available in a new macro called
9808 EVP_MD_md(). Change code that uses it and update docs.
9811 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9812 where the 'void *' argument is replaced by a function pointer argument.
9813 Previously 'void *' was abused to point to functions, which works on
9814 many platforms, but is not correct. As these functions are usually
9815 called by macros defined in OpenSSL header files, most source code
9816 should work without changes.
9819 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9820 sections with information on -D... compiler switches used for
9821 compiling the library so that applications can see them. To enable
9822 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9823 must be defined. E.g.,
9824 #define OPENSSL_ALGORITHM_DEFINES
9825 #include <openssl/opensslconf.h>
9826 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9827 [Richard Levitte, Ulf and Bodo Möller]
9829 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9833 *) Change the 'other' type in certificate aux info to a STACK_OF
9834 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9835 the required ASN1 format: arbitrary types determined by an OID.
9838 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9839 argument to 'req'. This is not because the function is newer or
9840 better than others it just uses the work 'NEW' in the certificate
9841 request header lines. Some software needs this.
9844 *) Reorganise password command line arguments: now passwords can be
9845 obtained from various sources. Delete the PEM_cb function and make
9846 it the default behaviour: i.e. if the callback is NULL and the
9847 usrdata argument is not NULL interpret it as a null terminated pass
9848 phrase. If usrdata and the callback are NULL then the pass phrase
9849 is prompted for as usual.
9852 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9853 the support is automatically enabled. The resulting binaries will
9854 autodetect the card and use it if present.
9855 [Ben Laurie and Compaq Inc.]
9857 *) Work around for Netscape hang bug. This sends certificate request
9858 and server done in one record. Since this is perfectly legal in the
9859 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9860 the bugs/SSLv3 entry for more info.
9863 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9866 *) Add -rand argument to smime and pkcs12 applications and read/write
9870 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9873 *) Add command line password options to the remaining applications.
9876 *) Bug fix for BN_div_recp() for numerators with an even number of
9880 *) More tests in bntest.c, and changed test_bn output.
9883 *) ./config recognizes MacOS X now.
9886 *) Bug fix for BN_div() when the first words of num and divsor are
9887 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9890 *) Add support for various broken PKCS#8 formats, and command line
9891 options to produce them.
9894 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9895 get temporary BIGNUMs from a BN_CTX.
9898 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9902 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9903 include a #define from the old name to the new. The original intent
9904 was that statically linked binaries could for example just call
9905 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9906 link with digests. This never worked because SSLeay_add_all_digests()
9907 and SSLeay_add_all_ciphers() were in the same source file so calling
9908 one would link with the other. They are now in separate source files.
9911 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9914 *) Use a less unusual form of the Miller-Rabin primality test (it used
9915 a binary algorithm for exponentiation integrated into the Miller-Rabin
9916 loop, our standard modexp algorithms are faster).
9919 *) Support for the EBCDIC character set completed.
9920 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9922 *) Source code cleanups: use const where appropriate, eliminate casts,
9923 use void * instead of char * in lhash.
9926 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9927 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9928 this the server could overwrite ephemeral keys that the client
9932 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9933 using 50 iterations of the Rabin-Miller test.
9935 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9936 iterations of the Rabin-Miller test as required by the appendix
9937 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9938 As BN_is_prime_fasttest includes trial division, DSA parameter
9939 generation becomes much faster.
9941 This implies a change for the callback functions in DSA_is_prime
9942 and DSA_generate_parameters: The callback function is called once
9943 for each positive witness in the Rabin-Miller test, not just
9944 occasionally in the inner loop; and the parameters to the
9945 callback function now provide an iteration count for the outer
9946 loop rather than for the current invocation of the inner loop.
9947 DSA_generate_parameters additionally can call the callback
9948 function with an 'iteration count' of -1, meaning that a
9949 candidate has passed the trial division test (when q is generated
9950 from an application-provided seed, trial division is skipped).
9953 *) New function BN_is_prime_fasttest that optionally does trial
9954 division before starting the Rabin-Miller test and has
9955 an additional BN_CTX * argument (whereas BN_is_prime always
9956 has to allocate at least one BN_CTX).
9957 'callback(1, -1, cb_arg)' is called when a number has passed the
9958 trial division stage.
9961 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9965 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9968 *) New function BN_pseudo_rand().
9971 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9972 bignum version of BN_from_montgomery() with the working code from
9973 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9977 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9978 made it impossible to use the same SSL_SESSION data structure in
9979 SSL2 clients in multiple threads.
9982 *) The return value of RAND_load_file() no longer counts bytes obtained
9983 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9984 to seed the PRNG (previously an explicit byte count was required).
9985 [Ulf Möller, Bodo Möller]
9987 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9988 used (char *) instead of (void *) and had casts all over the place.
9991 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9994 *) Retain source code compatibility for BN_prime_checks macro:
9995 BN_is_prime(..., BN_prime_checks, ...) now uses
9996 BN_prime_checks_for_size to determine the appropriate number of
9997 Rabin-Miller iterations.
10000 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10001 DH_CHECK_P_NOT_SAFE_PRIME.
10002 (Check if this is true? OpenPGP calls them "strong".)
10005 *) Merge the functionality of "dh" and "gendh" programs into a new program
10006 "dhparam". The old programs are retained for now but will handle DH keys
10007 (instead of parameters) in future.
10010 *) Make the ciphers, s_server and s_client programs check the return values
10011 when a new cipher list is set.
10014 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10015 ciphers. Before when the 56bit ciphers were enabled the sorting was
10018 The syntax for the cipher sorting has been extended to support sorting by
10019 cipher-strength (using the strength_bits hard coded in the tables).
10020 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10022 Fix a bug in the cipher-command parser: when supplying a cipher command
10023 string with an "undefined" symbol (neither command nor alphanumeric
10024 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10025 an error is flagged.
10027 Due to the strength-sorting extension, the code of the
10028 ssl_create_cipher_list() function was completely rearranged. I hope that
10029 the readability was also increased :-)
10030 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10032 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10033 for the first serial number and places 2 in the serial number file. This
10034 avoids problems when the root CA is created with serial number zero and
10035 the first user certificate has the same issuer name and serial number
10039 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10040 the new code. Add documentation for this stuff.
10043 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10044 X509_*() to X509at_*() on the grounds that they don't handle X509
10045 structures and behave in an analogous way to the X509v3 functions:
10046 they shouldn't be called directly but wrapper functions should be used
10049 So we also now have some wrapper functions that call the X509at functions
10050 when passed certificate requests. (TO DO: similar things can be done with
10051 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10052 things. Some of these need some d2i or i2d and print functionality
10053 because they handle more complex structures.)
10056 *) Add missing #ifndefs that caused missing symbols when building libssl
10057 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10058 NO_RSA in ssl/s2*.c.
10059 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10061 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10062 has a return value which indicates the quality of the random data
10063 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10064 error queue. New function RAND_pseudo_bytes() generates output that is
10065 guaranteed to be unique but not unpredictable. RAND_add is like
10066 RAND_seed, but takes an extra argument for an entropy estimate
10067 (RAND_seed always assumes full entropy).
10070 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10071 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10072 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10073 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10074 false-positive rate of at most 2^-80 for random input.
10077 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10080 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10081 in the 0.9.5 release), this returns the chain
10082 from an X509_CTX structure with a dup of the stack and all
10083 the X509 reference counts upped: so the stack will exist
10084 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10087 Also make SSL_SESSION_print() print out the verify return
10091 *) Add manpage for the pkcs12 command. Also change the default
10092 behaviour so MAC iteration counts are used unless the new
10093 -nomaciter option is used. This improves file security and
10094 only older versions of MSIE (4.0 for example) need it.
10097 *) Honor the no-xxx Configure options when creating .DEF files.
10100 *) Add PKCS#10 attributes to field table: challengePassword,
10101 unstructuredName and unstructuredAddress. These are taken from
10102 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10103 international characters are used.
10105 More changes to X509_ATTRIBUTE code: allow the setting of types
10106 based on strings. Remove the 'loc' parameter when adding
10107 attributes because these will be a SET OF encoding which is sorted
10111 *) Initial changes to the 'req' utility to allow request generation
10112 automation. This will allow an application to just generate a template
10113 file containing all the field values and have req construct the
10116 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10117 used all over the place including certificate requests and PKCS#7
10118 structures. They are currently handled manually where necessary with
10119 some primitive wrappers for PKCS#7. The new functions behave in a
10120 manner analogous to the X509 extension functions: they allow
10121 attributes to be looked up by NID and added.
10123 Later something similar to the X509V3 code would be desirable to
10124 automatically handle the encoding, decoding and printing of the
10125 more complex types. The string types like challengePassword can
10126 be handled by the string table functions.
10128 Also modified the multi byte string table handling. Now there is
10129 a 'global mask' which masks out certain types. The table itself
10130 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10131 is useful when for example there is only one permissible type
10132 (as in countryName) and using the mask might result in no valid
10136 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10137 SSL_get_peer_finished to allow applications to obtain the latest
10138 Finished messages sent to the peer or expected from the peer,
10139 respectively. (SSL_get_peer_finished is usually the Finished message
10140 actually received from the peer, otherwise the protocol will be aborted.)
10142 As the Finished message are message digests of the complete handshake
10143 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10144 be used for external authentication procedures when the authentication
10145 provided by SSL/TLS is not desired or is not enough.
10148 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10149 the host supports BWX extension and if Compaq C is present on the
10150 $PATH. Just exploiting of the BWX extension results in 20-30%
10151 performance kick for some algorithms, e.g. DES and RC4 to mention
10152 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10156 *) Add support for MS "fast SGC". This is arguably a violation of the
10157 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10158 weak crypto and after checking the certificate is SGC a second one
10159 with strong crypto. MS SGC stops the first handshake after receiving
10160 the server certificate message and sends a second client hello. Since
10161 a server will typically do all the time consuming operations before
10162 expecting any further messages from the client (server key exchange
10163 is the most expensive) there is little difference between the two.
10165 To get OpenSSL to support MS SGC we have to permit a second client
10166 hello message after we have sent server done. In addition we have to
10167 reset the MAC if we do get this second client hello.
10170 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10171 if a DER encoded private key is RSA or DSA traditional format. Changed
10172 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10173 format DER encoded private key. Newer code should use PKCS#8 format which
10174 has the key type encoded in the ASN1 structure. Added DER private key
10175 support to pkcs8 application.
10178 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10179 ciphersuites has been selected (as required by the SSL 3/TLS 1
10180 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10181 is set, we interpret this as a request to violate the specification
10182 (the worst that can happen is a handshake failure, and 'correct'
10183 behaviour would result in a handshake failure anyway).
10186 *) In SSL_CTX_add_session, take into account that there might be multiple
10187 SSL_SESSION structures with the same session ID (e.g. when two threads
10188 concurrently obtain them from an external cache).
10189 The internal cache can handle only one SSL_SESSION with a given ID,
10190 so if there's a conflict, we now throw out the old one to achieve
10194 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10195 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10196 some routines that use cipher OIDs: some ciphers do not have OIDs
10197 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10201 *) Simplify the trust setting structure and code. Now we just have
10202 two sequences of OIDs for trusted and rejected settings. These will
10203 typically have values the same as the extended key usage extension
10204 and any application specific purposes.
10206 The trust checking code now has a default behaviour: it will just
10207 check for an object with the same NID as the passed id. Functions can
10208 be provided to override either the default behaviour or the behaviour
10209 for a given id. SSL client, server and email already have functions
10210 in place for compatibility: they check the NID and also return "trusted"
10211 if the certificate is self signed.
10214 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10215 traditional format into an EVP_PKEY structure.
10218 *) Add a password callback function PEM_cb() which either prompts for
10219 a password if usr_data is NULL or otherwise assumes it is a null
10220 terminated password. Allow passwords to be passed on command line
10221 environment or config files in a few more utilities.
10224 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10225 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10226 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10227 Update documentation.
10230 *) Support for ASN1 "NULL" type. This could be handled before by using
10231 ASN1_TYPE but there wasn't any function that would try to read a NULL
10232 and produce an error if it couldn't. For compatibility we also have
10233 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10234 don't allocate anything because they don't need to.
10237 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10239 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10241 *) Rebuild of the memory allocation routines used by OpenSSL code and
10242 possibly others as well. The purpose is to make an interface that
10243 provide hooks so anyone can build a separate set of allocation and
10244 deallocation routines to be used by OpenSSL, for example memory
10245 pool implementations, or something else, which was previously hard
10246 since Malloc(), Realloc() and Free() were defined as macros having
10247 the values malloc, realloc and free, respectively (except for Win32
10248 compilations). The same is provided for memory debugging code.
10249 OpenSSL already comes with functionality to find memory leaks, but
10250 this gives people a chance to debug other memory problems.
10252 With these changes, a new set of functions and macros have appeared:
10254 CRYPTO_set_mem_debug_functions() [F]
10255 CRYPTO_get_mem_debug_functions() [F]
10256 CRYPTO_dbg_set_options() [F]
10257 CRYPTO_dbg_get_options() [F]
10258 CRYPTO_malloc_debug_init() [M]
10260 The memory debug functions are NULL by default, unless the library
10261 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10262 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10263 gives the standard debugging functions that come with OpenSSL) or
10264 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10265 provided by the library user) must be used. When the standard
10266 debugging functions are used, CRYPTO_dbg_set_options can be used to
10267 request additional information:
10268 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10269 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10271 Also, things like CRYPTO_set_mem_functions will always give the
10272 expected result (the new set of functions is used for allocation
10273 and deallocation) at all times, regardless of platform and compiler
10276 To finish it up, some functions that were never use in any other
10277 way than through macros have a new API and new semantic:
10279 CRYPTO_dbg_malloc()
10280 CRYPTO_dbg_realloc()
10283 All macros of value have retained their old syntax.
10284 [Richard Levitte and Bodo Moeller]
10286 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10287 ordering of SMIMECapabilities wasn't in "strength order" and there
10288 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10292 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10293 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10294 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10296 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10297 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10298 functionality to handle multipart/signed properly) and a utility
10299 called 'smime' to call all this stuff. This is based on code I
10300 originally wrote for Celo who have kindly allowed it to be
10301 included in OpenSSL.
10304 *) Add variants des_set_key_checked and des_set_key_unchecked of
10305 des_set_key (aka des_key_sched). Global variable des_check_key
10306 decides which of these is called by des_set_key; this way
10307 des_check_key behaves as it always did, but applications and
10308 the library itself, which was buggy for des_check_key == 1,
10309 have a cleaner way to pick the version they need.
10312 *) New function PKCS12_newpass() which changes the password of a
10316 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10317 dynamic mix. In both cases the ids can be used as an index into the
10318 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10319 functions so they accept a list of the field values and the
10320 application doesn't need to directly manipulate the X509_TRUST
10324 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10328 *) Modify the way the V3 extension code looks up extensions. This now
10329 works in a similar way to the object code: we have some "standard"
10330 extensions in a static table which is searched with OBJ_bsearch()
10331 and the application can add dynamic ones if needed. The file
10332 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10333 updated whenever a new extension is added to the core code and kept
10334 in ext_nid order. There is a simple program 'tabtest.c' which checks
10335 this. New extensions are not added too often so this file can readily
10336 be maintained manually.
10338 There are two big advantages in doing things this way. The extensions
10339 can be looked up immediately and no longer need to be "added" using
10340 X509V3_add_standard_extensions(): this function now does nothing.
10341 [Side note: I get *lots* of email saying the extension code doesn't
10342 work because people forget to call this function]
10343 Also no dynamic allocation is done unless new extensions are added:
10344 so if we don't add custom extensions there is no need to call
10345 X509V3_EXT_cleanup().
10348 *) Modify enc utility's salting as follows: make salting the default. Add a
10349 magic header, so unsalted files fail gracefully instead of just decrypting
10350 to garbage. This is because not salting is a big security hole, so people
10351 should be discouraged from doing it.
10354 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10355 digest to be passed on the command line but it only used this
10356 parameter when signing a certificate. Modified so all relevant
10357 operations are affected by the digest parameter including the
10358 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10359 DSA key was used because it didn't fix the digest.
10362 *) Initial certificate chain verify code. Currently tests the untrusted
10363 certificates for consistency with the verify purpose (which is set
10364 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10366 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10367 this is because it will reject chains with invalid extensions whereas
10368 every previous version of OpenSSL and SSLeay made no checks at all.
10370 Trust code: checks the root CA for the relevant trust settings. Trust
10371 settings have an initial value consistent with the verify purpose: e.g.
10372 if the verify purpose is for SSL client use it expects the CA to be
10373 trusted for SSL client use. However the default value can be changed to
10374 permit custom trust settings: one example of this would be to only trust
10375 certificates from a specific "secure" set of CAs.
10377 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10378 which should be used for version portability: especially since the
10379 verify structure is likely to change more often now.
10381 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10382 to set them. If not set then assume SSL clients will verify SSL servers
10385 Two new options to the verify program: -untrusted allows a set of
10386 untrusted certificates to be passed in and -purpose which sets the
10387 intended purpose of the certificate. If a purpose is set then the
10388 new chain verify code is used to check extension consistency.
10391 *) Support for the authority information access extension.
10394 *) Modify RSA and DSA PEM read routines to transparently handle
10395 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10396 public keys in a format compatible with certificate
10397 SubjectPublicKeyInfo structures. Unfortunately there were already
10398 functions called *_PublicKey_* which used various odd formats so
10399 these are retained for compatibility: however the DSA variants were
10400 never in a public release so they have been deleted. Changed dsa/rsa
10401 utilities to handle the new format: note no releases ever handled public
10402 keys so we should be OK.
10404 The primary motivation for this change is to avoid the same fiasco
10405 that dogs private keys: there are several incompatible private key
10406 formats some of which are standard and some OpenSSL specific and
10407 require various evil hacks to allow partial transparent handling and
10408 even then it doesn't work with DER formats. Given the option anything
10409 other than PKCS#8 should be dumped: but the other formats have to
10410 stay in the name of compatibility.
10412 With public keys and the benefit of hindsight one standard format
10413 is used which works with EVP_PKEY, RSA or DSA structures: though
10414 it clearly returns an error if you try to read the wrong kind of key.
10416 Added a -pubkey option to the 'x509' utility to output the public key.
10417 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10418 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10419 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10420 that do the same as the EVP_PKEY_assign_*() except they up the
10421 reference count of the added key (they don't "swallow" the
10425 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10426 CRLs would fail if the file contained no certificates or no CRLs:
10427 added a new function to read in both types and return the number
10428 read: this means that if none are read it will be an error. The
10429 DER versions of the certificate and CRL reader would always fail
10430 because it isn't possible to mix certificates and CRLs in DER format
10431 without choking one or the other routine. Changed this to just read
10432 a certificate: this is the best we can do. Also modified the code
10433 in apps/verify.c to take notice of return codes: it was previously
10434 attempting to read in certificates from NULL pointers and ignoring
10435 any errors: this is one reason why the cert and CRL reader seemed
10436 to work. It doesn't check return codes from the default certificate
10437 routines: these may well fail if the certificates aren't installed.
10440 *) Code to support otherName option in GeneralName.
10443 *) First update to verify code. Change the verify utility
10444 so it warns if it is passed a self signed certificate:
10445 for consistency with the normal behaviour. X509_verify
10446 has been modified to it will now verify a self signed
10447 certificate if *exactly* the same certificate appears
10448 in the store: it was previously impossible to trust a
10449 single self signed certificate. This means that:
10450 openssl verify ss.pem
10451 now gives a warning about a self signed certificate but
10452 openssl verify -CAfile ss.pem ss.pem
10456 *) For servers, store verify_result in SSL_SESSION data structure
10457 (and add it to external session representation).
10458 This is needed when client certificate verifications fails,
10459 but an application-provided verification callback (set by
10460 SSL_CTX_set_cert_verify_callback) allows accepting the session
10461 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10462 but returns 1): When the session is reused, we have to set
10463 ssl->verify_result to the appropriate error code to avoid
10465 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10467 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10468 case in PKCS7_dataInit() where the signed PKCS7 structure
10469 didn't contain any existing data because it was being created.
10470 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10472 *) Add a salt to the key derivation routines in enc.c. This
10473 forms the first 8 bytes of the encrypted file. Also add a
10474 -S option to allow a salt to be input on the command line.
10477 *) New function X509_cmp(). Oddly enough there wasn't a function
10478 to compare two certificates. We do this by working out the SHA1
10479 hash and comparing that. X509_cmp() will be needed by the trust
10483 *) SSL_get1_session() is like SSL_get_session(), but increments
10484 the reference count in the SSL_SESSION returned.
10485 [Geoff Thorpe <geoff@eu.c2.net>]
10487 *) Fix for 'req': it was adding a null to request attributes.
10488 Also change the X509_LOOKUP and X509_INFO code to handle
10489 certificate auxiliary information.
10492 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10496 *) Add the possibility to add extra information to the memory leak
10497 detecting output, to form tracebacks, showing from where each
10498 allocation was originated: CRYPTO_push_info("constant string") adds
10499 the string plus current file name and line number to a per-thread
10500 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10501 is like calling CYRPTO_pop_info() until the stack is empty.
10502 Also updated memory leak detection code to be multi-thread-safe.
10505 *) Add options -text and -noout to pkcs7 utility and delete the
10506 encryption options which never did anything. Update docs.
10509 *) Add options to some of the utilities to allow the pass phrase
10510 to be included on either the command line (not recommended on
10511 OSes like Unix) or read from the environment. Update the
10512 manpages and fix a few bugs.
10515 *) Add a few manpages for some of the openssl commands.
10518 *) Fix the -revoke option in ca. It was freeing up memory twice,
10519 leaking and not finding already revoked certificates.
10522 *) Extensive changes to support certificate auxiliary information.
10523 This involves the use of X509_CERT_AUX structure and X509_AUX
10524 functions. An X509_AUX function such as PEM_read_X509_AUX()
10525 can still read in a certificate file in the usual way but it
10526 will also read in any additional "auxiliary information". By
10527 doing things this way a fair degree of compatibility can be
10528 retained: existing certificates can have this information added
10529 using the new 'x509' options.
10531 Current auxiliary information includes an "alias" and some trust
10532 settings. The trust settings will ultimately be used in enhanced
10533 certificate chain verification routines: currently a certificate
10534 can only be trusted if it is self signed and then it is trusted
10538 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10539 The problem was that one of the replacement routines had not been working
10540 since SSLeay releases. For now the offending routine has been replaced
10541 with non-optimised assembler. Even so, this now gives around 95%
10542 performance improvement for 1024 bit RSA signs.
10545 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10546 handling. Most clients have the effective key size in bits equal to
10547 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10548 A few however don't do this and instead use the size of the decrypted key
10549 to determine the RC2 key length and the AlgorithmIdentifier to determine
10550 the effective key length. In this case the effective key length can still
10551 be 40 bits but the key length can be 168 bits for example. This is fixed
10552 by manually forcing an RC2 key into the EVP_PKEY structure because the
10553 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10554 the key length and effective key length are equal.
10557 *) Add a bunch of functions that should simplify the creation of
10558 X509_NAME structures. Now you should be able to do:
10559 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10560 and have it automatically work out the correct field type and fill in
10561 the structures. The more adventurous can try:
10562 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10563 and it will (hopefully) work out the correct multibyte encoding.
10566 *) Change the 'req' utility to use the new field handling and multibyte
10567 copy routines. Before the DN field creation was handled in an ad hoc
10568 way in req, ca, and x509 which was rather broken and didn't support
10569 BMPStrings or UTF8Strings. Since some software doesn't implement
10570 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10571 using the dirstring_type option. See the new comment in the default
10572 openssl.cnf for more info.
10575 *) Make crypto/rand/md_rand.c more robust:
10576 - Assure unique random numbers after fork().
10577 - Make sure that concurrent threads access the global counter and
10578 md serializably so that we never lose entropy in them
10579 or use exactly the same state in multiple threads.
10580 Access to the large state is not always serializable because
10581 the additional locking could be a performance killer, and
10582 md should be large enough anyway.
10585 *) New file apps/app_rand.c with commonly needed functionality
10586 for handling the random seed file.
10588 Use the random seed file in some applications that previously did not:
10590 dsaparam -genkey (which also ignored its '-rand' option),
10593 x509 (when signing).
10594 Except on systems with /dev/urandom, it is crucial to have a random
10595 seed file at least for key creation, DSA signing, and for DH exchanges;
10596 for RSA signatures we could do without one.
10598 gendh and gendsa (unlike genrsa) used to read only the first byte
10599 of each file listed in the '-rand' option. The function as previously
10600 found in genrsa is now in app_rand.c and is used by all programs
10601 that support '-rand'.
10604 *) In RAND_write_file, use mode 0600 for creating files;
10605 don't just chmod when it may be too late.
10608 *) Report an error from X509_STORE_load_locations
10609 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10612 *) New function ASN1_mbstring_copy() this copies a string in either
10613 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10614 into an ASN1_STRING type. A mask of permissible types is passed
10615 and it chooses the "minimal" type to use or an error if not type
10619 *) Add function equivalents to the various macros in asn1.h. The old
10620 macros are retained with an M_ prefix. Code inside the library can
10621 use the M_ macros. External code (including the openssl utility)
10622 should *NOT* in order to be "shared library friendly".
10625 *) Add various functions that can check a certificate's extensions
10626 to see if it usable for various purposes such as SSL client,
10627 server or S/MIME and CAs of these types. This is currently
10628 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10629 verification. Also added a -purpose flag to x509 utility to
10630 print out all the purposes.
10633 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10637 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10638 for, obtain and decode and extension and obtain its critical flag.
10639 This allows all the necessary extension code to be handled in a
10640 single function call.
10643 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10644 platforms. See crypto/rc4/rc4_enc.c for further details.
10647 *) New -noout option to asn1parse. This causes no output to be produced
10648 its main use is when combined with -strparse and -out to extract data
10649 from a file (which may not be in ASN.1 format).
10652 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10653 when producing the local key id.
10654 [Richard Levitte <levitte@stacken.kth.se>]
10656 *) New option -dhparam in s_server. This allows a DH parameter file to be
10657 stated explicitly. If it is not stated then it tries the first server
10658 certificate file. The previous behaviour hard coded the filename
10662 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10663 a public key to be input or output. For example:
10664 openssl rsa -in key.pem -pubout -out pubkey.pem
10665 Also added necessary DSA public key functions to handle this.
10668 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10669 in the message. This was handled by allowing
10670 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10671 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10673 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10674 to the end of the strings whereas this didn't. This would cause problems
10675 if strings read with d2i_ASN1_bytes() were later modified.
10676 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10678 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10679 data and it contains EOF it will end up returning an error. This is
10680 caused by input 46 bytes long. The cause is due to the way base64
10681 BIOs find the start of base64 encoded data. They do this by trying a
10682 trial decode on each line until they find one that works. When they
10683 do a flag is set and it starts again knowing it can pass all the
10684 data directly through the decoder. Unfortunately it doesn't reset
10685 the context it uses. This means that if EOF is reached an attempt
10686 is made to pass two EOFs through the context and this causes the
10687 resulting error. This can also cause other problems as well. As is
10688 usual with these problems it takes *ages* to find and the fix is
10689 trivial: move one line.
10690 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10692 *) Ugly workaround to get s_client and s_server working under Windows. The
10693 old code wouldn't work because it needed to select() on sockets and the
10694 tty (for keypresses and to see if data could be written). Win32 only
10695 supports select() on sockets so we select() with a 1s timeout on the
10696 sockets and then see if any characters are waiting to be read, if none
10697 are present then we retry, we also assume we can always write data to
10698 the tty. This isn't nice because the code then blocks until we've
10699 received a complete line of data and it is effectively polling the
10700 keyboard at 1s intervals: however it's quite a bit better than not
10701 working at all :-) A dedicated Windows application might handle this
10702 with an event loop for example.
10705 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10706 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10707 will be called when RSA_sign() and RSA_verify() are used. This is useful
10708 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10709 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10710 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10711 This necessitated the support of an extra signature type NID_md5_sha1
10712 for SSL signatures and modifications to the SSL library to use it instead
10713 of calling RSA_public_decrypt() and RSA_private_encrypt().
10716 *) Add new -verify -CAfile and -CApath options to the crl program, these
10717 will lookup a CRL issuers certificate and verify the signature in a
10718 similar way to the verify program. Tidy up the crl program so it
10719 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10720 less strict. It will now permit CRL extensions even if it is not
10721 a V2 CRL: this will allow it to tolerate some broken CRLs.
10724 *) Initialize all non-automatic variables each time one of the openssl
10725 sub-programs is started (this is necessary as they may be started
10726 multiple times from the "OpenSSL>" prompt).
10727 [Lennart Bang, Bodo Moeller]
10729 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10730 removing all other RSA functionality (this is what NO_RSA does). This
10731 is so (for example) those in the US can disable those operations covered
10732 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10736 *) Non-copying interface to BIO pairs.
10737 (still largely untested)
10740 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10741 ASCII string. This was handled independently in various places before.
10744 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10745 UTF8 strings a character at a time.
10748 *) Use client_version from client hello to select the protocol
10749 (s23_srvr.c) and for RSA client key exchange verification
10750 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10753 *) Add various utility functions to handle SPKACs, these were previously
10754 handled by poking round in the structure internals. Added new function
10755 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10756 print, verify and generate SPKACs. Based on an original idea from
10757 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10760 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10763 *) Allow the config file extension section to be overwritten on the
10764 command line. Based on an original idea from Massimiliano Pala
10765 <madwolf@comune.modena.it>. The new option is called -extensions
10766 and can be applied to ca, req and x509. Also -reqexts to override
10767 the request extensions in req and -crlexts to override the crl extensions
10771 *) Add new feature to the SPKAC handling in ca. Now you can include
10772 the same field multiple times by preceding it by "XXXX." for example:
10775 this is the same syntax as used in the req config file.
10778 *) Allow certificate extensions to be added to certificate requests. These
10779 are specified in a 'req_extensions' option of the req section of the
10780 config file. They can be printed out with the -text option to req but
10781 are otherwise ignored at present.
10784 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10785 data read consists of only the final block it would not decrypted because
10786 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10787 A misplaced 'break' also meant the decrypted final block might not be
10788 copied until the next read.
10791 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10792 a few extra parameters to the DH structure: these will be useful if
10793 for example we want the value of 'q' or implement X9.42 DH.
10796 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10797 provides hooks that allow the default DSA functions or functions on a
10798 "per key" basis to be replaced. This allows hardware acceleration and
10799 hardware key storage to be handled without major modification to the
10800 library. Also added low level modexp hooks and CRYPTO_EX structure and
10801 associated functions.
10804 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10805 as "read only": it can't be written to and the buffer it points to will
10806 not be freed. Reading from a read only BIO is much more efficient than
10807 a normal memory BIO. This was added because there are several times when
10808 an area of memory needs to be read from a BIO. The previous method was
10809 to create a memory BIO and write the data to it, this results in two
10810 copies of the data and an O(n^2) reading algorithm. There is a new
10811 function BIO_new_mem_buf() which creates a read only memory BIO from
10812 an area of memory. Also modified the PKCS#7 routines to use read only
10816 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10817 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10818 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10819 but a retry condition occurred while trying to read the rest.
10822 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10823 NID_pkcs7_encrypted by default: this was wrong since this should almost
10824 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10825 the encrypted data type: this is a more sensible place to put it and it
10826 allows the PKCS#12 code to be tidied up that duplicated this
10830 *) Changed obj_dat.pl script so it takes its input and output files on
10831 the command line. This should avoid shell escape redirection problems
10835 *) Initial support for certificate extension requests, these are included
10836 in things like Xenroll certificate requests. Included functions to allow
10837 extensions to be obtained and added.
10840 *) -crlf option to s_client and s_server for sending newlines as
10841 CRLF (as required by many protocols).
10844 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10846 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10847 [Ralf S. Engelschall]
10849 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10850 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10852 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10856 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10857 DH parameters/keys (q is lost during that conversion, but the resulting
10858 DH parameters contain its length).
10860 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10861 much faster than DH_generate_parameters (which creates parameters
10862 where p = 2*q + 1), and also the smaller q makes DH computations
10863 much more efficient (160-bit exponentiation instead of 1024-bit
10864 exponentiation); so this provides a convenient way to support DHE
10865 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10866 utter importance to use
10867 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10869 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10870 when such DH parameters are used, because otherwise small subgroup
10871 attacks may become possible!
10874 *) Avoid memory leak in i2d_DHparams.
10877 *) Allow the -k option to be used more than once in the enc program:
10878 this allows the same encrypted message to be read by multiple recipients.
10881 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10882 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10883 it will always use the numerical form of the OID, even if it has a short
10887 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10888 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10889 otherwise bn_mod_exp was called. In the case of hardware keys for example
10890 no private key components need be present and it might store extra data
10891 in the RSA structure, which cannot be accessed from bn_mod_exp.
10892 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10893 private key operations.
10896 *) Added support for SPARC Linux.
10899 *) pem_password_cb function type incompatibly changed from
10900 typedef int pem_password_cb(char *buf, int size, int rwflag);
10902 ....(char *buf, int size, int rwflag, void *userdata);
10903 so that applications can pass data to their callbacks:
10904 The PEM[_ASN1]_{read,write}... functions and macros now take an
10905 additional void * argument, which is just handed through whenever
10906 the password callback is called.
10907 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10909 New function SSL_CTX_set_default_passwd_cb_userdata.
10911 Compatibility note: As many C implementations push function arguments
10912 onto the stack in reverse order, the new library version is likely to
10913 interoperate with programs that have been compiled with the old
10914 pem_password_cb definition (PEM_whatever takes some data that
10915 happens to be on the stack as its last argument, and the callback
10916 just ignores this garbage); but there is no guarantee whatsoever that
10919 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10920 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10921 problems not only on Windows, but also on some Unix platforms.
10922 To avoid problematic command lines, these definitions are now in an
10923 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10924 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10927 *) MIPS III/IV assembler module is reimplemented.
10930 *) More DES library cleanups: remove references to srand/rand and
10931 delete an unused file.
10934 *) Add support for the the free Netwide assembler (NASM) under Win32,
10935 since not many people have MASM (ml) and it can be hard to obtain.
10936 This is currently experimental but it seems to work OK and pass all
10937 the tests. Check out INSTALL.W32 for info.
10940 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10941 without temporary keys kept an extra copy of the server key,
10942 and connections with temporary keys did not free everything in case
10946 *) New function RSA_check_key and new openssl rsa option -check
10947 for verifying the consistency of RSA keys.
10948 [Ulf Moeller, Bodo Moeller]
10950 *) Various changes to make Win32 compile work:
10951 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10952 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10953 comparison" warnings.
10954 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10957 *) Add a debugging option to PKCS#5 v2 key generation function: when
10958 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10959 derived keys are printed to stderr.
10962 *) Copy the flags in ASN1_STRING_dup().
10963 [Roman E. Pavlov <pre@mo.msk.ru>]
10965 *) The x509 application mishandled signing requests containing DSA
10966 keys when the signing key was also DSA and the parameters didn't match.
10968 It was supposed to omit the parameters when they matched the signing key:
10969 the verifying software was then supposed to automatically use the CA's
10970 parameters if they were absent from the end user certificate.
10972 Omitting parameters is no longer recommended. The test was also
10973 the wrong way round! This was probably due to unusual behaviour in
10974 EVP_cmp_parameters() which returns 1 if the parameters match.
10975 This meant that parameters were omitted when they *didn't* match and
10976 the certificate was useless. Certificates signed with 'ca' didn't have
10978 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10980 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10981 The interface is as follows:
10982 Applications can use
10983 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10984 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10985 "off" is now the default.
10986 The library internally uses
10987 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10988 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10989 to disable memory-checking temporarily.
10991 Some inconsistent states that previously were possible (and were
10992 even the default) are now avoided.
10994 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10995 with each memory chunk allocated; this is occasionally more helpful
10996 than just having a counter.
10998 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11000 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11004 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11005 which largely parallels "options", but is for changing API behaviour,
11006 whereas "options" are about protocol behaviour.
11007 Initial "mode" flags are:
11009 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11010 a single record has been written.
11011 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11012 retries use the same buffer location.
11013 (But all of the contents must be
11017 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11020 *) Fix problems with no-hmac etc.
11021 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11023 *) New functions RSA_get_default_method(), RSA_set_method() and
11024 RSA_get_method(). These allows replacement of RSA_METHODs without having
11025 to mess around with the internals of an RSA structure.
11028 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11029 Also really enable memory leak checks in openssl.c and in some
11031 [Chad C. Mulligan, Bodo Moeller]
11033 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11034 up the length of negative integers. This has now been simplified to just
11035 store the length when it is first determined and use it later, rather
11036 than trying to keep track of where data is copied and updating it to
11038 [Steve Henson, reported by Brien Wheeler
11039 <bwheeler@authentica-security.com>]
11041 *) Add a new function PKCS7_signatureVerify. This allows the verification
11042 of a PKCS#7 signature but with the signing certificate passed to the
11043 function itself. This contrasts with PKCS7_dataVerify which assumes the
11044 certificate is present in the PKCS#7 structure. This isn't always the
11045 case: certificates can be omitted from a PKCS#7 structure and be
11046 distributed by "out of band" means (such as a certificate database).
11049 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11050 function prototypes in pem.h, also change util/mkdef.pl to add the
11051 necessary function names.
11054 *) mk1mf.pl (used by Windows builds) did not properly read the
11055 options set by Configure in the top level Makefile, and Configure
11056 was not even able to write more than one option correctly.
11057 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11060 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11061 file to be loaded from a BIO or FILE pointer. The BIO version will
11062 for example allow memory BIOs to contain config info.
11065 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11066 Whoever hopes to achieve shared-library compatibility across versions
11067 must use this, not the compile-time macro.
11068 (Exercise 0.9.4: Which is the minimum library version required by
11070 Note: All this applies only to multi-threaded programs, others don't
11074 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11075 through a BIO pair triggered the default case, i.e.
11076 SSLerr(...,SSL_R_UNKNOWN_STATE).
11079 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11080 can use the SSL library even if none of the specific BIOs is
11084 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11085 for the encoded length.
11086 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11088 *) Add initial documentation of the X509V3 functions.
11091 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11092 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11093 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11094 secure PKCS#8 private key format with a high iteration count.
11097 *) Fix determination of Perl interpreter: A perl or perl5
11098 _directory_ in $PATH was also accepted as the interpreter.
11099 [Ralf S. Engelschall]
11101 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11102 wrong with it but it was very old and did things like calling
11103 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11104 unusual formatting.
11107 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11108 to use the new extension code.
11111 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11112 with macros. This should make it easier to change their form, add extra
11113 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11117 *) Add to configuration table a new entry that can specify an alternative
11118 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11119 according to Mark Crispin <MRC@Panda.COM>.
11123 *) DES CBC did not update the IV. Weird.
11126 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11127 Changing the behaviour of the former might break existing programs --
11128 where IV updating is needed, des_ncbc_encrypt can be used.
11131 *) When bntest is run from "make test" it drives bc to check its
11132 calculations, as well as internally checking them. If an internal check
11133 fails, it needs to cause bc to give a non-zero result or make test carries
11134 on without noticing the failure. Fixed.
11137 *) DES library cleanups.
11140 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11141 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11142 ciphers. NOTE: although the key derivation function has been verified
11143 against some published test vectors it has not been extensively tested
11144 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11148 *) Instead of "mkdir -p", which is not fully portable, use new
11149 Perl script "util/mkdir-p.pl".
11152 *) Rewrite the way password based encryption (PBE) is handled. It used to
11153 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11154 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11155 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11156 the 'parameter' field of the AlgorithmIdentifier is passed to the
11157 underlying key generation function so it must do its own ASN1 parsing.
11158 This has also changed the EVP_PBE_CipherInit() function which now has a
11159 'parameter' argument instead of literal salt and iteration count values
11160 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11163 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11164 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11165 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11166 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11167 value was just used as a "magic string" and not used directly its
11168 value doesn't matter.
11171 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11175 *) "linux-sparc64" configuration (ultrapenguin).
11176 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11177 "linux-sparc" configuration.
11178 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11180 *) config now generates no-xxx options for missing ciphers.
11183 *) Support the EBCDIC character set (work in progress).
11184 File ebcdic.c not yet included because it has a different license.
11185 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11187 *) Support BS2000/OSD-POSIX.
11188 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11190 *) Make callbacks for key generation use void * instead of char *.
11193 *) Make S/MIME samples compile (not yet tested).
11196 *) Additional typesafe stacks.
11199 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11203 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11205 *) New configuration variant "sco5-gcc".
11207 *) Updated some demos.
11208 [Sean O Riordain, Wade Scholine]
11210 *) Add missing BIO_free at exit of pkcs12 application.
11213 *) Fix memory leak in conf.c.
11216 *) Updates for Win32 to assembler version of MD5.
11219 *) Set #! path to perl in apps/der_chop to where we found it
11220 instead of using a fixed path.
11223 *) SHA library changes for irix64-mips4-cc.
11226 *) Improvements for VMS support.
11230 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11232 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11233 This also avoids the problems with SC4.2 and unpatched SC5.
11234 [Andy Polyakov <appro@fy.chalmers.se>]
11236 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11237 These are required because of the typesafe stack would otherwise break
11238 existing code. If old code used a structure member which used to be STACK
11239 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11240 sk_num or sk_value it would produce an error because the num, data members
11241 are not present in STACK_OF. Now it just produces a warning. sk_set
11242 replaces the old method of assigning a value to sk_value
11243 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11244 that does this will no longer work (and should use sk_set instead) but
11245 this could be regarded as a "questionable" behaviour anyway.
11248 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11249 correctly handle encrypted S/MIME data.
11252 *) Change type of various DES function arguments from des_cblock
11253 (which means, in function argument declarations, pointer to char)
11254 to des_cblock * (meaning pointer to array with 8 char elements),
11255 which allows the compiler to do more typechecking; it was like
11256 that back in SSLeay, but with lots of ugly casts.
11258 Introduce new type const_des_cblock.
11261 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11262 problems: find RecipientInfo structure that matches recipient certificate
11263 and initialise the ASN1 structures properly based on passed cipher.
11266 *) Belatedly make the BN tests actually check the results.
11269 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11270 to and from BNs: it was completely broken. New compilation option
11271 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11272 key elements as negative integers.
11275 *) Reorganize and speed up MD5.
11276 [Andy Polyakov <appro@fy.chalmers.se>]
11279 [Richard Levitte <richard@levitte.org>]
11281 *) New option -out to asn1parse to allow the parsed structure to be
11282 output to a file. This is most useful when combined with the -strparse
11283 option to examine the output of things like OCTET STRINGS.
11286 *) Make SSL library a little more fool-proof by not requiring any longer
11287 that SSL_set_{accept,connect}_state be called before
11288 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11289 in many applications because usually everything *appeared* to work as
11290 intended anyway -- now it really works as intended).
11293 *) Move openssl.cnf out of lib/.
11296 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11297 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11298 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11299 [Ralf S. Engelschall]
11301 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11302 handle PKCS#7 enveloped data properly.
11303 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11305 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11306 copying pointers. The cert_st handling is changed by this in
11307 various ways (and thus what used to be known as ctx->default_cert
11308 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11309 any longer when s->cert does not give us what we need).
11310 ssl_cert_instantiate becomes obsolete by this change.
11311 As soon as we've got the new code right (possibly it already is?),
11312 we have solved a couple of bugs of the earlier code where s->cert
11313 was used as if it could not have been shared with other SSL structures.
11315 Note that using the SSL API in certain dirty ways now will result
11316 in different behaviour than observed with earlier library versions:
11317 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11318 does not influence s as it used to.
11320 In order to clean up things more thoroughly, inside SSL_SESSION
11321 we don't use CERT any longer, but a new structure SESS_CERT
11322 that holds per-session data (if available); currently, this is
11323 the peer's certificate chain and, for clients, the server's certificate
11324 and temporary key. CERT holds only those values that can have
11325 meaningful defaults in an SSL_CTX.
11328 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11329 from the internal representation. Various PKCS#7 fixes: remove some
11330 evil casts and set the enc_dig_alg field properly based on the signing
11334 *) Allow PKCS#12 password to be set from the command line or the
11335 environment. Let 'ca' get its config file name from the environment
11336 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11340 *) Allow certificate policies extension to use an IA5STRING for the
11341 organization field. This is contrary to the PKIX definition but
11342 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11346 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11347 without disallowing inline assembler and the like for non-pedantic builds.
11350 *) Support Borland C++ builder.
11351 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11353 *) Support Mingw32.
11356 *) SHA-1 cleanups and performance enhancements.
11357 [Andy Polyakov <appro@fy.chalmers.se>]
11359 *) Sparc v8plus assembler for the bignum library.
11360 [Andy Polyakov <appro@fy.chalmers.se>]
11362 *) Accept any -xxx and +xxx compiler options in Configure.
11365 *) Update HPUX configuration.
11368 *) Add missing sk_<type>_unshift() function to safestack.h
11369 [Ralf S. Engelschall]
11371 *) New function SSL_CTX_use_certificate_chain_file that sets the
11372 "extra_cert"s in addition to the certificate. (This makes sense
11373 only for "PEM" format files, as chains as a whole are not
11377 *) Support verify_depth from the SSL API.
11378 x509_vfy.c had what can be considered an off-by-one-error:
11379 Its depth (which was not part of the external interface)
11380 was actually counting the number of certificates in a chain;
11381 now it really counts the depth.
11384 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11385 instead of X509err, which often resulted in confusing error
11386 messages since the error codes are not globally unique
11387 (e.g. an alleged error in ssl3_accept when a certificate
11388 didn't match the private key).
11390 *) New function SSL_CTX_set_session_id_context that allows to set a default
11391 value (so that you don't need SSL_set_session_id_context for each
11392 connection using the SSL_CTX).
11395 *) OAEP decoding bug fix.
11398 *) Support INSTALL_PREFIX for package builders, as proposed by
11402 *) New Configure options "threads" and "no-threads". For systems
11403 where the proper compiler options are known (currently Solaris
11404 and Linux), "threads" is the default.
11407 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11410 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11411 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11412 such as /usr/local/bin.
11415 *) "make linux-shared" to build shared libraries.
11416 [Niels Poppe <niels@netbox.org>]
11418 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11421 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11422 extension adding in x509 utility.
11425 *) Remove NOPROTO sections and error code comments.
11428 *) Partial rewrite of the DEF file generator to now parse the ANSI
11432 *) New Configure options --prefix=DIR and --openssldir=DIR.
11435 *) Complete rewrite of the error code script(s). It is all now handled
11436 by one script at the top level which handles error code gathering,
11437 header rewriting and C source file generation. It should be much better
11438 than the old method: it now uses a modified version of Ulf's parser to
11439 read the ANSI prototypes in all header files (thus the old K&R definitions
11440 aren't needed for error creation any more) and do a better job of
11441 translating function codes into names. The old 'ASN1 error code imbedded
11442 in a comment' is no longer necessary and it doesn't use .err files which
11443 have now been deleted. Also the error code call doesn't have to appear all
11444 on one line (which resulted in some large lines...).
11447 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11450 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11451 0 (which usually indicates a closed connection), but continue reading.
11454 *) Fix some race conditions.
11457 *) Add support for CRL distribution points extension. Add Certificate
11458 Policies and CRL distribution points documentation.
11461 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11464 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11465 8 of keying material. Merlin has also confirmed interop with this fix
11466 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11467 [Merlin Hughes <merlin@baltimore.ie>]
11469 *) Fix lots of warnings.
11470 [Richard Levitte <levitte@stacken.kth.se>]
11472 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11473 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11474 [Richard Levitte <levitte@stacken.kth.se>]
11476 *) Fix problems with sizeof(long) == 8.
11477 [Andy Polyakov <appro@fy.chalmers.se>]
11479 *) Change functions to ANSI C.
11482 *) Fix typos in error codes.
11483 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11485 *) Remove defunct assembler files from Configure.
11488 *) SPARC v8 assembler BIGNUM implementation.
11489 [Andy Polyakov <appro@fy.chalmers.se>]
11491 *) Support for Certificate Policies extension: both print and set.
11492 Various additions to support the r2i method this uses.
11495 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11496 return a const string when you are expecting an allocated buffer.
11499 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11500 types DirectoryString and DisplayText.
11503 *) Add code to allow r2i extensions to access the configuration database,
11504 add an LHASH database driver and add several ctx helper functions.
11507 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11508 fail when they extended the size of a BIGNUM.
11511 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11512 support typesafe stack.
11515 *) Fix typo in SSL_[gs]et_options().
11516 [Nils Frostberg <nils@medcom.se>]
11518 *) Delete various functions and files that belonged to the (now obsolete)
11519 old X509V3 handling code.
11522 *) New Configure option "rsaref".
11525 *) Don't auto-generate pem.h.
11528 *) Introduce type-safe ASN.1 SETs.
11531 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11532 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11534 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11535 that links with OpenSSL (well at least cause lots of warnings), but fear
11536 not: the conversion is trivial, and it eliminates loads of evil casts. A
11537 few STACKed things have been converted already. Feel free to convert more.
11538 In the fullness of time, I'll do away with the STACK type altogether.
11541 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11542 specified in <certfile> by updating the entry in the index.txt file.
11543 This way one no longer has to edit the index.txt file manually for
11544 revoking a certificate. The -revoke option does the gory details now.
11545 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11547 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11548 `-text' option at all and this way the `-noout -text' combination was
11549 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11550 [Ralf S. Engelschall]
11552 *) Make sure a corresponding plain text error message exists for the
11553 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11554 verify callback function determined that a certificate was revoked.
11555 [Ralf S. Engelschall]
11557 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11558 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11559 all available cipers including rc5, which was forgotten until now.
11560 In order to let the testing shell script know which algorithms
11561 are available, a new (up to now undocumented) command
11562 "openssl list-cipher-commands" is used.
11565 *) Bugfix: s_client occasionally would sleep in select() when
11566 it should have checked SSL_pending() first.
11569 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11570 the raw DSA values prior to ASN.1 encoding.
11573 *) Tweaks to Configure
11574 [Niels Poppe <niels@netbox.org>]
11576 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11580 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11583 *) New config option to avoid instructions that are illegal on the 80386.
11584 The default code is faster, but requires at least a 486.
11587 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11588 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11589 same as SSL2_VERSION anyway.
11592 *) New "-showcerts" option for s_client.
11595 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11596 application. Various cleanups and fixes.
11599 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11600 modify error routines to work internally. Add error codes and PBE init
11601 to library startup routines.
11604 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11605 packing functions to asn1 and evp. Changed function names and error
11606 codes along the way.
11609 *) PKCS12 integration: and so it begins... First of several patches to
11610 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11611 objects to objects.h
11614 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11615 and display support for Thawte strong extranet extension.
11618 *) Add LinuxPPC support.
11619 [Jeff Dubrule <igor@pobox.org>]
11621 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11622 bn_div_words in alpha.s.
11623 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11625 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11626 OAEP isn't supported when OpenSSL is built with RSAref.
11627 [Ulf Moeller <ulf@fitug.de>]
11629 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11630 so they no longer are missing under -DNOPROTO.
11631 [Soren S. Jorvang <soren@t.dk>]
11634 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11636 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11637 doesn't work when the session is reused. Coming soon!
11640 *) Fix a security hole, that allows sessions to be reused in the wrong
11641 context thus bypassing client cert protection! All software that uses
11642 client certs and session caches in multiple contexts NEEDS PATCHING to
11643 allow session reuse! A fuller solution is in the works.
11644 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11646 *) Some more source tree cleanups (removed obsolete files
11647 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11648 permission on "config" script to be executable) and a fix for the INSTALL
11650 [Ulf Moeller <ulf@fitug.de>]
11652 *) Remove some legacy and erroneous uses of malloc, free instead of
11654 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11656 *) Make rsa_oaep_test return non-zero on error.
11657 [Ulf Moeller <ulf@fitug.de>]
11659 *) Add support for native Solaris shared libraries. Configure
11660 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11661 if someone would make that last step automatic.
11662 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11664 *) ctx_size was not built with the right compiler during "make links". Fixed.
11667 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11668 except NULL ciphers". This means the default cipher list will no longer
11669 enable NULL ciphers. They need to be specifically enabled e.g. with
11670 the string "DEFAULT:eNULL".
11673 *) Fix to RSA private encryption routines: if p < q then it would
11674 occasionally produce an invalid result. This will only happen with
11675 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11678 *) Be less restrictive and allow also `perl util/perlpath.pl
11679 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11680 because this way one can also use an interpreter named `perl5' (which is
11681 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11682 installed as `perl').
11683 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11685 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11686 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11688 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11689 advapi32.lib to Win32 build and change the pem test comparison
11690 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11691 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11692 and crypto/des/ede_cbcm_enc.c.
11695 *) DES quad checksum was broken on big-endian architectures. Fixed.
11698 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11699 Win32 test batch file so it (might) work again. The Win32 test batch file
11700 is horrible: I feel ill....
11703 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11704 in e_os.h. Audit of header files to check ANSI and non ANSI
11705 sections: 10 functions were absent from non ANSI section and not exported
11706 from Windows DLLs. Fixed up libeay.num for new functions.
11709 *) Make `openssl version' output lines consistent.
11710 [Ralf S. Engelschall]
11712 *) Fix Win32 symbol export lists for BIO functions: Added
11713 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11714 to ms/libeay{16,32}.def.
11715 [Ralf S. Engelschall]
11717 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11718 fine under Unix and passes some trivial tests I've now added. But the
11719 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11720 added to make sure no one expects that this stuff really works in the
11721 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11722 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11724 [Ralf S. Engelschall]
11726 *) Fix the generation of two part addresses in perl.
11727 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11729 *) Add config entry for Linux on MIPS.
11730 [John Tobey <jtobey@channel1.com>]
11732 *) Make links whenever Configure is run, unless we are on Windoze.
11735 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11736 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11740 *) Add a useful kludge to allow package maintainers to specify compiler and
11741 other platforms details on the command line without having to patch the
11742 Configure script everytime: One now can use ``perl Configure
11743 <id>:<details>'', i.e. platform ids are allowed to have details appended
11744 to them (separated by colons). This is treated as there would be a static
11745 pre-configured entry in Configure's %table under key <id> with value
11746 <details> and ``perl Configure <id>'' is called. So, when you want to
11747 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11748 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11749 now, which overrides the FreeBSD-elf entry on-the-fly.
11750 [Ralf S. Engelschall]
11752 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11755 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11756 on the `perl Configure ...' command line. This way one can compile
11757 OpenSSL libraries with Position Independent Code (PIC) which is needed
11758 for linking it into DSOs.
11759 [Ralf S. Engelschall]
11761 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11765 *) Cleaned up the LICENSE document: The official contact for any license
11766 questions now is the OpenSSL core team under openssl-core@openssl.org.
11767 And add a paragraph about the dual-license situation to make sure people
11768 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11769 to the OpenSSL toolkit.
11770 [Ralf S. Engelschall]
11772 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11773 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11774 Additionally cleaned up the `make links' target: Remove unnecessary
11775 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11776 to speed processing and no longer clutter the display with confusing
11777 stuff. Instead only the actually done links are displayed.
11778 [Ralf S. Engelschall]
11780 *) Permit null encryption ciphersuites, used for authentication only. It used
11781 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11782 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11786 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11787 signed attributes when verifying signatures (this would break them),
11788 the detached data encoding was wrong and public keys obtained using
11789 X509_get_pubkey() weren't freed.
11792 *) Add text documentation for the BUFFER functions. Also added a work around
11793 to a Win95 console bug. This was triggered by the password read stuff: the
11794 last character typed gets carried over to the next fread(). If you were
11795 generating a new cert request using 'req' for example then the last
11796 character of the passphrase would be CR which would then enter the first
11800 *) Added the new `Includes OpenSSL Cryptography Software' button as
11801 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11802 button and can be used by applications based on OpenSSL to show the
11803 relationship to the OpenSSL project.
11804 [Ralf S. Engelschall]
11806 *) Remove confusing variables in function signatures in files
11807 ssl/ssl_lib.c and ssl/ssl.h.
11808 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11810 *) Don't install bss_file.c under PREFIX/include/
11811 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11813 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11814 functions that return function pointers and has support for NT specific
11815 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11816 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11817 unsigned to signed types: this was killing the Win32 compile.
11820 *) Add new certificate file to stack functions,
11821 SSL_add_dir_cert_subjects_to_stack() and
11822 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11823 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11824 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11825 This means that Apache-SSL and similar packages don't have to mess around
11826 to add as many CAs as they want to the preferred list.
11829 *) Experiment with doxygen documentation. Currently only partially applied to
11831 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11832 openssl.doxy as the configuration file.
11835 *) Get rid of remaining C++-style comments which strict C compilers hate.
11836 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11838 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11839 compiled in by default: it has problems with large keys.
11842 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11843 DH private keys and/or callback functions which directly correspond to
11844 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11845 is needed for applications which have to configure certificates on a
11846 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11848 For the RSA certificate situation is makes no difference, but
11849 for the DSA certificate situation this fixes the "no shared cipher"
11850 problem where the OpenSSL cipher selection procedure failed because the
11851 temporary keys were not overtaken from the context and the API provided
11852 no way to reconfigure them.
11853 The new functions now let applications reconfigure the stuff and they
11854 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11855 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11856 non-public-API function ssl_cert_instantiate() is used as a helper
11857 function and also to reduce code redundancy inside ssl_rsa.c.
11858 [Ralf S. Engelschall]
11860 *) Move s_server -dcert and -dkey options out of the undocumented feature
11861 area because they are useful for the DSA situation and should be
11862 recognized by the users.
11863 [Ralf S. Engelschall]
11865 *) Fix the cipher decision scheme for export ciphers: the export bits are
11866 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11867 SSL_EXP_MASK. So, the original variable has to be used instead of the
11868 already masked variable.
11869 [Richard Levitte <levitte@stacken.kth.se>]
11871 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11872 [Richard Levitte <levitte@stacken.kth.se>]
11874 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11875 from `int' to `unsigned int' because it's a length and initialized by
11876 EVP_DigestFinal() which expects an `unsigned int *'.
11877 [Richard Levitte <levitte@stacken.kth.se>]
11879 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11880 script. Instead use the usual Shell->Perl transition trick.
11881 [Ralf S. Engelschall]
11883 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11884 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11885 -noout -modulus' as it's already the case for `openssl rsa -noout
11886 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11887 currently the public key is printed (a decision which was already done by
11888 `openssl dsa -modulus' in the past) which serves a similar purpose.
11889 Additionally the NO_RSA no longer completely removes the whole -modulus
11890 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11892 [Ralf S. Engelschall]
11894 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11895 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11896 [Arne Ansper <arne@ats.cyber.ee>]
11898 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11899 to be added. Now both 'req' and 'ca' can use new objects defined in the
11903 *) Add cool BIO that does syslog (or event log on NT).
11904 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11906 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11907 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11908 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11909 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11912 *) Add preliminary config info for new extension code.
11915 *) Make RSA_NO_PADDING really use no padding.
11916 [Ulf Moeller <ulf@fitug.de>]
11918 *) Generate errors when private/public key check is done.
11921 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11922 for some CRL extensions and new objects added.
11925 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11926 key usage extension and fuller support for authority key id.
11929 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11930 padding method for RSA, which is recommended for new applications in PKCS
11931 #1 v2.0 (RFC 2437, October 1998).
11932 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11933 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11934 against Bleichbacher's attack on RSA.
11935 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11938 *) Updates to the new SSL compression code
11939 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11941 *) Fix so that the version number in the master secret, when passed
11942 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11943 (because the server will not accept higher), that the version number
11944 is 0x03,0x01, not 0x03,0x00
11945 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11947 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11948 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11949 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11952 *) Support for RAW extensions where an arbitrary extension can be
11953 created by including its DER encoding. See apps/openssl.cnf for
11957 *) Make sure latest Perl versions don't interpret some generated C array
11958 code as Perl array code in the crypto/err/err_genc.pl script.
11959 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11961 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11962 not many people have the assembler. Various Win32 compilation fixes and
11963 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11964 build instructions.
11967 *) Modify configure script 'Configure' to automatically create crypto/date.h
11968 file under Win32 and also build pem.h from pem.org. New script
11969 util/mkfiles.pl to create the MINFO file on environments that can't do a
11970 'make files': perl util/mkfiles.pl >MINFO should work.
11973 *) Major rework of DES function declarations, in the pursuit of correctness
11974 and purity. As a result, many evil casts evaporated, and some weirdness,
11975 too. You may find this causes warnings in your code. Zapping your evil
11976 casts will probably fix them. Mostly.
11979 *) Fix for a typo in asn1.h. Bug fix to object creation script
11980 obj_dat.pl. It considered a zero in an object definition to mean
11981 "end of object": none of the objects in objects.h have any zeros
11982 so it wasn't spotted.
11983 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11985 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11986 Masking (CBCM). In the absence of test vectors, the best I have been able
11987 to do is check that the decrypt undoes the encrypt, so far. Send me test
11988 vectors if you have them.
11991 *) Correct calculation of key length for export ciphers (too much space was
11992 allocated for null ciphers). This has not been tested!
11995 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11996 message is now correct (it understands "crypto" and "ssl" on its
11997 command line). There is also now an "update" option. This will update
11998 the util/ssleay.num and util/libeay.num files with any new functions.
12000 perl util/mkdef.pl crypto ssl update
12001 it will update them.
12004 *) Overhauled the Perl interface (perl/*):
12005 - ported BN stuff to OpenSSL's different BN library
12006 - made the perl/ source tree CVS-aware
12007 - renamed the package from SSLeay to OpenSSL (the files still contain
12008 their history because I've copied them in the repository)
12009 - removed obsolete files (the test scripts will be replaced
12010 by better Test::Harness variants in the future)
12011 [Ralf S. Engelschall]
12013 *) First cut for a very conservative source tree cleanup:
12014 1. merge various obsolete readme texts into doc/ssleay.txt
12015 where we collect the old documents and readme texts.
12016 2. remove the first part of files where I'm already sure that we no
12017 longer need them because of three reasons: either they are just temporary
12018 files which were left by Eric or they are preserved original files where
12019 I've verified that the diff is also available in the CVS via "cvs diff
12020 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12021 the crypto/md/ stuff).
12022 [Ralf S. Engelschall]
12024 *) More extension code. Incomplete support for subject and issuer alt
12025 name, issuer and authority key id. Change the i2v function parameters
12026 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12027 what that's for :-) Fix to ASN1 macro which messed up
12028 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12031 *) Preliminary support for ENUMERATED type. This is largely copied from the
12035 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12036 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12038 *) Make sure `make rehash' target really finds the `openssl' program.
12039 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12041 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12042 like to hear about it if this slows down other processors.
12045 *) Add CygWin32 platform information to Configure script.
12046 [Alan Batie <batie@aahz.jf.intel.com>]
12048 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12049 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12051 *) New program nseq to manipulate netscape certificate sequences
12054 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12058 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12059 but the BN code had some problems that would cause failures when
12060 doing certificate verification and some other functions.
12061 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12063 *) Add ASN1 and PEM code to support netscape certificate sequences.
12066 *) Add ASN1 and PEM code to support netscape certificate sequences.
12069 *) Add several PKIX and private extended key usage OIDs.
12072 *) Modify the 'ca' program to handle the new extension code. Modify
12073 openssl.cnf for new extension format, add comments.
12076 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12077 and add a sample to openssl.cnf so req -x509 now adds appropriate
12081 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12082 error code, add initial support to X509_print() and x509 application.
12085 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12086 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12087 stuff is currently isolated and isn't even compiled yet.
12090 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12091 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12092 Removed the versions check from X509 routines when loading extensions:
12093 this allows certain broken certificates that don't set the version
12094 properly to be processed.
12097 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12098 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12099 can still be regenerated with "make depend".
12102 *) Spelling mistake in C version of CAST-128.
12103 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12105 *) Changes to the error generation code. The perl script err-code.pl
12106 now reads in the old error codes and retains the old numbers, only
12107 adding new ones if necessary. It also only changes the .err files if new
12108 codes are added. The makefiles have been modified to only insert errors
12109 when needed (to avoid needlessly modifying header files). This is done
12110 by only inserting errors if the .err file is newer than the auto generated
12111 C file. To rebuild all the error codes from scratch (the old behaviour)
12112 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12113 or delete all the .err files.
12116 *) CAST-128 was incorrectly implemented for short keys. The C version has
12117 been fixed, but is untested. The assembler versions are also fixed, but
12118 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12119 to regenerate it if needed.
12120 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12121 Hagino <itojun@kame.net>]
12123 *) File was opened incorrectly in randfile.c.
12124 [Ulf Möller <ulf@fitug.de>]
12126 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12127 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12128 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12129 al: it's just almost always a UTCTime. Note this patch adds new error
12130 codes so do a "make errors" if there are problems.
12133 *) Correct Linux 1 recognition in config.
12134 [Ulf Möller <ulf@fitug.de>]
12136 *) Remove pointless MD5 hash when using DSA keys in ca.
12137 [Anonymous <nobody@replay.com>]
12139 *) Generate an error if given an empty string as a cert directory. Also
12140 generate an error if handed NULL (previously returned 0 to indicate an
12141 error, but didn't set one).
12142 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12144 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12147 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12148 parameters. This was causing a warning which killed off the Win32 compile.
12151 *) Remove C++ style comments from crypto/bn/bn_local.h.
12152 [Neil Costigan <neil.costigan@celocom.com>]
12154 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12155 based on a text string, looking up short and long names and finally
12156 "dot" format. The "dot" format stuff didn't work. Added new function
12157 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12158 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12159 OID is not part of the table.
12162 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12163 X509_LOOKUP_by_alias().
12166 *) Sort openssl functions by name.
12169 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12170 encryption from sample DSA keys (in case anyone is interested the password
12174 *) Make _all_ *_free functions accept a NULL pointer.
12175 [Frans Heymans <fheymans@isaserver.be>]
12177 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12179 [Anonymous <nobody@replay.com>]
12181 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12182 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12184 *) Don't blow it for numeric -newkey arguments to apps/req.
12185 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12187 *) Temp key "for export" tests were wrong in s3_srvr.c.
12188 [Anonymous <nobody@replay.com>]
12190 *) Add prototype for temp key callback functions
12191 SSL_CTX_set_tmp_{rsa,dh}_callback().
12194 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12195 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12198 *) X509_name_add_entry() freed the wrong thing after an error.
12199 [Arne Ansper <arne@ats.cyber.ee>]
12201 *) rsa_eay.c would attempt to free a NULL context.
12202 [Arne Ansper <arne@ats.cyber.ee>]
12204 *) BIO_s_socket() had a broken should_retry() on Windoze.
12205 [Arne Ansper <arne@ats.cyber.ee>]
12207 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12208 [Arne Ansper <arne@ats.cyber.ee>]
12210 *) Make sure the already existing X509_STORE->depth variable is initialized
12211 in X509_STORE_new(), but document the fact that this variable is still
12212 unused in the certificate verification process.
12213 [Ralf S. Engelschall]
12215 *) Fix the various library and apps files to free up pkeys obtained from
12216 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12219 *) Fix reference counting in X509_PUBKEY_get(). This makes
12220 demos/maurice/example2.c work, amongst others, probably.
12221 [Steve Henson and Ben Laurie]
12223 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12224 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12225 are no longer created. This way we have a single and consistent command
12226 line interface `openssl <command>', similar to `cvs <command>'.
12227 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12229 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12230 BIT STRING wrapper always have zero unused bits.
12233 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12236 *) Make the top-level INSTALL documentation easier to understand.
12239 *) Makefiles updated to exit if an error occurs in a sub-directory
12240 make (including if user presses ^C) [Paul Sutton]
12242 *) Make Montgomery context stuff explicit in RSA data structure.
12245 *) Fix build order of pem and err to allow for generated pem.h.
12248 *) Fix renumbering bug in X509_NAME_delete_entry().
12251 *) Enhanced the err-ins.pl script so it makes the error library number
12252 global and can add a library name. This is needed for external ASN1 and
12253 other error libraries.
12256 *) Fixed sk_insert which never worked properly.
12259 *) Fix ASN1 macros so they can handle indefinite length constructed
12260 EXPLICIT tags. Some non standard certificates use these: they can now
12264 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12265 into a single doc/ssleay.txt bundle. This way the information is still
12266 preserved but no longer messes up this directory. Now it's new room for
12267 the new set of documentation files.
12268 [Ralf S. Engelschall]
12270 *) SETs were incorrectly DER encoded. This was a major pain, because they
12271 shared code with SEQUENCEs, which aren't coded the same. This means that
12272 almost everything to do with SETs or SEQUENCEs has either changed name or
12273 number of arguments.
12274 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12276 *) Fix test data to work with the above.
12279 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12280 was already fixed by Eric for 0.9.1 it seems.
12281 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12283 *) Autodetect FreeBSD3.
12286 *) Fix various bugs in Configure. This affects the following platforms:
12290 unixware-2.0-pentium
12294 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12295 before they are needed.
12298 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12302 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12304 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12305 changed SSLeay to OpenSSL in version strings.
12306 [Ralf S. Engelschall]
12308 *) Some fixups to the top-level documents.
12311 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12312 because the symlink to include/ was missing.
12313 [Ralf S. Engelschall]
12315 *) Incorporated the popular no-RSA/DSA-only patches
12316 which allow to compile a RSA-free SSLeay.
12317 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12319 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12320 when "ssleay" is still not found.
12321 [Ralf S. Engelschall]
12323 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12324 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12326 *) Updated the README file.
12327 [Ralf S. Engelschall]
12329 *) Added various .cvsignore files in the CVS repository subdirs
12330 to make a "cvs update" really silent.
12331 [Ralf S. Engelschall]
12333 *) Recompiled the error-definition header files and added
12334 missing symbols to the Win32 linker tables.
12335 [Ralf S. Engelschall]
12337 *) Cleaned up the top-level documents;
12338 o new files: CHANGES and LICENSE
12339 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12340 o merged COPYRIGHT into LICENSE
12341 o removed obsolete TODO file
12342 o renamed MICROSOFT to INSTALL.W32
12343 [Ralf S. Engelschall]
12345 *) Removed dummy files from the 0.9.1b source tree:
12346 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12347 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12348 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12349 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12350 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12351 [Ralf S. Engelschall]
12353 *) Added various platform portability fixes.
12356 *) The Genesis of the OpenSSL rpject:
12357 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12358 Young and Tim J. Hudson created while they were working for C2Net until
12360 [The OpenSSL Project]
12363 Changes between 0.9.0b and 0.9.1b [not released]
12365 *) Updated a few CA certificates under certs/
12368 *) Changed some BIGNUM api stuff.
12371 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12372 DGUX x86, Linux Alpha, etc.
12375 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12376 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12380 *) Add -strparse option to asn1pars program which parses nested
12382 [Dr Stephen Henson <shenson@bigfoot.com>]
12384 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12387 *) DSA fix for "ca" program.
12390 *) Added "-genkey" option to "dsaparam" program.
12393 *) Added RIPE MD160 (rmd160) message digest.
12396 *) Added -a (all) option to "ssleay version" command.
12399 *) Added PLATFORM define which is the id given to Configure.
12402 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12405 *) Extended the ASN.1 parser routines.
12408 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12411 *) Added a BN_CTX to the BN library.
12414 *) Fixed the weak key values in DES library
12417 *) Changed API in EVP library for cipher aliases.
12420 *) Added support for RC2/64bit cipher.
12423 *) Converted the lhash library to the crypto/mem.c functions.
12426 *) Added more recognized ASN.1 object ids.
12429 *) Added more RSA padding checks for SSL/TLS.
12432 *) Added BIO proxy/filter functionality.
12435 *) Added extra_certs to SSL_CTX which can be used
12436 send extra CA certificates to the client in the CA cert chain sending
12437 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12440 *) Now Fortezza is denied in the authentication phase because
12441 this is key exchange mechanism is not supported by SSLeay at all.
12444 *) Additional PKCS1 checks.
12447 *) Support the string "TLSv1" for all TLS v1 ciphers.
12450 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12451 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12454 *) Fixed a few memory leaks.
12457 *) Fixed various code and comment typos.
12460 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12461 bytes sent in the client random.
12462 [Edward Bishop <ebishop@spyglass.com>]