5 Changes between 1.1.0e and 1.1.0f [xx XXX xxxx]
9 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
10 platform rather than 'mingw'.
13 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
14 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
15 which is the minimum version we support.
18 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
20 *) Encrypt-Then-Mac renegotiation crash
22 During a renegotiation handshake if the Encrypt-Then-Mac extension is
23 negotiated where it was not in the original handshake (or vice-versa) then
24 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
25 and servers are affected.
27 This issue was reported to OpenSSL by Joe Orton (Red Hat).
31 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
33 *) Truncated packet could crash via OOB read
35 If one side of an SSL/TLS path is running on a 32-bit host and a specific
36 cipher is being used, then a truncated packet can cause that host to
37 perform an out-of-bounds read, usually resulting in a crash.
39 This issue was reported to OpenSSL by Robert Święcki of Google.
43 *) Bad (EC)DHE parameters cause a client crash
45 If a malicious server supplies bad parameters for a DHE or ECDHE key
46 exchange then this can result in the client attempting to dereference a
47 NULL pointer leading to a client crash. This could be exploited in a Denial
50 This issue was reported to OpenSSL by Guido Vranken.
54 *) BN_mod_exp may produce incorrect results on x86_64
56 There is a carry propagating bug in the x86_64 Montgomery squaring
57 procedure. No EC algorithms are affected. Analysis suggests that attacks
58 against RSA and DSA as a result of this defect would be very difficult to
59 perform and are not believed likely. Attacks against DH are considered just
60 feasible (although very difficult) because most of the work necessary to
61 deduce information about a private key may be performed offline. The amount
62 of resources required for such an attack would be very significant and
63 likely only accessible to a limited number of attackers. An attacker would
64 additionally need online access to an unpatched system using the target
65 private key in a scenario with persistent DH parameters and a private
66 key that is shared between multiple clients. For example this can occur by
67 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
68 similar to CVE-2015-3193 but must be treated as a separate problem.
70 This issue was reported to OpenSSL by the OSS-Fuzz project.
74 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
76 *) ChaCha20/Poly1305 heap-buffer-overflow
78 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
79 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
80 crash. This issue is not considered to be exploitable beyond a DoS.
82 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
86 *) CMS Null dereference
88 Applications parsing invalid CMS structures can crash with a NULL pointer
89 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
90 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
91 structure callback if an attempt is made to free certain invalid encodings.
92 Only CHOICE structures using a callback which do not handle NULL value are
95 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
99 *) Montgomery multiplication may produce incorrect results
101 There is a carry propagating bug in the Broadwell-specific Montgomery
102 multiplication procedure that handles input lengths divisible by, but
103 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
104 and DH private keys are impossible. This is because the subroutine in
105 question is not used in operations with the private key itself and an input
106 of the attacker's direct choice. Otherwise the bug can manifest itself as
107 transient authentication and key negotiation failures or reproducible
108 erroneous outcome of public-key operations with specially crafted input.
109 Among EC algorithms only Brainpool P-512 curves are affected and one
110 presumably can attack ECDH key negotiation. Impact was not analyzed in
111 detail, because pre-requisites for attack are considered unlikely. Namely
112 multiple clients have to choose the curve in question and the server has to
113 share the private key among them, neither of which is default behaviour.
114 Even then only clients that chose the curve will be affected.
116 This issue was publicly reported as transient failures and was not
117 initially recognized as a security issue. Thanks to Richard Morgan for
118 providing reproducible case.
122 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
123 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
124 prevent issues where no progress is being made and the peer continually
125 sends unrecognised record types, using up resources processing them.
128 *) Removed automatic addition of RPATH in shared libraries and executables,
129 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
132 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
134 *) Fix Use After Free for large message sizes
136 The patch applied to address CVE-2016-6307 resulted in an issue where if a
137 message larger than approx 16k is received then the underlying buffer to
138 store the incoming message is reallocated and moved. Unfortunately a
139 dangling pointer to the old location is left which results in an attempt to
140 write to the previously freed location. This is likely to result in a
141 crash, however it could potentially lead to execution of arbitrary code.
143 This issue only affects OpenSSL 1.1.0a.
145 This issue was reported to OpenSSL by Robert Święcki.
149 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
151 *) OCSP Status Request extension unbounded memory growth
153 A malicious client can send an excessively large OCSP Status Request
154 extension. If that client continually requests renegotiation, sending a
155 large OCSP Status Request extension each time, then there will be unbounded
156 memory growth on the server. This will eventually lead to a Denial Of
157 Service attack through memory exhaustion. Servers with a default
158 configuration are vulnerable even if they do not support OCSP. Builds using
159 the "no-ocsp" build time option are not affected.
161 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
165 *) SSL_peek() hang on empty record
167 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
168 sends an empty record. This could be exploited by a malicious peer in a
169 Denial Of Service attack.
171 This issue was reported to OpenSSL by Alex Gaynor.
175 *) Excessive allocation of memory in tls_get_message_header() and
176 dtls1_preprocess_fragment()
178 A (D)TLS message includes 3 bytes for its length in the header for the
179 message. This would allow for messages up to 16Mb in length. Messages of
180 this length are excessive and OpenSSL includes a check to ensure that a
181 peer is sending reasonably sized messages in order to avoid too much memory
182 being consumed to service a connection. A flaw in the logic of version
183 1.1.0 means that memory for the message is allocated too early, prior to
184 the excessive message length check. Due to way memory is allocated in
185 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
186 to service a connection. This could lead to a Denial of Service through
187 memory exhaustion. However, the excessive message length check still takes
188 place, and this would cause the connection to immediately fail. Assuming
189 that the application calls SSL_free() on the failed connection in a timely
190 manner then the 21Mb of allocated memory will then be immediately freed
191 again. Therefore the excessive memory allocation will be transitory in
192 nature. This then means that there is only a security impact if:
194 1) The application does not call SSL_free() in a timely manner in the event
195 that the connection fails
197 2) The application is working in a constrained environment where there is
198 very little free memory
200 3) The attacker initiates multiple connection attempts such that there are
201 multiple connections in a state where memory has been allocated for the
202 connection; SSL_free() has not yet been called; and there is insufficient
203 memory to service the multiple requests.
205 Except in the instance of (1) above any Denial Of Service is likely to be
206 transitory because as soon as the connection fails the memory is
207 subsequently freed again in the SSL_free() call. However there is an
208 increased risk during this period of application crashes due to the lack of
209 memory - which would then mean a more serious Denial of Service.
211 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
212 (CVE-2016-6307 and CVE-2016-6308)
215 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
216 had to be removed. Primary reason is that vendor assembler can't
217 assemble our modules with -KPIC flag. As result it, assembly
218 support, was not even available as option. But its lack means
219 lack of side-channel resistant code, which is incompatible with
220 security by todays standards. Fortunately gcc is readily available
221 prepackaged option, which we firmly point at...
224 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
226 *) Windows command-line tool supports UTF-8 opt-in option for arguments
227 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
228 (to any value) allows Windows user to access PKCS#12 file generated
229 with Windows CryptoAPI and protected with non-ASCII password, as well
230 as files generated under UTF-8 locale on Linux also protected with
234 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
235 have been disabled by default and removed from DEFAULT, just like RC4.
236 See the RC4 item below to re-enable both.
239 *) The method for finding the storage location for the Windows RAND seed file
240 has changed. First we check %RANDFILE%. If that is not set then we check
241 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
242 all else fails we fall back to C:\.
245 *) The EVP_EncryptUpdate() function has had its return type changed from void
246 to int. A return of 0 indicates and error while a return of 1 indicates
250 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
251 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
252 off the constant time implementation for RSA, DSA and DH have been made
253 no-ops and deprecated.
256 *) Windows RAND implementation was simplified to only get entropy by
257 calling CryptGenRandom(). Various other RAND-related tickets
259 [Joseph Wylie Yandle, Rich Salz]
261 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
262 and OPENSSL_LH_, respectively. The old names are available
263 with API compatibility. They new names are now completely documented.
266 *) Unify TYPE_up_ref(obj) methods signature.
267 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
268 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
269 int (instead of void) like all others TYPE_up_ref() methods.
270 So now these methods also check the return value of CRYPTO_atomic_add(),
271 and the validity of object reference counter.
272 [fdasilvayy@gmail.com]
274 *) With Windows Visual Studio builds, the .pdb files are installed
275 alongside the installed libraries and executables. For a static
276 library installation, ossl_static.pdb is the associate compiler
277 generated .pdb file to be used when linking programs.
280 *) Remove openssl.spec. Packaging files belong with the packagers.
283 *) Automatic Darwin/OSX configuration has had a refresh, it will now
284 recognise x86_64 architectures automatically. You can still decide
285 to build for a different bitness with the environment variable
286 KERNEL_BITS (can be 32 or 64), for example:
288 KERNEL_BITS=32 ./config
292 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
293 256 bit AES and HMAC with SHA256.
296 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
299 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
302 *) To enable users to have their own config files and build file templates,
303 Configure looks in the directory indicated by the environment variable
304 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
305 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
306 name and is used as is.
309 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
310 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
311 X509_CERT_FILE_CTX was removed.
314 *) "shared" builds are now the default. To create only static libraries use
315 the "no-shared" Configure option.
318 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
319 All of these option have not worked for some while and are fundamental
323 *) Make various cleanup routines no-ops and mark them as deprecated. Most
324 global cleanup functions are no longer required because they are handled
325 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
326 Explicitly de-initing can cause problems (e.g. where a library that uses
327 OpenSSL de-inits, but an application is still using it). The affected
328 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
329 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
330 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
334 *) --strict-warnings no longer enables runtime debugging options
335 such as REF_DEBUG. Instead, debug options are automatically
336 enabled with '--debug' builds.
337 [Andy Polyakov, Emilia Käsper]
339 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
340 have been moved out of the public header files. New functions for managing
341 these have been added.
344 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
345 objects have been moved out of the public header files. New
346 functions for managing these have been added.
349 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
350 have been moved out of the public header files. New functions for managing
351 these have been added.
354 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
355 moved out of the public header files. New functions for managing these
359 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
362 *) Removed the mk1mf build scripts.
365 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
366 it is always safe to #include a header now.
369 *) Removed the aged BC-32 config and all its supporting scripts
372 *) Removed support for Ultrix, Netware, and OS/2.
375 *) Add support for HKDF.
378 *) Add support for blake2b and blake2s
381 *) Added support for "pipelining". Ciphers that have the
382 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
383 encryptions/decryptions simultaneously. There are currently no built-in
384 ciphers with this property but the expectation is that engines will be able
385 to offer it to significantly improve throughput. Support has been extended
386 into libssl so that multiple records for a single connection can be
387 processed in one go (for >=TLS 1.1).
390 *) Added the AFALG engine. This is an async capable engine which is able to
391 offload work to the Linux kernel. In this initial version it only supports
392 AES128-CBC. The kernel must be version 4.1.0 or greater.
395 *) OpenSSL now uses a new threading API. It is no longer necessary to
396 set locking callbacks to use OpenSSL in a multi-threaded environment. There
397 are two supported threading models: pthreads and windows threads. It is
398 also possible to configure OpenSSL at compile time for "no-threads". The
399 old threading API should no longer be used. The functions have been
400 replaced with "no-op" compatibility macros.
401 [Alessandro Ghedini, Matt Caswell]
403 *) Modify behavior of ALPN to invoke callback after SNI/servername
404 callback, such that updates to the SSL_CTX affect ALPN.
407 *) Add SSL_CIPHER queries for authentication and key-exchange.
410 *) Changes to the DEFAULT cipherlist:
411 - Prefer (EC)DHE handshakes over plain RSA.
412 - Prefer AEAD ciphers over legacy ciphers.
413 - Prefer ECDSA over RSA when both certificates are available.
414 - Prefer TLSv1.2 ciphers/PRF.
415 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
419 *) Change the ECC default curve list to be this, in order: x25519,
420 secp256r1, secp521r1, secp384r1.
423 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
424 disabled by default. They can be re-enabled using the
425 enable-weak-ssl-ciphers option to Configure.
428 *) If the server has ALPN configured, but supports no protocols that the
429 client advertises, send a fatal "no_application_protocol" alert.
430 This behaviour is SHALL in RFC 7301, though it isn't universally
431 implemented by other servers.
434 *) Add X25519 support.
435 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
436 for public and private key encoding using the format documented in
437 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
438 key generation and key derivation.
440 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
444 *) Deprecate SRP_VBASE_get_by_user.
445 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
446 In order to fix an unavoidable memory leak (CVE-2016-0798),
447 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
448 seed, even if the seed is configured.
450 Users should use SRP_VBASE_get1_by_user instead. Note that in
451 SRP_VBASE_get1_by_user, caller must free the returned value. Note
452 also that even though configuring the SRP seed attempts to hide
453 invalid usernames by continuing the handshake with fake
454 credentials, this behaviour is not constant time and no strong
455 guarantees are made that the handshake is indistinguishable from
456 that of a valid user.
459 *) Configuration change; it's now possible to build dynamic engines
460 without having to build shared libraries and vice versa. This
461 only applies to the engines in engines/, those in crypto/engine/
462 will always be built into libcrypto (i.e. "static").
464 Building dynamic engines is enabled by default; to disable, use
465 the configuration option "disable-dynamic-engine".
467 The only requirements for building dynamic engines are the
468 presence of the DSO module and building with position independent
469 code, so they will also automatically be disabled if configuring
470 with "disable-dso" or "disable-pic".
472 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
473 are also taken away from openssl/opensslconf.h, as they are
477 *) Configuration change; if there is a known flag to compile
478 position independent code, it will always be applied on the
479 libcrypto and libssl object files, and never on the application
480 object files. This means other libraries that use routines from
481 libcrypto / libssl can be made into shared libraries regardless
482 of how OpenSSL was configured.
484 If this isn't desirable, the configuration options "disable-pic"
485 or "no-pic" can be used to disable the use of PIC. This will
486 also disable building shared libraries and dynamic engines.
489 *) Removed JPAKE code. It was experimental and has no wide use.
492 *) The INSTALL_PREFIX Makefile variable has been renamed to
493 DESTDIR. That makes for less confusion on what this variable
494 is for. Also, the configuration option --install_prefix is
498 *) Heartbeat for TLS has been removed and is disabled by default
499 for DTLS; configure with enable-heartbeats. Code that uses the
500 old #define's might need to be updated.
501 [Emilia Käsper, Rich Salz]
503 *) Rename REF_CHECK to REF_DEBUG.
506 *) New "unified" build system
508 The "unified" build system is aimed to be a common system for all
509 platforms we support. With it comes new support for VMS.
511 This system builds supports building in a different directory tree
512 than the source tree. It produces one Makefile (for unix family
513 or lookalikes), or one descrip.mms (for VMS).
515 The source of information to make the Makefile / descrip.mms is
516 small files called 'build.info', holding the necessary
517 information for each directory with source to compile, and a
518 template in Configurations, like unix-Makefile.tmpl or
521 With this change, the library names were also renamed on Windows
522 and on VMS. They now have names that are closer to the standard
523 on Unix, and include the major version number, and in certain
524 cases, the architecture they are built for. See "Notes on shared
525 libraries" in INSTALL.
527 We rely heavily on the perl module Text::Template.
530 *) Added support for auto-initialisation and de-initialisation of the library.
531 OpenSSL no longer requires explicit init or deinit routines to be called,
532 except in certain circumstances. See the OPENSSL_init_crypto() and
533 OPENSSL_init_ssl() man pages for further information.
536 *) The arguments to the DTLSv1_listen function have changed. Specifically the
537 "peer" argument is now expected to be a BIO_ADDR object.
539 *) Rewrite of BIO networking library. The BIO library lacked consistent
540 support of IPv6, and adding it required some more extensive
541 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
542 which hold all types of addresses and chains of address information.
543 It also introduces a new API, with functions like BIO_socket,
544 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
545 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
546 have been adapted accordingly.
549 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
553 *) CRIME protection: disable compression by default, even if OpenSSL is
554 compiled with zlib enabled. Applications can still enable compression
555 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
556 using the SSL_CONF library to configure compression.
559 *) The signature of the session callback configured with
560 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
561 was explicitly marked as 'const unsigned char*' instead of
565 *) Always DPURIFY. Remove the use of uninitialized memory in the
566 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
569 *) Removed many obsolete configuration items, including
570 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
571 MD2_CHAR, MD2_INT, MD2_LONG
573 IDEA_SHORT, IDEA_LONG
574 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
575 [Rich Salz, with advice from Andy Polyakov]
577 *) Many BN internals have been moved to an internal header file.
578 [Rich Salz with help from Andy Polyakov]
580 *) Configuration and writing out the results from it has changed.
581 Files such as Makefile include/openssl/opensslconf.h and are now
582 produced through general templates, such as Makefile.in and
583 crypto/opensslconf.h.in and some help from the perl module
586 Also, the center of configuration information is no longer
587 Makefile. Instead, Configure produces a perl module in
588 configdata.pm which holds most of the config data (in the hash
589 table %config), the target data that comes from the target
590 configuration in one of the Configurations/*.conf files (in
594 *) To clarify their intended purposes, the Configure options
595 --prefix and --openssldir change their semantics, and become more
596 straightforward and less interdependent.
598 --prefix shall be used exclusively to give the location INSTALLTOP
599 where programs, scripts, libraries, include files and manuals are
600 going to be installed. The default is now /usr/local.
602 --openssldir shall be used exclusively to give the default
603 location OPENSSLDIR where certificates, private keys, CRLs are
604 managed. This is also where the default openssl.cnf gets
606 If the directory given with this option is a relative path, the
607 values of both the --prefix value and the --openssldir value will
608 be combined to become OPENSSLDIR.
609 The default for --openssldir is INSTALLTOP/ssl.
611 Anyone who uses --openssldir to specify where OpenSSL is to be
612 installed MUST change to use --prefix instead.
615 *) The GOST engine was out of date and therefore it has been removed. An up
616 to date GOST engine is now being maintained in an external repository.
617 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
618 support for GOST ciphersuites (these are only activated if a GOST engine
622 *) EGD is no longer supported by default; use enable-egd when
624 [Ben Kaduk and Rich Salz]
626 *) The distribution now has Makefile.in files, which are used to
627 create Makefile's when Configure is run. *Configure must be run
628 before trying to build now.*
631 *) The return value for SSL_CIPHER_description() for error conditions
635 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
637 Obtaining and performing DNSSEC validation of TLSA records is
638 the application's responsibility. The application provides
639 the TLSA records of its choice to OpenSSL, and these are then
640 used to authenticate the peer.
642 The TLSA records need not even come from DNS. They can, for
643 example, be used to implement local end-entity certificate or
644 trust-anchor "pinning", where the "pin" data takes the form
645 of TLSA records, which can augment or replace verification
646 based on the usual WebPKI public certification authorities.
649 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
650 continues to support deprecated interfaces in default builds.
651 However, applications are strongly advised to compile their
652 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
653 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
654 or the 1.1.0 releases.
656 In environments in which all applications have been ported to
657 not use any deprecated interfaces OpenSSL's Configure script
658 should be used with the --api=1.1.0 option to entirely remove
659 support for the deprecated features from the library and
660 unconditionally disable them in the installed headers.
661 Essentially the same effect can be achieved with the "no-deprecated"
662 argument to Configure, except that this will always restrict
663 the build to just the latest API, rather than a fixed API
666 As applications are ported to future revisions of the API,
667 they should update their compile-time OPENSSL_API_COMPAT define
668 accordingly, but in most cases should be able to continue to
669 compile with later releases.
671 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
672 0x10000000L and 0x00908000L, respectively. However those
673 versions did not support the OPENSSL_API_COMPAT feature, and
674 so applications are not typically tested for explicit support
675 of just the undeprecated features of either release.
678 *) Add support for setting the minimum and maximum supported protocol.
679 It can bet set via the SSL_set_min_proto_version() and
680 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
681 MaxProtcol. It's recommended to use the new APIs to disable
682 protocols instead of disabling individual protocols using
683 SSL_set_options() or SSL_CONF's Protocol. This change also
684 removes support for disabling TLS 1.2 in the OpenSSL TLS
685 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
688 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
691 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
692 and integrates ECDSA and ECDH functionality into EC. Implementations can
693 now redirect key generation and no longer need to convert to or from
696 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
697 include the ec.h header file instead.
700 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
701 ciphers who are no longer supported and drops support the ephemeral RSA key
702 exchange. The LOW ciphers currently doesn't have any ciphers in it.
705 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
706 opaque. For HMAC_CTX, the following constructors and destructors
709 HMAC_CTX *HMAC_CTX_new(void);
710 void HMAC_CTX_free(HMAC_CTX *ctx);
712 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
713 destroy such methods has been added. See EVP_MD_meth_new(3) and
714 EVP_CIPHER_meth_new(3) for documentation.
717 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
718 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
719 EVP_MD_CTX_reset() should be called instead to reinitialise
720 an already created structure.
721 2) For consistency with the majority of our object creators and
722 destructors, EVP_MD_CTX_(create|destroy) were renamed to
723 EVP_MD_CTX_(new|free). The old names are retained as macros
724 for deprecated builds.
727 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
728 cryptographic operations to be performed asynchronously as long as an
729 asynchronous capable engine is used. See the ASYNC_start_job() man page for
730 further details. Libssl has also had this capability integrated with the
731 introduction of the new mode SSL_MODE_ASYNC and associated error
732 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
733 pages. This work was developed in partnership with Intel Corp.
736 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
737 always enabled now. If you want to disable the support you should
738 exclude it using the list of supported ciphers. This also means that the
739 "-no_ecdhe" option has been removed from s_server.
742 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
743 SSL_{CTX_}set1_curves() which can set a list.
746 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
747 curve you want to support using SSL_{CTX_}set1_curves().
750 *) State machine rewrite. The state machine code has been significantly
751 refactored in order to remove much duplication of code and solve issues
752 with the old code (see ssl/statem/README for further details). This change
753 does have some associated API changes. Notably the SSL_state() function
754 has been removed and replaced by SSL_get_state which now returns an
755 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
756 altogether. The previous handshake states defined in ssl.h and ssl3.h have
760 *) All instances of the string "ssleay" in the public API were replaced
761 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
762 Some error codes related to internal RSA_eay API's were renamed.
765 *) The demo files in crypto/threads were moved to demo/threads.
768 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
770 [Matt Caswell, Rich Salz]
772 *) New ASN.1 embed macro.
774 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
775 structure is not allocated: it is part of the parent. That is instead of
783 This reduces memory fragmentation and make it impossible to accidentally
784 set a mandatory field to NULL.
786 This currently only works for some fields specifically a SEQUENCE, CHOICE,
787 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
788 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
792 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
795 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
796 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
797 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
798 DES and RC4 ciphersuites.
801 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
802 This changes the decoding behaviour for some invalid messages,
803 though the change is mostly in the more lenient direction, and
804 legacy behaviour is preserved as much as possible.
807 *) Fix no-stdio build.
808 [ David Woodhouse <David.Woodhouse@intel.com> and also
809 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
811 *) New testing framework
812 The testing framework has been largely rewritten and is now using
813 perl and the perl modules Test::Harness and an extended variant of
814 Test::More called OpenSSL::Test to do its work. All test scripts in
815 test/ have been rewritten into test recipes, and all direct calls to
816 executables in test/Makefile have become individual recipes using the
817 simplified testing OpenSSL::Test::Simple.
819 For documentation on our testing modules, do:
821 perldoc test/testlib/OpenSSL/Test/Simple.pm
822 perldoc test/testlib/OpenSSL/Test.pm
826 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
827 are used; the latter aborts on memory leaks (usually checked on exit).
828 Some undocumented "set malloc, etc., hooks" functions were removed
829 and others were changed. All are now documented.
832 *) In DSA_generate_parameters_ex, if the provided seed is too short,
834 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
836 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
837 from RFC4279, RFC4785, RFC5487, RFC5489.
839 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
840 original RSA_PSK patch.
843 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
844 era flag was never set throughout the codebase (only read). Also removed
845 SSL3_FLAGS_POP_BUFFER which was only used if
846 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
849 *) Changed the default name options in the "ca", "crl", "req" and "x509"
850 to be "oneline" instead of "compat".
853 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
854 not aware of clients that still exhibit this bug, and the workaround
855 hasn't been working properly for a while.
858 *) The return type of BIO_number_read() and BIO_number_written() as well as
859 the corresponding num_read and num_write members in the BIO structure has
860 changed from unsigned long to uint64_t. On platforms where an unsigned
861 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
865 *) Given the pervasive nature of TLS extensions it is inadvisable to run
866 OpenSSL without support for them. It also means that maintaining
867 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
868 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
871 *) Removed support for the two export grade static DH ciphersuites
872 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
873 were newly added (along with a number of other static DH ciphersuites) to
874 1.0.2. However the two export ones have *never* worked since they were
875 introduced. It seems strange in any case to be adding new export
876 ciphersuites, and given "logjam" it also does not seem correct to fix them.
879 *) Version negotiation has been rewritten. In particular SSLv23_method(),
880 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
881 and turned into macros which simply call the new preferred function names
882 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
883 should use the new names instead. Also as part of this change the ssl23.h
884 header file has been removed.
887 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
888 code and the associated standard is no longer considered fit-for-purpose.
891 *) RT2547 was closed. When generating a private key, try to make the
892 output file readable only by the owner. This behavior change might
893 be noticeable when interacting with other software.
895 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
899 *) Added HTTP GET support to the ocsp command.
902 *) Changed default digest for the dgst and enc commands from MD5 to
906 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
909 *) Added support for TLS extended master secret from
910 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
911 initial patch which was a great help during development.
914 *) All libssl internal structures have been removed from the public header
915 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
916 now redundant). Users should not attempt to access internal structures
917 directly. Instead they should use the provided API functions.
920 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
921 Access to deprecated functions can be re-enabled by running config with
922 "enable-deprecated". In addition applications wishing to use deprecated
923 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
924 will, by default, disable some transitive includes that previously existed
925 in the header files (e.g. ec.h will no longer, by default, include bn.h)
928 *) Added support for OCB mode. OpenSSL has been granted a patent license
929 compatible with the OpenSSL license for use of OCB. Details are available
930 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
931 for OCB can be removed by calling config with no-ocb.
934 *) SSLv2 support has been removed. It still supports receiving a SSLv2
935 compatible client hello.
938 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
939 done while fixing the error code for the key-too-small case.
940 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
942 *) CA.sh has been removed; use CA.pl instead.
945 *) Removed old DES API.
948 *) Remove various unsupported platforms:
954 Sinix/ReliantUNIX RM400
959 16-bit platforms such as WIN16
962 *) Clean up OPENSSL_NO_xxx #define's
963 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
964 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
965 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
966 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
967 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
968 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
969 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
970 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
971 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
972 Remove MS_STATIC; it's a relic from platforms <32 bits.
975 *) Cleaned up dead code
976 Remove all but one '#ifdef undef' which is to be looked at.
979 *) Clean up calling of xxx_free routines.
980 Just like free(), fix most of the xxx_free routines to accept
981 NULL. Remove the non-null checks from callers. Save much code.
984 *) Add secure heap for storage of private keys (when possible).
985 Add BIO_s_secmem(), CBIGNUM, etc.
986 Contributed by Akamai Technologies under our Corporate CLA.
989 *) Experimental support for a new, fast, unbiased prime candidate generator,
990 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
991 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
993 *) New output format NSS in the sess_id command line tool. This allows
994 exporting the session id and the master key in NSS keylog format.
995 [Martin Kaiser <martin@kaiser.cx>]
997 *) Harmonize version and its documentation. -f flag is used to display
999 [mancha <mancha1@zoho.com>]
1001 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1002 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1003 [mancha <mancha1@zoho.com>]
1005 *) Fix some double frees. These are not thought to be exploitable.
1006 [mancha <mancha1@zoho.com>]
1008 *) A missing bounds check in the handling of the TLS heartbeat extension
1009 can be used to reveal up to 64k of memory to a connected client or
1012 Thanks for Neel Mehta of Google Security for discovering this bug and to
1013 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1014 preparing the fix (CVE-2014-0160)
1015 [Adam Langley, Bodo Moeller]
1017 *) Fix for the attack described in the paper "Recovering OpenSSL
1018 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1019 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1020 http://eprint.iacr.org/2014/140
1022 Thanks to Yuval Yarom and Naomi Benger for discovering this
1023 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1024 [Yuval Yarom and Naomi Benger]
1026 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1027 this fixes a limitation in previous versions of OpenSSL.
1030 *) Experimental encrypt-then-mac support.
1032 Experimental support for encrypt then mac from
1033 draft-gutmann-tls-encrypt-then-mac-02.txt
1035 To enable it set the appropriate extension number (0x42 for the test
1036 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1038 For non-compliant peers (i.e. just about everything) this should have no
1041 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1045 *) Add EVP support for key wrapping algorithms, to avoid problems with
1046 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1047 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1048 algorithms and include tests cases.
1051 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1055 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1056 MGF1 digest and OAEP label.
1059 *) Make openssl verify return errors.
1060 [Chris Palmer <palmer@google.com> and Ben Laurie]
1062 *) New function ASN1_TIME_diff to calculate the difference between two
1063 ASN1_TIME structures or one structure and the current time.
1066 *) Update fips_test_suite to support multiple command line options. New
1067 test to induce all self test errors in sequence and check expected
1071 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1072 sign or verify all in one operation.
1075 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1076 test programs and fips_test_suite. Includes functionality to parse
1077 the minimal script output of fipsalgest.pl directly.
1080 *) Add authorisation parameter to FIPS_module_mode_set().
1083 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1086 *) Use separate DRBG fields for internal and external flags. New function
1087 FIPS_drbg_health_check() to perform on demand health checking. Add
1088 generation tests to fips_test_suite with reduced health check interval to
1089 demonstrate periodic health checking. Add "nodh" option to
1090 fips_test_suite to skip very slow DH test.
1093 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1097 *) More extensive health check for DRBG checking many more failure modes.
1098 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1099 combination: call this in fips_test_suite.
1102 *) Add support for canonical generation of DSA parameter 'g'. See
1105 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1106 POST to handle HMAC cases.
1109 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1110 to return numerical and string versions of the FIPS module number.
1113 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1114 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1115 outside the validated module in the FIPS capable OpenSSL.
1118 *) Minor change to DRBG entropy callback semantics. In some cases
1119 there is no multiple of the block length between min_len and
1120 max_len. Allow the callback to return more than max_len bytes
1121 of entropy but discard any extra: it is the callback's responsibility
1122 to ensure that the extra data discarded does not impact the
1123 requested amount of entropy.
1126 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1127 information in FIPS186-3, SP800-57 and SP800-131A.
1130 *) CCM support via EVP. Interface is very similar to GCM case except we
1131 must supply all data in one chunk (i.e. no update, final) and the
1132 message length must be supplied if AAD is used. Add algorithm test
1136 *) Initial version of POST overhaul. Add POST callback to allow the status
1137 of POST to be monitored and/or failures induced. Modify fips_test_suite
1138 to use callback. Always run all selftests even if one fails.
1141 *) XTS support including algorithm test driver in the fips_gcmtest program.
1142 Note: this does increase the maximum key length from 32 to 64 bytes but
1143 there should be no binary compatibility issues as existing applications
1144 will never use XTS mode.
1147 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1148 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1149 performs algorithm blocking for unapproved PRNG types. Also do not
1150 set PRNG type in FIPS_mode_set(): leave this to the application.
1151 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1152 the standard OpenSSL PRNG: set additional data to a date time vector.
1155 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1156 This shouldn't present any incompatibility problems because applications
1157 shouldn't be using these directly and any that are will need to rethink
1158 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1161 *) Extensive self tests and health checking required by SP800-90 DRBG.
1162 Remove strength parameter from FIPS_drbg_instantiate and always
1163 instantiate at maximum supported strength.
1166 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1169 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1172 *) New function DH_compute_key_padded() to compute a DH key and pad with
1173 leading zeroes if needed: this complies with SP800-56A et al.
1176 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1177 anything, incomplete, subject to change and largely untested at present.
1180 *) Modify fipscanisteronly build option to only build the necessary object
1181 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1184 *) Add experimental option FIPSSYMS to give all symbols in
1185 fipscanister.o and FIPS or fips prefix. This will avoid
1186 conflicts with future versions of OpenSSL. Add perl script
1187 util/fipsas.pl to preprocess assembly language source files
1188 and rename any affected symbols.
1191 *) Add selftest checks and algorithm block of non-fips algorithms in
1192 FIPS mode. Remove DES2 from selftests.
1195 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1196 return internal method without any ENGINE dependencies. Add new
1197 tiny fips sign and verify functions.
1200 *) New build option no-ec2m to disable characteristic 2 code.
1203 *) New build option "fipscanisteronly". This only builds fipscanister.o
1204 and (currently) associated fips utilities. Uses the file Makefile.fips
1205 instead of Makefile.org as the prototype.
1208 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1209 Update fips_gcmtest to use IV generator.
1212 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1213 setting output buffer to NULL. The *Final function must be
1214 called although it will not retrieve any additional data. The tag
1215 can be set or retrieved with a ctrl. The IV length is by default 12
1216 bytes (96 bits) but can be set to an alternative value. If the IV
1217 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1221 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1222 underlying do_cipher function handles all cipher semantics itself
1223 including padding and finalisation. This is useful if (for example)
1224 an ENGINE cipher handles block padding itself. The behaviour of
1225 do_cipher is subtly changed if this flag is set: the return value
1226 is the number of characters written to the output buffer (zero is
1227 no longer an error code) or a negative error code. Also if the
1228 input buffer is NULL and length 0 finalisation should be performed.
1231 *) If a candidate issuer certificate is already part of the constructed
1232 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1235 *) Improve forward-security support: add functions
1237 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1238 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1240 for use by SSL/TLS servers; the callback function will be called whenever a
1241 new session is created, and gets to decide whether the session may be
1242 cached to make it resumable (return 0) or not (return 1). (As by the
1243 SSL/TLS protocol specifications, the session_id sent by the server will be
1244 empty to indicate that the session is not resumable; also, the server will
1245 not generate RFC 4507 (RFC 5077) session tickets.)
1247 A simple reasonable callback implementation is to return is_forward_secure.
1248 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1249 by the SSL/TLS server library, indicating whether it can provide forward
1251 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1253 *) New -verify_name option in command line utilities to set verification
1257 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1258 Add CMAC pkey methods.
1261 *) Experimental renegotiation in s_server -www mode. If the client
1262 browses /reneg connection is renegotiated. If /renegcert it is
1263 renegotiated requesting a certificate.
1266 *) Add an "external" session cache for debugging purposes to s_server. This
1267 should help trace issues which normally are only apparent in deployed
1268 multi-process servers.
1271 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1272 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1273 BIO_set_cipher() and some obscure PEM functions were changed so they
1274 can now return an error. The RAND changes required a change to the
1275 RAND_METHOD structure.
1278 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1279 a gcc attribute to warn if the result of a function is ignored. This
1280 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1281 whose return value is often ignored.
1284 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1285 These allow SCTs (signed certificate timestamps) to be requested and
1286 validated when establishing a connection.
1287 [Rob Percival <robpercival@google.com>]
1289 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1291 *) Prevent padding oracle in AES-NI CBC MAC check
1293 A MITM attacker can use a padding oracle attack to decrypt traffic
1294 when the connection uses an AES CBC cipher and the server support
1297 This issue was introduced as part of the fix for Lucky 13 padding
1298 attack (CVE-2013-0169). The padding check was rewritten to be in
1299 constant time by making sure that always the same bytes are read and
1300 compared against either the MAC or padding bytes. But it no longer
1301 checked that there was enough data to have both the MAC and padding
1304 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1308 *) Fix EVP_EncodeUpdate overflow
1310 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1311 Base64 encoding of binary data. If an attacker is able to supply very large
1312 amounts of input data then a length check can overflow resulting in a heap
1315 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1316 the PEM_write_bio* family of functions. These are mainly used within the
1317 OpenSSL command line applications, so any application which processes data
1318 from an untrusted source and outputs it as a PEM file should be considered
1319 vulnerable to this issue. User applications that call these APIs directly
1320 with large amounts of untrusted data may also be vulnerable.
1322 This issue was reported by Guido Vranken.
1326 *) Fix EVP_EncryptUpdate overflow
1328 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1329 is able to supply very large amounts of input data after a previous call to
1330 EVP_EncryptUpdate() with a partial block then a length check can overflow
1331 resulting in a heap corruption. Following an analysis of all OpenSSL
1332 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1333 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1334 the first called function after an EVP_EncryptInit(), and therefore that
1335 specific call must be safe. The second form is where the length passed to
1336 EVP_EncryptUpdate() can be seen from the code to be some small value and
1337 therefore there is no possibility of an overflow. Since all instances are
1338 one of these two forms, it is believed that there can be no overflows in
1339 internal code due to this problem. It should be noted that
1340 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1341 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1342 of these calls have also been analysed too and it is believed there are no
1343 instances in internal usage where an overflow could occur.
1345 This issue was reported by Guido Vranken.
1349 *) Prevent ASN.1 BIO excessive memory allocation
1351 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1352 a short invalid encoding can cause allocation of large amounts of memory
1353 potentially consuming excessive resources or exhausting memory.
1355 Any application parsing untrusted data through d2i BIO functions is
1356 affected. The memory based functions such as d2i_X509() are *not* affected.
1357 Since the memory based functions are used by the TLS library, TLS
1358 applications are not affected.
1360 This issue was reported by Brian Carpenter.
1366 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1367 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1368 in arbitrary stack data being returned in the buffer.
1370 This issue was reported by Guido Vranken.
1374 *) Modify behavior of ALPN to invoke callback after SNI/servername
1375 callback, such that updates to the SSL_CTX affect ALPN.
1378 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1382 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1383 methods are enabled and ssl2 is disabled the methods return NULL.
1386 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1388 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1389 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1390 provide any "EXPORT" or "LOW" strength ciphers.
1393 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1394 is by default disabled at build-time. Builds that are not configured with
1395 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1396 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1397 will need to explicitly call either of:
1399 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1401 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1403 as appropriate. Even if either of those is used, or the application
1404 explicitly uses the version-specific SSLv2_method() or its client and
1405 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1406 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1407 ciphers, and SSLv2 56-bit DES are no longer available.
1411 *) Fix a double-free in DSA code
1413 A double free bug was discovered when OpenSSL parses malformed DSA private
1414 keys and could lead to a DoS attack or memory corruption for applications
1415 that receive DSA private keys from untrusted sources. This scenario is
1418 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1423 *) Disable SRP fake user seed to address a server memory leak.
1425 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1427 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1428 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1429 was changed to ignore the "fake user" SRP seed, even if the seed
1432 Users should use SRP_VBASE_get1_by_user instead. Note that in
1433 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1434 also that even though configuring the SRP seed attempts to hide
1435 invalid usernames by continuing the handshake with fake
1436 credentials, this behaviour is not constant time and no strong
1437 guarantees are made that the handshake is indistinguishable from
1438 that of a valid user.
1442 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1444 In the BN_hex2bn function the number of hex digits is calculated using an
1445 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1446 large values of |i| this can result in |bn_expand| not allocating any
1447 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1448 field as NULL leading to a subsequent NULL ptr deref. For very large values
1449 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1450 In this case memory is allocated to the internal BIGNUM data field, but it
1451 is insufficiently sized leading to heap corruption. A similar issue exists
1452 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1453 is ever called by user applications with very large untrusted hex/dec data.
1454 This is anticipated to be a rare occurrence.
1456 All OpenSSL internal usage of these functions use data that is not expected
1457 to be untrusted, e.g. config file data or application command line
1458 arguments. If user developed applications generate config file data based
1459 on untrusted data then it is possible that this could also lead to security
1460 consequences. This is also anticipated to be rare.
1462 This issue was reported to OpenSSL by Guido Vranken.
1466 *) Fix memory issues in BIO_*printf functions
1468 The internal |fmtstr| function used in processing a "%s" format string in
1469 the BIO_*printf functions could overflow while calculating the length of a
1470 string and cause an OOB read when printing very long strings.
1472 Additionally the internal |doapr_outch| function can attempt to write to an
1473 OOB memory location (at an offset from the NULL pointer) in the event of a
1474 memory allocation failure. In 1.0.2 and below this could be caused where
1475 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1476 could be in processing a very long "%s" format string. Memory leaks can
1479 The first issue may mask the second issue dependent on compiler behaviour.
1480 These problems could enable attacks where large amounts of untrusted data
1481 is passed to the BIO_*printf functions. If applications use these functions
1482 in this way then they could be vulnerable. OpenSSL itself uses these
1483 functions when printing out human-readable dumps of ASN.1 data. Therefore
1484 applications that print this data could be vulnerable if the data is from
1485 untrusted sources. OpenSSL command line applications could also be
1486 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1487 as command line arguments.
1489 Libssl is not considered directly vulnerable. Additionally certificates etc
1490 received via remote connections via libssl are also unlikely to be able to
1491 trigger these issues because of message size limits enforced within libssl.
1493 This issue was reported to OpenSSL Guido Vranken.
1497 *) Side channel attack on modular exponentiation
1499 A side-channel attack was found which makes use of cache-bank conflicts on
1500 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1501 of RSA keys. The ability to exploit this issue is limited as it relies on
1502 an attacker who has control of code in a thread running on the same
1503 hyper-threaded core as the victim thread which is performing decryptions.
1505 This issue was reported to OpenSSL by Yuval Yarom, The University of
1506 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1507 Nadia Heninger, University of Pennsylvania with more information at
1508 http://cachebleed.info.
1512 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1513 if no keysize is specified with default_bits. This fixes an
1514 omission in an earlier change that changed all RSA/DSA key generation
1515 apps to use 2048 bits by default.
1518 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1519 *) DH small subgroups
1521 Historically OpenSSL only ever generated DH parameters based on "safe"
1522 primes. More recently (in version 1.0.2) support was provided for
1523 generating X9.42 style parameter files such as those required for RFC 5114
1524 support. The primes used in such files may not be "safe". Where an
1525 application is using DH configured with parameters based on primes that are
1526 not "safe" then an attacker could use this fact to find a peer's private
1527 DH exponent. This attack requires that the attacker complete multiple
1528 handshakes in which the peer uses the same private DH exponent. For example
1529 this could be used to discover a TLS server's private DH exponent if it's
1530 reusing the private DH exponent or it's using a static DH ciphersuite.
1532 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1533 TLS. It is not on by default. If the option is not set then the server
1534 reuses the same private DH exponent for the life of the server process and
1535 would be vulnerable to this attack. It is believed that many popular
1536 applications do set this option and would therefore not be at risk.
1538 The fix for this issue adds an additional check where a "q" parameter is
1539 available (as is the case in X9.42 based parameters). This detects the
1540 only known attack, and is the only possible defense for static DH
1541 ciphersuites. This could have some performance impact.
1543 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1544 default and cannot be disabled. This could have some performance impact.
1546 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1550 *) SSLv2 doesn't block disabled ciphers
1552 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1553 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1554 been disabled, provided that the SSLv2 protocol was not also disabled via
1557 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1558 and Sebastian Schinzel.
1562 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1564 *) BN_mod_exp may produce incorrect results on x86_64
1566 There is a carry propagating bug in the x86_64 Montgomery squaring
1567 procedure. No EC algorithms are affected. Analysis suggests that attacks
1568 against RSA and DSA as a result of this defect would be very difficult to
1569 perform and are not believed likely. Attacks against DH are considered just
1570 feasible (although very difficult) because most of the work necessary to
1571 deduce information about a private key may be performed offline. The amount
1572 of resources required for such an attack would be very significant and
1573 likely only accessible to a limited number of attackers. An attacker would
1574 additionally need online access to an unpatched system using the target
1575 private key in a scenario with persistent DH parameters and a private
1576 key that is shared between multiple clients. For example this can occur by
1577 default in OpenSSL DHE based SSL/TLS ciphersuites.
1579 This issue was reported to OpenSSL by Hanno Böck.
1583 *) Certificate verify crash with missing PSS parameter
1585 The signature verification routines will crash with a NULL pointer
1586 dereference if presented with an ASN.1 signature using the RSA PSS
1587 algorithm and absent mask generation function parameter. Since these
1588 routines are used to verify certificate signature algorithms this can be
1589 used to crash any certificate verification operation and exploited in a
1590 DoS attack. Any application which performs certificate verification is
1591 vulnerable including OpenSSL clients and servers which enable client
1594 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1598 *) X509_ATTRIBUTE memory leak
1600 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1601 memory. This structure is used by the PKCS#7 and CMS routines so any
1602 application which reads PKCS#7 or CMS data from untrusted sources is
1603 affected. SSL/TLS is not affected.
1605 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1610 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1611 This changes the decoding behaviour for some invalid messages,
1612 though the change is mostly in the more lenient direction, and
1613 legacy behaviour is preserved as much as possible.
1616 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1618 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1620 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1622 *) Alternate chains certificate forgery
1624 During certificate verification, OpenSSL will attempt to find an
1625 alternative certificate chain if the first attempt to build such a chain
1626 fails. An error in the implementation of this logic can mean that an
1627 attacker could cause certain checks on untrusted certificates to be
1628 bypassed, such as the CA flag, enabling them to use a valid leaf
1629 certificate to act as a CA and "issue" an invalid certificate.
1631 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1635 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1637 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1638 incompatibility in the handling of HMAC. The previous ABI has now been
1642 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1644 *) Malformed ECParameters causes infinite loop
1646 When processing an ECParameters structure OpenSSL enters an infinite loop
1647 if the curve specified is over a specially malformed binary polynomial
1650 This can be used to perform denial of service against any
1651 system which processes public keys, certificate requests or
1652 certificates. This includes TLS clients and TLS servers with
1653 client authentication enabled.
1655 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1659 *) Exploitable out-of-bounds read in X509_cmp_time
1661 X509_cmp_time does not properly check the length of the ASN1_TIME
1662 string and can read a few bytes out of bounds. In addition,
1663 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1666 An attacker can use this to craft malformed certificates and CRLs of
1667 various sizes and potentially cause a segmentation fault, resulting in
1668 a DoS on applications that verify certificates or CRLs. TLS clients
1669 that verify CRLs are affected. TLS clients and servers with client
1670 authentication enabled may be affected if they use custom verification
1673 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1674 independently by Hanno Böck.
1678 *) PKCS7 crash with missing EnvelopedContent
1680 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1681 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1682 with missing content and trigger a NULL pointer dereference on parsing.
1684 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1685 structures from untrusted sources are affected. OpenSSL clients and
1686 servers are not affected.
1688 This issue was reported to OpenSSL by Michal Zalewski (Google).
1692 *) CMS verify infinite loop with unknown hash function
1694 When verifying a signedData message the CMS code can enter an infinite loop
1695 if presented with an unknown hash function OID. This can be used to perform
1696 denial of service against any system which verifies signedData messages using
1698 This issue was reported to OpenSSL by Johannes Bauer.
1702 *) Race condition handling NewSessionTicket
1704 If a NewSessionTicket is received by a multi-threaded client when attempting to
1705 reuse a previous ticket then a race condition can occur potentially leading to
1706 a double free of the ticket data.
1710 *) Only support 256-bit or stronger elliptic curves with the
1711 'ecdh_auto' setting (server) or by default (client). Of supported
1712 curves, prefer P-256 (both).
1715 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1717 *) ClientHello sigalgs DoS fix
1719 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1720 invalid signature algorithms extension a NULL pointer dereference will
1721 occur. This can be exploited in a DoS attack against the server.
1723 This issue was was reported to OpenSSL by David Ramos of Stanford
1726 [Stephen Henson and Matt Caswell]
1728 *) Multiblock corrupted pointer fix
1730 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1731 feature only applies on 64 bit x86 architecture platforms that support AES
1732 NI instructions. A defect in the implementation of "multiblock" can cause
1733 OpenSSL's internal write buffer to become incorrectly set to NULL when
1734 using non-blocking IO. Typically, when the user application is using a
1735 socket BIO for writing, this will only result in a failed connection.
1736 However if some other BIO is used then it is likely that a segmentation
1737 fault will be triggered, thus enabling a potential DoS attack.
1739 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1743 *) Segmentation fault in DTLSv1_listen fix
1745 The DTLSv1_listen function is intended to be stateless and processes the
1746 initial ClientHello from many peers. It is common for user code to loop
1747 over the call to DTLSv1_listen until a valid ClientHello is received with
1748 an associated cookie. A defect in the implementation of DTLSv1_listen means
1749 that state is preserved in the SSL object from one invocation to the next
1750 that can lead to a segmentation fault. Errors processing the initial
1751 ClientHello can trigger this scenario. An example of such an error could be
1752 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1755 This issue was reported to OpenSSL by Per Allansson.
1759 *) Segmentation fault in ASN1_TYPE_cmp fix
1761 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1762 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1763 certificate signature algorithm consistency this can be used to crash any
1764 certificate verification operation and exploited in a DoS attack. Any
1765 application which performs certificate verification is vulnerable including
1766 OpenSSL clients and servers which enable client authentication.
1770 *) Segmentation fault for invalid PSS parameters fix
1772 The signature verification routines will crash with a NULL pointer
1773 dereference if presented with an ASN.1 signature using the RSA PSS
1774 algorithm and invalid parameters. Since these routines are used to verify
1775 certificate signature algorithms this can be used to crash any
1776 certificate verification operation and exploited in a DoS attack. Any
1777 application which performs certificate verification is vulnerable including
1778 OpenSSL clients and servers which enable client authentication.
1780 This issue was was reported to OpenSSL by Brian Carpenter.
1784 *) ASN.1 structure reuse memory corruption fix
1786 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1787 memory corruption via an invalid write. Such reuse is and has been
1788 strongly discouraged and is believed to be rare.
1790 Applications that parse structures containing CHOICE or ANY DEFINED BY
1791 components may be affected. Certificate parsing (d2i_X509 and related
1792 functions) are however not affected. OpenSSL clients and servers are
1797 *) PKCS7 NULL pointer dereferences fix
1799 The PKCS#7 parsing code does not handle missing outer ContentInfo
1800 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1801 missing content and trigger a NULL pointer dereference on parsing.
1803 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1804 otherwise parse PKCS#7 structures from untrusted sources are
1805 affected. OpenSSL clients and servers are not affected.
1807 This issue was reported to OpenSSL by Michal Zalewski (Google).
1811 *) DoS via reachable assert in SSLv2 servers fix
1813 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1814 servers that both support SSLv2 and enable export cipher suites by sending
1815 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1817 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1818 (OpenSSL development team).
1822 *) Empty CKE with client auth and DHE fix
1824 If client auth is used then a server can seg fault in the event of a DHE
1825 ciphersuite being selected and a zero length ClientKeyExchange message
1826 being sent by the client. This could be exploited in a DoS attack.
1830 *) Handshake with unseeded PRNG fix
1832 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1833 with an unseeded PRNG. The conditions are:
1834 - The client is on a platform where the PRNG has not been seeded
1835 automatically, and the user has not seeded manually
1836 - A protocol specific client method version has been used (i.e. not
1837 SSL_client_methodv23)
1838 - A ciphersuite is used that does not require additional random data from
1839 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1841 If the handshake succeeds then the client random that has been used will
1842 have been generated from a PRNG with insufficient entropy and therefore the
1843 output may be predictable.
1845 For example using the following command with an unseeded openssl will
1846 succeed on an unpatched platform:
1848 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1852 *) Use After Free following d2i_ECPrivatekey error fix
1854 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1855 could cause a use after free condition. This, in turn, could cause a double
1856 free in several private key parsing functions (such as d2i_PrivateKey
1857 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1858 for applications that receive EC private keys from untrusted
1859 sources. This scenario is considered rare.
1861 This issue was discovered by the BoringSSL project and fixed in their
1866 *) X509_to_X509_REQ NULL pointer deref fix
1868 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1869 the certificate key is invalid. This function is rarely used in practice.
1871 This issue was discovered by Brian Carpenter.
1875 *) Removed the export ciphers from the DEFAULT ciphers
1878 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1880 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1881 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1882 So far those who have to target multiple platforms would compromise
1883 and argue that binary targeting say ARMv5 would still execute on
1884 ARMv8. "Universal" build resolves this compromise by providing
1885 near-optimal performance even on newer platforms.
1888 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1889 (other platforms pending).
1890 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1892 *) Add support for the SignedCertificateTimestampList certificate and
1893 OCSP response extensions from RFC6962.
1896 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1897 for corner cases. (Certain input points at infinity could lead to
1898 bogus results, with non-infinity inputs mapped to infinity too.)
1901 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1902 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1903 common cases are optimized and there still is room for further
1904 improvements. Vector Permutation AES for Altivec is also added.
1907 *) Add support for little-endian ppc64 Linux target.
1908 [Marcelo Cerri (IBM)]
1910 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1911 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1912 are optimized and there still is room for further improvements.
1913 Both 32- and 64-bit modes are supported.
1914 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1916 *) Improved ARMv7 NEON support.
1919 *) Support for SPARC Architecture 2011 crypto extensions, first
1920 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1921 SHA256/512, MD5, GHASH and modular exponentiation.
1922 [Andy Polyakov, David Miller]
1924 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1926 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1928 *) Support for new and upcoming Intel processors, including AVX2,
1929 BMI and SHA ISA extensions. This includes additional "stitched"
1930 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1933 This work was sponsored by Intel Corp.
1936 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1937 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1938 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1941 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1942 this fixes a limitation in previous versions of OpenSSL.
1945 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1946 MGF1 digest and OAEP label.
1949 *) Add EVP support for key wrapping algorithms, to avoid problems with
1950 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1951 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1952 algorithms and include tests cases.
1955 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1957 [Douglas E. Engert, Steve Henson]
1959 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1960 difference in days and seconds between two tm or ASN1_TIME structures.
1963 *) Add -rev test option to s_server to just reverse order of characters
1964 received by client and send back to server. Also prints an abbreviated
1965 summary of the connection parameters.
1968 *) New option -brief for s_client and s_server to print out a brief summary
1969 of connection parameters.
1972 *) Add callbacks for arbitrary TLS extensions.
1973 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1975 *) New option -crl_download in several openssl utilities to download CRLs
1976 from CRLDP extension in certificates.
1979 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1982 *) New function X509_CRL_diff to generate a delta CRL from the difference
1983 of two full CRLs. Add support to "crl" utility.
1986 *) New functions to set lookup_crls function and to retrieve
1987 X509_STORE from X509_STORE_CTX.
1990 *) Print out deprecated issuer and subject unique ID fields in
1994 *) Extend OCSP I/O functions so they can be used for simple general purpose
1995 HTTP as well as OCSP. New wrapper function which can be used to download
1996 CRLs using the OCSP API.
1999 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2002 *) SSL_CONF* functions. These provide a common framework for application
2003 configuration using configuration files or command lines.
2006 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2007 message callback and prints the results. Needs compile time option
2008 "enable-ssl-trace". New options to s_client and s_server to enable
2012 *) New ctrl and macro to retrieve supported points extensions.
2013 Print out extension in s_server and s_client.
2016 *) New functions to retrieve certificate signature and signature
2020 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2024 *) New Suite B modes for TLS code. These use and enforce the requirements
2025 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2026 only use Suite B curves. The Suite B modes can be set by using the
2027 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2030 *) New chain verification flags for Suite B levels of security. Check
2031 algorithms are acceptable when flags are set in X509_verify_cert.
2034 *) Make tls1_check_chain return a set of flags indicating checks passed
2035 by a certificate chain. Add additional tests to handle client
2036 certificates: checks for matching certificate type and issuer name
2040 *) If an attempt is made to use a signature algorithm not in the peer
2041 preference list abort the handshake. If client has no suitable
2042 signature algorithms in response to a certificate request do not
2043 use the certificate.
2046 *) If server EC tmp key is not in client preference list abort handshake.
2049 *) Add support for certificate stores in CERT structure. This makes it
2050 possible to have different stores per SSL structure or one store in
2051 the parent SSL_CTX. Include distinct stores for certificate chain
2052 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2053 to build and store a certificate chain in CERT structure: returning
2054 an error if the chain cannot be built: this will allow applications
2055 to test if a chain is correctly configured.
2057 Note: if the CERT based stores are not set then the parent SSL_CTX
2058 store is used to retain compatibility with existing behaviour.
2062 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2063 mask based on the current session, check mask when sending client
2064 hello and checking the requested ciphersuite.
2067 *) New ctrls to retrieve and set certificate types in a certificate
2068 request message. Print out received values in s_client. If certificate
2069 types is not set with custom values set sensible values based on
2070 supported signature algorithms.
2073 *) Support for distinct client and server supported signature algorithms.
2076 *) Add certificate callback. If set this is called whenever a certificate
2077 is required by client or server. An application can decide which
2078 certificate chain to present based on arbitrary criteria: for example
2079 supported signature algorithms. Add very simple example to s_server.
2080 This fixes many of the problems and restrictions of the existing client
2081 certificate callback: for example you can now clear an existing
2082 certificate and specify the whole chain.
2085 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2086 the certificate can be used for (if anything). Set valid_flags field
2087 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2088 to have similar checks in it.
2090 Add new "cert_flags" field to CERT structure and include a "strict mode".
2091 This enforces some TLS certificate requirements (such as only permitting
2092 certificate signature algorithms contained in the supported algorithms
2093 extension) which some implementations ignore: this option should be used
2094 with caution as it could cause interoperability issues.
2097 *) Update and tidy signature algorithm extension processing. Work out
2098 shared signature algorithms based on preferences and peer algorithms
2099 and print them out in s_client and s_server. Abort handshake if no
2100 shared signature algorithms.
2103 *) Add new functions to allow customised supported signature algorithms
2104 for SSL and SSL_CTX structures. Add options to s_client and s_server
2108 *) New function SSL_certs_clear() to delete all references to certificates
2109 from an SSL structure. Before this once a certificate had been added
2110 it couldn't be removed.
2113 *) Integrate hostname, email address and IP address checking with certificate
2114 verification. New verify options supporting checking in openssl utility.
2117 *) Fixes and wildcard matching support to hostname and email checking
2118 functions. Add manual page.
2119 [Florian Weimer (Red Hat Product Security Team)]
2121 *) New functions to check a hostname email or IP address against a
2122 certificate. Add options x509 utility to print results of checks against
2126 *) Fix OCSP checking.
2127 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2129 *) Initial experimental support for explicitly trusted non-root CAs.
2130 OpenSSL still tries to build a complete chain to a root but if an
2131 intermediate CA has a trust setting included that is used. The first
2132 setting is used: whether to trust (e.g., -addtrust option to the x509
2136 *) Add -trusted_first option which attempts to find certificates in the
2137 trusted store even if an untrusted chain is also supplied.
2140 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2141 platform support for Linux and Android.
2144 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2147 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2148 When in FIPS mode the approved implementations are used as normal,
2149 when not in FIPS mode the internal unapproved versions are used instead.
2150 This means that the FIPS capable OpenSSL isn't forced to use the
2151 (often lower performance) FIPS implementations outside FIPS mode.
2154 *) Transparently support X9.42 DH parameters when calling
2155 PEM_read_bio_DHparameters. This means existing applications can handle
2156 the new parameter format automatically.
2159 *) Initial experimental support for X9.42 DH parameter format: mainly
2160 to support use of 'q' parameter for RFC5114 parameters.
2163 *) Add DH parameters from RFC5114 including test data to dhtest.
2166 *) Support for automatic EC temporary key parameter selection. If enabled
2167 the most preferred EC parameters are automatically used instead of
2168 hardcoded fixed parameters. Now a server just has to call:
2169 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2170 support ECDH and use the most appropriate parameters.
2173 *) Enhance and tidy EC curve and point format TLS extension code. Use
2174 static structures instead of allocation if default values are used.
2175 New ctrls to set curves we wish to support and to retrieve shared curves.
2176 Print out shared curves in s_server. New options to s_server and s_client
2177 to set list of supported curves.
2180 *) New ctrls to retrieve supported signature algorithms and
2181 supported curve values as an array of NIDs. Extend openssl utility
2182 to print out received values.
2185 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2186 between NIDs and the more common NIST names such as "P-256". Enhance
2187 ecparam utility and ECC method to recognise the NIST names for curves.
2190 *) Enhance SSL/TLS certificate chain handling to support different
2191 chains for each certificate instead of one chain in the parent SSL_CTX.
2194 *) Support for fixed DH ciphersuite client authentication: where both
2195 server and client use DH certificates with common parameters.
2198 *) Support for fixed DH ciphersuites: those requiring DH server
2202 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2204 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2205 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2206 X509_CINF_get_signature were reverted post internal team review.
2208 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2210 *) Build fixes for the Windows and OpenVMS platforms
2211 [Matt Caswell and Richard Levitte]
2213 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2215 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2216 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2217 dereference. This could lead to a Denial Of Service attack. Thanks to
2218 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2222 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2223 dtls1_buffer_record function under certain conditions. In particular this
2224 could occur if an attacker sent repeated DTLS records with the same
2225 sequence number but for the next epoch. The memory leak could be exploited
2226 by an attacker in a Denial of Service attack through memory exhaustion.
2227 Thanks to Chris Mueller for reporting this issue.
2231 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2232 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2233 method would be set to NULL which could later result in a NULL pointer
2234 dereference. Thanks to Frank Schmirler for reporting this issue.
2238 *) Abort handshake if server key exchange message is omitted for ephemeral
2241 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2242 reporting this issue.
2246 *) Remove non-export ephemeral RSA code on client and server. This code
2247 violated the TLS standard by allowing the use of temporary RSA keys in
2248 non-export ciphersuites and could be used by a server to effectively
2249 downgrade the RSA key length used to a value smaller than the server
2250 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2251 INRIA or reporting this issue.
2255 *) Fixed issue where DH client certificates are accepted without verification.
2256 An OpenSSL server will accept a DH certificate for client authentication
2257 without the certificate verify message. This effectively allows a client to
2258 authenticate without the use of a private key. This only affects servers
2259 which trust a client certificate authority which issues certificates
2260 containing DH keys: these are extremely rare and hardly ever encountered.
2261 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2266 *) Ensure that the session ID context of an SSL is updated when its
2267 SSL_CTX is updated via SSL_set_SSL_CTX.
2269 The session ID context is typically set from the parent SSL_CTX,
2270 and can vary with the CTX.
2273 *) Fix various certificate fingerprint issues.
2275 By using non-DER or invalid encodings outside the signed portion of a
2276 certificate the fingerprint can be changed without breaking the signature.
2277 Although no details of the signed portion of the certificate can be changed
2278 this can cause problems with some applications: e.g. those using the
2279 certificate fingerprint for blacklists.
2281 1. Reject signatures with non zero unused bits.
2283 If the BIT STRING containing the signature has non zero unused bits reject
2284 the signature. All current signature algorithms require zero unused bits.
2286 2. Check certificate algorithm consistency.
2288 Check the AlgorithmIdentifier inside TBS matches the one in the
2289 certificate signature. NB: this will result in signature failure
2290 errors for some broken certificates.
2292 Thanks to Konrad Kraszewski from Google for reporting this issue.
2294 3. Check DSA/ECDSA signatures use DER.
2296 Re-encode DSA/ECDSA signatures and compare with the original received
2297 signature. Return an error if there is a mismatch.
2299 This will reject various cases including garbage after signature
2300 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2301 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2302 (negative or with leading zeroes).
2304 Further analysis was conducted and fixes were developed by Stephen Henson
2305 of the OpenSSL core team.
2310 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2311 results on some platforms, including x86_64. This bug occurs at random
2312 with a very low probability, and is not known to be exploitable in any
2313 way, though its exact impact is difficult to determine. Thanks to Pieter
2314 Wuille (Blockstream) who reported this issue and also suggested an initial
2315 fix. Further analysis was conducted by the OpenSSL development team and
2316 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2317 the OpenSSL core team.
2321 *) Do not resume sessions on the server if the negotiated protocol
2322 version does not match the session's version. Resuming with a different
2323 version, while not strictly forbidden by the RFC, is of questionable
2324 sanity and breaks all known clients.
2325 [David Benjamin, Emilia Käsper]
2327 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2328 early CCS messages during renegotiation. (Note that because
2329 renegotiation is encrypted, this early CCS was not exploitable.)
2332 *) Tighten client-side session ticket handling during renegotiation:
2333 ensure that the client only accepts a session ticket if the server sends
2334 the extension anew in the ServerHello. Previously, a TLS client would
2335 reuse the old extension state and thus accept a session ticket if one was
2336 announced in the initial ServerHello.
2338 Similarly, ensure that the client requires a session ticket if one
2339 was advertised in the ServerHello. Previously, a TLS client would
2340 ignore a missing NewSessionTicket message.
2343 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2345 *) SRTP Memory Leak.
2347 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2348 sends a carefully crafted handshake message, to cause OpenSSL to fail
2349 to free up to 64k of memory causing a memory leak. This could be
2350 exploited in a Denial Of Service attack. This issue affects OpenSSL
2351 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2352 whether SRTP is used or configured. Implementations of OpenSSL that
2353 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2355 The fix was developed by the OpenSSL team.
2359 *) Session Ticket Memory Leak.
2361 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2362 integrity of that ticket is first verified. In the event of a session
2363 ticket integrity check failing, OpenSSL will fail to free memory
2364 causing a memory leak. By sending a large number of invalid session
2365 tickets an attacker could exploit this issue in a Denial Of Service
2370 *) Build option no-ssl3 is incomplete.
2372 When OpenSSL is configured with "no-ssl3" as a build option, servers
2373 could accept and complete a SSL 3.0 handshake, and clients could be
2374 configured to send them.
2376 [Akamai and the OpenSSL team]
2378 *) Add support for TLS_FALLBACK_SCSV.
2379 Client applications doing fallback retries should call
2380 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2382 [Adam Langley, Bodo Moeller]
2384 *) Add additional DigestInfo checks.
2386 Re-encode DigestInto in DER and check against the original when
2387 verifying RSA signature: this will reject any improperly encoded
2388 DigestInfo structures.
2390 Note: this is a precautionary measure and no attacks are currently known.
2394 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2396 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2397 SRP code can be overrun an internal buffer. Add sanity check that
2398 g, A, B < N to SRP code.
2400 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2401 Group for discovering this issue.
2405 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2406 TLS 1.0 instead of higher protocol versions when the ClientHello message
2407 is badly fragmented. This allows a man-in-the-middle attacker to force a
2408 downgrade to TLS 1.0 even if both the server and the client support a
2409 higher protocol version, by modifying the client's TLS records.
2411 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2412 researching this issue.
2416 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2417 to a denial of service attack. A malicious server can crash the client
2418 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2419 ciphersuite and sending carefully crafted handshake messages.
2421 Thanks to Felix Gröbert (Google) for discovering and researching this
2426 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2427 to leak memory. This can be exploited through a Denial of Service attack.
2428 Thanks to Adam Langley for discovering and researching this issue.
2432 *) An attacker can force openssl to consume large amounts of memory whilst
2433 processing DTLS handshake messages. This can be exploited through a
2434 Denial of Service attack.
2435 Thanks to Adam Langley for discovering and researching this issue.
2439 *) An attacker can force an error condition which causes openssl to crash
2440 whilst processing DTLS packets due to memory being freed twice. This
2441 can be exploited through a Denial of Service attack.
2442 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2447 *) If a multithreaded client connects to a malicious server using a resumed
2448 session and the server sends an ec point format extension it could write
2449 up to 255 bytes to freed memory.
2451 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2456 *) A malicious server can crash an OpenSSL client with a null pointer
2457 dereference (read) by specifying an SRP ciphersuite even though it was not
2458 properly negotiated with the client. This can be exploited through a
2459 Denial of Service attack.
2461 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2462 discovering and researching this issue.
2466 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2467 X509_name_oneline, X509_name_print_ex et al. to leak some information
2468 from the stack. Applications may be affected if they echo pretty printing
2469 output to the attacker.
2471 Thanks to Ivan Fratric (Google) for discovering this issue.
2473 [Emilia Käsper, and Steve Henson]
2475 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2476 for corner cases. (Certain input points at infinity could lead to
2477 bogus results, with non-infinity inputs mapped to infinity too.)
2480 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2482 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2483 handshake can force the use of weak keying material in OpenSSL
2484 SSL/TLS clients and servers.
2486 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2487 researching this issue. (CVE-2014-0224)
2488 [KIKUCHI Masashi, Steve Henson]
2490 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2491 OpenSSL DTLS client the code can be made to recurse eventually crashing
2494 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2496 [Imre Rad, Steve Henson]
2498 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2499 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2500 client or server. This is potentially exploitable to run arbitrary
2501 code on a vulnerable client or server.
2503 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2504 [Jüri Aedla, Steve Henson]
2506 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2507 are subject to a denial of service attack.
2509 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2510 this issue. (CVE-2014-3470)
2511 [Felix Gröbert, Ivan Fratric, Steve Henson]
2513 *) Harmonize version and its documentation. -f flag is used to display
2515 [mancha <mancha1@zoho.com>]
2517 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2518 in i2d_ECPrivateKey.
2519 [mancha <mancha1@zoho.com>]
2521 *) Fix some double frees. These are not thought to be exploitable.
2522 [mancha <mancha1@zoho.com>]
2524 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2526 *) A missing bounds check in the handling of the TLS heartbeat extension
2527 can be used to reveal up to 64k of memory to a connected client or
2530 Thanks for Neel Mehta of Google Security for discovering this bug and to
2531 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2532 preparing the fix (CVE-2014-0160)
2533 [Adam Langley, Bodo Moeller]
2535 *) Fix for the attack described in the paper "Recovering OpenSSL
2536 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2537 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2538 http://eprint.iacr.org/2014/140
2540 Thanks to Yuval Yarom and Naomi Benger for discovering this
2541 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2542 [Yuval Yarom and Naomi Benger]
2544 *) TLS pad extension: draft-agl-tls-padding-03
2546 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2547 TLS client Hello record length value would otherwise be > 255 and
2548 less that 512 pad with a dummy extension containing zeroes so it
2549 is at least 512 bytes long.
2551 [Adam Langley, Steve Henson]
2553 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2555 *) Fix for TLS record tampering bug. A carefully crafted invalid
2556 handshake could crash OpenSSL with a NULL pointer exception.
2557 Thanks to Anton Johansson for reporting this issues.
2560 *) Keep original DTLS digest and encryption contexts in retransmission
2561 structures so we can use the previous session parameters if they need
2562 to be resent. (CVE-2013-6450)
2565 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2566 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2567 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2568 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2569 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2570 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2571 [Rob Stradling, Adam Langley]
2573 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2575 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2576 supporting platforms or when small records were transferred.
2577 [Andy Polyakov, Steve Henson]
2579 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2581 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2583 This addresses the flaw in CBC record processing discovered by
2584 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2585 at: http://www.isg.rhul.ac.uk/tls/
2587 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2588 Security Group at Royal Holloway, University of London
2589 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2590 Emilia Käsper for the initial patch.
2592 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2594 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2595 ciphersuites which can be exploited in a denial of service attack.
2596 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2597 and detecting this bug and to Wolfgang Ettlinger
2598 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2602 *) Return an error when checking OCSP signatures when key is NULL.
2603 This fixes a DoS attack. (CVE-2013-0166)
2606 *) Make openssl verify return errors.
2607 [Chris Palmer <palmer@google.com> and Ben Laurie]
2609 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2610 the right response is stapled. Also change SSL_get_certificate()
2611 so it returns the certificate actually sent.
2612 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2613 [Rob Stradling <rob.stradling@comodo.com>]
2615 *) Fix possible deadlock when decoding public keys.
2618 *) Don't use TLS 1.0 record version number in initial client hello
2622 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2624 *) Sanity check record length before skipping explicit IV in TLS
2625 1.2, 1.1 and DTLS to fix DoS attack.
2627 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2628 fuzzing as a service testing platform.
2632 *) Initialise tkeylen properly when encrypting CMS messages.
2633 Thanks to Solar Designer of Openwall for reporting this issue.
2636 *) In FIPS mode don't try to use composite ciphers as they are not
2640 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2642 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2643 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2644 mean any application compiled against OpenSSL 1.0.0 headers setting
2645 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2646 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2647 0x10000000L Any application which was previously compiled against
2648 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2649 will need to be recompiled as a result. Letting be results in
2650 inability to disable specifically TLS 1.1 and in client context,
2651 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2654 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2655 disable just protocol X, but all protocols above X *if* there are
2656 protocols *below* X still enabled. In more practical terms it means
2657 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2658 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2659 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2663 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2665 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2666 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2667 in CRYPTO_realloc_clean.
2669 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2670 issue and to Adam Langley <agl@chromium.org> for fixing it.
2672 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2674 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2677 *) Workarounds for some broken servers that "hang" if a client hello
2678 record length exceeds 255 bytes.
2680 1. Do not use record version number > TLS 1.0 in initial client
2681 hello: some (but not all) hanging servers will now work.
2682 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2683 the number of ciphers sent in the client hello. This should be
2684 set to an even number, such as 50, for example by passing:
2685 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2686 Most broken servers should now work.
2687 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2688 TLS 1.2 client support entirely.
2691 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2694 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2696 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2697 STRING form instead of a DigestInfo.
2700 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2701 and the RSA_sign/RSA_verify functions. This was made more apparent when
2702 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2703 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2704 the correct format in RSA_verify so both forms transparently work.
2707 *) Some servers which support TLS 1.0 can choke if we initially indicate
2708 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2709 encrypted premaster secret. As a workaround use the maximum permitted
2710 client version in client hello, this should keep such servers happy
2711 and still work with previous versions of OpenSSL.
2714 *) Add support for TLS/DTLS heartbeats.
2715 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2717 *) Add support for SCTP.
2718 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2720 *) Improved PRNG seeding for VOS.
2721 [Paul Green <Paul.Green@stratus.com>]
2723 *) Extensive assembler packs updates, most notably:
2725 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2726 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2727 - x86_64: bit-sliced AES implementation;
2728 - ARM: NEON support, contemporary platforms optimizations;
2729 - s390x: z196 support;
2730 - *: GHASH and GF(2^m) multiplication implementations;
2734 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2735 (removal of unnecessary code)
2736 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2738 *) Add TLS key material exporter from RFC 5705.
2741 *) Add DTLS-SRTP negotiation from RFC 5764.
2744 *) Add Next Protocol Negotiation,
2745 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2746 disabled with a no-npn flag to config or Configure. Code donated
2748 [Adam Langley <agl@google.com> and Ben Laurie]
2750 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2751 NIST-P256, NIST-P521, with constant-time single point multiplication on
2752 typical inputs. Compiler support for the nonstandard type __uint128_t is
2753 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2754 Code made available under Apache License version 2.0.
2756 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2757 line to include this in your build of OpenSSL, and run "make depend" (or
2758 "make update"). This enables the following EC_METHODs:
2760 EC_GFp_nistp224_method()
2761 EC_GFp_nistp256_method()
2762 EC_GFp_nistp521_method()
2764 EC_GROUP_new_by_curve_name() will automatically use these (while
2765 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2767 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2769 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2770 all platforms. Move ssize_t definition from e_os.h to the public
2771 header file e_os2.h as it now appears in public header file cms.h
2774 *) New -sigopt option to the ca, req and x509 utilities. Additional
2775 signature parameters can be passed using this option and in
2779 *) Add RSA PSS signing function. This will generate and set the
2780 appropriate AlgorithmIdentifiers for PSS based on those in the
2781 corresponding EVP_MD_CTX structure. No application support yet.
2784 *) Support for companion algorithm specific ASN1 signing routines.
2785 New function ASN1_item_sign_ctx() signs a pre-initialised
2786 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2787 the appropriate parameters.
2790 *) Add new algorithm specific ASN1 verification initialisation function
2791 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2792 handling will be the same no matter what EVP_PKEY_METHOD is used.
2793 Add a PSS handler to support verification of PSS signatures: checked
2794 against a number of sample certificates.
2797 *) Add signature printing for PSS. Add PSS OIDs.
2798 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2800 *) Add algorithm specific signature printing. An individual ASN1 method
2801 can now print out signatures instead of the standard hex dump.
2803 More complex signatures (e.g. PSS) can print out more meaningful
2804 information. Include DSA version that prints out the signature
2808 *) Password based recipient info support for CMS library: implementing
2812 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2813 neatly separates the code into cipher and PBE sections and is required
2814 for some algorithms that split PBES2 into separate pieces (such as
2815 password based CMS).
2818 *) Session-handling fixes:
2819 - Fix handling of connections that are resuming with a session ID,
2820 but also support Session Tickets.
2821 - Fix a bug that suppressed issuing of a new ticket if the client
2822 presented a ticket with an expired session.
2823 - Try to set the ticket lifetime hint to something reasonable.
2824 - Make tickets shorter by excluding irrelevant information.
2825 - On the client side, don't ignore renewed tickets.
2826 [Adam Langley, Bodo Moeller (Google)]
2828 *) Fix PSK session representation.
2831 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2833 This work was sponsored by Intel.
2836 *) Add GCM support to TLS library. Some custom code is needed to split
2837 the IV between the fixed (from PRF) and explicit (from TLS record)
2838 portions. This adds all GCM ciphersuites supported by RFC5288 and
2839 RFC5289. Generalise some AES* cipherstrings to include GCM and
2840 add a special AESGCM string for GCM only.
2843 *) Expand range of ctrls for AES GCM. Permit setting invocation
2844 field on decrypt and retrieval of invocation field only on encrypt.
2847 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2848 As required by RFC5289 these ciphersuites cannot be used if for
2849 versions of TLS earlier than 1.2.
2852 *) For FIPS capable OpenSSL interpret a NULL default public key method
2853 as unset and return the appropriate default but do *not* set the default.
2854 This means we can return the appropriate method in applications that
2855 switch between FIPS and non-FIPS modes.
2858 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2859 ENGINE is used then we cannot handle that in the FIPS module so we
2860 keep original code iff non-FIPS operations are allowed.
2863 *) Add -attime option to openssl utilities.
2864 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2866 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2869 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2870 FIPS EC methods unconditionally for now.
2873 *) New build option no-ec2m to disable characteristic 2 code.
2876 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2877 all cases can be covered as some introduce binary incompatibilities.
2880 *) Redirect RSA operations to FIPS module including keygen,
2881 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2884 *) Add similar low level API blocking to ciphers.
2887 *) Low level digest APIs are not approved in FIPS mode: any attempt
2888 to use these will cause a fatal error. Applications that *really* want
2889 to use them can use the private_* version instead.
2892 *) Redirect cipher operations to FIPS module for FIPS builds.
2895 *) Redirect digest operations to FIPS module for FIPS builds.
2898 *) Update build system to add "fips" flag which will link in fipscanister.o
2899 for static and shared library builds embedding a signature if needed.
2902 *) Output TLS supported curves in preference order instead of numerical
2903 order. This is currently hardcoded for the highest order curves first.
2904 This should be configurable so applications can judge speed vs strength.
2907 *) Add TLS v1.2 server support for client authentication.
2910 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2914 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2915 FIPS modules versions.
2918 *) Add TLS v1.2 client side support for client authentication. Keep cache
2919 of handshake records longer as we don't know the hash algorithm to use
2920 until after the certificate request message is received.
2923 *) Initial TLS v1.2 client support. Add a default signature algorithms
2924 extension including all the algorithms we support. Parse new signature
2925 format in client key exchange. Relax some ECC signing restrictions for
2926 TLS v1.2 as indicated in RFC5246.
2929 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2930 to new signature format when needed using client digest preference.
2931 All server ciphersuites should now work correctly in TLS v1.2. No client
2932 support yet and no support for client certificates.
2935 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2936 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2937 ciphersuites. At present only RSA key exchange ciphersuites work with
2938 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2939 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2940 and version checking.
2943 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2944 with this defined it will not be affected by any changes to ssl internal
2945 structures. Add several utility functions to allow openssl application
2946 to work with OPENSSL_NO_SSL_INTERN defined.
2950 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2952 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2955 *) Permit abbreviated handshakes when renegotiating using the function
2956 SSL_renegotiate_abbreviated().
2957 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2959 *) Add call to ENGINE_register_all_complete() to
2960 ENGINE_load_builtin_engines(), so some implementations get used
2961 automatically instead of needing explicit application support.
2964 *) Add support for TLS key exporter as described in RFC5705.
2965 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2967 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2968 a few changes are required:
2970 Add SSL_OP_NO_TLSv1_1 flag.
2971 Add TLSv1_1 methods.
2972 Update version checking logic to handle version 1.1.
2973 Add explicit IV handling (ported from DTLS code).
2974 Add command line options to s_client/s_server.
2977 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2979 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2980 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2981 content decryption and always return the same error. Note: this attack
2982 needs on average 2^20 messages so it only affects automated senders. The
2983 old behaviour can be re-enabled in the CMS code by setting the
2984 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2985 an MMA defence is not necessary.
2986 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2987 this issue. (CVE-2012-0884)
2990 *) Fix CVE-2011-4619: make sure we really are receiving a
2991 client hello before rejecting multiple SGC restarts. Thanks to
2992 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2995 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2997 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2998 Thanks to Antonio Martin, Enterprise Secure Access Research and
2999 Development, Cisco Systems, Inc. for discovering this bug and
3000 preparing a fix. (CVE-2012-0050)
3003 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3005 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3006 of the Vaudenay padding oracle attack on CBC mode encryption
3007 which enables an efficient plaintext recovery attack against
3008 the OpenSSL implementation of DTLS. Their attack exploits timing
3009 differences arising during decryption processing. A research
3010 paper describing this attack can be found at:
3011 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3012 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3013 Security Group at Royal Holloway, University of London
3014 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3015 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3016 for preparing the fix. (CVE-2011-4108)
3017 [Robin Seggelmann, Michael Tuexen]
3019 *) Clear bytes used for block padding of SSL 3.0 records.
3021 [Adam Langley (Google)]
3023 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3024 Kadianakis <desnacked@gmail.com> for discovering this issue and
3025 Adam Langley for preparing the fix. (CVE-2011-4619)
3026 [Adam Langley (Google)]
3028 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3029 [Andrey Kulikov <amdeich@gmail.com>]
3031 *) Prevent malformed RFC3779 data triggering an assertion failure.
3032 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3033 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3034 [Rob Austein <sra@hactrn.net>]
3036 *) Improved PRNG seeding for VOS.
3037 [Paul Green <Paul.Green@stratus.com>]
3039 *) Fix ssl_ciph.c set-up race.
3040 [Adam Langley (Google)]
3042 *) Fix spurious failures in ecdsatest.c.
3043 [Emilia Käsper (Google)]
3045 *) Fix the BIO_f_buffer() implementation (which was mixing different
3046 interpretations of the '..._len' fields).
3047 [Adam Langley (Google)]
3049 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3050 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3051 threads won't reuse the same blinding coefficients.
3053 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3054 lock to call BN_BLINDING_invert_ex, and avoids one use of
3055 BN_BLINDING_update for each BN_BLINDING structure (previously,
3056 the last update always remained unused).
3057 [Emilia Käsper (Google)]
3059 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3060 [Bob Buckholz (Google)]
3062 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3064 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3065 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3066 [Kaspar Brand <ossl@velox.ch>]
3068 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3069 for multi-threaded use of ECDH. (CVE-2011-3210)
3070 [Adam Langley (Google)]
3072 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3075 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3076 signature public key algorithm by using OID xref utilities instead.
3077 Before this you could only use some ECC ciphersuites with SHA1 only.
3080 *) Add protection against ECDSA timing attacks as mentioned in the paper
3081 by Billy Bob Brumley and Nicola Tuveri, see:
3083 http://eprint.iacr.org/2011/232.pdf
3085 [Billy Bob Brumley and Nicola Tuveri]
3087 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3089 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3090 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3092 *) Fix bug in string printing code: if *any* escaping is enabled we must
3093 escape the escape character (backslash) or the resulting string is
3097 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3099 *) Disable code workaround for ancient and obsolete Netscape browsers
3100 and servers: an attacker can use it in a ciphersuite downgrade attack.
3101 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3104 *) Fixed J-PAKE implementation error, originally discovered by
3105 Sebastien Martini, further info and confirmation from Stefan
3106 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3109 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3111 *) Fix extension code to avoid race conditions which can result in a buffer
3112 overrun vulnerability: resumed sessions must not be modified as they can
3113 be shared by multiple threads. CVE-2010-3864
3116 *) Fix WIN32 build system to correctly link an ENGINE directory into
3120 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3122 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3124 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3126 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3128 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3129 context. The operation can be customised via the ctrl mechanism in
3130 case ENGINEs want to include additional functionality.
3133 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3136 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3137 output hashes compatible with older versions of OpenSSL.
3138 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3140 *) Fix compression algorithm handling: if resuming a session use the
3141 compression algorithm of the resumed session instead of determining
3142 it from client hello again. Don't allow server to change algorithm.
3145 *) Add load_crls() function to apps tidying load_certs() too. Add option
3146 to verify utility to allow additional CRLs to be included.
3149 *) Update OCSP request code to permit adding custom headers to the request:
3150 some responders need this.
3153 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3155 [Julia Lawall <julia@diku.dk>]
3157 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3158 needlessly dereferenced structures, used obsolete functions and
3159 didn't handle all updated verify codes correctly.
3162 *) Disable MD2 in the default configuration.
3165 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3166 indicate the initial BIO being pushed or popped. This makes it possible
3167 to determine whether the BIO is the one explicitly called or as a result
3168 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3169 it handles reference counts correctly and doesn't zero out the I/O bio
3170 when it is not being explicitly popped. WARNING: applications which
3171 included workarounds for the old buggy behaviour will need to be modified
3172 or they could free up already freed BIOs.
3175 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3176 renaming to all platforms (within the 0.9.8 branch, this was
3177 done conditionally on Netware platforms to avoid a name clash).
3178 [Guenter <lists@gknw.net>]
3180 *) Add ECDHE and PSK support to DTLS.
3181 [Michael Tuexen <tuexen@fh-muenster.de>]
3183 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3187 *) Add "missing" function EVP_MD_flags() (without this the only way to
3188 retrieve a digest flags is by accessing the structure directly. Update
3189 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3190 or cipher is registered as in the "from" argument. Print out all
3191 registered digests in the dgst usage message instead of manually
3192 attempting to work them out.
3195 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3196 this allows the use of compression and extensions. Change default cipher
3197 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3198 by default unless an application cipher string requests it.
3201 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3202 key ids to find matching certificates and keys but some PKCS#12 files
3203 don't follow the (somewhat unwritten) rules and this strategy fails.
3204 Now just gather all certificates together and the first private key
3205 then look for the first certificate that matches the key.
3208 *) Support use of registered digest and cipher names for dgst and cipher
3209 commands instead of having to add each one as a special case. So now
3216 openssl dgst -sha256 foo
3218 and this works for ENGINE based algorithms too.
3222 *) Update Gost ENGINE to support parameter files.
3223 [Victor B. Wagner <vitus@cryptocom.ru>]
3225 *) Support GeneralizedTime in ca utility.
3226 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3228 *) Enhance the hash format used for certificate directory links. The new
3229 form uses the canonical encoding (meaning equivalent names will work
3230 even if they aren't identical) and uses SHA1 instead of MD5. This form
3231 is incompatible with the older format and as a result c_rehash should
3232 be used to rebuild symbolic links.
3235 *) Make PKCS#8 the default write format for private keys, replacing the
3236 traditional format. This form is standardised, more secure and doesn't
3237 include an implicit MD5 dependency.
3240 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3241 committed to OpenSSL should pass this lot as a minimum.
3244 *) Add session ticket override functionality for use by EAP-FAST.
3245 [Jouni Malinen <j@w1.fi>]
3247 *) Modify HMAC functions to return a value. Since these can be implemented
3248 in an ENGINE errors can occur.
3251 *) Type-checked OBJ_bsearch_ex.
3254 *) Type-checked OBJ_bsearch. Also some constification necessitated
3255 by type-checking. Still to come: TXT_DB, bsearch(?),
3256 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3260 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3261 seconds to a tm structure directly, instead of going through OS
3262 specific date routines. This avoids any issues with OS routines such
3263 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3264 and X509_time_adj_ex() to cover the extended range. The existing
3265 X509_time_adj() is still usable and will no longer have any date issues.
3268 *) Delta CRL support. New use deltas option which will attempt to locate
3269 and search any appropriate delta CRLs available.
3271 This work was sponsored by Google.
3274 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3275 code and add additional score elements. Validate alternate CRL paths
3276 as part of the CRL checking and indicate a new error "CRL path validation
3277 error" in this case. Applications wanting additional details can use
3278 the verify callback and check the new "parent" field. If this is not
3279 NULL CRL path validation is taking place. Existing applications won't
3280 see this because it requires extended CRL support which is off by
3283 This work was sponsored by Google.
3286 *) Support for freshest CRL extension.
3288 This work was sponsored by Google.
3291 *) Initial indirect CRL support. Currently only supported in the CRLs
3292 passed directly and not via lookup. Process certificate issuer
3293 CRL entry extension and lookup CRL entries by bother issuer name
3294 and serial number. Check and process CRL issuer entry in IDP extension.
3296 This work was sponsored by Google.
3299 *) Add support for distinct certificate and CRL paths. The CRL issuer
3300 certificate is validated separately in this case. Only enabled if
3301 an extended CRL support flag is set: this flag will enable additional
3302 CRL functionality in future.
3304 This work was sponsored by Google.
3307 *) Add support for policy mappings extension.
3309 This work was sponsored by Google.
3312 *) Fixes to pathlength constraint, self issued certificate handling,
3313 policy processing to align with RFC3280 and PKITS tests.
3315 This work was sponsored by Google.
3318 *) Support for name constraints certificate extension. DN, email, DNS
3319 and URI types are currently supported.
3321 This work was sponsored by Google.
3324 *) To cater for systems that provide a pointer-based thread ID rather
3325 than numeric, deprecate the current numeric thread ID mechanism and
3326 replace it with a structure and associated callback type. This
3327 mechanism allows a numeric "hash" to be extracted from a thread ID in
3328 either case, and on platforms where pointers are larger than 'long',
3329 mixing is done to help ensure the numeric 'hash' is usable even if it
3330 can't be guaranteed unique. The default mechanism is to use "&errno"
3331 as a pointer-based thread ID to distinguish between threads.
3333 Applications that want to provide their own thread IDs should now use
3334 CRYPTO_THREADID_set_callback() to register a callback that will call
3335 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3337 Note that ERR_remove_state() is now deprecated, because it is tied
3338 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3339 to free the current thread's error state should be replaced by
3340 ERR_remove_thread_state(NULL).
3342 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3343 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3344 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3345 application was previously providing a numeric thread callback that
3346 was inappropriate for distinguishing threads, then uniqueness might
3347 have been obtained with &errno that happened immediately in the
3348 intermediate development versions of OpenSSL; this is no longer the
3349 case, the numeric thread callback will now override the automatic use
3351 [Geoff Thorpe, with help from Bodo Moeller]
3353 *) Initial support for different CRL issuing certificates. This covers a
3354 simple case where the self issued certificates in the chain exist and
3355 the real CRL issuer is higher in the existing chain.
3357 This work was sponsored by Google.
3360 *) Removed effectively defunct crypto/store from the build.
3363 *) Revamp of STACK to provide stronger type-checking. Still to come:
3364 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3365 ASN1_STRING, CONF_VALUE.
3368 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3369 RAM on SSL connections. This option can save about 34k per idle SSL.
3372 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3373 STACK, TXT_DB, bsearch, qsort.
3376 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3377 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3378 support for data, signedData, compressedData, digestedData and
3379 encryptedData, envelopedData types included. Scripts to check against
3380 RFC4134 examples draft and interop and consistency checks of many
3381 content types and variants.
3384 *) Add options to enc utility to support use of zlib compression BIO.
3387 *) Extend mk1mf to support importing of options and assembly language
3388 files from Configure script, currently only included in VC-WIN32.
3389 The assembly language rules can now optionally generate the source
3390 files from the associated perl scripts.
3393 *) Implement remaining functionality needed to support GOST ciphersuites.
3394 Interop testing has been performed using CryptoPro implementations.
3395 [Victor B. Wagner <vitus@cryptocom.ru>]
3397 *) s390x assembler pack.
3400 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3404 *) Implement Opaque PRF Input TLS extension as specified in
3405 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3406 official specification yet and no extension type assignment by
3407 IANA exists, this extension (for now) will have to be explicitly
3408 enabled when building OpenSSL by providing the extension number
3409 to use. For example, specify an option
3411 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3413 to the "config" or "Configure" script to enable the extension,
3414 assuming extension number 0x9527 (which is a completely arbitrary
3415 and unofficial assignment based on the MD5 hash of the Internet
3416 Draft). Note that by doing so, you potentially lose
3417 interoperability with other TLS implementations since these might
3418 be using the same extension number for other purposes.
3420 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3421 opaque PRF input value to use in the handshake. This will create
3422 an interal copy of the length-'len' string at 'src', and will
3423 return non-zero for success.
3425 To get more control and flexibility, provide a callback function
3428 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3429 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3433 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3436 Callback function 'cb' will be called in handshakes, and is
3437 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3438 Argument 'arg' is for application purposes (the value as given to
3439 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3440 be provided to the callback function). The callback function
3441 has to return non-zero to report success: usually 1 to use opaque
3442 PRF input just if possible, or 2 to enforce use of the opaque PRF
3443 input. In the latter case, the library will abort the handshake
3444 if opaque PRF input is not successfully negotiated.
3446 Arguments 'peerinput' and 'len' given to the callback function
3447 will always be NULL and 0 in the case of a client. A server will
3448 see the client's opaque PRF input through these variables if
3449 available (NULL and 0 otherwise). Note that if the server
3450 provides an opaque PRF input, the length must be the same as the
3451 length of the client's opaque PRF input.
3453 Note that the callback function will only be called when creating
3454 a new session (session resumption can resume whatever was
3455 previously negotiated), and will not be called in SSL 2.0
3456 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3457 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3458 for applications that need to enforce opaque PRF input.
3462 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3465 [Victor B. Wagner <vitus@cryptocom.ru>]
3467 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3468 RFC4507bis. The encrypted ticket format is an encrypted encoded
3469 SSL_SESSION structure, that way new session features are automatically
3472 If a client application caches session in an SSL_SESSION structure
3473 support is transparent because tickets are now stored in the encoded
3476 The SSL_CTX structure automatically generates keys for ticket
3477 protection in servers so again support should be possible
3478 with no application modification.
3480 If a client or server wishes to disable RFC4507 support then the option
3481 SSL_OP_NO_TICKET can be set.
3483 Add a TLS extension debugging callback to allow the contents of any client
3484 or server extensions to be examined.
3486 This work was sponsored by Google.
3489 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3490 OpenSSL should now compile cleanly on gcc 4.2
3491 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3493 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3494 support including streaming MAC support: this is required for GOST
3495 ciphersuite support.
3496 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3498 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3499 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3500 to output in BER and PEM format.
3503 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3504 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3505 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3506 ENGINE support for HMAC keys which are unextractable. New -mac and
3507 -macopt options to dgst utility.
3510 *) New option -sigopt to dgst utility. Update dgst to use
3511 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3512 alternative signing parameters such as X9.31 or PSS in the dgst
3516 *) Change ssl_cipher_apply_rule(), the internal function that does
3517 the work each time a ciphersuite string requests enabling
3518 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3519 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3520 the order of disabled ciphersuites such that those ciphersuites
3521 that most recently went from enabled to disabled not only stay
3522 in order with respect to each other, but also have higher priority
3523 than other disabled ciphersuites the next time ciphersuites are
3526 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3527 the same ciphersuites as with "HIGH" alone, but in a specific
3528 order where the PSK ciphersuites come first (since they are the
3529 most recently disabled ciphersuites when "HIGH" is parsed).
3531 Also, change ssl_create_cipher_list() (using this new
3532 funcionality) such that between otherwise identical
3533 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3537 *) Change ssl_create_cipher_list() so that it automatically
3538 arranges the ciphersuites in reasonable order before starting
3539 to process the rule string. Thus, the definition for "DEFAULT"
3540 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3541 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3542 This makes it much easier to arrive at a reasonable default order
3543 in applications for which anonymous ciphers are OK (meaning
3544 that you can't actually use DEFAULT).
3545 [Bodo Moeller; suggested by Victor Duchovni]
3547 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3548 processing) into multiple integers instead of setting
3549 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3550 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3551 (These masks as well as the individual bit definitions are hidden
3552 away into the non-exported interface ssl/ssl_locl.h, so this
3553 change to the definition of the SSL_CIPHER structure shouldn't
3554 affect applications.) This give us more bits for each of these
3555 categories, so there is no longer a need to coagulate AES128 and
3556 AES256 into a single algorithm bit, and to coagulate Camellia128
3557 and Camellia256 into a single algorithm bit, which has led to all
3560 Thus, among other things, the kludge introduced in 0.9.7m and
3561 0.9.8e for masking out AES256 independently of AES128 or masking
3562 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3564 With the change, we also introduce new ciphersuite aliases that
3565 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3569 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3570 Use the leftmost N bytes of the signature input if the input is
3571 larger than the prime q (with N being the size in bytes of q).
3574 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3575 it yet and it is largely untested.
3578 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3581 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3582 some compilers (gcc 4.2 and later) reject their use. Safestack is
3583 reimplemented. Update ASN1 to avoid use of legacy functions.
3586 *) Win32/64 targets are linked with Winsock2.
3589 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3590 to external functions. This can be used to increase CRL handling
3591 efficiency especially when CRLs are very large by (for example) storing
3592 the CRL revoked certificates in a database.
3595 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3596 new CRLs added to a directory can be used. New command line option
3597 -verify_return_error to s_client and s_server. This causes real errors
3598 to be returned by the verify callback instead of carrying on no matter
3599 what. This reflects the way a "real world" verify callback would behave.
3602 *) GOST engine, supporting several GOST algorithms and public key formats.
3603 Kindly donated by Cryptocom.
3606 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3607 partitioned by DP are handled but no indirect CRL or reason partitioning
3608 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3609 selected via a scoring technique which handles IDP and AKID in CRLs.
3612 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3613 will ultimately be used for all verify operations: this will remove the
3614 X509_STORE dependency on certificate verification and allow alternative
3615 lookup methods. X509_STORE based implementations of these two callbacks.
3618 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3619 Modify get_crl() to find a valid (unexpired) CRL if possible.
3622 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3623 this would be called X509_CRL_cmp() but that name is already used by
3624 a function that just compares CRL issuer names. Cache several CRL
3625 extensions in X509_CRL structure and cache CRLDP in X509.
3628 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3629 this maps equivalent X509_NAME structures into a consistent structure.
3630 Name comparison can then be performed rapidly using memcmp().
3633 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3637 *) Allow digests to supply their own micalg string for S/MIME type using
3638 the ctrl EVP_MD_CTRL_MICALG.
3641 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3642 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3643 ctrl. It can then customise the structure before and/or after signing
3647 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3648 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3649 to free up any added signature OIDs.
3652 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3653 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3654 digest and cipher tables. New options added to openssl utility:
3655 list-message-digest-algorithms and list-cipher-algorithms.
3658 *) Change the array representation of binary polynomials: the list
3659 of degrees of non-zero coefficients is now terminated with -1.
3660 Previously it was terminated with 0, which was also part of the
3661 value; thus, the array representation was not applicable to
3662 polynomials where t^0 has coefficient zero. This change makes
3663 the array representation useful in a more general context.
3666 *) Various modifications and fixes to SSL/TLS cipher string
3667 handling. For ECC, the code now distinguishes between fixed ECDH
3668 with RSA certificates on the one hand and with ECDSA certificates
3669 on the other hand, since these are separate ciphersuites. The
3670 unused code for Fortezza ciphersuites has been removed.
3672 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3673 (not "ECDHE"). For consistency with the code for DH
3674 certificates, use of ECDH certificates is now considered ECDH
3675 authentication, not RSA or ECDSA authentication (the latter is
3676 merely the CA's signing algorithm and not actively used in the
3679 The temporary ciphersuite alias "ECCdraft" is no longer
3680 available, and ECC ciphersuites are no longer excluded from "ALL"
3681 and "DEFAULT". The following aliases now exist for RFC 4492
3682 ciphersuites, most of these by analogy with the DH case:
3684 kECDHr - ECDH cert, signed with RSA
3685 kECDHe - ECDH cert, signed with ECDSA
3686 kECDH - ECDH cert (signed with either RSA or ECDSA)
3687 kEECDH - ephemeral ECDH
3688 ECDH - ECDH cert or ephemeral ECDH
3694 AECDH - anonymous ECDH
3695 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3699 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3700 Use correct micalg parameters depending on digest(s) in signed message.
3703 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3704 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3707 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3708 an engine to register a method. Add ENGINE lookups for methods and
3709 functional reference processing.
3712 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3713 EVP_{Sign,Verify}* which allow an application to customise the signature
3717 *) New -resign option to smime utility. This adds one or more signers
3718 to an existing PKCS#7 signedData structure. Also -md option to use an
3719 alternative message digest algorithm for signing.
3722 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3723 create PKCS7 structures containing multiple signers. Update smime
3724 application to support multiple signers.
3727 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3731 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3732 Reorganize PBE internals to lookup from a static table using NIDs,
3733 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3734 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3735 PRF which will be automatically used with PBES2.
3738 *) Replace the algorithm specific calls to generate keys in "req" with the
3742 *) Update PKCS#7 enveloped data routines to use new API. This is now
3743 supported by any public key method supporting the encrypt operation. A
3744 ctrl is added to allow the public key algorithm to examine or modify
3745 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3749 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3750 a default digest type to use. In most cases this will be SHA1 but some
3751 algorithms (such as GOST) need to specify an alternative digest. The
3752 return value indicates how strong the preference is 1 means optional and
3753 2 is mandatory (that is it is the only supported type). Modify
3754 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3755 use the default md. Update openssl utilities to use the default digest
3756 type for signing if it is not explicitly indicated.
3759 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3760 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3761 signing method from the key type. This effectively removes the link
3762 between digests and public key types.
3765 *) Add an OID cross reference table and utility functions. Its purpose is to
3766 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3767 rsaEncryption. This will allow some of the algorithm specific hackery
3768 needed to use the correct OID to be removed.
3771 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3772 structures for PKCS7_sign(). They are now set up by the relevant public
3776 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3779 *) Add support for key derivation (agreement) in the API, DH method and
3783 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3784 public and private key formats. As a side effect these add additional
3785 command line functionality not previously available: DSA signatures can be
3786 generated and verified using pkeyutl and DH key support and generation in
3791 [Oliver Tappe <zooey@hirschkaefer.de>]
3793 *) New make target "install_html_docs" installs HTML renditions of the
3795 [Oliver Tappe <zooey@hirschkaefer.de>]
3797 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3798 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3799 support key and parameter generation and add initial key generation
3800 functionality for RSA.
3803 *) Add functions for main EVP_PKEY_method operations. The undocumented
3804 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3805 EVP_PKEY_{encrypt,decrypt}_old.
3808 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3809 key API, doesn't do much yet.
3812 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3813 public key algorithms. New option to openssl utility:
3814 "list-public-key-algorithms" to print out info.
3817 *) Implement the Supported Elliptic Curves Extension for
3818 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3821 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3822 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3825 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3826 utilities such as rsa, dsa, dsaparam etc except they process any key
3830 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3831 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3832 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3836 *) Initial support for pluggable public key ASN1.
3837 De-spaghettify the public key ASN1 handling. Move public and private
3838 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3839 algorithm specific handling to a single module within the relevant
3840 algorithm directory. Add functions to allow (near) opaque processing
3841 of public and private key structures.
3844 *) Implement the Supported Point Formats Extension for
3845 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3848 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3849 for the psk identity [hint] and the psk callback functions to the
3850 SSL_SESSION, SSL and SSL_CTX structure.
3853 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3857 SSL_CTX_use_psk_identity_hint
3858 SSL_get_psk_identity_hint
3859 SSL_get_psk_identity
3860 SSL_use_psk_identity_hint
3862 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3864 *) Add RFC 3161 compliant time stamp request creation, response generation
3865 and response verification functionality.
3866 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3868 *) Add initial support for TLS extensions, specifically for the server_name
3869 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3870 have new members for a host name. The SSL data structure has an
3871 additional member SSL_CTX *initial_ctx so that new sessions can be
3872 stored in that context to allow for session resumption, even after the
3873 SSL has been switched to a new SSL_CTX in reaction to a client's
3874 server_name extension.
3876 New functions (subject to change):
3878 SSL_get_servername()
3879 SSL_get_servername_type()
3882 New CTRL codes and macros (subject to change):
3884 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3885 - SSL_CTX_set_tlsext_servername_callback()
3886 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3887 - SSL_CTX_set_tlsext_servername_arg()
3888 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3890 openssl s_client has a new '-servername ...' option.
3892 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3893 '-key2 ...', '-servername_fatal' (subject to change). This allows
3894 testing the HostName extension for a specific single host name ('-cert'
3895 and '-key' remain fallbacks for handshakes without HostName
3896 negotiation). If the unrecognized_name alert has to be sent, this by
3897 default is a warning; it becomes fatal with the '-servername_fatal'
3900 [Peter Sylvester, Remy Allais, Christophe Renou]
3902 *) Whirlpool hash implementation is added.
3905 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3906 bn(64,32). Because of instruction set limitations it doesn't have
3907 any negative impact on performance. This was done mostly in order
3908 to make it possible to share assembler modules, such as bn_mul_mont
3909 implementations, between 32- and 64-bit builds without hassle.
3912 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3913 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3917 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3918 dedicated Montgomery multiplication procedure, is introduced.
3919 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3920 "64-bit" performance on certain 32-bit targets.
3923 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3924 in SSL structures. New SSL ctrl to set maximum send fragment size.
3925 Save memory by seeting the I/O buffer sizes dynamically instead of
3926 using the maximum available value.
3929 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3930 in addition to the text details.
3933 *) Very, very preliminary EXPERIMENTAL support for printing of general
3934 ASN1 structures. This currently produces rather ugly output and doesn't
3935 handle several customised structures at all.
3938 *) Integrated support for PVK file format and some related formats such
3939 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3940 these in the 'rsa' and 'dsa' utilities.
3943 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3946 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3947 place for the (very old) "NETSCAPE" format certificates which are now
3948 handled using new ASN1 code equivalents.
3951 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3952 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3953 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3956 *) Modify CRL distribution points extension code to print out previously
3957 unsupported fields. Enhance extension setting code to allow setting of
3961 *) Add print and set support for Issuing Distribution Point CRL extension.
3964 *) Change 'Configure' script to enable Camellia by default.
3967 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3969 *) When rejecting SSL/TLS records due to an incorrect version number, never
3970 update s->server with a new major version number. As of
3971 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3972 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3973 the previous behavior could result in a read attempt at NULL when
3974 receiving specific incorrect SSL/TLS records once record payload
3975 protection is active. (CVE-2010-0740)
3976 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3978 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3979 could be crashed if the relevant tables were not present (e.g. chrooted).
3980 [Tomas Hoger <thoger@redhat.com>]
3982 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3984 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3985 [Martin Olsson, Neel Mehta]
3987 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3988 accommodate for stack sorting, always a write lock!).
3991 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3992 excessive delays in the RAND_poll(): over a minute. As a workaround
3993 include a time check in the inner Heap32Next loop too.
3996 *) The code that handled flushing of data in SSL/TLS originally used the
3997 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3998 the problem outlined in PR#1949. The fix suggested there however can
3999 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4000 of Apache). So instead simplify the code to flush unconditionally.
4001 This should be fine since flushing with no data to flush is a no op.
4004 *) Handle TLS versions 2.0 and later properly and correctly use the
4005 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4006 off ancient servers have a habit of sticking around for a while...
4009 *) Modify compression code so it frees up structures without using the
4010 ex_data callbacks. This works around a problem where some applications
4011 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4012 restarting) then use compression (e.g. SSL with compression) later.
4013 This results in significant per-connection memory leaks and
4014 has caused some security issues including CVE-2008-1678 and
4018 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4019 change when encrypting or decrypting.
4022 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4023 connect and renegotiate with servers which do not support RI.
4024 Until RI is more widely deployed this option is enabled by default.
4027 *) Add "missing" ssl ctrls to clear options and mode.
4030 *) If client attempts to renegotiate and doesn't support RI respond with
4031 a no_renegotiation alert as required by RFC5746. Some renegotiating
4032 TLS clients will continue a connection gracefully when they receive
4033 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4034 waiting for a server hello which it will never receive. Now we treat a
4035 received no_renegotiation alert as a fatal error. This is because
4036 applications requesting a renegotiation might well expect it to succeed
4037 and would have no code in place to handle the server denying it so the
4038 only safe thing to do is to terminate the connection.
4041 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4042 peer supports secure renegotiation and 0 otherwise. Print out peer
4043 renegotiation support in s_client/s_server.
4046 *) Replace the highly broken and deprecated SPKAC certification method with
4047 the updated NID creation version. This should correctly handle UTF8.
4050 *) Implement RFC5746. Re-enable renegotiation but require the extension
4051 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4052 turns out to be a bad idea. It has been replaced by
4053 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4054 SSL_CTX_set_options(). This is really not recommended unless you
4055 know what you are doing.
4056 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4058 *) Fixes to stateless session resumption handling. Use initial_ctx when
4059 issuing and attempting to decrypt tickets in case it has changed during
4060 servername handling. Use a non-zero length session ID when attempting
4061 stateless session resumption: this makes it possible to determine if
4062 a resumption has occurred immediately after receiving server hello
4063 (several places in OpenSSL subtly assume this) instead of later in
4067 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4068 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4069 fixes for a few places where the return code is not checked
4071 [Julia Lawall <julia@diku.dk>]
4073 *) Add --strict-warnings option to Configure script to include devteam
4074 warnings in other configurations.
4077 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4078 makes it possible to install openssl libraries in locations which
4079 have names other than "lib", for example "/usr/lib64" which some
4081 [Steve Henson, based on patch from Jeremy Utley]
4083 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4084 X690 8.9.12 and can produce some misleading textual output of OIDs.
4085 [Steve Henson, reported by Dan Kaminsky]
4087 *) Delete MD2 from algorithm tables. This follows the recommendation in
4088 several standards that it is not used in new applications due to
4089 several cryptographic weaknesses. For binary compatibility reasons
4090 the MD2 API is still compiled in by default.
4093 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4097 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4098 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4100 [Guenter <lists@gknw.net>]
4102 *) Fix the server certificate chain building code to use X509_verify_cert(),
4103 it used to have an ad-hoc builder which was unable to cope with anything
4104 other than a simple chain.
4105 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4107 *) Don't check self signed certificate signatures in X509_verify_cert()
4108 by default (a flag can override this): it just wastes time without
4109 adding any security. As a useful side effect self signed root CAs
4110 with non-FIPS digests are now usable in FIPS mode.
4113 *) In dtls1_process_out_of_seq_message() the check if the current message
4114 is already buffered was missing. For every new message was memory
4115 allocated, allowing an attacker to perform an denial of service attack
4116 with sending out of seq handshake messages until there is no memory
4117 left. Additionally every future messege was buffered, even if the
4118 sequence number made no sense and would be part of another handshake.
4119 So only messages with sequence numbers less than 10 in advance will be
4120 buffered. (CVE-2009-1378)
4121 [Robin Seggelmann, discovered by Daniel Mentz]
4123 *) Records are buffered if they arrive with a future epoch to be
4124 processed after finishing the corresponding handshake. There is
4125 currently no limitation to this buffer allowing an attacker to perform
4126 a DOS attack with sending records with future epochs until there is no
4127 memory left. This patch adds the pqueue_size() function to determine
4128 the size of a buffer and limits the record buffer to 100 entries.
4130 [Robin Seggelmann, discovered by Daniel Mentz]
4132 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4133 parent structure is freed. (CVE-2009-1379)
4136 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4137 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4140 [Ilya O. <vrghost@gmail.com>]
4142 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4144 *) Disable renegotiation completely - this fixes a severe security
4145 problem (CVE-2009-3555) at the cost of breaking all
4146 renegotiation. Renegotiation can be re-enabled by setting
4147 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4148 run-time. This is really not recommended unless you know what
4152 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4154 *) Don't set val to NULL when freeing up structures, it is freed up by
4155 underlying code. If sizeof(void *) > sizeof(long) this can result in
4156 zeroing past the valid field. (CVE-2009-0789)
4157 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4159 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4160 checked correctly. This would allow some invalid signed attributes to
4161 appear to verify correctly. (CVE-2009-0591)
4162 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4164 *) Reject UniversalString and BMPString types with invalid lengths. This
4165 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4166 a legal length. (CVE-2009-0590)
4169 *) Set S/MIME signing as the default purpose rather than setting it
4170 unconditionally. This allows applications to override it at the store
4174 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4175 to handle some structures.
4178 *) Improve efficiency of mem_gets: don't search whole buffer each time
4180 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4182 *) New -hex option for openssl rand.
4185 *) Print out UTF8String and NumericString when parsing ASN1.
4188 *) Support NumericString type for name components.
4191 *) Allow CC in the environment to override the automatically chosen
4192 compiler. Note that nothing is done to ensure flags work with the
4196 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4198 *) Properly check EVP_VerifyFinal() and similar return values
4200 [Ben Laurie, Bodo Moeller, Google Security Team]
4202 *) Enable TLS extensions by default.
4205 *) Allow the CHIL engine to be loaded, whether the application is
4206 multithreaded or not. (This does not release the developer from the
4207 obligation to set up the dynamic locking callbacks.)
4208 [Sander Temme <sander@temme.net>]
4210 *) Use correct exit code if there is an error in dgst command.
4211 [Steve Henson; problem pointed out by Roland Dirlewanger]
4213 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4214 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4217 *) Add experimental JPAKE support, including demo authentication in
4218 s_client and s_server.
4221 *) Set the comparison function in v3_addr_canonize().
4222 [Rob Austein <sra@hactrn.net>]
4224 *) Add support for XMPP STARTTLS in s_client.
4225 [Philip Paeps <philip@freebsd.org>]
4227 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4228 to ensure that even with this option, only ciphersuites in the
4229 server's preference list will be accepted. (Note that the option
4230 applies only when resuming a session, so the earlier behavior was
4231 just about the algorithm choice for symmetric cryptography.)
4234 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4236 *) Fix NULL pointer dereference if a DTLS server received
4237 ChangeCipherSpec as first record (CVE-2009-1386).
4240 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4241 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4244 *) The fix in 0.9.8c that supposedly got rid of unsafe
4245 double-checked locking was incomplete for RSA blinding,
4246 addressing just one layer of what turns out to have been
4247 doubly unsafe triple-checked locking.
4249 So now fix this for real by retiring the MONT_HELPER macro
4250 in crypto/rsa/rsa_eay.c.
4252 [Bodo Moeller; problem pointed out by Marius Schilder]
4254 *) Various precautionary measures:
4256 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4258 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4259 (NB: This would require knowledge of the secret session ticket key
4260 to exploit, in which case you'd be SOL either way.)
4262 - Change bn_nist.c so that it will properly handle input BIGNUMs
4263 outside the expected range.
4265 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4268 [Neel Mehta, Bodo Moeller]
4270 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4271 the load fails. Useful for distros.
4272 [Ben Laurie and the FreeBSD team]
4274 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4277 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4280 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4282 This work was sponsored by Logica.
4285 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4286 keystores. Support for SSL/TLS client authentication too.
4287 Not compiled unless enable-capieng specified to Configure.
4289 This work was sponsored by Logica.
4292 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4293 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4294 attribute creation routines such as certificate requests and PKCS#12
4298 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4300 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4301 handshake which could lead to a cilent crash as found using the
4302 Codenomicon TLS test suite (CVE-2008-1672)
4303 [Steve Henson, Mark Cox]
4305 *) Fix double free in TLS server name extensions which could lead to
4306 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4309 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4311 Clear the error queue to ensure that error entries left from
4312 older function calls do not interfere with the correct operation.
4313 [Lutz Jaenicke, Erik de Castro Lopo]
4315 *) Remove root CA certificates of commercial CAs:
4317 The OpenSSL project does not recommend any specific CA and does not
4318 have any policy with respect to including or excluding any CA.
4319 Therefore it does not make any sense to ship an arbitrary selection
4320 of root CA certificates with the OpenSSL software.
4323 *) RSA OAEP patches to fix two separate invalid memory reads.
4324 The first one involves inputs when 'lzero' is greater than
4325 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4326 before the beginning of from). The second one involves inputs where
4327 the 'db' section contains nothing but zeroes (there is a one-byte
4328 invalid read after the end of 'db').
4329 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4331 *) Partial backport from 0.9.9-dev:
4333 Introduce bn_mul_mont (dedicated Montgomery multiplication
4334 procedure) as a candidate for BIGNUM assembler implementation.
4335 While 0.9.9-dev uses assembler for various architectures, only
4336 x86_64 is available by default here in the 0.9.8 branch, and
4337 32-bit x86 is available through a compile-time setting.
4339 To try the 32-bit x86 assembler implementation, use Configure
4340 option "enable-montasm" (which exists only for this backport).
4342 As "enable-montasm" for 32-bit x86 disclaims code stability
4343 anyway, in this constellation we activate additional code
4344 backported from 0.9.9-dev for further performance improvements,
4345 namely BN_from_montgomery_word. (To enable this otherwise,
4346 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4348 [Andy Polyakov (backport partially by Bodo Moeller)]
4350 *) Add TLS session ticket callback. This allows an application to set
4351 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4352 values. This is useful for key rollover for example where several key
4353 sets may exist with different names.
4356 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4357 This was broken until now in 0.9.8 releases, such that the only way
4358 a registered ENGINE could be used (assuming it initialises
4359 successfully on the host) was to explicitly set it as the default
4360 for the relevant algorithms. This is in contradiction with 0.9.7
4361 behaviour and the documentation. With this fix, when an ENGINE is
4362 registered into a given algorithm's table of implementations, the
4363 'uptodate' flag is reset so that auto-discovery will be used next
4364 time a new context for that algorithm attempts to select an
4366 [Ian Lister (tweaked by Geoff Thorpe)]
4368 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4369 implementation in the following ways:
4371 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4374 Lack of BER streaming support means one pass streaming processing is
4375 only supported if data is detached: setting the streaming flag is
4376 ignored for embedded content.
4378 CMS support is disabled by default and must be explicitly enabled
4379 with the enable-cms configuration option.
4382 *) Update the GMP engine glue to do direct copies between BIGNUM and
4383 mpz_t when openssl and GMP use the same limb size. Otherwise the
4384 existing "conversion via a text string export" trick is still used.
4385 [Paul Sheer <paulsheer@gmail.com>]
4387 *) Zlib compression BIO. This is a filter BIO which compressed and
4388 uncompresses any data passed through it.
4391 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4392 RFC3394 compatible AES key wrapping.
4395 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4396 sets string data without copying. X509_ALGOR_set0() and
4397 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4398 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4399 from an X509_ATTRIBUTE structure optionally checking it occurs only
4400 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4404 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4405 to get the expected BN_FLG_CONSTTIME behavior.
4406 [Bodo Moeller (Google)]
4410 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4411 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4412 - added some more tests to do_tests.pl
4413 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4414 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4415 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4416 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4417 - various changes to netware.pl to enable gcc-cross builds on Win32
4419 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4420 - various changes to fix missing prototype warnings
4421 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4422 - added AES, WHIRLPOOL and CPUID assembler code to build files
4423 - added missing AES assembler make rules to mk1mf.pl
4424 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4425 [Guenter Knauf <eflash@gmx.net>]
4427 *) Implement certificate status request TLS extension defined in RFC3546.
4428 A client can set the appropriate parameters and receive the encoded
4429 OCSP response via a callback. A server can query the supplied parameters
4430 and set the encoded OCSP response in the callback. Add simplified examples
4431 to s_client and s_server.
4434 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4436 *) Fix various bugs:
4437 + Binary incompatibility of ssl_ctx_st structure
4438 + DTLS interoperation with non-compliant servers
4439 + Don't call get_session_cb() without proposed session
4440 + Fix ia64 assembler code
4441 [Andy Polyakov, Steve Henson]
4443 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4445 *) DTLS Handshake overhaul. There were longstanding issues with
4446 OpenSSL DTLS implementation, which were making it impossible for
4447 RFC 4347 compliant client to communicate with OpenSSL server.
4448 Unfortunately just fixing these incompatibilities would "cut off"
4449 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4450 server keeps tolerating non RFC compliant syntax. The opposite is
4451 not true, 0.9.8f client can not communicate with earlier server.
4452 This update even addresses CVE-2007-4995.
4455 *) Changes to avoid need for function casts in OpenSSL: some compilers
4456 (gcc 4.2 and later) reject their use.
4457 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4460 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4461 RFC4507bis. The encrypted ticket format is an encrypted encoded
4462 SSL_SESSION structure, that way new session features are automatically
4465 If a client application caches session in an SSL_SESSION structure
4466 support is transparent because tickets are now stored in the encoded
4469 The SSL_CTX structure automatically generates keys for ticket
4470 protection in servers so again support should be possible
4471 with no application modification.
4473 If a client or server wishes to disable RFC4507 support then the option
4474 SSL_OP_NO_TICKET can be set.
4476 Add a TLS extension debugging callback to allow the contents of any client
4477 or server extensions to be examined.
4479 This work was sponsored by Google.
4482 *) Add initial support for TLS extensions, specifically for the server_name
4483 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4484 have new members for a host name. The SSL data structure has an
4485 additional member SSL_CTX *initial_ctx so that new sessions can be
4486 stored in that context to allow for session resumption, even after the
4487 SSL has been switched to a new SSL_CTX in reaction to a client's
4488 server_name extension.
4490 New functions (subject to change):
4492 SSL_get_servername()
4493 SSL_get_servername_type()
4496 New CTRL codes and macros (subject to change):
4498 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4499 - SSL_CTX_set_tlsext_servername_callback()
4500 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4501 - SSL_CTX_set_tlsext_servername_arg()
4502 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4504 openssl s_client has a new '-servername ...' option.
4506 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4507 '-key2 ...', '-servername_fatal' (subject to change). This allows
4508 testing the HostName extension for a specific single host name ('-cert'
4509 and '-key' remain fallbacks for handshakes without HostName
4510 negotiation). If the unrecognized_name alert has to be sent, this by
4511 default is a warning; it becomes fatal with the '-servername_fatal'
4514 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4516 *) Add AES and SSE2 assembly language support to VC++ build.
4519 *) Mitigate attack on final subtraction in Montgomery reduction.
4522 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4523 (which previously caused an internal error).
4526 *) Squeeze another 10% out of IGE mode when in != out.
4529 *) AES IGE mode speedup.
4530 [Dean Gaudet (Google)]
4532 *) Add the Korean symmetric 128-bit cipher SEED (see
4533 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4534 add SEED ciphersuites from RFC 4162:
4536 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4537 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4538 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4539 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4541 To minimize changes between patchlevels in the OpenSSL 0.9.8
4542 series, SEED remains excluded from compilation unless OpenSSL
4543 is configured with 'enable-seed'.
4544 [KISA, Bodo Moeller]
4546 *) Mitigate branch prediction attacks, which can be practical if a
4547 single processor is shared, allowing a spy process to extract
4548 information. For detailed background information, see
4549 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4550 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4551 and Necessary Software Countermeasures"). The core of the change
4552 are new versions BN_div_no_branch() and
4553 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4554 respectively, which are slower, but avoid the security-relevant
4555 conditional branches. These are automatically called by BN_div()
4556 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4557 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4558 remove a conditional branch.
4560 BN_FLG_CONSTTIME is the new name for the previous
4561 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4562 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4563 in the exponent causes BN_mod_exp_mont() to use the alternative
4564 implementation in BN_mod_exp_mont_consttime().) The old name
4565 remains as a deprecated alias.
4567 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4568 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4569 constant-time implementations for more than just exponentiation.
4570 Here too the old name is kept as a deprecated alias.
4572 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4573 the BN_BLINDING structure gets an independent copy of the
4574 modulus. This means that the previous "BIGNUM *m" argument to
4575 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4576 essentially becomes "const BIGNUM *m", although we can't actually
4577 change this in the header file before 0.9.9. It allows
4578 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4579 enable BN_FLG_CONSTTIME.
4581 [Matthew D Wood (Intel Corp)]
4583 *) In the SSL/TLS server implementation, be strict about session ID
4584 context matching (which matters if an application uses a single
4585 external cache for different purposes). Previously,
4586 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4587 set. This did ensure strict client verification, but meant that,
4588 with applications using a single external cache for quite
4589 different requirements, clients could circumvent ciphersuite
4590 restrictions for a given session ID context by starting a session
4591 in a different context.
4594 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4595 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4596 authentication-only ciphersuites.
4599 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4600 not complete and could lead to a possible single byte overflow
4601 (CVE-2007-5135) [Ben Laurie]
4603 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4605 *) Since AES128 and AES256 (and similarly Camellia128 and
4606 Camellia256) share a single mask bit in the logic of
4607 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4608 kludge to work properly if AES128 is available and AES256 isn't
4609 (or if Camellia128 is available and Camellia256 isn't).
4612 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4613 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4614 When a point or a seed is encoded in a BIT STRING, we need to
4615 prevent the removal of trailing zero bits to get the proper DER
4616 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4617 of a NamedBitList, for which trailing 0 bits need to be removed.)
4620 *) Have SSL/TLS server implementation tolerate "mismatched" record
4621 protocol version while receiving ClientHello even if the
4622 ClientHello is fragmented. (The server can't insist on the
4623 particular protocol version it has chosen before the ServerHello
4624 message has informed the client about his choice.)
4627 *) Add RFC 3779 support.
4628 [Rob Austein for ARIN, Ben Laurie]
4630 *) Load error codes if they are not already present instead of using a
4631 static variable. This allows them to be cleanly unloaded and reloaded.
4632 Improve header file function name parsing.
4635 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4636 or CAPABILITY handshake as required by RFCs.
4639 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4641 *) Introduce limits to prevent malicious keys being able to
4642 cause a denial of service. (CVE-2006-2940)
4643 [Steve Henson, Bodo Moeller]
4645 *) Fix ASN.1 parsing of certain invalid structures that can result
4646 in a denial of service. (CVE-2006-2937) [Steve Henson]
4648 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4649 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4651 *) Fix SSL client code which could crash if connecting to a
4652 malicious SSLv2 server. (CVE-2006-4343)
4653 [Tavis Ormandy and Will Drewry, Google Security Team]
4655 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4656 match only those. Before that, "AES256-SHA" would be interpreted
4657 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4658 the same strength classification in 0.9.7h) as we currently only
4659 have a single AES bit in the ciphersuite description bitmap.
4660 That change, however, also applied to ciphersuite strings such as
4661 "RC4-MD5" that intentionally matched multiple ciphersuites --
4662 namely, SSL 2.0 ciphersuites in addition to the more common ones
4663 from SSL 3.0/TLS 1.0.
4665 So we change the selection algorithm again: Naming an explicit
4666 ciphersuite selects this one ciphersuite, and any other similar
4667 ciphersuite (same bitmap) from *other* protocol versions.
4668 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4669 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4671 Since SSL 2.0 does not have any ciphersuites for which the
4672 128/256 bit distinction would be relevant, this works for now.
4673 The proper fix will be to use different bits for AES128 and
4674 AES256, which would have avoided the problems from the beginning;
4675 however, bits are scarce, so we can only do this in a new release
4676 (not just a patchlevel) when we can change the SSL_CIPHER
4677 definition to split the single 'unsigned long mask' bitmap into
4678 multiple values to extend the available space.
4682 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4684 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4685 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4687 *) Add AES IGE and biIGE modes.
4690 *) Change the Unix randomness entropy gathering to use poll() when
4691 possible instead of select(), since the latter has some
4692 undesirable limitations.
4693 [Darryl Miles via Richard Levitte and Bodo Moeller]
4695 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4696 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4697 cannot be implicitly activated as part of, e.g., the "AES" alias.
4698 However, please upgrade to OpenSSL 0.9.9[-dev] for
4699 non-experimental use of the ECC ciphersuites to get TLS extension
4700 support, which is required for curve and point format negotiation
4701 to avoid potential handshake problems.
4704 *) Disable rogue ciphersuites:
4706 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4707 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4708 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4710 The latter two were purportedly from
4711 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4714 Also deactivate the remaining ciphersuites from
4715 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4716 unofficial, and the ID has long expired.
4719 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4720 dual-core machines) and other potential thread-safety issues.
4723 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4724 versions), which is now available for royalty-free use
4725 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4726 Also, add Camellia TLS ciphersuites from RFC 4132.
4728 To minimize changes between patchlevels in the OpenSSL 0.9.8
4729 series, Camellia remains excluded from compilation unless OpenSSL
4730 is configured with 'enable-camellia'.
4733 *) Disable the padding bug check when compression is in use. The padding
4734 bug check assumes the first packet is of even length, this is not
4735 necessarily true if compresssion is enabled and can result in false
4736 positives causing handshake failure. The actual bug test is ancient
4737 code so it is hoped that implementations will either have fixed it by
4738 now or any which still have the bug do not support compression.
4741 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4743 *) When applying a cipher rule check to see if string match is an explicit
4744 cipher suite and only match that one cipher suite if it is.
4747 *) Link in manifests for VC++ if needed.
4748 [Austin Ziegler <halostatue@gmail.com>]
4750 *) Update support for ECC-based TLS ciphersuites according to
4751 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4752 TLS extensions, which are supported starting with the 0.9.9
4753 branch, not in the OpenSSL 0.9.8 branch).
4756 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4757 opaque EVP_CIPHER_CTX handling.
4760 *) Fixes and enhancements to zlib compression code. We now only use
4761 "zlib1.dll" and use the default __cdecl calling convention on Win32
4762 to conform with the standards mentioned here:
4763 http://www.zlib.net/DLL_FAQ.txt
4764 Static zlib linking now works on Windows and the new --with-zlib-include
4765 --with-zlib-lib options to Configure can be used to supply the location
4766 of the headers and library. Gracefully handle case where zlib library
4770 *) Several fixes and enhancements to the OID generation code. The old code
4771 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4772 handle numbers larger than ULONG_MAX, truncated printing and had a
4773 non standard OBJ_obj2txt() behaviour.
4776 *) Add support for building of engines under engine/ as shared libraries
4777 under VC++ build system.
4780 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4781 Hopefully, we will not see any false combination of paths any more.
4784 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4786 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4787 (part of SSL_OP_ALL). This option used to disable the
4788 countermeasure against man-in-the-middle protocol-version
4789 rollback in the SSL 2.0 server implementation, which is a bad
4790 idea. (CVE-2005-2969)
4792 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4793 for Information Security, National Institute of Advanced Industrial
4794 Science and Technology [AIST], Japan)]
4796 *) Add two function to clear and return the verify parameter flags.
4799 *) Keep cipherlists sorted in the source instead of sorting them at
4800 runtime, thus removing the need for a lock.
4803 *) Avoid some small subgroup attacks in Diffie-Hellman.
4804 [Nick Mathewson and Ben Laurie]
4806 *) Add functions for well-known primes.
4809 *) Extended Windows CE support.
4810 [Satoshi Nakamura and Andy Polyakov]
4812 *) Initialize SSL_METHOD structures at compile time instead of during
4813 runtime, thus removing the need for a lock.
4816 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4817 attempting to decrypt each encrypted key in turn. Add support to
4821 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4823 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4826 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4829 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4830 key into the same file any more.
4833 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4836 *) Add -utf8 command line and config file option to 'ca'.
4837 [Stefan <stf@udoma.org]
4839 *) Removed the macro des_crypt(), as it seems to conflict with some
4840 libraries. Use DES_crypt().
4843 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4844 involves renaming the source and generated shared-libs for
4845 both. The engines will accept the corrected or legacy ids
4846 ('ncipher' and '4758_cca' respectively) when binding. NB,
4847 this only applies when building 'shared'.
4848 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4850 *) Add attribute functions to EVP_PKEY structure. Modify
4851 PKCS12_create() to recognize a CSP name attribute and
4852 use it. Make -CSP option work again in pkcs12 utility.
4855 *) Add new functionality to the bn blinding code:
4856 - automatic re-creation of the BN_BLINDING parameters after
4857 a fixed number of uses (currently 32)
4858 - add new function for parameter creation
4859 - introduce flags to control the update behaviour of the
4860 BN_BLINDING parameters
4861 - hide BN_BLINDING structure
4862 Add a second BN_BLINDING slot to the RSA structure to improve
4863 performance when a single RSA object is shared among several
4867 *) Add support for DTLS.
4868 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4870 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4871 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4874 *) Remove buggy and incomplete DH cert support from
4875 ssl/ssl_rsa.c and ssl/s3_both.c
4878 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4879 the apps/openssl applications.
4882 *) Compile clean with "-Wall -Wmissing-prototypes
4883 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4884 DEBUG_SAFESTACK must also be set.
4887 *) Change ./Configure so that certain algorithms can be disabled by default.
4888 The new counterpiece to "no-xxx" is "enable-xxx".
4890 The patented RC5 and MDC2 algorithms will now be disabled unless
4891 "enable-rc5" and "enable-mdc2", respectively, are specified.
4893 (IDEA remains enabled despite being patented. This is because IDEA
4894 is frequently required for interoperability, and there is no license
4895 fee for non-commercial use. As before, "no-idea" can be used to
4896 avoid this algorithm.)
4900 *) Add processing of proxy certificates (see RFC 3820). This work was
4901 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4902 EGEE (Enabling Grids for E-science in Europe).
4905 *) RC4 performance overhaul on modern architectures/implementations, such
4906 as Intel P4, IA-64 and AMD64.
4909 *) New utility extract-section.pl. This can be used specify an alternative
4910 section number in a pod file instead of having to treat each file as
4911 a separate case in Makefile. This can be done by adding two lines to the
4914 =for comment openssl_section:XXX
4916 The blank line is mandatory.
4920 *) New arguments -certform, -keyform and -pass for s_client and s_server
4921 to allow alternative format key and certificate files and passphrase
4925 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4926 update associated structures and add various utility functions.
4928 Add new policy related verify parameters, include policy checking in
4929 standard verify code. Enhance 'smime' application with extra parameters
4930 to support policy checking and print out.
4933 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4934 Nehemiah processors. These extensions support AES encryption in hardware
4935 as well as RNG (though RNG support is currently disabled).
4936 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4938 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4941 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4942 [Andy Polyakov and a number of other people]
4944 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4945 implementation contributed by IBM.
4946 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4948 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4949 exponent rather than 'unsigned long'. There is a corresponding change to
4950 the new 'rsa_keygen' element of the RSA_METHOD structure.
4951 [Jelte Jansen, Geoff Thorpe]
4953 *) Functionality for creating the initial serial number file is now
4954 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4956 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4957 number file to 1, which is bound to cause problems. To avoid
4958 the problems while respecting compatibility between different 0.9.7
4959 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4960 CA.pl for serial number initialization. With the new release 0.9.8,
4961 we can fix the problem directly in the 'ca' utility.)
4964 *) Reduced header interdepencies by declaring more opaque objects in
4965 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4966 give fewer recursive includes, which could break lazy source code - so
4967 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4968 developers should define this symbol when building and using openssl to
4969 ensure they track the recommended behaviour, interfaces, [etc], but
4970 backwards-compatible behaviour prevails when this isn't defined.
4973 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4976 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4977 This will generate a random key of the appropriate length based on the
4978 cipher context. The EVP_CIPHER can provide its own random key generation
4979 routine to support keys of a specific form. This is used in the des and
4980 3des routines to generate a key of the correct parity. Update S/MIME
4981 code to use new functions and hence generate correct parity DES keys.
4982 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4983 valid (weak or incorrect parity).
4986 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4987 as looking them up. This is useful when the verified structure may contain
4988 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4989 present unless the new PKCS7_NO_CRL flag is asserted.
4992 *) Extend ASN1 oid configuration module. It now additionally accepts the
4995 shortName = some long name, 1.2.3.4
4998 *) Reimplemented the BN_CTX implementation. There is now no more static
4999 limitation on the number of variables it can handle nor the depth of the
5000 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5001 information can now expand as required, and rather than having a single
5002 static array of bignums, BN_CTX now uses a linked-list of such arrays
5003 allowing it to expand on demand whilst maintaining the usefulness of
5004 BN_CTX's "bundling".
5007 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5008 to allow all RSA operations to function using a single BN_CTX.
5011 *) Preliminary support for certificate policy evaluation and checking. This
5012 is initially intended to pass the tests outlined in "Conformance Testing
5013 of Relying Party Client Certificate Path Processing Logic" v1.07.
5016 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5017 remained unused and not that useful. A variety of other little bignum
5018 tweaks and fixes have also been made continuing on from the audit (see
5022 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5023 associated ASN1, EVP and SSL functions and old ASN1 macros.
5026 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5027 and this should never fail. So the return value from the use of
5028 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5029 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5032 *) BN_CTX_get() should return zero-valued bignums, providing the same
5033 initialised value as BN_new().
5034 [Geoff Thorpe, suggested by Ulf Möller]
5036 *) Support for inhibitAnyPolicy certificate extension.
5039 *) An audit of the BIGNUM code is underway, for which debugging code is
5040 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5041 is considered valid when processing BIGNUMs, and causes execution to
5042 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5043 further steps are taken to deliberately pollute unused data in BIGNUM
5044 structures to try and expose faulty code further on. For now, openssl will
5045 (in its default mode of operation) continue to tolerate the inconsistent
5046 forms that it has tolerated in the past, but authors and packagers should
5047 consider trying openssl and their own applications when compiled with
5048 these debugging symbols defined. It will help highlight potential bugs in
5049 their own code, and will improve the test coverage for OpenSSL itself. At
5050 some point, these tighter rules will become openssl's default to improve
5051 maintainability, though the assert()s and other overheads will remain only
5052 in debugging configurations. See bn.h for more details.
5053 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5055 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5056 that can only be obtained through BN_CTX_new() (which implicitly
5057 initialises it). The presence of this function only made it possible
5058 to overwrite an existing structure (and cause memory leaks).
5061 *) Because of the callback-based approach for implementing LHASH as a
5062 template type, lh_insert() adds opaque objects to hash-tables and
5063 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5064 to clean up those corresponding objects before destroying the hash table
5065 (and losing the object pointers). So some over-zealous constifications in
5066 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5067 objects as "const" and the lh_doall[_arg] callback wrappers are not
5068 prototyped to have "const" restrictions on the object pointers they are
5069 given (and so aren't required to cast them away any more).
5072 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5073 (speed) prefers to use its own implementation. The two implementations
5074 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5075 its object type properly exposed (MS_TM) instead of casting to/from "char
5076 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5077 aren't necessarily the greatest nomenclatures - but this is what was used
5078 internally to the implementation so I've used that for now.
5081 *) Ensure that deprecated functions do not get compiled when
5082 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5083 the self-tests were still using deprecated key-generation functions so
5084 these have been updated also.
5087 *) Reorganise PKCS#7 code to separate the digest location functionality
5088 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5089 New function PKCS7_set_digest() to set the digest type for PKCS#7
5090 digestedData type. Add additional code to correctly generate the
5091 digestedData type and add support for this type in PKCS7 initialization
5095 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5096 structure of type "other".
5099 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5100 sure the loop does correctly stop and breaking ("division by zero")
5101 modulus operations are not performed. The (pre-generated) prime
5102 table crypto/bn/bn_prime.h was already correct, but it could not be
5103 re-generated on some platforms because of the "division by zero"
5104 situation in the script.
5105 [Ralf S. Engelschall]
5107 *) Update support for ECC-based TLS ciphersuites according to
5108 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5109 SHA-1 now is only used for "small" curves (where the
5110 representation of a field element takes up to 24 bytes); for
5111 larger curves, the field element resulting from ECDH is directly
5112 used as premaster secret.
5113 [Douglas Stebila (Sun Microsystems Laboratories)]
5115 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5116 curve secp160r1 to the tests.
5117 [Douglas Stebila (Sun Microsystems Laboratories)]
5119 *) Add the possibility to load symbols globally with DSO.
5120 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5122 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5123 control of the error stack.
5126 *) Add support for STORE in ENGINE.
5129 *) Add the STORE type. The intention is to provide a common interface
5130 to certificate and key stores, be they simple file-based stores, or
5131 HSM-type store, or LDAP stores, or...
5132 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5135 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5136 pass a list of arguments to any function as well as provide a way
5137 for a function to pass data back to the caller.
5140 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5141 works like BUF_strdup() but can be used to duplicate a portion of
5142 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5146 *) Add the function sk_find_ex() which works like sk_find(), but will
5147 return an index to an element even if an exact match couldn't be
5148 found. The index is guaranteed to point at the element where the
5149 searched-for key would be inserted to preserve sorting order.
5152 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5153 takes an extra flags argument for optional functionality. Currently,
5154 the following flags are defined:
5156 OBJ_BSEARCH_VALUE_ON_NOMATCH
5157 This one gets OBJ_bsearch_ex() to return a pointer to the first
5158 element where the comparing function returns a negative or zero
5161 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5162 This one gets OBJ_bsearch_ex() to return a pointer to the first
5163 element where the comparing function returns zero. This is useful
5164 if there are more than one element where the comparing function
5168 *) Make it possible to create self-signed certificates with 'openssl ca'
5169 in such a way that the self-signed certificate becomes part of the
5170 CA database and uses the same mechanisms for serial number generation
5171 as all other certificate signing. The new flag '-selfsign' enables
5172 this functionality. Adapt CA.sh and CA.pl.in.
5175 *) Add functionality to check the public key of a certificate request
5176 against a given private. This is useful to check that a certificate
5177 request can be signed by that key (self-signing).
5180 *) Make it possible to have multiple active certificates with the same
5181 subject in the CA index file. This is done only if the keyword
5182 'unique_subject' is set to 'no' in the main CA section (default
5183 if 'CA_default') of the configuration file. The value is saved
5184 with the database itself in a separate index attribute file,
5185 named like the index file with '.attr' appended to the name.
5188 *) Generate muti valued AVAs using '+' notation in config files for
5192 *) Support for nameConstraints certificate extension.
5195 *) Support for policyConstraints certificate extension.
5198 *) Support for policyMappings certificate extension.
5201 *) Make sure the default DSA_METHOD implementation only uses its
5202 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5203 and change its own handlers to be NULL so as to remove unnecessary
5204 indirection. This lets alternative implementations fallback to the
5205 default implementation more easily.
5208 *) Support for directoryName in GeneralName related extensions
5212 *) Make it possible to link applications using Makefile.shared.
5213 Make that possible even when linking against static libraries!
5216 *) Support for single pass processing for S/MIME signing. This now
5217 means that S/MIME signing can be done from a pipe, in addition
5218 cleartext signing (multipart/signed type) is effectively streaming
5219 and the signed data does not need to be all held in memory.
5221 This is done with a new flag PKCS7_STREAM. When this flag is set
5222 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5223 is done after the data is output (and digests calculated) in
5224 SMIME_write_PKCS7().
5227 *) Add full support for -rpath/-R, both in shared libraries and
5228 applications, at least on the platforms where it's known how
5232 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5233 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5234 will now compute a table of multiples of the generator that
5235 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5236 faster (notably in the case of a single point multiplication,
5237 scalar * generator).
5238 [Nils Larsch, Bodo Moeller]
5240 *) IPv6 support for certificate extensions. The various extensions
5241 which use the IP:a.b.c.d can now take IPv6 addresses using the
5242 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5246 *) Added an ENGINE that implements RSA by performing private key
5247 exponentiations with the GMP library. The conversions to and from
5248 GMP's mpz_t format aren't optimised nor are any montgomery forms
5249 cached, and on x86 it appears OpenSSL's own performance has caught up.
5250 However there are likely to be other architectures where GMP could
5251 provide a boost. This ENGINE is not built in by default, but it can be
5252 specified at Configure time and should be accompanied by the necessary
5253 linker additions, eg;
5254 ./config -DOPENSSL_USE_GMP -lgmp
5257 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5258 testing availability of engines with "-t" - the old behaviour is
5259 produced by increasing the feature's verbosity with "-tt".
5262 *) ECDSA routines: under certain error conditions uninitialized BN objects
5263 could be freed. Solution: make sure initialization is performed early
5264 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5268 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5269 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5270 software implementations. For DSA and DH, parameter generation can
5271 also be overridden by providing the appropriate method callbacks.
5274 *) Change the "progress" mechanism used in key-generation and
5275 primality testing to functions that take a new BN_GENCB pointer in
5276 place of callback/argument pairs. The new API functions have "_ex"
5277 postfixes and the older functions are reimplemented as wrappers for
5278 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5279 declarations of the old functions to help (graceful) attempts to
5280 migrate to the new functions. Also, the new key-generation API
5281 functions operate on a caller-supplied key-structure and return
5282 success/failure rather than returning a key or NULL - this is to
5283 help make "keygen" another member function of RSA_METHOD etc.
5285 Example for using the new callback interface:
5287 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5291 BN_GENCB_set(&my_cb, my_callback, my_arg);
5293 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5294 /* For the meaning of a, b in calls to my_callback(), see the
5295 * documentation of the function that calls the callback.
5296 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5297 * my_callback should return 1 if it wants BN_is_prime_ex()
5298 * to continue, or 0 to stop.
5303 *) Change the ZLIB compression method to be stateful, and make it
5304 available to TLS with the number defined in
5305 draft-ietf-tls-compression-04.txt.
5308 *) Add the ASN.1 structures and functions for CertificatePair, which
5309 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5311 CertificatePair ::= SEQUENCE {
5312 forward [0] Certificate OPTIONAL,
5313 reverse [1] Certificate OPTIONAL,
5314 -- at least one of the pair shall be present -- }
5316 Also implement the PEM functions to read and write certificate
5317 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5319 This needed to be defined, mostly for the sake of the LDAP
5320 attribute crossCertificatePair, but may prove useful elsewhere as
5324 *) Make it possible to inhibit symlinking of shared libraries in
5325 Makefile.shared, for Cygwin's sake.
5328 *) Extend the BIGNUM API by creating a function
5329 void BN_set_negative(BIGNUM *a, int neg);
5330 and a macro that behave like
5331 int BN_is_negative(const BIGNUM *a);
5333 to avoid the need to access 'a->neg' directly in applications.
5336 *) Implement fast modular reduction for pseudo-Mersenne primes
5337 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5338 EC_GROUP_new_curve_GFp() will now automatically use this
5340 [Nils Larsch <nla@trustcenter.de>]
5342 *) Add new lock type (CRYPTO_LOCK_BN).
5345 *) Change the ENGINE framework to automatically load engines
5346 dynamically from specific directories unless they could be
5347 found to already be built in or loaded. Move all the
5348 current engines except for the cryptodev one to a new
5350 The engines in engines/ are built as shared libraries if
5351 the "shared" options was given to ./Configure or ./config.
5352 Otherwise, they are inserted in libcrypto.a.
5353 /usr/local/ssl/engines is the default directory for dynamic
5354 engines, but that can be overridden at configure time through
5355 the usual use of --prefix and/or --openssldir, and at run
5356 time with the environment variable OPENSSL_ENGINES.
5357 [Geoff Thorpe and Richard Levitte]
5359 *) Add Makefile.shared, a helper makefile to build shared
5360 libraries. Adapt Makefile.org.
5363 *) Add version info to Win32 DLLs.
5364 [Peter 'Luna' Runestig" <peter@runestig.com>]
5366 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5367 can be added using this API to created arbitrary PKCS#12
5368 files while avoiding the low level API.
5370 New options to PKCS12_create(), key or cert can be NULL and
5371 will then be omitted from the output file. The encryption
5372 algorithm NIDs can be set to -1 for no encryption, the mac
5373 iteration count can be set to 0 to omit the mac.
5375 Enhance pkcs12 utility by making the -nokeys and -nocerts
5376 options work when creating a PKCS#12 file. New option -nomac
5377 to omit the mac, NONE can be set for an encryption algorithm.
5378 New code is modified to use the enhanced PKCS12_create()
5379 instead of the low level API.
5382 *) Extend ASN1 encoder to support indefinite length constructed
5383 encoding. This can output sequences tags and octet strings in
5384 this form. Modify pk7_asn1.c to support indefinite length
5385 encoding. This is experimental and needs additional code to
5386 be useful, such as an ASN1 bio and some enhanced streaming
5389 Extend template encode functionality so that tagging is passed
5390 down to the template encoder.
5393 *) Let 'openssl req' fail if an argument to '-newkey' is not
5394 recognized instead of using RSA as a default.
5397 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5398 As these are not official, they are not included in "ALL";
5399 the "ECCdraft" ciphersuite group alias can be used to select them.
5400 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5402 *) Add ECDH engine support.
5403 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5405 *) Add ECDH in new directory crypto/ecdh/.
5406 [Douglas Stebila (Sun Microsystems Laboratories)]
5408 *) Let BN_rand_range() abort with an error after 100 iterations
5409 without success (which indicates a broken PRNG).
5412 *) Change BN_mod_sqrt() so that it verifies that the input value
5413 is really the square of the return value. (Previously,
5414 BN_mod_sqrt would show GIGO behaviour.)
5417 *) Add named elliptic curves over binary fields from X9.62, SECG,
5418 and WAP/WTLS; add OIDs that were still missing.
5420 [Sheueling Chang Shantz and Douglas Stebila
5421 (Sun Microsystems Laboratories)]
5423 *) Extend the EC library for elliptic curves over binary fields
5424 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5427 EC_GF2m_simple_method
5431 EC_GROUP_new_curve_GF2m
5432 EC_GROUP_set_curve_GF2m
5433 EC_GROUP_get_curve_GF2m
5434 EC_POINT_set_affine_coordinates_GF2m
5435 EC_POINT_get_affine_coordinates_GF2m
5436 EC_POINT_set_compressed_coordinates_GF2m
5438 Point compression for binary fields is disabled by default for
5439 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5442 As binary polynomials are represented as BIGNUMs, various members
5443 of the EC_GROUP and EC_POINT data structures can be shared
5444 between the implementations for prime fields and binary fields;
5445 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5446 are essentially identical to their ..._GFp counterparts.
5447 (For simplicity, the '..._GFp' prefix has been dropped from
5448 various internal method names.)
5450 An internal 'field_div' method (similar to 'field_mul' and
5451 'field_sqr') has been added; this is used only for binary fields.
5453 [Sheueling Chang Shantz and Douglas Stebila
5454 (Sun Microsystems Laboratories)]
5456 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5457 through methods ('mul', 'precompute_mult').
5459 The generic implementations (now internally called 'ec_wNAF_mul'
5460 and 'ec_wNAF_precomputed_mult') remain the default if these
5461 methods are undefined.
5463 [Sheueling Chang Shantz and Douglas Stebila
5464 (Sun Microsystems Laboratories)]
5466 *) New function EC_GROUP_get_degree, which is defined through
5467 EC_METHOD. For curves over prime fields, this returns the bit
5468 length of the modulus.
5470 [Sheueling Chang Shantz and Douglas Stebila
5471 (Sun Microsystems Laboratories)]
5473 *) New functions EC_GROUP_dup, EC_POINT_dup.
5474 (These simply call ..._new and ..._copy).
5476 [Sheueling Chang Shantz and Douglas Stebila
5477 (Sun Microsystems Laboratories)]
5479 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5480 Polynomials are represented as BIGNUMs (where the sign bit is not
5481 used) in the following functions [macros]:
5484 BN_GF2m_sub [= BN_GF2m_add]
5485 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5486 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5487 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5489 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5490 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5491 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5492 BN_GF2m_cmp [= BN_ucmp]
5494 (Note that only the 'mod' functions are actually for fields GF(2^m).
5495 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5497 For some functions, an the irreducible polynomial defining a
5498 field can be given as an 'unsigned int[]' with strictly
5499 decreasing elements giving the indices of those bits that are set;
5500 i.e., p[] represents the polynomial
5501 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5503 p[0] > p[1] > ... > p[k] = 0.
5504 This applies to the following functions:
5509 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5510 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5512 BN_GF2m_mod_sqrt_arr
5513 BN_GF2m_mod_solve_quad_arr
5517 Conversion can be performed by the following functions:
5522 bntest.c has additional tests for binary polynomial arithmetic.
5524 Two implementations for BN_GF2m_mod_div() are available.
5525 The default algorithm simply uses BN_GF2m_mod_inv() and
5526 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5527 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5528 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5530 [Sheueling Chang Shantz and Douglas Stebila
5531 (Sun Microsystems Laboratories)]
5533 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5534 functionality is disabled at compile-time.
5535 [Douglas Stebila <douglas.stebila@sun.com>]
5537 *) Change default behaviour of 'openssl asn1parse' so that more
5538 information is visible when viewing, e.g., a certificate:
5540 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5541 mode the content of non-printable OCTET STRINGs is output in a
5542 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5543 avoid the appearance of a printable string.
5544 [Nils Larsch <nla@trustcenter.de>]
5546 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5548 EC_GROUP_set_asn1_flag()
5549 EC_GROUP_get_asn1_flag()
5550 EC_GROUP_set_point_conversion_form()
5551 EC_GROUP_get_point_conversion_form()
5552 These control ASN1 encoding details:
5553 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5554 has been set to OPENSSL_EC_NAMED_CURVE.
5555 - Points are encoded in uncompressed form by default; options for
5556 asn1_for are as for point2oct, namely
5557 POINT_CONVERSION_COMPRESSED
5558 POINT_CONVERSION_UNCOMPRESSED
5559 POINT_CONVERSION_HYBRID
5561 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5564 EC_GROUP_get0_seed()
5565 EC_GROUP_get_seed_len()
5566 This is used only for ASN1 purposes (so far).
5567 [Nils Larsch <nla@trustcenter.de>]
5569 *) Add 'field_type' member to EC_METHOD, which holds the NID
5570 of the appropriate field type OID. The new function
5571 EC_METHOD_get_field_type() returns this value.
5572 [Nils Larsch <nla@trustcenter.de>]
5577 EC_POINT_point2hex()
5578 EC_POINT_hex2point()
5579 providing useful interfaces to EC_POINT_point2oct() and
5580 EC_POINT_oct2point().
5581 [Nils Larsch <nla@trustcenter.de>]
5583 *) Change internals of the EC library so that the functions
5584 EC_GROUP_set_generator()
5585 EC_GROUP_get_generator()
5586 EC_GROUP_get_order()
5587 EC_GROUP_get_cofactor()
5588 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5589 to methods, which would lead to unnecessary code duplication when
5590 adding different types of curves.
5591 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5593 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5594 arithmetic, and such that modified wNAFs are generated
5595 (which avoid length expansion in many cases).
5598 *) Add a function EC_GROUP_check_discriminant() (defined via
5599 EC_METHOD) that verifies that the curve discriminant is non-zero.
5601 Add a function EC_GROUP_check() that makes some sanity tests
5602 on a EC_GROUP, its generator and order. This includes
5603 EC_GROUP_check_discriminant().
5604 [Nils Larsch <nla@trustcenter.de>]
5606 *) Add ECDSA in new directory crypto/ecdsa/.
5608 Add applications 'openssl ecparam' and 'openssl ecdsa'
5609 (these are based on 'openssl dsaparam' and 'openssl dsa').
5611 ECDSA support is also included in various other files across the
5612 library. Most notably,
5613 - 'openssl req' now has a '-newkey ecdsa:file' option;
5614 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5615 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5616 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5617 them suitable for ECDSA where domain parameters must be
5618 extracted before the specific public key;
5619 - ECDSA engine support has been added.
5620 [Nils Larsch <nla@trustcenter.de>]
5622 *) Include some named elliptic curves, and add OIDs from X9.62,
5623 SECG, and WAP/WTLS. Each curve can be obtained from the new
5625 EC_GROUP_new_by_curve_name(),
5626 and the list of available named curves can be obtained with
5627 EC_get_builtin_curves().
5628 Also add a 'curve_name' member to EC_GROUP objects, which can be
5630 EC_GROUP_set_curve_name()
5631 EC_GROUP_get_curve_name()
5632 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5634 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5635 was actually never needed) and in BN_mul(). The removal in BN_mul()
5636 required a small change in bn_mul_part_recursive() and the addition
5637 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5638 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5639 bn_sub_words() and bn_add_words() except they take arrays with
5643 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5645 *) Cleanse PEM buffers before freeing them since they may contain
5647 [Benjamin Bennett <ben@psc.edu>]
5649 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5650 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5651 authentication-only ciphersuites.
5654 *) Since AES128 and AES256 share a single mask bit in the logic of
5655 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5656 kludge to work properly if AES128 is available and AES256 isn't.
5659 *) Expand security boundary to match 1.1.1 module.
5662 *) Remove redundant features: hash file source, editing of test vectors
5663 modify fipsld to use external fips_premain.c signature.
5666 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5667 run algorithm test programs.
5670 *) Make algorithm test programs more tolerant of whitespace.
5673 *) Have SSL/TLS server implementation tolerate "mismatched" record
5674 protocol version while receiving ClientHello even if the
5675 ClientHello is fragmented. (The server can't insist on the
5676 particular protocol version it has chosen before the ServerHello
5677 message has informed the client about his choice.)
5680 *) Load error codes if they are not already present instead of using a
5681 static variable. This allows them to be cleanly unloaded and reloaded.
5684 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5686 *) Introduce limits to prevent malicious keys being able to
5687 cause a denial of service. (CVE-2006-2940)
5688 [Steve Henson, Bodo Moeller]
5690 *) Fix ASN.1 parsing of certain invalid structures that can result
5691 in a denial of service. (CVE-2006-2937) [Steve Henson]
5693 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5694 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5696 *) Fix SSL client code which could crash if connecting to a
5697 malicious SSLv2 server. (CVE-2006-4343)
5698 [Tavis Ormandy and Will Drewry, Google Security Team]
5700 *) Change ciphersuite string processing so that an explicit
5701 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5702 will no longer include "AES128-SHA"), and any other similar
5703 ciphersuite (same bitmap) from *other* protocol versions (so that
5704 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5705 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5706 changes from 0.9.8b and 0.9.8d.
5709 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5711 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5712 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5714 *) Change the Unix randomness entropy gathering to use poll() when
5715 possible instead of select(), since the latter has some
5716 undesirable limitations.
5717 [Darryl Miles via Richard Levitte and Bodo Moeller]
5719 *) Disable rogue ciphersuites:
5721 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5722 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5723 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5725 The latter two were purportedly from
5726 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5729 Also deactive the remaining ciphersuites from
5730 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5731 unofficial, and the ID has long expired.
5734 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5735 dual-core machines) and other potential thread-safety issues.
5738 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5740 *) Adapt fipsld and the build system to link against the validated FIPS
5741 module in FIPS mode.
5744 *) Fixes for VC++ 2005 build under Windows.
5747 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5748 from a Windows bash shell such as MSYS. It is autodetected from the
5749 "config" script when run from a VC++ environment. Modify standard VC++
5750 build to use fipscanister.o from the GNU make build.
5753 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5755 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5756 The value now differs depending on if you build for FIPS or not.
5757 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5758 safely run with a non-FIPSed libcrypto, as it may crash because of
5759 the difference induced by this change.
5762 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5764 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5765 (part of SSL_OP_ALL). This option used to disable the
5766 countermeasure against man-in-the-middle protocol-version
5767 rollback in the SSL 2.0 server implementation, which is a bad
5768 idea. (CVE-2005-2969)
5770 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5771 for Information Security, National Institute of Advanced Industrial
5772 Science and Technology [AIST], Japan)]
5774 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5775 mainly for FIPS compliance and not fully integrated at this stage.
5778 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5779 the exponentiation using a fixed-length exponent. (Otherwise,
5780 the information leaked through timing could expose the secret key
5781 after many signatures; cf. Bleichenbacher's attack on DSA with
5785 *) Make a new fixed-window mod_exp implementation the default for
5786 RSA, DSA, and DH private-key operations so that the sequence of
5787 squares and multiplies and the memory access pattern are
5788 independent of the particular secret key. This will mitigate
5789 cache-timing and potential related attacks.
5791 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5792 and this is automatically used by BN_mod_exp_mont() if the new flag
5793 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5794 will use this BN flag for private exponents unless the flag
5795 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5796 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5798 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5800 *) Change the client implementation for SSLv23_method() and
5801 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5802 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5803 (Previously, the SSL 2.0 backwards compatible Client Hello
5804 message format would be used even with SSL_OP_NO_SSLv2.)
5807 *) Add support for smime-type MIME parameter in S/MIME messages which some
5811 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5812 a threadsafe manner. Modify rsa code to use new function and add calls
5813 to dsa and dh code (which had race conditions before).
5816 *) Include the fixed error library code in the C error file definitions
5817 instead of fixing them up at runtime. This keeps the error code
5818 structures constant.
5821 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5823 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5826 *) Fixes for newer kerberos headers. NB: the casts are needed because
5827 the 'length' field is signed on one version and unsigned on another
5828 with no (?) obvious way to tell the difference, without these VC++
5829 complains. Also the "definition" of FAR (blank) is no longer included
5830 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5831 some needed definitions.
5834 *) Undo Cygwin change.
5837 *) Added support for proxy certificates according to RFC 3820.
5838 Because they may be a security thread to unaware applications,
5839 they must be explicitly allowed in run-time. See
5840 docs/HOWTO/proxy_certificates.txt for further information.
5843 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5845 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5846 server and client random values. Previously
5847 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5848 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5850 This change has negligible security impact because:
5852 1. Server and client random values still have 24 bytes of pseudo random
5855 2. Server and client random values are sent in the clear in the initial
5858 3. The master secret is derived using the premaster secret (48 bytes in
5859 size for static RSA ciphersuites) as well as client server and random
5862 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5865 [Stephen Henson, reported by UK NISCC]
5867 *) Use Windows randomness collection on Cygwin.
5870 *) Fix hang in EGD/PRNGD query when communication socket is closed
5871 prematurely by EGD/PRNGD.
5872 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5874 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5877 *) Back-port of selected performance improvements from development
5878 branch, as well as improved support for PowerPC platforms.
5881 *) Add lots of checks for memory allocation failure, error codes to indicate
5882 failure and freeing up memory if a failure occurs.
5883 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5885 *) Add new -passin argument to dgst.
5888 *) Perform some character comparisons of different types in X509_NAME_cmp:
5889 this is needed for some certificates that re-encode DNs into UTF8Strings
5890 (in violation of RFC3280) and can't or won't issue name rollover
5894 *) Make an explicit check during certificate validation to see that
5895 the CA setting in each certificate on the chain is correct. As a
5896 side effect always do the following basic checks on extensions,
5897 not just when there's an associated purpose to the check:
5899 - if there is an unhandled critical extension (unless the user
5900 has chosen to ignore this fault)
5901 - if the path length has been exceeded (if one is set at all)
5902 - that certain extensions fit the associated purpose (if one has
5906 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5908 *) Avoid a race condition when CRLs are checked in a multi threaded
5909 environment. This would happen due to the reordering of the revoked
5910 entries during signature checking and serial number lookup. Now the
5911 encoding is cached and the serial number sort performed under a lock.
5912 Add new STACK function sk_is_sorted().
5915 *) Add Delta CRL to the extension code.
5918 *) Various fixes to s3_pkt.c so alerts are sent properly.
5919 [David Holmes <d.holmes@f5.com>]
5921 *) Reduce the chances of duplicate issuer name and serial numbers (in
5922 violation of RFC3280) using the OpenSSL certificate creation utilities.
5923 This is done by creating a random 64 bit value for the initial serial
5924 number when a serial number file is created or when a self signed
5925 certificate is created using 'openssl req -x509'. The initial serial
5926 number file is created using 'openssl x509 -next_serial' in CA.pl
5927 rather than being initialized to 1.
5930 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5932 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5933 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5934 [Joe Orton, Steve Henson]
5936 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5938 [Joe Orton, Steve Henson]
5940 *) Make it possible to have multiple active certificates with the same
5941 subject in the CA index file. This is done only if the keyword
5942 'unique_subject' is set to 'no' in the main CA section (default
5943 if 'CA_default') of the configuration file. The value is saved
5944 with the database itself in a separate index attribute file,
5945 named like the index file with '.attr' appended to the name.
5948 *) X509 verify fixes. Disable broken certificate workarounds when
5949 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5950 keyUsage extension present. Don't accept CRLs with unhandled critical
5951 extensions: since verify currently doesn't process CRL extensions this
5952 rejects a CRL with *any* critical extensions. Add new verify error codes
5956 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5957 A clarification of RFC2560 will require the use of OCTET STRINGs and
5958 some implementations cannot handle the current raw format. Since OpenSSL
5959 copies and compares OCSP nonces as opaque blobs without any attempt at
5960 parsing them this should not create any compatibility issues.
5963 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5964 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5965 this HMAC (and other) operations are several times slower than OpenSSL
5969 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5970 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5972 *) Use the correct content when signing type "other".
5975 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5977 *) Fix various bugs revealed by running the NISCC test suite:
5979 Stop out of bounds reads in the ASN1 code when presented with
5980 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5982 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5984 If verify callback ignores invalid public key errors don't try to check
5985 certificate signature with the NULL public key.
5989 *) New -ignore_err option in ocsp application to stop the server
5990 exiting on the first error in a request.
5993 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5994 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5998 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5999 extra data after the compression methods not only for TLS 1.0
6000 but also for SSL 3.0 (as required by the specification).
6001 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6003 *) Change X509_certificate_type() to mark the key as exported/exportable
6004 when it's 512 *bits* long, not 512 bytes.
6007 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6008 blocks during encryption.
6011 *) Various fixes to base64 BIO and non blocking I/O. On write
6012 flushes were not handled properly if the BIO retried. On read
6013 data was not being buffered properly and had various logic bugs.
6014 This also affects blocking I/O when the data being decoded is a
6018 *) Various S/MIME bugfixes and compatibility changes:
6019 output correct application/pkcs7 MIME type if
6020 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6021 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6022 of files as .eml work). Correctly handle very long lines in MIME
6026 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6028 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6029 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6030 a protocol version number mismatch like a decryption error
6031 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6034 *) Turn on RSA blinding by default in the default implementation
6035 to avoid a timing attack. Applications that don't want it can call
6036 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6037 They would be ill-advised to do so in most cases.
6038 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6040 *) Change RSA blinding code so that it works when the PRNG is not
6041 seeded (in this case, the secret RSA exponent is abused as
6042 an unpredictable seed -- if it is not unpredictable, there
6043 is no point in blinding anyway). Make RSA blinding thread-safe
6044 by remembering the creator's thread ID in rsa->blinding and
6045 having all other threads use local one-time blinding factors
6046 (this requires more computation than sharing rsa->blinding, but
6047 avoids excessive locking; and if an RSA object is not shared
6048 between threads, blinding will still be very fast).
6051 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6052 ENGINE as defaults for all supported algorithms irrespective of
6053 the 'flags' parameter. 'flags' is now honoured, so applications
6054 should make sure they are passing it correctly.
6057 *) Target "mingw" now allows native Windows code to be generated in
6058 the Cygwin environment as well as with the MinGW compiler.
6061 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6063 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6064 via timing by performing a MAC computation even if incorrect
6065 block cipher padding has been found. This is a countermeasure
6066 against active attacks where the attacker has to distinguish
6067 between bad padding and a MAC verification error. (CVE-2003-0078)
6069 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6070 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6071 Martin Vuagnoux (EPFL, Ilion)]
6073 *) Make the no-err option work as intended. The intention with no-err
6074 is not to have the whole error stack handling routines removed from
6075 libcrypto, it's only intended to remove all the function name and
6076 reason texts, thereby removing some of the footprint that may not
6077 be interesting if those errors aren't displayed anyway.
6079 NOTE: it's still possible for any application or module to have it's
6080 own set of error texts inserted. The routines are there, just not
6081 used by default when no-err is given.
6084 *) Add support for FreeBSD on IA64.
6085 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6087 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6088 Kerberos function mit_des_cbc_cksum(). Before this change,
6089 the value returned by DES_cbc_cksum() was like the one from
6090 mit_des_cbc_cksum(), except the bytes were swapped.
6091 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6093 *) Allow an application to disable the automatic SSL chain building.
6094 Before this a rather primitive chain build was always performed in
6095 ssl3_output_cert_chain(): an application had no way to send the
6096 correct chain if the automatic operation produced an incorrect result.
6098 Now the chain builder is disabled if either:
6100 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6102 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6104 The reasoning behind this is that an application would not want the
6105 auto chain building to take place if extra chain certificates are
6106 present and it might also want a means of sending no additional
6107 certificates (for example the chain has two certificates and the
6111 *) Add the possibility to build without the ENGINE framework.
6112 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6114 *) Under Win32 gmtime() can return NULL: check return value in
6115 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6118 *) DSA routines: under certain error conditions uninitialized BN objects
6119 could be freed. Solution: make sure initialization is performed early
6120 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6121 Nils Larsch <nla@trustcenter.de> via PR#459)
6124 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6125 checked on reconnect on the client side, therefore session resumption
6126 could still fail with a "ssl session id is different" error. This
6127 behaviour is masked when SSL_OP_ALL is used due to
6128 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6129 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6130 followup to PR #377.
6133 *) IA-32 assembler support enhancements: unified ELF targets, support
6134 for SCO/Caldera platforms, fix for Cygwin shared build.
6137 *) Add support for FreeBSD on sparc64. As a consequence, support for
6138 FreeBSD on non-x86 processors is separate from x86 processors on
6139 the config script, much like the NetBSD support.
6140 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6142 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6144 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6147 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6148 code (06) was taken as the first octet of the session ID and the last
6149 octet was ignored consequently. As a result SSLv2 client side session
6150 caching could not have worked due to the session ID mismatch between
6152 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6156 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6157 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6161 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6162 seems that in spite of existing for more than a year, many application
6163 author have done nothing to provide the necessary callbacks, which
6164 means that this particular engine will not work properly anywhere.
6165 This is a very unfortunate situation which forces us, in the name
6166 of usability, to give the hw_ncipher.c a static lock, which is part
6168 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6169 appear in 0.9.8 or later. We EXPECT application authors to have
6170 dealt properly with this when 0.9.8 is released (unless we actually
6171 make such changes in the libcrypto locking code that changes will
6172 have to be made anyway).
6175 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6176 octets have been read, EOF or an error occurs. Without this change
6177 some truncated ASN1 structures will not produce an error.
6180 *) Disable Heimdal support, since it hasn't been fully implemented.
6181 Still give the possibility to force the use of Heimdal, but with
6182 warnings and a request that patches get sent to openssl-dev.
6185 *) Add the VC-CE target, introduce the WINCE sysname, and add
6186 INSTALL.WCE and appropriate conditionals to make it build.
6187 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6189 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6190 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6191 edit numbers of the version.
6192 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6194 *) Introduce safe string copy and catenation functions
6195 (BUF_strlcpy() and BUF_strlcat()).
6196 [Ben Laurie (CHATS) and Richard Levitte]
6198 *) Avoid using fixed-size buffers for one-line DNs.
6199 [Ben Laurie (CHATS)]
6201 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6202 resizing buffers containing secrets, and use where appropriate.
6203 [Ben Laurie (CHATS)]
6205 *) Avoid using fixed size buffers for configuration file location.
6206 [Ben Laurie (CHATS)]
6208 *) Avoid filename truncation for various CA files.
6209 [Ben Laurie (CHATS)]
6211 *) Use sizeof in preference to magic numbers.
6212 [Ben Laurie (CHATS)]
6214 *) Avoid filename truncation in cert requests.
6215 [Ben Laurie (CHATS)]
6217 *) Add assertions to check for (supposedly impossible) buffer
6219 [Ben Laurie (CHATS)]
6221 *) Don't cache truncated DNS entries in the local cache (this could
6222 potentially lead to a spoofing attack).
6223 [Ben Laurie (CHATS)]
6225 *) Fix various buffers to be large enough for hex/decimal
6226 representations in a platform independent manner.
6227 [Ben Laurie (CHATS)]
6229 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6230 resizing buffers containing secrets, and use where appropriate.
6231 [Ben Laurie (CHATS)]
6233 *) Add BIO_indent() to avoid much slightly worrying code to do
6235 [Ben Laurie (CHATS)]
6237 *) Convert sprintf()/BIO_puts() to BIO_printf().
6238 [Ben Laurie (CHATS)]
6240 *) buffer_gets() could terminate with the buffer only half
6242 [Ben Laurie (CHATS)]
6244 *) Add assertions to prevent user-supplied crypto functions from
6245 overflowing internal buffers by having large block sizes, etc.
6246 [Ben Laurie (CHATS)]
6248 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6250 [Ben Laurie (CHATS)]
6252 *) Eliminate unused copy of key in RC4.
6253 [Ben Laurie (CHATS)]
6255 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6256 [Ben Laurie (CHATS)]
6258 *) Fix off-by-one error in EGD path.
6259 [Ben Laurie (CHATS)]
6261 *) If RANDFILE path is too long, ignore instead of truncating.
6262 [Ben Laurie (CHATS)]
6264 *) Eliminate unused and incorrectly sized X.509 structure
6266 [Ben Laurie (CHATS)]
6268 *) Eliminate unused and dangerous function knumber().
6269 [Ben Laurie (CHATS)]
6271 *) Eliminate unused and dangerous structure, KSSL_ERR.
6272 [Ben Laurie (CHATS)]
6274 *) Protect against overlong session ID context length in an encoded
6275 session object. Since these are local, this does not appear to be
6277 [Ben Laurie (CHATS)]
6279 *) Change from security patch (see 0.9.6e below) that did not affect
6280 the 0.9.6 release series:
6282 Remote buffer overflow in SSL3 protocol - an attacker could
6283 supply an oversized master key in Kerberos-enabled versions.
6285 [Ben Laurie (CHATS)]
6287 *) Change the SSL kerb5 codes to match RFC 2712.
6290 *) Make -nameopt work fully for req and add -reqopt switch.
6291 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6293 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6294 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6296 *) Make sure tests can be performed even if the corresponding algorithms
6297 have been removed entirely. This was also the last step to make
6298 OpenSSL compilable with DJGPP under all reasonable conditions.
6299 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6301 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6302 to allow version independent disabling of normally unselected ciphers,
6303 which may be activated as a side-effect of selecting a single cipher.
6305 (E.g., cipher list string "RSA" enables ciphersuites that are left
6306 out of "ALL" because they do not provide symmetric encryption.
6307 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6308 [Lutz Jaenicke, Bodo Moeller]
6310 *) Add appropriate support for separate platform-dependent build
6311 directories. The recommended way to make a platform-dependent
6312 build directory is the following (tested on Linux), maybe with
6315 # Place yourself outside of the OpenSSL source tree. In
6316 # this example, the environment variable OPENSSL_SOURCE
6317 # is assumed to contain the absolute OpenSSL source directory.
6318 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6319 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6320 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6321 mkdir -p `dirname $F`
6322 ln -s $OPENSSL_SOURCE/$F $F
6325 To be absolutely sure not to disturb the source tree, a "make clean"
6326 is a good thing. If it isn't successful, don't worry about it,
6327 it probably means the source directory is very clean.
6330 *) Make sure any ENGINE control commands make local copies of string
6331 pointers passed to them whenever necessary. Otherwise it is possible
6332 the caller may have overwritten (or deallocated) the original string
6333 data when a later ENGINE operation tries to use the stored values.
6334 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6336 *) Improve diagnostics in file reading and command-line digests.
6337 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6339 *) Add AES modes CFB and OFB to the object database. Correct an
6340 error in AES-CFB decryption.
6343 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6344 allows existing EVP_CIPHER_CTX structures to be reused after
6345 calling EVP_*Final(). This behaviour is used by encryption
6346 BIOs and some applications. This has the side effect that
6347 applications must explicitly clean up cipher contexts with
6348 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6351 *) Check the values of dna and dnb in bn_mul_recursive before calling
6352 bn_mul_comba (a non zero value means the a or b arrays do not contain
6353 n2 elements) and fallback to bn_mul_normal if either is not zero.
6356 *) Fix escaping of non-ASCII characters when using the -subj option
6357 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6360 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6361 form for "surname", serialNumber has no short form.
6362 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6363 therefore remove "mail" short name for "internet 7".
6364 The OID for unique identifiers in X509 certificates is
6365 x500UniqueIdentifier, not uniqueIdentifier.
6366 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6369 *) Add an "init" command to the ENGINE config module and auto initialize
6370 ENGINEs. Without any "init" command the ENGINE will be initialized
6371 after all ctrl commands have been executed on it. If init=1 the
6372 ENGINE is initailized at that point (ctrls before that point are run
6373 on the uninitialized ENGINE and after on the initialized one). If
6374 init=0 then the ENGINE will not be iniatialized at all.
6377 *) Fix the 'app_verify_callback' interface so that the user-defined
6378 argument is actually passed to the callback: In the
6379 SSL_CTX_set_cert_verify_callback() prototype, the callback
6380 declaration has been changed from
6383 int (*cb)(X509_STORE_CTX *,void *);
6384 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6385 i=s->ctx->app_verify_callback(&ctx)
6386 has been changed into
6387 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6389 To update applications using SSL_CTX_set_cert_verify_callback(),
6390 a dummy argument can be added to their callback functions.
6391 [D. K. Smetters <smetters@parc.xerox.com>]
6393 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6394 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6396 *) Add and OPENSSL_LOAD_CONF define which will cause
6397 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6398 This allows older applications to transparently support certain
6399 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6400 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6401 load the config file and OPENSSL_add_all_algorithms_conf() which will
6402 always load it have also been added.
6405 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6406 Adjust NIDs and EVP layer.
6407 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6409 *) Config modules support in openssl utility.
6411 Most commands now load modules from the config file,
6412 though in a few (such as version) this isn't done
6413 because it couldn't be used for anything.
6415 In the case of ca and req the config file used is
6416 the same as the utility itself: that is the -config
6417 command line option can be used to specify an
6421 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6422 use "openssl_conf" if filename is NULL use default openssl config file.
6425 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6426 config section name. Add a new flag to tolerate a missing config file
6427 and move code to CONF_modules_load_file().
6430 *) Support for crypto accelerator cards from Accelerated Encryption
6431 Processing, www.aep.ie. (Use engine 'aep')
6432 The support was copied from 0.9.6c [engine] and adapted/corrected
6433 to work with the new engine framework.
6434 [AEP Inc. and Richard Levitte]
6436 *) Support for SureWare crypto accelerator cards from Baltimore
6437 Technologies. (Use engine 'sureware')
6438 The support was copied from 0.9.6c [engine] and adapted
6439 to work with the new engine framework.
6442 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6443 make the newer ENGINE framework commands for the CHIL engine work.
6444 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6446 *) Make it possible to produce shared libraries on ReliantUNIX.
6447 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6449 *) Add the configuration target debug-linux-ppro.
6450 Make 'openssl rsa' use the general key loading routines
6451 implemented in apps.c, and make those routines able to
6452 handle the key format FORMAT_NETSCAPE and the variant
6454 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6456 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6457 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6459 *) Add -keyform to rsautl, and document -engine.
6460 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6462 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6463 BIO_R_NO_SUCH_FILE error code rather than the generic
6464 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6467 *) Add new functions
6469 ERR_peek_last_error_line
6470 ERR_peek_last_error_line_data.
6471 These are similar to
6474 ERR_peek_error_line_data,
6475 but report on the latest error recorded rather than the first one
6476 still in the error queue.
6477 [Ben Laurie, Bodo Moeller]
6479 *) default_algorithms option in ENGINE config module. This allows things
6481 default_algorithms = ALL
6482 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6485 *) Preliminary ENGINE config module.
6488 *) New experimental application configuration code.
6491 *) Change the AES code to follow the same name structure as all other
6492 symmetric ciphers, and behave the same way. Move everything to
6493 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6494 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6496 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6497 [Ben Laurie and Theo de Raadt]
6499 *) Add option to output public keys in req command.
6500 [Massimiliano Pala madwolf@openca.org]
6502 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6503 (up to about 10% better than before for P-192 and P-224).
6506 *) New functions/macros
6508 SSL_CTX_set_msg_callback(ctx, cb)
6509 SSL_CTX_set_msg_callback_arg(ctx, arg)
6510 SSL_set_msg_callback(ssl, cb)
6511 SSL_set_msg_callback_arg(ssl, arg)
6513 to request calling a callback function
6515 void cb(int write_p, int version, int content_type,
6516 const void *buf, size_t len, SSL *ssl, void *arg)
6518 whenever a protocol message has been completely received
6519 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6520 protocol version according to which the SSL library interprets
6521 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6522 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6523 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6524 specification (change_cipher_spec(20), alert(21), handshake(22)).
6525 'buf' and 'len' point to the actual message, 'ssl' to the
6526 SSL object, and 'arg' is the application-defined value set by
6527 SSL[_CTX]_set_msg_callback_arg().
6529 'openssl s_client' and 'openssl s_server' have new '-msg' options
6530 to enable a callback that displays all protocol messages.
6533 *) Change the shared library support so shared libraries are built as
6534 soon as the corresponding static library is finished, and thereby get
6535 openssl and the test programs linked against the shared library.
6536 This still only happens when the keyword "shard" has been given to
6537 the configuration scripts.
6539 NOTE: shared library support is still an experimental thing, and
6540 backward binary compatibility is still not guaranteed.
6541 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6543 *) Add support for Subject Information Access extension.
6544 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6546 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6547 additional bytes when new memory had to be allocated, not just
6548 when reusing an existing buffer.
6551 *) New command line and configuration option 'utf8' for the req command.
6552 This allows field values to be specified as UTF8 strings.
6555 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6556 runs for the former and machine-readable output for the latter.
6559 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6560 of the e-mail address in the DN (i.e., it will go into a certificate
6561 extension only). The new configuration file option 'email_in_dn = no'
6562 has the same effect.
6563 [Massimiliano Pala madwolf@openca.org]
6565 *) Change all functions with names starting with des_ to be starting
6566 with DES_ instead. Add wrappers that are compatible with libdes,
6567 but are named _ossl_old_des_*. Finally, add macros that map the
6568 des_* symbols to the corresponding _ossl_old_des_* if libdes
6569 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6570 desired, the des_* symbols will be mapped to DES_*, with one
6573 Since we provide two compatibility mappings, the user needs to
6574 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6575 compatibility is desired. The default (i.e., when that macro
6576 isn't defined) is OpenSSL 0.9.6c compatibility.
6578 There are also macros that enable and disable the support of old
6579 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6580 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6581 are defined, the default will apply: to support the old des routines.
6583 In either case, one must include openssl/des.h to get the correct
6584 definitions. Do not try to just include openssl/des_old.h, that
6587 NOTE: This is a major break of an old API into a new one. Software
6588 authors are encouraged to switch to the DES_ style functions. Some
6589 time in the future, des_old.h and the libdes compatibility functions
6590 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6591 default), and then completely removed.
6594 *) Test for certificates which contain unsupported critical extensions.
6595 If such a certificate is found during a verify operation it is
6596 rejected by default: this behaviour can be overridden by either
6597 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6598 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6599 X509_supported_extension() has also been added which returns 1 if a
6600 particular extension is supported.
6603 *) Modify the behaviour of EVP cipher functions in similar way to digests
6604 to retain compatibility with existing code.
6607 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6608 compatibility with existing code. In particular the 'ctx' parameter does
6609 not have to be to be initialized before the call to EVP_DigestInit() and
6610 it is tidied up after a call to EVP_DigestFinal(). New function
6611 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6612 EVP_MD_CTX_copy() changed to not require the destination to be
6613 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6614 requires the destination to be valid.
6616 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6617 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6620 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6621 so that complete 'Handshake' protocol structures are kept in memory
6622 instead of overwriting 'msg_type' and 'length' with 'body' data.
6625 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6626 [Massimo Santin via Richard Levitte]
6628 *) Major restructuring to the underlying ENGINE code. This includes
6629 reduction of linker bloat, separation of pure "ENGINE" manipulation
6630 (initialisation, etc) from functionality dealing with implementations
6631 of specific crypto iterfaces. This change also introduces integrated
6632 support for symmetric ciphers and digest implementations - so ENGINEs
6633 can now accelerate these by providing EVP_CIPHER and EVP_MD
6634 implementations of their own. This is detailed in crypto/engine/README
6635 as it couldn't be adequately described here. However, there are a few
6636 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6637 were changed in the original introduction of ENGINE code have now
6638 reverted back - the hooking from this code to ENGINE is now a good
6639 deal more passive and at run-time, operations deal directly with
6640 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6641 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6642 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6643 they were not being used by the framework as there is no concept of a
6644 BIGNUM_METHOD and they could not be generalised to the new
6645 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6646 ENGINE_cpy() has been removed as it cannot be consistently defined in
6650 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6653 *) Change mkdef.pl to sort symbols that get the same entry number,
6654 and make sure the automatically generated functions ERR_load_*
6655 become part of libeay.num as well.
6658 *) New function SSL_renegotiate_pending(). This returns true once
6659 renegotiation has been requested (either SSL_renegotiate() call
6660 or HelloRequest/ClientHello received from the peer) and becomes
6661 false once a handshake has been completed.
6662 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6663 sends a HelloRequest, but does not ensure that a handshake takes
6664 place. SSL_renegotiate_pending() is useful for checking if the
6665 client has followed the request.)
6668 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6669 By default, clients may request session resumption even during
6670 renegotiation (if session ID contexts permit); with this option,
6671 session resumption is possible only in the first handshake.
6673 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6674 more bits available for options that should not be part of
6675 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6678 *) Add some demos for certificate and certificate request creation.
6681 *) Make maximum certificate chain size accepted from the peer application
6682 settable (SSL*_get/set_max_cert_list()), as proposed by
6683 "Douglas E. Engert" <deengert@anl.gov>.
6686 *) Add support for shared libraries for Unixware-7
6687 (Boyd Lynn Gerber <gerberb@zenez.com>).
6690 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6691 be done prior to destruction. Use this to unload error strings from
6692 ENGINEs that load their own error strings. NB: This adds two new API
6693 functions to "get" and "set" this destroy handler in an ENGINE.
6696 *) Alter all existing ENGINE implementations (except "openssl" and
6697 "openbsd") to dynamically instantiate their own error strings. This
6698 makes them more flexible to be built both as statically-linked ENGINEs
6699 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6700 Also, add stub code to each that makes building them as self-contained
6701 shared-libraries easier (see README.ENGINE).
6704 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6705 implementations into applications that are completely implemented in
6706 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6707 commands that can be used to configure what shared-library to load and
6708 to control aspects of the way it is handled. Also, made an update to
6709 the README.ENGINE file that brings its information up-to-date and
6710 provides some information and instructions on the "dynamic" ENGINE
6711 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6714 *) Make it possible to unload ranges of ERR strings with a new
6715 "ERR_unload_strings" function.
6718 *) Add a copy() function to EVP_MD.
6721 *) Make EVP_MD routines take a context pointer instead of just the
6722 md_data void pointer.
6725 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6726 that the digest can only process a single chunk of data
6727 (typically because it is provided by a piece of
6728 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6729 is only going to provide a single chunk of data, and hence the
6730 framework needn't accumulate the data for oneshot drivers.
6733 *) As with "ERR", make it possible to replace the underlying "ex_data"
6734 functions. This change also alters the storage and management of global
6735 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6736 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6737 index counters. The API functions that use this state have been changed
6738 to take a "class_index" rather than pointers to the class's local STACK
6739 and counter, and there is now an API function to dynamically create new
6740 classes. This centralisation allows us to (a) plug a lot of the
6741 thread-safety problems that existed, and (b) makes it possible to clean
6742 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6743 such data would previously have always leaked in application code and
6744 workarounds were in place to make the memory debugging turn a blind eye
6745 to it. Application code that doesn't use this new function will still
6746 leak as before, but their memory debugging output will announce it now
6747 rather than letting it slide.
6749 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6750 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6751 has a return value to indicate success or failure.
6754 *) Make it possible to replace the underlying "ERR" functions such that the
6755 global state (2 LHASH tables and 2 locks) is only used by the "default"
6756 implementation. This change also adds two functions to "get" and "set"
6757 the implementation prior to it being automatically set the first time
6758 any other ERR function takes place. Ie. an application can call "get",
6759 pass the return value to a module it has just loaded, and that module
6760 can call its own "set" function using that value. This means the
6761 module's "ERR" operations will use (and modify) the error state in the
6762 application and not in its own statically linked copy of OpenSSL code.
6765 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6766 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6767 the operation, and provides a more encapsulated way for external code
6768 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6769 to use these functions rather than manually incrementing the counts.
6771 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6774 *) Add EVP test program.
6777 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6780 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6781 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6782 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6783 These allow a CRL to be built without having to access X509_CRL fields
6784 directly. Modify 'ca' application to use new functions.
6787 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6788 bug workarounds. Rollback attack detection is a security feature.
6789 The problem will only arise on OpenSSL servers when TLSv1 is not
6790 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6791 Software authors not wanting to support TLSv1 will have special reasons
6792 for their choice and can explicitly enable this option.
6793 [Bodo Moeller, Lutz Jaenicke]
6795 *) Rationalise EVP so it can be extended: don't include a union of
6796 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6797 (similar to those existing for EVP_CIPHER_CTX).
6802 EVP_MD_CTX_init(&md); /* new function call */
6803 EVP_DigestInit(&md, EVP_sha1());
6804 EVP_DigestUpdate(&md, in, len);
6805 EVP_DigestFinal(&md, out, NULL);
6806 EVP_MD_CTX_cleanup(&md); /* new function call */
6810 *) Make DES key schedule conform to the usual scheme, as well as
6811 correcting its structure. This means that calls to DES functions
6812 now have to pass a pointer to a des_key_schedule instead of a
6813 plain des_key_schedule (which was actually always a pointer
6816 des_key_schedule ks;
6818 des_set_key_checked(..., &ks);
6819 des_ncbc_encrypt(..., &ks, ...);
6821 (Note that a later change renames 'des_...' into 'DES_...'.)
6824 *) Initial reduction of linker bloat: the use of some functions, such as
6825 PEM causes large amounts of unused functions to be linked in due to
6826 poor organisation. For example pem_all.c contains every PEM function
6827 which has a knock on effect of linking in large amounts of (unused)
6828 ASN1 code. Grouping together similar functions and splitting unrelated
6829 functions prevents this.
6832 *) Cleanup of EVP macros.
6835 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6836 correct _ecb suffix.
6839 *) Add initial OCSP responder support to ocsp application. The
6840 revocation information is handled using the text based index
6841 use by the ca application. The responder can either handle
6842 requests generated internally, supplied in files (for example
6843 via a CGI script) or using an internal minimal server.
6846 *) Add configuration choices to get zlib compression for TLS.
6849 *) Changes to Kerberos SSL for RFC 2712 compliance:
6850 1. Implemented real KerberosWrapper, instead of just using
6851 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6852 2. Implemented optional authenticator field of KerberosWrapper.
6854 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6855 and authenticator structs; see crypto/krb5/.
6857 Generalized Kerberos calls to support multiple Kerberos libraries.
6858 [Vern Staats <staatsvr@asc.hpc.mil>,
6859 Jeffrey Altman <jaltman@columbia.edu>
6860 via Richard Levitte]
6862 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6863 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6864 values for each of the key sizes rather than having just
6865 parameters (and 'speed' generating keys each time).
6868 *) Speed up EVP routines.
6871 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6872 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6873 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6874 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6876 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6877 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6878 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6881 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6883 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6886 *) Added the OS2-EMX target.
6887 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6889 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6890 to support NCONF routines in extension code. New function CONF_set_nconf()
6891 to allow functions which take an NCONF to also handle the old LHASH
6892 structure: this means that the old CONF compatible routines can be
6893 retained (in particular wrt extensions) without having to duplicate the
6894 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6897 *) Enhance the general user interface with mechanisms for inner control
6898 and with possibilities to have yes/no kind of prompts.
6901 *) Change all calls to low level digest routines in the library and
6902 applications to use EVP. Add missing calls to HMAC_cleanup() and
6903 don't assume HMAC_CTX can be copied using memcpy().
6904 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6906 *) Add the possibility to control engines through control names but with
6907 arbitrary arguments instead of just a string.
6908 Change the key loaders to take a UI_METHOD instead of a callback
6909 function pointer. NOTE: this breaks binary compatibility with earlier
6910 versions of OpenSSL [engine].
6911 Adapt the nCipher code for these new conditions and add a card insertion
6915 *) Enhance the general user interface with mechanisms to better support
6916 dialog box interfaces, application-defined prompts, the possibility
6917 to use defaults (for example default passwords from somewhere else)
6918 and interrupts/cancellations.
6921 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6922 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6925 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6926 tidy up some unnecessarily weird code in 'sk_new()').
6927 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6929 *) Change the key loading routines for ENGINEs to use the same kind
6930 callback (pem_password_cb) as all other routines that need this
6934 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6935 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6936 than this minimum value is recommended.
6939 *) New random seeder for OpenVMS, using the system process statistics
6940 that are easily reachable.
6943 *) Windows apparently can't transparently handle global
6944 variables defined in DLLs. Initialisations such as:
6946 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6948 won't compile. This is used by the any applications that need to
6949 declare their own ASN1 modules. This was fixed by adding the option
6950 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6951 needed for static libraries under Win32.
6954 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6955 setting of purpose and trust fields. New X509_STORE trust and
6956 purpose functions and tidy up setting in other SSL functions.
6959 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6960 structure. These are inherited by X509_STORE_CTX when it is
6961 initialised. This allows various defaults to be set in the
6962 X509_STORE structure (such as flags for CRL checking and custom
6963 purpose or trust settings) for functions which only use X509_STORE_CTX
6964 internally such as S/MIME.
6966 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6967 trust settings if they are not set in X509_STORE. This allows X509_STORE
6968 purposes and trust (in S/MIME for example) to override any set by default.
6970 Add command line options for CRL checking to smime, s_client and s_server
6974 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6975 are set then the CRL is looked up in the X509_STORE structure and
6976 its validity and signature checked, then if the certificate is found
6977 in the CRL the verify fails with a revoked error.
6979 Various new CRL related callbacks added to X509_STORE_CTX structure.
6981 Command line options added to 'verify' application to support this.
6983 This needs some additional work, such as being able to handle multiple
6984 CRLs with different times, extension based lookup (rather than just
6985 by subject name) and ultimately more complete V2 CRL extension
6989 *) Add a general user interface API (crypto/ui/). This is designed
6990 to replace things like des_read_password and friends (backward
6991 compatibility functions using this new API are provided).
6992 The purpose is to remove prompting functions from the DES code
6993 section as well as provide for prompting through dialog boxes in
6994 a window system and the like.
6997 *) Add "ex_data" support to ENGINE so implementations can add state at a
6998 per-structure level rather than having to store it globally.
7001 *) Make it possible for ENGINE structures to be copied when retrieved by
7002 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7003 This causes the "original" ENGINE structure to act like a template,
7004 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7005 operational state can be localised to each ENGINE structure, despite the
7006 fact they all share the same "methods". New ENGINE structures returned in
7007 this case have no functional references and the return value is the single
7008 structural reference. This matches the single structural reference returned
7009 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7013 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7014 needs to match any other type at all we need to manually clear the
7018 *) Changes to the "openssl engine" utility to include;
7019 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7020 about an ENGINE's available control commands.
7021 - executing control commands from command line arguments using the
7022 '-pre' and '-post' switches. '-post' is only used if '-t' is
7023 specified and the ENGINE is successfully initialised. The syntax for
7024 the individual commands are colon-separated, for example;
7025 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7028 *) New dynamic control command support for ENGINEs. ENGINEs can now
7029 declare their own commands (numbers), names (strings), descriptions,
7030 and input types for run-time discovery by calling applications. A
7031 subset of these commands are implicitly classed as "executable"
7032 depending on their input type, and only these can be invoked through
7033 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7034 can be based on user input, config files, etc). The distinction is
7035 that "executable" commands cannot return anything other than a boolean
7036 result and can only support numeric or string input, whereas some
7037 discoverable commands may only be for direct use through
7038 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7039 pointers, or other custom uses. The "executable" commands are to
7040 support parameterisations of ENGINE behaviour that can be
7041 unambiguously defined by ENGINEs and used consistently across any
7042 OpenSSL-based application. Commands have been added to all the
7043 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7044 control over shared-library paths without source code alterations.
7047 *) Changed all ENGINE implementations to dynamically allocate their
7048 ENGINEs rather than declaring them statically. Apart from this being
7049 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7050 this also allows the implementations to compile without using the
7051 internal engine_int.h header.
7054 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7055 'const' value. Any code that should be able to modify a RAND_METHOD
7056 should already have non-const pointers to it (ie. they should only
7057 modify their own ones).
7060 *) Made a variety of little tweaks to the ENGINE code.
7061 - "atalla" and "ubsec" string definitions were moved from header files
7062 to C code. "nuron" string definitions were placed in variables
7063 rather than hard-coded - allowing parameterisation of these values
7064 later on via ctrl() commands.
7065 - Removed unused "#if 0"'d code.
7066 - Fixed engine list iteration code so it uses ENGINE_free() to release
7067 structural references.
7068 - Constified the RAND_METHOD element of ENGINE structures.
7069 - Constified various get/set functions as appropriate and added
7070 missing functions (including a catch-all ENGINE_cpy that duplicates
7071 all ENGINE values onto a new ENGINE except reference counts/state).
7072 - Removed NULL parameter checks in get/set functions. Setting a method
7073 or function to NULL is a way of cancelling out a previously set
7074 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7075 and doesn't justify the extra error symbols and code.
7076 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7077 flags from engine_int.h to engine.h.
7078 - Changed prototypes for ENGINE handler functions (init(), finish(),
7079 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7082 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7083 to the algorithm using long division. The binary algorithm can be
7084 used only if the modulus is odd. On 32-bit systems, it is faster
7085 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7086 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7087 up to 450 bits. In 64-bit environments, the binary algorithm
7088 appears to be advantageous for much longer moduli; here we use it
7089 for moduli up to 2048 bits.
7092 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7093 could not support the combine flag in choice fields.
7096 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7097 extensions from a certificate request to the certificate.
7100 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7101 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7102 file: this allows the display of the certificate about to be
7103 signed to be customised, to allow certain fields to be included
7104 or excluded and extension details. The old system didn't display
7105 multicharacter strings properly, omitted fields not in the policy
7106 and couldn't display additional details such as extensions.
7109 *) Function EC_POINTs_mul for multiple scalar multiplication
7110 of an arbitrary number of elliptic curve points
7111 \sum scalars[i]*points[i],
7112 optionally including the generator defined for the EC_GROUP:
7113 scalar*generator + \sum scalars[i]*points[i].
7115 EC_POINT_mul is a simple wrapper function for the typical case
7116 that the point list has just one item (besides the optional
7120 *) First EC_METHODs for curves over GF(p):
7122 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7123 operations and provides various method functions that can also
7124 operate with faster implementations of modular arithmetic.
7126 EC_GFp_mont_method() reuses most functions that are part of
7127 EC_GFp_simple_method, but uses Montgomery arithmetic.
7129 [Bodo Moeller; point addition and point doubling
7130 implementation directly derived from source code provided by
7131 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7133 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7134 crypto/ec/ec_lib.c):
7136 Curves are EC_GROUP objects (with an optional group generator)
7137 based on EC_METHODs that are built into the library.
7139 Points are EC_POINT objects based on EC_GROUP objects.
7141 Most of the framework would be able to handle curves over arbitrary
7142 finite fields, but as there are no obvious types for fields other
7143 than GF(p), some functions are limited to that for now.
7146 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7147 that the file contains a complete HTTP response.
7150 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7151 change the def and num file printf format specifier from "%-40sXXX"
7152 to "%-39s XXX". The latter will always guarantee a space after the
7153 field while the former will cause them to run together if the field
7154 is 40 of more characters long.
7157 *) Constify the cipher and digest 'method' functions and structures
7158 and modify related functions to take constant EVP_MD and EVP_CIPHER
7162 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7163 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7166 *) Modify EVP_Digest*() routines so they now return values. Although the
7167 internal software routines can never fail additional hardware versions
7171 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7173 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7174 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7177 ERR_R_NESTED_ASN1_ERROR
7179 ERR_R_MISSING_ASN1_EOS
7180 were 4 .. 9, conflicting with
7181 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7183 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7184 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7186 Add new error code 'ERR_R_INTERNAL_ERROR'.
7189 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7193 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7194 sets the subject name for a new request or supersedes the
7195 subject name in a given request. Formats that can be parsed are
7196 'CN=Some Name, OU=myOU, C=IT'
7198 'CN=Some Name/OU=myOU/C=IT'.
7200 Add options '-batch' and '-verbose' to 'openssl req'.
7201 [Massimiliano Pala <madwolf@hackmasters.net>]
7203 *) Introduce the possibility to access global variables through
7204 functions on platform were that's the best way to handle exporting
7205 global variables in shared libraries. To enable this functionality,
7206 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7207 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7208 is normally done by Configure or something similar).
7210 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7211 in the source file (foo.c) like this:
7213 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7214 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7216 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7217 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7219 OPENSSL_DECLARE_GLOBAL(int,foo);
7220 #define foo OPENSSL_GLOBAL_REF(foo)
7221 OPENSSL_DECLARE_GLOBAL(double,bar);
7222 #define bar OPENSSL_GLOBAL_REF(bar)
7224 The #defines are very important, and therefore so is including the
7225 header file everywhere where the defined globals are used.
7227 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7228 of ASN.1 items, but that structure is a bit different.
7230 The largest change is in util/mkdef.pl which has been enhanced with
7231 better and easier to understand logic to choose which symbols should
7232 go into the Windows .def files as well as a number of fixes and code
7233 cleanup (among others, algorithm keywords are now sorted
7234 lexicographically to avoid constant rewrites).
7237 *) In BN_div() keep a copy of the sign of 'num' before writing the
7238 result to 'rm' because if rm==num the value will be overwritten
7239 and produce the wrong result if 'num' is negative: this caused
7240 problems with BN_mod() and BN_nnmod().
7243 *) Function OCSP_request_verify(). This checks the signature on an
7244 OCSP request and verifies the signer certificate. The signer
7245 certificate is just checked for a generic purpose and OCSP request
7249 *) Add OCSP_check_validity() function to check the validity of OCSP
7250 responses. OCSP responses are prepared in real time and may only
7251 be a few seconds old. Simply checking that the current time lies
7252 between thisUpdate and nextUpdate max reject otherwise valid responses
7253 caused by either OCSP responder or client clock inaccuracy. Instead
7254 we allow thisUpdate and nextUpdate to fall within a certain period of
7255 the current time. The age of the response can also optionally be
7256 checked. Two new options -validity_period and -status_age added to
7260 *) If signature or public key algorithm is unrecognized print out its
7261 OID rather that just UNKNOWN.
7264 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7265 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7266 ID to be generated from the issuer certificate alone which can then be
7267 passed to OCSP_id_issuer_cmp().
7270 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7271 ASN1 modules to export functions returning ASN1_ITEM pointers
7272 instead of the ASN1_ITEM structures themselves. This adds several
7273 new macros which allow the underlying ASN1 function/structure to
7274 be accessed transparently. As a result code should not use ASN1_ITEM
7275 references directly (such as &X509_it) but instead use the relevant
7276 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7277 use of the new ASN1 code on platforms where exporting structures
7278 is problematical (for example in shared libraries) but exporting
7279 functions returning pointers to structures is not.
7282 *) Add support for overriding the generation of SSL/TLS session IDs.
7283 These callbacks can be registered either in an SSL_CTX or per SSL.
7284 The purpose of this is to allow applications to control, if they wish,
7285 the arbitrary values chosen for use as session IDs, particularly as it
7286 can be useful for session caching in multiple-server environments. A
7287 command-line switch for testing this (and any client code that wishes
7288 to use such a feature) has been added to "s_server".
7289 [Geoff Thorpe, Lutz Jaenicke]
7291 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7292 of the form '#if defined(...) || defined(...) || ...' and
7293 '#if !defined(...) && !defined(...) && ...'. This also avoids
7294 the growing number of special cases it was previously handling.
7297 *) Make all configuration macros available for application by making
7298 sure they are available in opensslconf.h, by giving them names starting
7299 with "OPENSSL_" to avoid conflicts with other packages and by making
7300 sure e_os2.h will cover all platform-specific cases together with
7302 Additionally, it is now possible to define configuration/platform-
7303 specific names (called "system identities"). In the C code, these
7304 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7305 macro with the name beginning with "OPENSSL_SYS_", which is determined
7306 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7310 *) New option -set_serial to 'req' and 'x509' this allows the serial
7311 number to use to be specified on the command line. Previously self
7312 signed certificates were hard coded with serial number 0 and the
7313 CA options of 'x509' had to use a serial number in a file which was
7317 *) New options to 'ca' utility to support V2 CRL entry extensions.
7318 Currently CRL reason, invalidity date and hold instruction are
7319 supported. Add new CRL extensions to V3 code and some new objects.
7322 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7323 disable standard block padding (aka PKCS#5 padding) in the EVP
7324 API, which was previously mandatory. This means that the data is
7325 not padded in any way and so the total length much be a multiple
7326 of the block size, otherwise an error occurs.
7329 *) Initial (incomplete) OCSP SSL support.
7332 *) New function OCSP_parse_url(). This splits up a URL into its host,
7333 port and path components: primarily to parse OCSP URLs. New -url
7334 option to ocsp utility.
7337 *) New nonce behavior. The return value of OCSP_check_nonce() now
7338 reflects the various checks performed. Applications can decide
7339 whether to tolerate certain situations such as an absent nonce
7340 in a response when one was present in a request: the ocsp application
7341 just prints out a warning. New function OCSP_add1_basic_nonce()
7342 this is to allow responders to include a nonce in a response even if
7343 the request is nonce-less.
7346 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7347 skipped when using openssl x509 multiple times on a single input file,
7348 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7351 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7352 set string type: to handle setting ASN1_TIME structures. Fix ca
7353 utility to correctly initialize revocation date of CRLs.
7356 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7357 the clients preferred ciphersuites and rather use its own preferences.
7358 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7359 Internet Explorer by ensuring unchanged hash method during stepup.
7360 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7363 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7364 to aes and add a new 'exist' option to print out symbols that don't
7368 *) Additional options to ocsp utility to allow flags to be set and
7369 additional certificates supplied.
7372 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7373 OCSP client a number of certificate to only verify the response
7377 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7378 handle the new API. Currently only ECB, CBC modes supported. Add new
7381 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7382 Encryption Standard (AES) Ciphersuites for Transport Layer
7383 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7384 not enabled by default and were not part of the "ALL" ciphersuite
7385 alias because they were not yet official; they could be
7386 explicitly requested by specifying the "AESdraft" ciphersuite
7387 group alias. In the final release of OpenSSL 0.9.7, the group
7388 alias is called "AES" and is part of "ALL".)
7389 [Ben Laurie, Steve Henson, Bodo Moeller]
7391 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7392 request to response.
7395 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7396 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7397 extract information from a certificate request. OCSP_response_create()
7398 creates a response and optionally adds a basic response structure.
7399 OCSP_basic_add1_status() adds a complete single response to a basic
7400 response and returns the OCSP_SINGLERESP structure just added (to allow
7401 extensions to be included for example). OCSP_basic_add1_cert() adds a
7402 certificate to a basic response and OCSP_basic_sign() signs a basic
7403 response with various flags. New helper functions ASN1_TIME_check()
7404 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7405 (converts ASN1_TIME to GeneralizedTime).
7408 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7409 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7410 structure from a certificate. X509_pubkey_digest() digests the public_key
7411 contents: this is used in various key identifiers.
7414 *) Make sk_sort() tolerate a NULL argument.
7415 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7417 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7418 passed by the function are trusted implicitly. If any of them signed the
7419 response then it is assumed to be valid and is not verified.
7422 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7423 to data. This was previously part of the PKCS7 ASN1 code. This
7424 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7425 [Steve Henson, reported by Kenneth R. Robinette
7426 <support@securenetterm.com>]
7428 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7429 routines: without these tracing memory leaks is very painful.
7430 Fix leaks in PKCS12 and PKCS7 routines.
7433 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7434 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7435 effectively meant GeneralizedTime would never be used. Now it
7436 is initialised to -1 but X509_time_adj() now has to check the value
7437 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7438 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7439 [Steve Henson, reported by Kenneth R. Robinette
7440 <support@securenetterm.com>]
7442 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7443 result in a zero length in the ASN1_INTEGER structure which was
7444 not consistent with the structure when d2i_ASN1_INTEGER() was used
7445 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7446 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7447 where it did not print out a minus for negative ASN1_INTEGER.
7450 *) Add summary printout to ocsp utility. The various functions which
7451 convert status values to strings have been renamed to:
7452 OCSP_response_status_str(), OCSP_cert_status_str() and
7453 OCSP_crl_reason_str() and are no longer static. New options
7454 to verify nonce values and to disable verification. OCSP response
7455 printout format cleaned up.
7458 *) Add additional OCSP certificate checks. These are those specified
7459 in RFC2560. This consists of two separate checks: the CA of the
7460 certificate being checked must either be the OCSP signer certificate
7461 or the issuer of the OCSP signer certificate. In the latter case the
7462 OCSP signer certificate must contain the OCSP signing extended key
7463 usage. This check is performed by attempting to match the OCSP
7464 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7465 in the OCSP_CERTID structures of the response.
7468 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7469 and related routines. This uses the standard OpenSSL certificate
7470 verify routines to perform initial checks (just CA validity) and
7471 to obtain the certificate chain. Then additional checks will be
7472 performed on the chain. Currently the root CA is checked to see
7473 if it is explicitly trusted for OCSP signing. This is used to set
7474 a root CA as a global signing root: that is any certificate that
7475 chains to that CA is an acceptable OCSP signing certificate.
7478 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7479 extensions from a separate configuration file.
7480 As when reading extensions from the main configuration file,
7481 the '-extensions ...' option may be used for specifying the
7483 [Massimiliano Pala <madwolf@comune.modena.it>]
7485 *) New OCSP utility. Allows OCSP requests to be generated or
7486 read. The request can be sent to a responder and the output
7487 parsed, outputed or printed in text form. Not complete yet:
7488 still needs to check the OCSP response validity.
7491 *) New subcommands for 'openssl ca':
7492 'openssl ca -status <serial>' prints the status of the cert with
7493 the given serial number (according to the index file).
7494 'openssl ca -updatedb' updates the expiry status of certificates
7496 [Massimiliano Pala <madwolf@comune.modena.it>]
7498 *) New '-newreq-nodes' command option to CA.pl. This is like
7499 '-newreq', but calls 'openssl req' with the '-nodes' option
7500 so that the resulting key is not encrypted.
7501 [Damien Miller <djm@mindrot.org>]
7503 *) New configuration for the GNU Hurd.
7504 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7506 *) Initial code to implement OCSP basic response verify. This
7507 is currently incomplete. Currently just finds the signer's
7508 certificate and verifies the signature on the response.
7511 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7512 value of OPENSSLDIR. This is available via the new '-d' option
7513 to 'openssl version', and is also included in 'openssl version -a'.
7516 *) Allowing defining memory allocation callbacks that will be given
7517 file name and line number information in additional arguments
7518 (a const char* and an int). The basic functionality remains, as
7519 well as the original possibility to just replace malloc(),
7520 realloc() and free() by functions that do not know about these
7521 additional arguments. To register and find out the current
7522 settings for extended allocation functions, the following
7523 functions are provided:
7525 CRYPTO_set_mem_ex_functions
7526 CRYPTO_set_locked_mem_ex_functions
7527 CRYPTO_get_mem_ex_functions
7528 CRYPTO_get_locked_mem_ex_functions
7530 These work the same way as CRYPTO_set_mem_functions and friends.
7531 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7532 extended allocation function is enabled.
7533 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7534 a conventional allocation function is enabled.
7535 [Richard Levitte, Bodo Moeller]
7537 *) Finish off removing the remaining LHASH function pointer casts.
7538 There should no longer be any prototype-casting required when using
7539 the LHASH abstraction, and any casts that remain are "bugs". See
7540 the callback types and macros at the head of lhash.h for details
7541 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7544 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7545 If /dev/[u]random devices are not available or do not return enough
7546 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7548 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7549 /etc/entropy will be queried once each in this sequence, querying stops
7550 when enough entropy was collected without querying more sockets.
7553 *) Change the Unix RAND_poll() variant to be able to poll several
7554 random devices, as specified by DEVRANDOM, until a sufficient amount
7555 of data has been collected. We spend at most 10 ms on each file
7556 (select timeout) and read in non-blocking mode. DEVRANDOM now
7557 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7558 (previously it was just the string "/dev/urandom"), so on typical
7559 platforms the 10 ms delay will never occur.
7560 Also separate out the Unix variant to its own file, rand_unix.c.
7561 For VMS, there's a currently-empty rand_vms.c.
7564 *) Move OCSP client related routines to ocsp_cl.c. These
7565 provide utility functions which an application needing
7566 to issue a request to an OCSP responder and analyse the
7567 response will typically need: as opposed to those which an
7568 OCSP responder itself would need which will be added later.
7570 OCSP_request_sign() signs an OCSP request with an API similar
7571 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7572 response. OCSP_response_get1_basic() extracts basic response
7573 from response. OCSP_resp_find_status(): finds and extracts status
7574 information from an OCSP_CERTID structure (which will be created
7575 when the request structure is built). These are built from lower
7576 level functions which work on OCSP_SINGLERESP structures but
7577 won't normally be used unless the application wishes to examine
7578 extensions in the OCSP response for example.
7580 Replace nonce routines with a pair of functions.
7581 OCSP_request_add1_nonce() adds a nonce value and optionally
7582 generates a random value. OCSP_check_nonce() checks the
7583 validity of the nonce in an OCSP response.
7586 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7587 This doesn't copy the supplied OCSP_CERTID and avoids the
7588 need to free up the newly created id. Change return type
7589 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7590 This can then be used to add extensions to the request.
7591 Deleted OCSP_request_new(), since most of its functionality
7592 is now in OCSP_REQUEST_new() (and the case insensitive name
7593 clash) apart from the ability to set the request name which
7594 will be added elsewhere.
7597 *) Update OCSP API. Remove obsolete extensions argument from
7598 various functions. Extensions are now handled using the new
7599 OCSP extension code. New simple OCSP HTTP function which
7600 can be used to send requests and parse the response.
7603 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7604 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7605 uses the special reorder version of SET OF to sort the attributes
7606 and reorder them to match the encoded order. This resolves a long
7607 standing problem: a verify on a PKCS7 structure just after signing
7608 it used to fail because the attribute order did not match the
7609 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7610 it uses the received order. This is necessary to tolerate some broken
7611 software that does not order SET OF. This is handled by encoding
7612 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7613 to produce the required SET OF.
7616 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7617 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7618 files to get correct declarations of the ASN.1 item variables.
7621 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7622 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7623 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7624 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7625 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7626 ASN1_ITEM and no wrapper functions.
7629 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7630 replace the old function pointer based I/O routines. Change most of
7631 the *_d2i_bio() and *_d2i_fp() functions to use these.
7634 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7635 lines, recognice more "algorithms" that can be deselected, and make
7636 it complain about algorithm deselection that isn't recognised.
7639 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7640 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7641 to use new functions. Add NO_ASN1_OLD which can be set to remove
7642 some old style ASN1 functions: this can be used to determine if old
7643 code will still work when these eventually go away.
7646 *) New extension functions for OCSP structures, these follow the
7647 same conventions as certificates and CRLs.
7650 *) New function X509V3_add1_i2d(). This automatically encodes and
7651 adds an extension. Its behaviour can be customised with various
7652 flags to append, replace or delete. Various wrappers added for
7653 certificates and CRLs.
7656 *) Fix to avoid calling the underlying ASN1 print routine when
7657 an extension cannot be parsed. Correct a typo in the
7658 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7661 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7662 entries for variables.
7665 *) Add functionality to apps/openssl.c for detecting locking
7666 problems: As the program is single-threaded, all we have
7667 to do is register a locking callback using an array for
7668 storing which locks are currently held by the program.
7671 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7672 SSL_get_ex_data_X509_STORE_idx(), which is used in
7673 ssl_verify_cert_chain() and thus can be called at any time
7674 during TLS/SSL handshakes so that thread-safety is essential.
7675 Unfortunately, the ex_data design is not at all suited
7676 for multi-threaded use, so it probably should be abolished.
7679 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7680 [Broadcom, tweaked and integrated by Geoff Thorpe]
7682 *) Move common extension printing code to new function
7683 X509V3_print_extensions(). Reorganise OCSP print routines and
7684 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7687 *) New function X509_signature_print() to remove duplication in some
7691 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7692 set (this was treated exactly the same as SET OF previously). This
7693 is used to reorder the STACK representing the structure to match the
7694 encoding. This will be used to get round a problem where a PKCS7
7695 structure which was signed could not be verified because the STACK
7696 order did not reflect the encoded order.
7699 *) Reimplement the OCSP ASN1 module using the new code.
7702 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7703 for its ASN1 operations. The old style function pointers still exist
7704 for now but they will eventually go away.
7707 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7708 completely replaces the old ASN1 functionality with a table driven
7709 encoder and decoder which interprets an ASN1_ITEM structure describing
7710 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7711 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7712 has also been converted to the new form.
7715 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7716 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7717 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7718 for negative moduli.
7721 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7722 of not touching the result's sign bit.
7725 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7729 *) Changed the LHASH code to use prototypes for callbacks, and created
7730 macros to declare and implement thin (optionally static) functions
7731 that provide type-safety and avoid function pointer casting for the
7732 type-specific callbacks.
7735 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7737 [Veers Staats <staatsvr@asc.hpc.mil>,
7738 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7740 *) Reformat the FAQ so the different questions and answers can be divided
7741 in sections depending on the subject.
7744 *) Have the zlib compression code load ZLIB.DLL dynamically under
7748 *) New function BN_mod_sqrt for computing square roots modulo a prime
7749 (using the probabilistic Tonelli-Shanks algorithm unless
7750 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7751 be handled deterministically).
7752 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7754 *) Make BN_mod_inverse faster by explicitly handling small quotients
7755 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7756 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7759 *) New function BN_kronecker.
7762 *) Fix BN_gcd so that it works on negative inputs; the result is
7763 positive unless both parameters are zero.
7764 Previously something reasonably close to an infinite loop was
7765 possible because numbers could be growing instead of shrinking
7766 in the implementation of Euclid's algorithm.
7769 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7770 sign of the number in question.
7772 Fix BN_is_word(a,w) to work correctly for w == 0.
7774 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7775 because its test if the absolute value of 'a' equals 'w'.
7776 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7777 it exists mostly for use in the implementations of BN_is_zero(),
7778 BN_is_one(), and BN_is_word().
7781 *) New function BN_swap.
7784 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7785 the exponentiation functions are more likely to produce reasonable
7786 results on negative inputs.
7789 *) Change BN_mod_mul so that the result is always non-negative.
7790 Previously, it could be negative if one of the factors was negative;
7791 I don't think anyone really wanted that behaviour.
7794 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7795 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7796 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7797 and add new functions:
7806 BN_mod_lshift1_quick
7810 These functions always generate non-negative results.
7812 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7813 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7815 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7816 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7817 be reduced modulo m.
7818 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7821 The following entry accidentally appeared in the CHANGES file
7822 distributed with OpenSSL 0.9.7. The modifications described in
7823 it do *not* apply to OpenSSL 0.9.7.
7825 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7826 was actually never needed) and in BN_mul(). The removal in BN_mul()
7827 required a small change in bn_mul_part_recursive() and the addition
7828 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7829 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7830 bn_sub_words() and bn_add_words() except they take arrays with
7835 *) In 'openssl passwd', verify passwords read from the terminal
7836 unless the '-salt' option is used (which usually means that
7837 verification would just waste user's time since the resulting
7838 hash is going to be compared with some given password hash)
7839 or the new '-noverify' option is used.
7841 This is an incompatible change, but it does not affect
7842 non-interactive use of 'openssl passwd' (passwords on the command
7843 line, '-stdin' option, '-in ...' option) and thus should not
7847 *) Remove all references to RSAref, since there's no more need for it.
7850 *) Make DSO load along a path given through an environment variable
7851 (SHLIB_PATH) with shl_load().
7854 *) Constify the ENGINE code as a result of BIGNUM constification.
7855 Also constify the RSA code and most things related to it. In a
7856 few places, most notable in the depth of the ASN.1 code, ugly
7857 casts back to non-const were required (to be solved at a later
7861 *) Make it so the openssl application has all engines loaded by default.
7864 *) Constify the BIGNUM routines a little more.
7867 *) Add the following functions:
7869 ENGINE_load_cswift()
7871 ENGINE_load_atalla()
7873 ENGINE_load_builtin_engines()
7875 That way, an application can itself choose if external engines that
7876 are built-in in OpenSSL shall ever be used or not. The benefit is
7877 that applications won't have to be linked with libdl or other dso
7878 libraries unless it's really needed.
7880 Changed 'openssl engine' to load all engines on demand.
7881 Changed the engine header files to avoid the duplication of some
7882 declarations (they differed!).
7885 *) 'openssl engine' can now list capabilities.
7888 *) Better error reporting in 'openssl engine'.
7891 *) Never call load_dh_param(NULL) in s_server.
7894 *) Add engine application. It can currently list engines by name and
7895 identity, and test if they are actually available.
7898 *) Improve RPM specification file by forcing symbolic linking and making
7899 sure the installed documentation is also owned by root.root.
7900 [Damien Miller <djm@mindrot.org>]
7902 *) Give the OpenSSL applications more possibilities to make use of
7903 keys (public as well as private) handled by engines.
7906 *) Add OCSP code that comes from CertCo.
7909 *) Add VMS support for the Rijndael code.
7912 *) Added untested support for Nuron crypto accelerator.
7915 *) Add support for external cryptographic devices. This code was
7916 previously distributed separately as the "engine" branch.
7917 [Geoff Thorpe, Richard Levitte]
7919 *) Rework the filename-translation in the DSO code. It is now possible to
7920 have far greater control over how a "name" is turned into a filename
7921 depending on the operating environment and any oddities about the
7922 different shared library filenames on each system.
7925 *) Support threads on FreeBSD-elf in Configure.
7928 *) Fix for SHA1 assembly problem with MASM: it produces
7929 warnings about corrupt line number information when assembling
7930 with debugging information. This is caused by the overlapping
7932 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7935 NCONF_get_number() has no error checking at all. As a replacement,
7936 NCONF_get_number_e() is defined (_e for "error checking") and is
7937 promoted strongly. The old NCONF_get_number is kept around for
7938 binary backward compatibility.
7939 Make it possible for methods to load from something other than a BIO,
7940 by providing a function pointer that is given a name instead of a BIO.
7941 For example, this could be used to load configuration data from an
7945 *) Fix for non blocking accept BIOs. Added new I/O special reason
7946 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7947 with non blocking I/O was not possible because no retry code was
7948 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7952 *) Added the beginnings of Rijndael support.
7955 *) Fix for bug in DirectoryString mask setting. Add support for
7956 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7957 to allow certificate printing to more controllable, additional
7958 'certopt' option to 'x509' to allow new printing options to be
7962 *) Clean old EAY MD5 hack from e_os.h.
7965 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7967 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7968 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7969 [Joe Orton, Steve Henson]
7971 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7973 *) Fix additional bug revealed by the NISCC test suite:
7975 Stop bug triggering large recursion when presented with
7976 certain ASN.1 tags (CVE-2003-0851)
7979 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7981 *) Fix various bugs revealed by running the NISCC test suite:
7983 Stop out of bounds reads in the ASN1 code when presented with
7984 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7986 If verify callback ignores invalid public key errors don't try to check
7987 certificate signature with the NULL public key.
7991 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7992 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7996 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7997 extra data after the compression methods not only for TLS 1.0
7998 but also for SSL 3.0 (as required by the specification).
7999 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8001 *) Change X509_certificate_type() to mark the key as exported/exportable
8002 when it's 512 *bits* long, not 512 bytes.
8005 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8007 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8008 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8009 a protocol version number mismatch like a decryption error
8010 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8013 *) Turn on RSA blinding by default in the default implementation
8014 to avoid a timing attack. Applications that don't want it can call
8015 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8016 They would be ill-advised to do so in most cases.
8017 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8019 *) Change RSA blinding code so that it works when the PRNG is not
8020 seeded (in this case, the secret RSA exponent is abused as
8021 an unpredictable seed -- if it is not unpredictable, there
8022 is no point in blinding anyway). Make RSA blinding thread-safe
8023 by remembering the creator's thread ID in rsa->blinding and
8024 having all other threads use local one-time blinding factors
8025 (this requires more computation than sharing rsa->blinding, but
8026 avoids excessive locking; and if an RSA object is not shared
8027 between threads, blinding will still be very fast).
8030 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8032 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8033 via timing by performing a MAC computation even if incorrrect
8034 block cipher padding has been found. This is a countermeasure
8035 against active attacks where the attacker has to distinguish
8036 between bad padding and a MAC verification error. (CVE-2003-0078)
8038 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8039 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8040 Martin Vuagnoux (EPFL, Ilion)]
8042 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8044 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8045 memory from it's contents. This is done with a counter that will
8046 place alternating values in each byte. This can be used to solve
8047 two issues: 1) the removal of calls to memset() by highly optimizing
8048 compilers, and 2) cleansing with other values than 0, since those can
8049 be read through on certain media, for example a swap space on disk.
8052 *) Bugfix: client side session caching did not work with external caching,
8053 because the session->cipher setting was not restored when reloading
8054 from the external cache. This problem was masked, when
8055 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8056 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8059 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8060 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8061 [Zeev Lieber <zeev-l@yahoo.com>]
8063 *) Undo an undocumented change introduced in 0.9.6e which caused
8064 repeated calls to OpenSSL_add_all_ciphers() and
8065 OpenSSL_add_all_digests() to be ignored, even after calling
8069 *) Change the default configuration reader to deal with last line not
8070 being properly terminated.
8073 *) Change X509_NAME_cmp() so it applies the special rules on handling
8074 DN values that are of type PrintableString, as well as RDNs of type
8075 emailAddress where the value has the type ia5String.
8076 [stefank@valicert.com via Richard Levitte]
8078 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8079 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8080 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8081 the bitwise-OR of the two for use by the majority of applications
8082 wanting this behaviour, and update the docs. The documented
8083 behaviour and actual behaviour were inconsistent and had been
8084 changing anyway, so this is more a bug-fix than a behavioural
8086 [Geoff Thorpe, diagnosed by Nadav Har'El]
8088 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8089 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8092 *) Fix initialization code race conditions in
8093 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8094 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8095 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8096 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8097 ssl2_get_cipher_by_char(),
8098 ssl3_get_cipher_by_char().
8099 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8101 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8102 the cached sessions are flushed, as the remove_cb() might use ex_data
8103 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8104 (see [openssl.org #212]).
8105 [Geoff Thorpe, Lutz Jaenicke]
8107 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8108 length, instead of the encoding length to d2i_ASN1_OBJECT.
8111 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8113 *) [In 0.9.6g-engine release:]
8114 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8115 [Lynn Gazis <lgazis@rainbow.com>]
8117 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8119 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8120 and get fix the header length calculation.
8121 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8122 Alon Kantor <alonk@checkpoint.com> (and others),
8125 *) Use proper error handling instead of 'assertions' in buffer
8126 overflow checks added in 0.9.6e. This prevents DoS (the
8127 assertions could call abort()).
8128 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8130 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8132 *) Add various sanity checks to asn1_get_length() to reject
8133 the ASN1 length bytes if they exceed sizeof(long), will appear
8134 negative or the content length exceeds the length of the
8136 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8138 *) Fix cipher selection routines: ciphers without encryption had no flags
8139 for the cipher strength set and where therefore not handled correctly
8140 by the selection routines (PR #130).
8143 *) Fix EVP_dsa_sha macro.
8147 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8148 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8149 that was added in OpenSSL 0.9.6d.
8151 As the countermeasure turned out to be incompatible with some
8152 broken SSL implementations, the new option is part of SSL_OP_ALL.
8153 SSL_OP_ALL is usually employed when compatibility with weird SSL
8154 implementations is desired (e.g. '-bugs' option to 's_client' and
8155 's_server'), so the new option is automatically set in many
8159 *) Changes in security patch:
8161 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8162 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8163 Air Force Materiel Command, USAF, under agreement number
8166 *) Add various sanity checks to asn1_get_length() to reject
8167 the ASN1 length bytes if they exceed sizeof(long), will appear
8168 negative or the content length exceeds the length of the
8169 supplied buffer. (CVE-2002-0659)
8170 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8172 *) Assertions for various potential buffer overflows, not known to
8174 [Ben Laurie (CHATS)]
8176 *) Various temporary buffers to hold ASCII versions of integers were
8177 too small for 64 bit platforms. (CVE-2002-0655)
8178 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8180 *) Remote buffer overflow in SSL3 protocol - an attacker could
8181 supply an oversized session ID to a client. (CVE-2002-0656)
8182 [Ben Laurie (CHATS)]
8184 *) Remote buffer overflow in SSL2 protocol - an attacker could
8185 supply an oversized client master key. (CVE-2002-0656)
8186 [Ben Laurie (CHATS)]
8188 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8190 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8191 encoded as NULL) with id-dsa-with-sha1.
8192 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8194 *) Check various X509_...() return values in apps/req.c.
8195 [Nils Larsch <nla@trustcenter.de>]
8197 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8198 an end-of-file condition would erroneously be flagged, when the CRLF
8199 was just at the end of a processed block. The bug was discovered when
8200 processing data through a buffering memory BIO handing the data to a
8201 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8202 <ptsekov@syntrex.com> and Nedelcho Stanev.
8205 *) Implement a countermeasure against a vulnerability recently found
8206 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8207 before application data chunks to avoid the use of known IVs
8208 with data potentially chosen by the attacker.
8211 *) Fix length checks in ssl3_get_client_hello().
8214 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8215 to prevent ssl3_read_internal() from incorrectly assuming that
8216 ssl3_read_bytes() found application data while handshake
8217 processing was enabled when in fact s->s3->in_read_app_data was
8218 merely automatically cleared during the initial handshake.
8219 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8221 *) Fix object definitions for Private and Enterprise: they were not
8222 recognized in their shortname (=lowercase) representation. Extend
8223 obj_dat.pl to issue an error when using undefined keywords instead
8224 of silently ignoring the problem (Svenning Sorensen
8225 <sss@sss.dnsalias.net>).
8228 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8229 generators, i.e. generators other than 2 and 5. (Previously, the
8230 code did not properly initialise the 'add' and 'rem' values to
8231 BN_generate_prime().)
8233 In the new general case, we do not insist that 'generator' is
8234 actually a primitive root: This requirement is rather pointless;
8235 a generator of the order-q subgroup is just as good, if not
8239 *) Map new X509 verification errors to alerts. Discovered and submitted by
8240 Tom Wu <tom@arcot.com>.
8243 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8244 returning non-zero before the data has been completely received
8245 when using non-blocking I/O.
8246 [Bodo Moeller; problem pointed out by John Hughes]
8248 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8249 [Ben Laurie, Lutz Jaenicke]
8251 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8252 Yoram Zahavi <YoramZ@gilian.com>).
8255 *) Add information about CygWin 1.3 and on, and preserve proper
8256 configuration for the versions before that.
8257 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8259 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8260 check whether we deal with a copy of a session and do not delete from
8261 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8262 <izhar@checkpoint.com>.
8265 *) Do not store session data into the internal session cache, if it
8266 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8267 flag is set). Proposed by Aslam <aslam@funk.com>.
8270 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8274 *) [In 0.9.6d-engine release:]
8275 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8276 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8278 *) Add the configuration target linux-s390x.
8279 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8281 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8282 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8283 variable as an indication that a ClientHello message has been
8284 received. As the flag value will be lost between multiple
8285 invocations of ssl3_accept when using non-blocking I/O, the
8286 function may not be aware that a handshake has actually taken
8287 place, thus preventing a new session from being added to the
8290 To avoid this problem, we now set s->new_session to 2 instead of
8291 using a local variable.
8292 [Lutz Jaenicke, Bodo Moeller]
8294 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8295 if the SSL_R_LENGTH_MISMATCH error is detected.
8296 [Geoff Thorpe, Bodo Moeller]
8298 *) New 'shared_ldflag' column in Configure platform table.
8301 *) Fix EVP_CIPHER_mode macro.
8302 ["Dan S. Camper" <dan@bti.net>]
8304 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8305 type, we must throw them away by setting rr->length to 0.
8306 [D P Chang <dpc@qualys.com>]
8308 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8310 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8311 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8312 worked incorrectly for those cases where range = 10..._2 and
8313 3*range is two bits longer than range.)
8316 *) Only add signing time to PKCS7 structures if it is not already
8320 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8321 OBJ_ld_ce should be OBJ_id_ce.
8322 Also some ip-pda OIDs in crypto/objects/objects.txt were
8323 incorrect (cf. RFC 3039).
8324 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8326 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8327 returns early because it has nothing to do.
8328 [Andy Schneider <andy.schneider@bjss.co.uk>]
8330 *) [In 0.9.6c-engine release:]
8331 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8332 [Andy Schneider <andy.schneider@bjss.co.uk>]
8334 *) [In 0.9.6c-engine release:]
8335 Add support for Cryptographic Appliance's keyserver technology.
8336 (Use engine 'keyclient')
8337 [Cryptographic Appliances and Geoff Thorpe]
8339 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8340 is called via tools/c89.sh because arguments have to be
8341 rearranged (all '-L' options must appear before the first object
8343 [Richard Shapiro <rshapiro@abinitio.com>]
8345 *) [In 0.9.6c-engine release:]
8346 Add support for Broadcom crypto accelerator cards, backported
8348 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8350 *) [In 0.9.6c-engine release:]
8351 Add support for SureWare crypto accelerator cards from
8352 Baltimore Technologies. (Use engine 'sureware')
8353 [Baltimore Technologies and Mark Cox]
8355 *) [In 0.9.6c-engine release:]
8356 Add support for crypto accelerator cards from Accelerated
8357 Encryption Processing, www.aep.ie. (Use engine 'aep')
8358 [AEP Inc. and Mark Cox]
8360 *) Add a configuration entry for gcc on UnixWare.
8361 [Gary Benson <gbenson@redhat.com>]
8363 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8364 messages are stored in a single piece (fixed-length part and
8365 variable-length part combined) and fix various bugs found on the way.
8368 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8369 instead. BIO_gethostbyname() does not know what timeouts are
8370 appropriate, so entries would stay in cache even when they have
8372 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8374 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8375 faced with a pathologically small ClientHello fragment that does
8376 not contain client_version: Instead of aborting with an error,
8377 simply choose the highest available protocol version (i.e.,
8378 TLS 1.0 unless it is disabled). In practice, ClientHello
8379 messages are never sent like this, but this change gives us
8380 strictly correct behaviour at least for TLS.
8383 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8384 never resets s->method to s->ctx->method when called from within
8385 one of the SSL handshake functions.
8386 [Bodo Moeller; problem pointed out by Niko Baric]
8388 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8389 (sent using the client's version number) if client_version is
8390 smaller than the protocol version in use. Also change
8391 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8392 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8393 the client will at least see that alert.
8396 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8400 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8401 client receives HelloRequest while in a handshake.
8402 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8404 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8405 should end in 'break', not 'goto end' which circumvents various
8406 cleanups done in state SSL_ST_OK. But session related stuff
8407 must be disabled for SSL_ST_OK in the case that we just sent a
8410 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8411 before just sending a HelloRequest.
8412 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8414 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8415 reveal whether illegal block cipher padding was found or a MAC
8416 verification error occurred. (Neither SSLerr() codes nor alerts
8417 are directly visible to potential attackers, but the information
8418 may leak via logfiles.)
8420 Similar changes are not required for the SSL 2.0 implementation
8421 because the number of padding bytes is sent in clear for SSL 2.0,
8422 and the extra bytes are just ignored. However ssl/s2_pkt.c
8423 failed to verify that the purported number of padding bytes is in
8427 *) Add OpenUNIX-8 support including shared libraries
8428 (Boyd Lynn Gerber <gerberb@zenez.com>).
8431 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8432 'wristwatch attack' using huge encoding parameters (cf.
8433 James H. Manger's CRYPTO 2001 paper). Note that the
8434 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8435 encoding parameters and hence was not vulnerable.
8438 *) BN_sqr() bug fix.
8439 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8441 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8442 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8443 followed by modular reduction.
8444 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8446 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8447 equivalent based on BN_pseudo_rand() instead of BN_rand().
8450 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8451 This function was broken, as the check for a new client hello message
8452 to handle SGC did not allow these large messages.
8453 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8456 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8459 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8460 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8463 *) Rework the configuration and shared library support for Tru64 Unix.
8464 The configuration part makes use of modern compiler features and
8465 still retains old compiler behavior for those that run older versions
8466 of the OS. The shared library support part includes a variant that
8467 uses the RPATH feature, and is available through the special
8468 configuration target "alpha-cc-rpath", which will never be selected
8470 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8472 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8473 with the same message size as in ssl3_get_certificate_request().
8474 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8475 messages might inadvertently be reject as too long.
8476 [Petr Lampa <lampa@fee.vutbr.cz>]
8478 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8481 *) Modified SSL library such that the verify_callback that has been set
8482 specificly for an SSL object with SSL_set_verify() is actually being
8483 used. Before the change, a verify_callback set with this function was
8484 ignored and the verify_callback() set in the SSL_CTX at the time of
8485 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8486 to allow the necessary settings.
8489 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8490 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8491 done automatically (in contradiction to the requirements of the C
8492 standard). This made problems when used from OpenSSH.
8495 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8496 dh->length and always used
8498 BN_rand_range(priv_key, dh->p).
8500 BN_rand_range() is not necessary for Diffie-Hellman, and this
8501 specific range makes Diffie-Hellman unnecessarily inefficient if
8502 dh->length (recommended exponent length) is much smaller than the
8503 length of dh->p. We could use BN_rand_range() if the order of
8504 the subgroup was stored in the DH structure, but we only have
8509 BN_rand(priv_key, l, ...)
8511 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8517 RSA_eay_public_encrypt
8518 RSA_eay_private_decrypt
8519 RSA_eay_private_encrypt (signing)
8520 RSA_eay_public_decrypt (signature verification)
8522 (default implementations for RSA_public_encrypt,
8523 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8524 always reject numbers >= n.
8527 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8528 to synchronize access to 'locking_thread'. This is necessary on
8529 systems where access to 'locking_thread' (an 'unsigned long'
8530 variable) is not atomic.
8533 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8534 *before* setting the 'crypto_lock_rand' flag. The previous code had
8535 a race condition if 0 is a valid thread ID.
8536 [Travis Vitek <vitek@roguewave.com>]
8538 *) Add support for shared libraries under Irix.
8539 [Albert Chin-A-Young <china@thewrittenword.com>]
8541 *) Add configuration option to build on Linux on both big-endian and
8543 [Ralf Baechle <ralf@uni-koblenz.de>]
8545 *) Add the possibility to create shared libraries on HP-UX.
8548 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8550 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8551 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8552 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8553 PRNG state recovery was possible based on the output of
8554 one PRNG request appropriately sized to gain knowledge on
8555 'md' followed by enough consecutive 1-byte PRNG requests
8556 to traverse all of 'state'.
8558 1. When updating 'md_local' (the current thread's copy of 'md')
8559 during PRNG output generation, hash all of the previous
8560 'md_local' value, not just the half used for PRNG output.
8562 2. Make the number of bytes from 'state' included into the hash
8563 independent from the number of PRNG bytes requested.
8565 The first measure alone would be sufficient to avoid
8566 Markku-Juhani's attack. (Actually it had never occurred
8567 to me that the half of 'md_local' used for chaining was the
8568 half from which PRNG output bytes were taken -- I had always
8569 assumed that the secret half would be used.) The second
8570 measure makes sure that additional data from 'state' is never
8571 mixed into 'md_local' in small portions; this heuristically
8572 further strengthens the PRNG.
8575 *) Fix crypto/bn/asm/mips3.s.
8578 *) When only the key is given to "enc", the IV is undefined. Print out
8579 an error message in this case.
8582 *) Handle special case when X509_NAME is empty in X509 printing routines.
8585 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8586 positive and less than q.
8589 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8590 used: it isn't thread safe and the add_lock_callback should handle
8592 [Paul Rose <Paul.Rose@bridge.com>]
8594 *) Verify that incoming data obeys the block size in
8595 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8599 [Ulf Möller, Bodo Möller]
8601 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8602 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8603 when fixing the server behaviour for backwards-compatible 'client
8604 hello' messages. (Note that the attack is impractical against
8605 SSL 3.0 and TLS 1.0 anyway because length and version checking
8606 means that the probability of guessing a valid ciphertext is
8607 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8610 Before 0.9.5, the countermeasure (hide the error by generating a
8611 random 'decryption result') did not work properly because
8612 ERR_clear_error() was missing, meaning that SSL_get_error() would
8613 detect the supposedly ignored error.
8615 Both problems are now fixed.
8618 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8619 (previously it was 1024).
8622 *) Fix for compatibility mode trust settings: ignore trust settings
8623 unless some valid trust or reject settings are present.
8626 *) Fix for blowfish EVP: its a variable length cipher.
8629 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8630 parameters in DSA public key structures and return an error in the
8631 DSA routines if parameters are absent.
8634 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8635 in the current directory if neither $RANDFILE nor $HOME was set.
8636 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8637 caused some confusion to Windows users who haven't defined $HOME.
8638 Thus RAND_file_name() is changed again: e_os.h can define a
8639 DEFAULT_HOME, which will be used if $HOME is not set.
8640 For Windows, we use "C:"; on other platforms, we still require
8641 environment variables.
8643 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8644 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8645 having multiple threads call RAND_poll() concurrently.
8648 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8649 combination of a flag and a thread ID variable.
8650 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8651 flag), *other* threads can enter ssleay_add_bytes without obeying
8652 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8653 that they do not hold after the first thread unsets add_do_not_lock).
8656 *) Change bctest again: '-x' expressions are not available in all
8660 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8662 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8663 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8665 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8666 the default extension for executables, if any. Also, make the perl
8667 scripts that use symlink() to test if it really exists and use "cp"
8668 if it doesn't. All this made OpenSSL compilable and installable in
8672 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8673 If SEQUENCE is length is indefinite just set c->slen to the total
8674 amount of data available.
8675 [Steve Henson, reported by shige@FreeBSD.org]
8676 [This change does not apply to 0.9.7.]
8678 *) Change bctest to avoid here-documents inside command substitution
8679 (workaround for FreeBSD /bin/sh bug).
8680 For compatibility with Ultrix, avoid shell functions (introduced
8681 in the bctest version that searches along $PATH).
8684 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8685 with des_encrypt() defined on some operating systems, like Solaris
8689 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8690 On the Importance of Eliminating Errors in Cryptographic
8691 Computations, J. Cryptology 14 (2001) 2, 101-119,
8692 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8695 *) MIPS assembler BIGNUM division bug fix.
8698 *) Disabled incorrect Alpha assembler code.
8701 *) Fix PKCS#7 decode routines so they correctly update the length
8702 after reading an EOC for the EXPLICIT tag.
8704 [This change does not apply to 0.9.7.]
8706 *) Fix bug in PKCS#12 key generation routines. This was triggered
8707 if a 3DES key was generated with a 0 initial byte. Include
8708 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8709 (but broken) behaviour.
8712 *) Enhance bctest to search for a working bc along $PATH and print
8714 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8716 *) Fix memory leaks in err.c: free err_data string if necessary;
8717 don't write to the wrong index in ERR_set_error_data.
8720 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8724 *) Replace rdtsc with _emit statements for VC++ version 5.
8725 [Jeremy Cooper <jeremy@baymoo.org>]
8727 *) Make it possible to reuse SSLv2 sessions.
8730 *) In copy_email() check for >= 0 as a return value for
8731 X509_NAME_get_index_by_NID() since 0 is a valid index.
8732 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8734 *) Avoid coredump with unsupported or invalid public keys by checking if
8735 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8736 PKCS7_verify() fails with non detached data.
8739 *) Don't use getenv in library functions when run as setuid/setgid.
8740 New function OPENSSL_issetugid().
8743 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8744 due to incorrect handling of multi-threading:
8746 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8748 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8750 3. Count how many times MemCheck_off() has been called so that
8751 nested use can be treated correctly. This also avoids
8752 inband-signalling in the previous code (which relied on the
8753 assumption that thread ID 0 is impossible).
8756 *) Add "-rand" option also to s_client and s_server.
8759 *) Fix CPU detection on Irix 6.x.
8760 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8761 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8763 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8766 [This change does not apply to 0.9.7.]
8768 *) Use the cached encoding of an X509_NAME structure rather than
8769 copying it. This is apparently the reason for the libsafe "errors"
8770 but the code is actually correct.
8773 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8774 Bleichenbacher's DSA attack.
8775 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8776 to be set and top=0 forces the highest bit to be set; top=-1 is new
8777 and leaves the highest bit random.
8778 [Ulf Moeller, Bodo Moeller]
8780 *) In the NCONF_...-based implementations for CONF_... queries
8781 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8782 a temporary CONF structure with the data component set to NULL
8783 (which gives segmentation faults in lh_retrieve).
8784 Instead, use NULL for the CONF pointer in CONF_get_string and
8785 CONF_get_number (which may use environment variables) and directly
8786 return NULL from CONF_get_section.
8789 *) Fix potential buffer overrun for EBCDIC.
8792 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8793 keyUsage if basicConstraints absent for a CA.
8796 *) Make SMIME_write_PKCS7() write mail header values with a format that
8797 is more generally accepted (no spaces before the semicolon), since
8798 some programs can't parse those values properly otherwise. Also make
8799 sure BIO's that break lines after each write do not create invalid
8803 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8804 macros previously used would not encode an empty SEQUENCE OF
8805 and break the signature.
8807 [This change does not apply to 0.9.7.]
8809 *) Zero the premaster secret after deriving the master secret in
8813 *) Add some EVP_add_digest_alias registrations (as found in
8814 OpenSSL_add_all_digests()) to SSL_library_init()
8815 aka OpenSSL_add_ssl_algorithms(). This provides improved
8816 compatibility with peers using X.509 certificates
8817 with unconventional AlgorithmIdentifier OIDs.
8820 *) Fix for Irix with NO_ASM.
8821 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8823 *) ./config script fixes.
8824 [Ulf Moeller, Richard Levitte]
8826 *) Fix 'openssl passwd -1'.
8829 *) Change PKCS12_key_gen_asc() so it can cope with non null
8830 terminated strings whose length is passed in the passlen
8831 parameter, for example from PEM callbacks. This was done
8832 by adding an extra length parameter to asc2uni().
8833 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8835 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8836 call failed, free the DSA structure.
8839 *) Fix to uni2asc() to cope with zero length Unicode strings.
8840 These are present in some PKCS#12 files.
8843 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8844 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8845 when writing a 32767 byte record.
8846 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8848 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8849 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8851 (RSA objects have a reference count access to which is protected
8852 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8853 so they are meant to be shared between threads.)
8854 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8855 "Reddie, Steven" <Steven.Reddie@ca.com>]
8857 *) Fix a deadlock in CRYPTO_mem_leaks().
8860 *) Use better test patterns in bntest.
8863 *) rand_win.c fix for Borland C.
8866 *) BN_rshift bugfix for n == 0.
8869 *) Add a 'bctest' script that checks for some known 'bc' bugs
8870 so that 'make test' does not abort just because 'bc' is broken.
8873 *) Store verify_result within SSL_SESSION also for client side to
8874 avoid potential security hole. (Re-used sessions on the client side
8875 always resulted in verify_result==X509_V_OK, not using the original
8876 result of the server certificate verification.)
8879 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8880 SSL3_RT_APPLICATION_DATA, return 0.
8881 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8885 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8886 releases, have been re-implemented by renaming the previous
8887 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8888 and ssl3_read_internal, respectively, and adding 'peek' parameters
8889 to them. The new ssl[23]_{read,peek} functions are calls to
8890 ssl[23]_read_internal with the 'peek' flag set appropriately.
8891 A 'peek' parameter has also been added to ssl3_read_bytes, which
8892 does the actual work for ssl3_read_internal.
8895 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8896 the method-specific "init()" handler. Also clean up ex_data after
8897 calling the method-specific "finish()" handler. Previously, this was
8898 happening the other way round.
8901 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8902 The previous value, 12, was not always sufficient for BN_mod_exp().
8905 *) Make sure that shared libraries get the internal name engine with
8906 the full version number and not just 0. This should mark the
8907 shared libraries as not backward compatible. Of course, this should
8908 be changed again when we can guarantee backward binary compatibility.
8911 *) Fix typo in get_cert_by_subject() in by_dir.c
8912 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8914 *) Rework the system to generate shared libraries:
8916 - Make note of the expected extension for the shared libraries and
8917 if there is a need for symbolic links from for example libcrypto.so.0
8918 to libcrypto.so.0.9.7. There is extended info in Configure for
8921 - Make as few rebuilds of the shared libraries as possible.
8923 - Still avoid linking the OpenSSL programs with the shared libraries.
8925 - When installing, install the shared libraries separately from the
8929 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8931 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8932 and not in SSL_clear because the latter is also used by the
8933 accept/connect functions; previously, the settings made by
8934 SSL_set_read_ahead would be lost during the handshake.
8935 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8937 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8938 Previously, it would create entries for disabled algorithms no
8942 *) Added several new manual pages for SSL_* function.
8945 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8947 *) In ssl23_get_client_hello, generate an error message when faced
8948 with an initial SSL 3.0/TLS record that is too small to contain the
8949 first two bytes of the ClientHello message, i.e. client_version.
8950 (Note that this is a pathologic case that probably has never happened
8951 in real life.) The previous approach was to use the version number
8952 from the record header as a substitute; but our protocol choice
8953 should not depend on that one because it is not authenticated
8954 by the Finished messages.
8957 *) More robust randomness gathering functions for Windows.
8958 [Jeffrey Altman <jaltman@columbia.edu>]
8960 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8961 not set then we don't setup the error code for issuer check errors
8962 to avoid possibly overwriting other errors which the callback does
8963 handle. If an application does set the flag then we assume it knows
8964 what it is doing and can handle the new informational codes
8968 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8969 a general "ANY" type, as such it should be able to decode anything
8970 including tagged types. However it didn't check the class so it would
8971 wrongly interpret tagged types in the same way as their universal
8972 counterpart and unknown types were just rejected. Changed so that the
8973 tagged and unknown types are handled in the same way as a SEQUENCE:
8974 that is the encoding is stored intact. There is also a new type
8975 "V_ASN1_OTHER" which is used when the class is not universal, in this
8976 case we have no idea what the actual type is so we just lump them all
8980 *) On VMS, stdout may very well lead to a file that is written to
8981 in a record-oriented fashion. That means that every write() will
8982 write a separate record, which will be read separately by the
8983 programs trying to read from it. This can be very confusing.
8985 The solution is to put a BIO filter in the way that will buffer
8986 text until a linefeed is reached, and then write everything a
8987 line at a time, so every record written will be an actual line,
8988 not chunks of lines and not (usually doesn't happen, but I've
8989 seen it once) several lines in one record. BIO_f_linebuffer() is
8992 Currently, it's a VMS-only method, because that's where it has
8993 been tested well enough.
8996 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8997 it can return incorrect results.
8998 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8999 but it was in 0.9.6-beta[12].)
9002 *) Disable the check for content being present when verifying detached
9003 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9004 include zero length content when signing messages.
9007 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9008 BIO_ctrl (for BIO pairs).
9011 *) Add DSO method for VMS.
9014 *) Bug fix: Montgomery multiplication could produce results with the
9018 *) Add RPM specification openssl.spec and modify it to build three
9019 packages. The default package contains applications, application
9020 documentation and run-time libraries. The devel package contains
9021 include files, static libraries and function documentation. The
9022 doc package contains the contents of the doc directory. The original
9023 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9026 *) Add a large number of documentation files for many SSL routines.
9027 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9029 *) Add a configuration entry for Sony News 4.
9030 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9032 *) Don't set the two most significant bits to one when generating a
9033 random number < q in the DSA library.
9036 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9037 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9038 the underlying transport is blocking) if a handshake took place.
9039 (The default behaviour is needed by applications such as s_client
9040 and s_server that use select() to determine when to use SSL_read;
9041 but for applications that know in advance when to expect data, it
9042 just makes things more complicated.)
9045 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9049 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9050 work better on such systems.
9051 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9053 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9054 Update PKCS12_parse() so it copies the friendlyName and the
9055 keyid to the certificates aux info.
9058 *) Fix bug in PKCS7_verify() which caused an infinite loop
9059 if there was more than one signature.
9060 [Sven Uszpelkat <su@celocom.de>]
9062 *) Major change in util/mkdef.pl to include extra information
9063 about each symbol, as well as presenting variables as well
9064 as functions. This change means that there's n more need
9065 to rebuild the .num files when some algorithms are excluded.
9068 *) Allow the verify time to be set by an application,
9069 rather than always using the current time.
9072 *) Phase 2 verify code reorganisation. The certificate
9073 verify code now looks up an issuer certificate by a
9074 number of criteria: subject name, authority key id
9075 and key usage. It also verifies self signed certificates
9076 by the same criteria. The main comparison function is
9077 X509_check_issued() which performs these checks.
9079 Lot of changes were necessary in order to support this
9080 without completely rewriting the lookup code.
9082 Authority and subject key identifier are now cached.
9084 The LHASH 'certs' is X509_STORE has now been replaced
9085 by a STACK_OF(X509_OBJECT). This is mainly because an
9086 LHASH can't store or retrieve multiple objects with
9087 the same hash value.
9089 As a result various functions (which were all internal
9090 use only) have changed to handle the new X509_STORE
9091 structure. This will break anything that messed round
9092 with X509_STORE internally.
9094 The functions X509_STORE_add_cert() now checks for an
9095 exact match, rather than just subject name.
9097 The X509_STORE API doesn't directly support the retrieval
9098 of multiple certificates matching a given criteria, however
9099 this can be worked round by performing a lookup first
9100 (which will fill the cache with candidate certificates)
9101 and then examining the cache for matches. This is probably
9102 the best we can do without throwing out X509_LOOKUP
9103 entirely (maybe later...).
9105 The X509_VERIFY_CTX structure has been enhanced considerably.
9107 All certificate lookup operations now go via a get_issuer()
9108 callback. Although this currently uses an X509_STORE it
9109 can be replaced by custom lookups. This is a simple way
9110 to bypass the X509_STORE hackery necessary to make this
9111 work and makes it possible to use more efficient techniques
9112 in future. A very simple version which uses a simple
9113 STACK for its trusted certificate store is also provided
9114 using X509_STORE_CTX_trusted_stack().
9116 The verify_cb() and verify() callbacks now have equivalents
9117 in the X509_STORE_CTX structure.
9119 X509_STORE_CTX also has a 'flags' field which can be used
9120 to customise the verify behaviour.
9123 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9124 excludes S/MIME capabilities.
9127 *) When a certificate request is read in keep a copy of the
9128 original encoding of the signed data and use it when outputting
9129 again. Signatures then use the original encoding rather than
9130 a decoded, encoded version which may cause problems if the
9131 request is improperly encoded.
9134 *) For consistency with other BIO_puts implementations, call
9135 buffer_write(b, ...) directly in buffer_puts instead of calling
9138 In BIO_puts, increment b->num_write as in BIO_write.
9139 [Peter.Sylvester@EdelWeb.fr]
9141 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9142 BN_zero, we may not return a BIGNUM with an array consisting of
9146 *) Avoid calling abort() from within the library when problems are
9147 detected, except if preprocessor symbols have been defined
9148 (such as REF_CHECK, BN_DEBUG etc.).
9151 *) New openssl application 'rsautl'. This utility can be
9152 used for low level RSA operations. DER public key
9153 BIO/fp routines also added.
9156 *) New Configure entry and patches for compiling on QNX 4.
9157 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9159 *) A demo state-machine implementation was sponsored by
9160 Nuron (http://www.nuron.com/) and is now available in
9161 demos/state_machine.
9164 *) New options added to the 'dgst' utility for signature
9165 generation and verification.
9168 *) Unrecognized PKCS#7 content types are now handled via a
9169 catch all ASN1_TYPE structure. This allows unsupported
9170 types to be stored as a "blob" and an application can
9171 encode and decode it manually.
9174 *) Fix various signed/unsigned issues to make a_strex.c
9176 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9178 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9179 length if passed a buffer. ASN1_INTEGER_to_BN failed
9180 if passed a NULL BN and its argument was negative.
9181 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9183 *) Modification to PKCS#7 encoding routines to output definite
9184 length encoding. Since currently the whole structures are in
9185 memory there's not real point in using indefinite length
9186 constructed encoding. However if OpenSSL is compiled with
9187 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9190 *) Added BIO_vprintf() and BIO_vsnprintf().
9193 *) Added more prefixes to parse for in the the strings written
9194 through a logging bio, to cover all the levels that are available
9195 through syslog. The prefixes are now:
9197 PANIC, EMERG, EMR => LOG_EMERG
9198 ALERT, ALR => LOG_ALERT
9199 CRIT, CRI => LOG_CRIT
9200 ERROR, ERR => LOG_ERR
9201 WARNING, WARN, WAR => LOG_WARNING
9202 NOTICE, NOTE, NOT => LOG_NOTICE
9203 INFO, INF => LOG_INFO
9204 DEBUG, DBG => LOG_DEBUG
9206 and as before, if none of those prefixes are present at the
9207 beginning of the string, LOG_ERR is chosen.
9209 On Win32, the LOG_* levels are mapped according to this:
9211 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9212 LOG_WARNING => EVENTLOG_WARNING_TYPE
9213 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9217 *) Made it possible to reconfigure with just the configuration
9218 argument "reconf" or "reconfigure". The command line arguments
9219 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9220 and are retrieved from there when reconfiguring.
9224 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9226 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9229 *) The obj_dat.pl script was messing up the sorting of object
9230 names. The reason was that it compared the quoted version
9231 of strings as a result "OCSP" > "OCSP Signing" because
9232 " > SPACE. Changed script to store unquoted versions of
9233 names and add quotes on output. It was also omitting some
9234 names from the lookup table if they were given a default
9235 value (that is if SN is missing it is given the same
9236 value as LN and vice versa), these are now added on the
9237 grounds that if an object has a name we should be able to
9238 look it up. Finally added warning output when duplicate
9239 short or long names are found.
9242 *) Changes needed for Tandem NSK.
9243 [Scott Uroff <scott@xypro.com>]
9245 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9246 RSA_padding_check_SSLv23(), special padding was never detected
9247 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9248 version rollback attacks was not effective.
9250 In s23_clnt.c, don't use special rollback-attack detection padding
9251 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9252 client; similarly, in s23_srvr.c, don't do the rollback check if
9253 SSL 2.0 is the only protocol enabled in the server.
9256 *) Make it possible to get hexdumps of unprintable data with 'openssl
9257 asn1parse'. By implication, the functions ASN1_parse_dump() and
9258 BIO_dump_indent() are added.
9261 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9262 these print out strings and name structures based on various
9263 flags including RFC2253 support and proper handling of
9264 multibyte characters. Added options to the 'x509' utility
9265 to allow the various flags to be set.
9268 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9269 Also change the functions X509_cmp_current_time() and
9270 X509_gmtime_adj() work with an ASN1_TIME structure,
9271 this will enable certificates using GeneralizedTime in validity
9272 dates to be checked.
9275 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9276 negative public key encodings) on by default,
9277 NO_NEG_PUBKEY_BUG can be set to disable it.
9280 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9281 content octets. An i2c_ASN1_OBJECT is unnecessary because
9282 the encoding can be trivially obtained from the structure.
9285 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9286 not read locks (CRYPTO_r_[un]lock).
9289 *) A first attempt at creating official support for shared
9290 libraries through configuration. I've kept it so the
9291 default is static libraries only, and the OpenSSL programs
9292 are always statically linked for now, but there are
9293 preparations for dynamic linking in place.
9294 This has been tested on Linux and Tru64.
9297 *) Randomness polling function for Win9x, as described in:
9298 Peter Gutmann, Software Generation of Practically Strong
9302 *) Fix so PRNG is seeded in req if using an already existing
9306 *) New options to smime application. -inform and -outform
9307 allow alternative formats for the S/MIME message including
9308 PEM and DER. The -content option allows the content to be
9309 specified separately. This should allow things like Netscape
9310 form signing output easier to verify.
9313 *) Fix the ASN1 encoding of tags using the 'long form'.
9316 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9317 STRING types. These convert content octets to and from the
9318 underlying type. The actual tag and length octets are
9319 already assumed to have been read in and checked. These
9320 are needed because all other string types have virtually
9321 identical handling apart from the tag. By having versions
9322 of the ASN1 functions that just operate on content octets
9323 IMPLICIT tagging can be handled properly. It also allows
9324 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9325 and ASN1_INTEGER are identical apart from the tag.
9328 *) Change the handling of OID objects as follows:
9330 - New object identifiers are inserted in objects.txt, following
9331 the syntax given in objects.README.
9332 - objects.pl is used to process obj_mac.num and create a new
9334 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9337 This is currently kind of a hack, and the perl code in objects.pl
9338 isn't very elegant, but it works as I intended. The simplest way
9339 to check that it worked correctly is to look in obj_dat.h and
9340 check the array nid_objs and make sure the objects haven't moved
9341 around (this is important!). Additions are OK, as well as
9342 consistent name changes.
9345 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9348 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9349 The given file adds to whatever has already been seeded into the
9350 random pool through the RANDFILE configuration file option or
9351 environment variable, or the default random state file.
9354 *) mkstack.pl now sorts each macro group into lexical order.
9355 Previously the output order depended on the order the files
9356 appeared in the directory, resulting in needless rewriting
9360 *) Patches to make OpenSSL compile under Win32 again. Mostly
9361 work arounds for the VC++ problem that it treats func() as
9362 func(void). Also stripped out the parts of mkdef.pl that
9363 added extra typesafe functions: these no longer exist.
9366 *) Reorganisation of the stack code. The macros are now all
9367 collected in safestack.h . Each macro is defined in terms of
9368 a "stack macro" of the form SKM_<name>(type, a, b). The
9369 DEBUG_SAFESTACK is now handled in terms of function casts,
9370 this has the advantage of retaining type safety without the
9371 use of additional functions. If DEBUG_SAFESTACK is not defined
9372 then the non typesafe macros are used instead. Also modified the
9373 mkstack.pl script to handle the new form. Needs testing to see
9374 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9375 the default if no major problems. Similar behaviour for ASN1_SET_OF
9376 and PKCS12_STACK_OF.
9379 *) When some versions of IIS use the 'NET' form of private key the
9380 key derivation algorithm is different. Normally MD5(password) is
9381 used as a 128 bit RC4 key. In the modified case
9382 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9383 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9384 as the old Netscape_RSA functions except they have an additional
9385 'sgckey' parameter which uses the modified algorithm. Also added
9386 an -sgckey command line option to the rsa utility. Thanks to
9387 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9388 algorithm to openssl-dev.
9391 *) The evp_local.h macros were using 'c.##kname' which resulted in
9392 invalid expansion on some systems (SCO 5.0.5 for example).
9393 Corrected to 'c.kname'.
9394 [Phillip Porch <root@theporch.com>]
9396 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9397 a STACK of email addresses from a certificate or request, these look
9398 in the subject name and the subject alternative name extensions and
9399 omit any duplicate addresses.
9402 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9403 This makes DSA verification about 2 % faster.
9406 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9407 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9408 plus overhead for 1024 bit moduli).
9409 This makes exponentiations about 0.5 % faster for 1024 bit
9410 exponents (as measured by "openssl speed rsa2048").
9413 *) Rename memory handling macros to avoid conflicts with other
9415 Malloc => OPENSSL_malloc
9416 Malloc_locked => OPENSSL_malloc_locked
9417 Realloc => OPENSSL_realloc
9418 Free => OPENSSL_free
9421 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9422 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9425 *) CygWin32 support.
9426 [John Jarvie <jjarvie@newsguy.com>]
9428 *) The type-safe stack code has been rejigged. It is now only compiled
9429 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9430 by default all type-specific stack functions are "#define"d back to
9431 standard stack functions. This results in more streamlined output
9432 but retains the type-safety checking possibilities of the original
9436 *) The STACK code has been cleaned up, and certain type declarations
9437 that didn't make a lot of sense have been brought in line. This has
9438 also involved a cleanup of sorts in safestack.h to more correctly
9439 map type-safe stack functions onto their plain stack counterparts.
9440 This work has also resulted in a variety of "const"ifications of
9441 lots of the code, especially "_cmp" operations which should normally
9442 be prototyped with "const" parameters anyway.
9445 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9446 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9447 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9448 where all of 'md' is used each time the PRNG is used, but 'state'
9449 is used only indexed by a cyclic counter. As entropy may not be
9450 well distributed from the beginning, 'md' is important as a
9451 chaining variable. However, the output function chains only half
9452 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9453 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9454 in all of 'state' being rewritten, with the new values depending
9455 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9458 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9459 the handshake is continued after ssl_verify_cert_chain();
9460 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9461 can lead to 'unexplainable' connection aborts later.
9462 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9464 *) Major EVP API cipher revision.
9465 Add hooks for extra EVP features. This allows various cipher
9466 parameters to be set in the EVP interface. Support added for variable
9467 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9468 setting of RC2 and RC5 parameters.
9470 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9473 Remove lots of duplicated code from the EVP library. For example *every*
9474 cipher init() function handles the 'iv' in the same way according to the
9475 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9476 for CFB and OFB modes they zero ctx->num.
9478 New functionality allows removal of S/MIME code RC2 hack.
9480 Most of the routines have the same form and so can be declared in terms
9483 By shifting this to the top level EVP_CipherInit() it can be removed from
9484 all individual ciphers. If the cipher wants to handle IVs or keys
9485 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9488 Change lots of functions like EVP_EncryptUpdate() to now return a
9489 value: although software versions of the algorithms cannot fail
9490 any installed hardware versions can.
9493 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9494 this option is set, tolerate broken clients that send the negotiated
9495 protocol version number instead of the requested protocol version
9499 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9500 i.e. non-zero for export ciphersuites, zero otherwise.
9501 Previous versions had this flag inverted, inconsistent with
9502 rsa_tmp_cb (..._TMP_RSA_CB).
9503 [Bodo Moeller; problem reported by Amit Chopra]
9505 *) Add missing DSA library text string. Work around for some IIS
9506 key files with invalid SEQUENCE encoding.
9509 *) Add a document (doc/standards.txt) that list all kinds of standards
9510 and so on that are implemented in OpenSSL.
9513 *) Enhance c_rehash script. Old version would mishandle certificates
9514 with the same subject name hash and wouldn't handle CRLs at all.
9515 Added -fingerprint option to crl utility, to support new c_rehash
9519 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9522 *) Fix for SSL server purpose checking. Server checking was
9523 rejecting certificates which had extended key usage present
9524 but no ssl client purpose.
9525 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9527 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9528 is a little unclear about how a blank password is handled.
9529 Since the password in encoded as a BMPString with terminating
9530 double NULL a zero length password would end up as just the
9531 double NULL. However no password at all is different and is
9532 handled differently in the PKCS#12 key generation code. NS
9533 treats a blank password as zero length. MSIE treats it as no
9534 password on export: but it will try both on import. We now do
9535 the same: PKCS12_parse() tries zero length and no password if
9536 the password is set to "" or NULL (NULL is now a valid password:
9537 it wasn't before) as does the pkcs12 application.
9540 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9541 perror when PEM_read_bio_X509_REQ fails, the error message must
9542 be obtained from the error queue.
9545 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9546 it in ERR_remove_state if appropriate, and change ERR_get_state
9547 accordingly to avoid race conditions (this is necessary because
9548 thread_hash is no longer constant once set).
9551 *) Bugfix for linux-elf makefile.one.
9554 *) RSA_get_default_method() will now cause a default
9555 RSA_METHOD to be chosen if one doesn't exist already.
9556 Previously this was only set during a call to RSA_new()
9557 or RSA_new_method(NULL) meaning it was possible for
9558 RSA_get_default_method() to return NULL.
9561 *) Added native name translation to the existing DSO code
9562 that will convert (if the flag to do so is set) filenames
9563 that are sufficiently small and have no path information
9564 into a canonical native form. Eg. "blah" converted to
9565 "libblah.so" or "blah.dll" etc.
9568 *) New function ERR_error_string_n(e, buf, len) which is like
9569 ERR_error_string(e, buf), but writes at most 'len' bytes
9570 including the 0 terminator. For ERR_error_string_n, 'buf'
9572 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9574 *) CONF library reworked to become more general. A new CONF
9575 configuration file reader "class" is implemented as well as a
9576 new functions (NCONF_*, for "New CONF") to handle it. The now
9577 old CONF_* functions are still there, but are reimplemented to
9578 work in terms of the new functions. Also, a set of functions
9579 to handle the internal storage of the configuration data is
9580 provided to make it easier to write new configuration file
9581 reader "classes" (I can definitely see something reading a
9582 configuration file in XML format, for example), called _CONF_*,
9583 or "the configuration storage API"...
9585 The new configuration file reading functions are:
9587 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9588 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9590 NCONF_default, NCONF_WIN32
9592 NCONF_dump_fp, NCONF_dump_bio
9594 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9595 NCONF_new creates a new CONF object. This works in the same way
9596 as other interfaces in OpenSSL, like the BIO interface.
9597 NCONF_dump_* dump the internal storage of the configuration file,
9598 which is useful for debugging. All other functions take the same
9599 arguments as the old CONF_* functions wth the exception of the
9600 first that must be a `CONF *' instead of a `LHASH *'.
9602 To make it easer to use the new classes with the old CONF_* functions,
9603 the function CONF_set_default_method is provided.
9606 *) Add '-tls1' option to 'openssl ciphers', which was already
9607 mentioned in the documentation but had not been implemented.
9608 (This option is not yet really useful because even the additional
9609 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9612 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9613 OpenSSL-based applications) load shared libraries and bind to
9614 them in a portable way.
9615 [Geoff Thorpe, with contributions from Richard Levitte]
9617 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9619 *) Make sure _lrotl and _lrotr are only used with MSVC.
9621 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9622 (the default implementation of RAND_status).
9624 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9625 to '-clrext' (= clear extensions), as intended and documented.
9626 [Bodo Moeller; inconsistency pointed out by Michael Attili
9629 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9630 was larger than the MD block size.
9631 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9633 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9634 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9635 using the passed key: if the passed key was a private key the result
9636 of X509_print(), for example, would be to print out all the private key
9640 *) des_quad_cksum() byte order bug fix.
9641 [Ulf Möller, using the problem description in krb4-0.9.7, where
9642 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9644 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9646 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9648 *) For easily testing in shell scripts whether some command
9649 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9650 returns with exit code 0 iff no command of the given name is available.
9651 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9652 the output goes to stdout and nothing is printed to stderr.
9653 Additional arguments are always ignored.
9655 Since for each cipher there is a command of the same name,
9656 the 'no-cipher' compilation switches can be tested this way.
9658 ('openssl no-XXX' is not able to detect pseudo-commands such
9659 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9662 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9665 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9666 is set; it will be thrown away anyway because each handshake creates
9668 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9669 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9670 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9671 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9674 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9675 'Q' and 'R' lose their special meanings (quit/renegotiate).
9676 This is part of what -quiet does; unlike -quiet, -ign_eof
9677 does not suppress any output.
9680 *) Add compatibility options to the purpose and trust code. The
9681 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9682 accepts a certificate or CA, this was the previous behaviour,
9683 with all the associated security issues.
9685 X509_TRUST_COMPAT is the old trust behaviour: only and
9686 automatically trust self signed roots in certificate store. A
9687 new trust setting X509_TRUST_DEFAULT is used to specify that
9688 a purpose has no associated trust setting and it should instead
9689 use the value in the default purpose.
9692 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9693 and fix a memory leak.
9696 *) In util/mkerr.pl (which implements 'make errors'), preserve
9697 reason strings from the previous version of the .c file, as
9698 the default to have only downcase letters (and digits) in
9699 automatically generated reasons codes is not always appropriate.
9702 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9703 using strerror. Previously, ERR_reason_error_string() returned
9704 library names as reason strings for SYSerr; but SYSerr is a special
9705 case where small numbers are errno values, not library numbers.
9708 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9709 converts DSA parameters into DH parameters. (When creating parameters,
9710 DSA_generate_parameters is used.)
9713 *) Include 'length' (recommended exponent length) in C code generated
9714 by 'openssl dhparam -C'.
9717 *) The second argument to set_label in perlasm was already being used
9718 so couldn't be used as a "file scope" flag. Moved to third argument
9722 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9723 instead of RAND_bytes for encryption IVs and salts.
9726 *) Include RAND_status() into RAND_METHOD instead of implementing
9727 it only for md_rand.c Otherwise replacing the PRNG by calling
9728 RAND_set_rand_method would be impossible.
9731 *) Don't let DSA_generate_key() enter an infinite loop if the random
9732 number generation fails.
9735 *) New 'rand' application for creating pseudo-random output.
9738 *) Added configuration support for Linux/IA64
9739 [Rolf Haberrecker <rolf@suse.de>]
9741 *) Assembler module support for Mingw32.
9744 *) Shared library support for HPUX (in shlib/).
9745 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9747 *) Shared library support for Solaris gcc.
9748 [Lutz Behnke <behnke@trustcenter.de>]
9750 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9752 *) PKCS7_encrypt() was adding text MIME headers twice because they
9753 were added manually and by SMIME_crlf_copy().
9756 *) In bntest.c don't call BN_rand with zero bits argument.
9757 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9759 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9760 case was implemented. This caused BN_div_recp() to fail occasionally.
9763 *) Add an optional second argument to the set_label() in the perl
9764 assembly language builder. If this argument exists and is set
9765 to 1 it signals that the assembler should use a symbol whose
9766 scope is the entire file, not just the current function. This
9767 is needed with MASM which uses the format label:: for this scope.
9768 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9770 *) Change the ASN1 types so they are typedefs by default. Before
9771 almost all types were #define'd to ASN1_STRING which was causing
9772 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9776 *) Change names of new functions to the new get1/get0 naming
9777 convention: After 'get1', the caller owns a reference count
9778 and has to call ..._free; 'get0' returns a pointer to some
9779 data structure without incrementing reference counters.
9780 (Some of the existing 'get' functions increment a reference
9781 counter, some don't.)
9782 Similarly, 'set1' and 'add1' functions increase reference
9783 counters or duplicate objects.
9786 *) Allow for the possibility of temp RSA key generation failure:
9787 the code used to assume it always worked and crashed on failure.
9790 *) Fix potential buffer overrun problem in BIO_printf().
9791 [Ulf Möller, using public domain code by Patrick Powell; problem
9792 pointed out by David Sacerdote <das33@cornell.edu>]
9794 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9795 RAND_egd() and RAND_status(). In the command line application,
9796 the EGD socket can be specified like a seed file using RANDFILE
9800 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9801 Some CAs (e.g. Verisign) distribute certificates in this form.
9804 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9805 list to exclude them. This means that no special compilation option
9806 is needed to use anonymous DH: it just needs to be included in the
9810 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9811 EVP_MD_type. The old functionality is available in a new macro called
9812 EVP_MD_md(). Change code that uses it and update docs.
9815 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9816 where the 'void *' argument is replaced by a function pointer argument.
9817 Previously 'void *' was abused to point to functions, which works on
9818 many platforms, but is not correct. As these functions are usually
9819 called by macros defined in OpenSSL header files, most source code
9820 should work without changes.
9823 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9824 sections with information on -D... compiler switches used for
9825 compiling the library so that applications can see them. To enable
9826 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9827 must be defined. E.g.,
9828 #define OPENSSL_ALGORITHM_DEFINES
9829 #include <openssl/opensslconf.h>
9830 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9831 [Richard Levitte, Ulf and Bodo Möller]
9833 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9837 *) Change the 'other' type in certificate aux info to a STACK_OF
9838 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9839 the required ASN1 format: arbitrary types determined by an OID.
9842 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9843 argument to 'req'. This is not because the function is newer or
9844 better than others it just uses the work 'NEW' in the certificate
9845 request header lines. Some software needs this.
9848 *) Reorganise password command line arguments: now passwords can be
9849 obtained from various sources. Delete the PEM_cb function and make
9850 it the default behaviour: i.e. if the callback is NULL and the
9851 usrdata argument is not NULL interpret it as a null terminated pass
9852 phrase. If usrdata and the callback are NULL then the pass phrase
9853 is prompted for as usual.
9856 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9857 the support is automatically enabled. The resulting binaries will
9858 autodetect the card and use it if present.
9859 [Ben Laurie and Compaq Inc.]
9861 *) Work around for Netscape hang bug. This sends certificate request
9862 and server done in one record. Since this is perfectly legal in the
9863 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9864 the bugs/SSLv3 entry for more info.
9867 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9870 *) Add -rand argument to smime and pkcs12 applications and read/write
9874 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9877 *) Add command line password options to the remaining applications.
9880 *) Bug fix for BN_div_recp() for numerators with an even number of
9884 *) More tests in bntest.c, and changed test_bn output.
9887 *) ./config recognizes MacOS X now.
9890 *) Bug fix for BN_div() when the first words of num and divsor are
9891 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9894 *) Add support for various broken PKCS#8 formats, and command line
9895 options to produce them.
9898 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9899 get temporary BIGNUMs from a BN_CTX.
9902 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9906 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9907 include a #define from the old name to the new. The original intent
9908 was that statically linked binaries could for example just call
9909 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9910 link with digests. This never worked because SSLeay_add_all_digests()
9911 and SSLeay_add_all_ciphers() were in the same source file so calling
9912 one would link with the other. They are now in separate source files.
9915 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9918 *) Use a less unusual form of the Miller-Rabin primality test (it used
9919 a binary algorithm for exponentiation integrated into the Miller-Rabin
9920 loop, our standard modexp algorithms are faster).
9923 *) Support for the EBCDIC character set completed.
9924 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9926 *) Source code cleanups: use const where appropriate, eliminate casts,
9927 use void * instead of char * in lhash.
9930 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9931 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9932 this the server could overwrite ephemeral keys that the client
9936 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9937 using 50 iterations of the Rabin-Miller test.
9939 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9940 iterations of the Rabin-Miller test as required by the appendix
9941 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9942 As BN_is_prime_fasttest includes trial division, DSA parameter
9943 generation becomes much faster.
9945 This implies a change for the callback functions in DSA_is_prime
9946 and DSA_generate_parameters: The callback function is called once
9947 for each positive witness in the Rabin-Miller test, not just
9948 occasionally in the inner loop; and the parameters to the
9949 callback function now provide an iteration count for the outer
9950 loop rather than for the current invocation of the inner loop.
9951 DSA_generate_parameters additionally can call the callback
9952 function with an 'iteration count' of -1, meaning that a
9953 candidate has passed the trial division test (when q is generated
9954 from an application-provided seed, trial division is skipped).
9957 *) New function BN_is_prime_fasttest that optionally does trial
9958 division before starting the Rabin-Miller test and has
9959 an additional BN_CTX * argument (whereas BN_is_prime always
9960 has to allocate at least one BN_CTX).
9961 'callback(1, -1, cb_arg)' is called when a number has passed the
9962 trial division stage.
9965 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9969 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9972 *) New function BN_pseudo_rand().
9975 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9976 bignum version of BN_from_montgomery() with the working code from
9977 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9981 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9982 made it impossible to use the same SSL_SESSION data structure in
9983 SSL2 clients in multiple threads.
9986 *) The return value of RAND_load_file() no longer counts bytes obtained
9987 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9988 to seed the PRNG (previously an explicit byte count was required).
9989 [Ulf Möller, Bodo Möller]
9991 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9992 used (char *) instead of (void *) and had casts all over the place.
9995 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9998 *) Retain source code compatibility for BN_prime_checks macro:
9999 BN_is_prime(..., BN_prime_checks, ...) now uses
10000 BN_prime_checks_for_size to determine the appropriate number of
10001 Rabin-Miller iterations.
10004 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10005 DH_CHECK_P_NOT_SAFE_PRIME.
10006 (Check if this is true? OpenPGP calls them "strong".)
10009 *) Merge the functionality of "dh" and "gendh" programs into a new program
10010 "dhparam". The old programs are retained for now but will handle DH keys
10011 (instead of parameters) in future.
10014 *) Make the ciphers, s_server and s_client programs check the return values
10015 when a new cipher list is set.
10018 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10019 ciphers. Before when the 56bit ciphers were enabled the sorting was
10022 The syntax for the cipher sorting has been extended to support sorting by
10023 cipher-strength (using the strength_bits hard coded in the tables).
10024 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10026 Fix a bug in the cipher-command parser: when supplying a cipher command
10027 string with an "undefined" symbol (neither command nor alphanumeric
10028 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10029 an error is flagged.
10031 Due to the strength-sorting extension, the code of the
10032 ssl_create_cipher_list() function was completely rearranged. I hope that
10033 the readability was also increased :-)
10034 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10036 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10037 for the first serial number and places 2 in the serial number file. This
10038 avoids problems when the root CA is created with serial number zero and
10039 the first user certificate has the same issuer name and serial number
10043 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10044 the new code. Add documentation for this stuff.
10047 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10048 X509_*() to X509at_*() on the grounds that they don't handle X509
10049 structures and behave in an analogous way to the X509v3 functions:
10050 they shouldn't be called directly but wrapper functions should be used
10053 So we also now have some wrapper functions that call the X509at functions
10054 when passed certificate requests. (TO DO: similar things can be done with
10055 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10056 things. Some of these need some d2i or i2d and print functionality
10057 because they handle more complex structures.)
10060 *) Add missing #ifndefs that caused missing symbols when building libssl
10061 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10062 NO_RSA in ssl/s2*.c.
10063 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10065 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10066 has a return value which indicates the quality of the random data
10067 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10068 error queue. New function RAND_pseudo_bytes() generates output that is
10069 guaranteed to be unique but not unpredictable. RAND_add is like
10070 RAND_seed, but takes an extra argument for an entropy estimate
10071 (RAND_seed always assumes full entropy).
10074 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10075 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10076 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10077 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10078 false-positive rate of at most 2^-80 for random input.
10081 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10084 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10085 in the 0.9.5 release), this returns the chain
10086 from an X509_CTX structure with a dup of the stack and all
10087 the X509 reference counts upped: so the stack will exist
10088 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10091 Also make SSL_SESSION_print() print out the verify return
10095 *) Add manpage for the pkcs12 command. Also change the default
10096 behaviour so MAC iteration counts are used unless the new
10097 -nomaciter option is used. This improves file security and
10098 only older versions of MSIE (4.0 for example) need it.
10101 *) Honor the no-xxx Configure options when creating .DEF files.
10104 *) Add PKCS#10 attributes to field table: challengePassword,
10105 unstructuredName and unstructuredAddress. These are taken from
10106 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10107 international characters are used.
10109 More changes to X509_ATTRIBUTE code: allow the setting of types
10110 based on strings. Remove the 'loc' parameter when adding
10111 attributes because these will be a SET OF encoding which is sorted
10115 *) Initial changes to the 'req' utility to allow request generation
10116 automation. This will allow an application to just generate a template
10117 file containing all the field values and have req construct the
10120 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10121 used all over the place including certificate requests and PKCS#7
10122 structures. They are currently handled manually where necessary with
10123 some primitive wrappers for PKCS#7. The new functions behave in a
10124 manner analogous to the X509 extension functions: they allow
10125 attributes to be looked up by NID and added.
10127 Later something similar to the X509V3 code would be desirable to
10128 automatically handle the encoding, decoding and printing of the
10129 more complex types. The string types like challengePassword can
10130 be handled by the string table functions.
10132 Also modified the multi byte string table handling. Now there is
10133 a 'global mask' which masks out certain types. The table itself
10134 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10135 is useful when for example there is only one permissible type
10136 (as in countryName) and using the mask might result in no valid
10140 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10141 SSL_get_peer_finished to allow applications to obtain the latest
10142 Finished messages sent to the peer or expected from the peer,
10143 respectively. (SSL_get_peer_finished is usually the Finished message
10144 actually received from the peer, otherwise the protocol will be aborted.)
10146 As the Finished message are message digests of the complete handshake
10147 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10148 be used for external authentication procedures when the authentication
10149 provided by SSL/TLS is not desired or is not enough.
10152 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10153 the host supports BWX extension and if Compaq C is present on the
10154 $PATH. Just exploiting of the BWX extension results in 20-30%
10155 performance kick for some algorithms, e.g. DES and RC4 to mention
10156 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10160 *) Add support for MS "fast SGC". This is arguably a violation of the
10161 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10162 weak crypto and after checking the certificate is SGC a second one
10163 with strong crypto. MS SGC stops the first handshake after receiving
10164 the server certificate message and sends a second client hello. Since
10165 a server will typically do all the time consuming operations before
10166 expecting any further messages from the client (server key exchange
10167 is the most expensive) there is little difference between the two.
10169 To get OpenSSL to support MS SGC we have to permit a second client
10170 hello message after we have sent server done. In addition we have to
10171 reset the MAC if we do get this second client hello.
10174 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10175 if a DER encoded private key is RSA or DSA traditional format. Changed
10176 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10177 format DER encoded private key. Newer code should use PKCS#8 format which
10178 has the key type encoded in the ASN1 structure. Added DER private key
10179 support to pkcs8 application.
10182 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10183 ciphersuites has been selected (as required by the SSL 3/TLS 1
10184 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10185 is set, we interpret this as a request to violate the specification
10186 (the worst that can happen is a handshake failure, and 'correct'
10187 behaviour would result in a handshake failure anyway).
10190 *) In SSL_CTX_add_session, take into account that there might be multiple
10191 SSL_SESSION structures with the same session ID (e.g. when two threads
10192 concurrently obtain them from an external cache).
10193 The internal cache can handle only one SSL_SESSION with a given ID,
10194 so if there's a conflict, we now throw out the old one to achieve
10198 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10199 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10200 some routines that use cipher OIDs: some ciphers do not have OIDs
10201 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10205 *) Simplify the trust setting structure and code. Now we just have
10206 two sequences of OIDs for trusted and rejected settings. These will
10207 typically have values the same as the extended key usage extension
10208 and any application specific purposes.
10210 The trust checking code now has a default behaviour: it will just
10211 check for an object with the same NID as the passed id. Functions can
10212 be provided to override either the default behaviour or the behaviour
10213 for a given id. SSL client, server and email already have functions
10214 in place for compatibility: they check the NID and also return "trusted"
10215 if the certificate is self signed.
10218 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10219 traditional format into an EVP_PKEY structure.
10222 *) Add a password callback function PEM_cb() which either prompts for
10223 a password if usr_data is NULL or otherwise assumes it is a null
10224 terminated password. Allow passwords to be passed on command line
10225 environment or config files in a few more utilities.
10228 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10229 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10230 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10231 Update documentation.
10234 *) Support for ASN1 "NULL" type. This could be handled before by using
10235 ASN1_TYPE but there wasn't any function that would try to read a NULL
10236 and produce an error if it couldn't. For compatibility we also have
10237 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10238 don't allocate anything because they don't need to.
10241 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10243 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10245 *) Rebuild of the memory allocation routines used by OpenSSL code and
10246 possibly others as well. The purpose is to make an interface that
10247 provide hooks so anyone can build a separate set of allocation and
10248 deallocation routines to be used by OpenSSL, for example memory
10249 pool implementations, or something else, which was previously hard
10250 since Malloc(), Realloc() and Free() were defined as macros having
10251 the values malloc, realloc and free, respectively (except for Win32
10252 compilations). The same is provided for memory debugging code.
10253 OpenSSL already comes with functionality to find memory leaks, but
10254 this gives people a chance to debug other memory problems.
10256 With these changes, a new set of functions and macros have appeared:
10258 CRYPTO_set_mem_debug_functions() [F]
10259 CRYPTO_get_mem_debug_functions() [F]
10260 CRYPTO_dbg_set_options() [F]
10261 CRYPTO_dbg_get_options() [F]
10262 CRYPTO_malloc_debug_init() [M]
10264 The memory debug functions are NULL by default, unless the library
10265 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10266 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10267 gives the standard debugging functions that come with OpenSSL) or
10268 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10269 provided by the library user) must be used. When the standard
10270 debugging functions are used, CRYPTO_dbg_set_options can be used to
10271 request additional information:
10272 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10273 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10275 Also, things like CRYPTO_set_mem_functions will always give the
10276 expected result (the new set of functions is used for allocation
10277 and deallocation) at all times, regardless of platform and compiler
10280 To finish it up, some functions that were never use in any other
10281 way than through macros have a new API and new semantic:
10283 CRYPTO_dbg_malloc()
10284 CRYPTO_dbg_realloc()
10287 All macros of value have retained their old syntax.
10288 [Richard Levitte and Bodo Moeller]
10290 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10291 ordering of SMIMECapabilities wasn't in "strength order" and there
10292 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10296 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10297 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10298 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10300 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10301 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10302 functionality to handle multipart/signed properly) and a utility
10303 called 'smime' to call all this stuff. This is based on code I
10304 originally wrote for Celo who have kindly allowed it to be
10305 included in OpenSSL.
10308 *) Add variants des_set_key_checked and des_set_key_unchecked of
10309 des_set_key (aka des_key_sched). Global variable des_check_key
10310 decides which of these is called by des_set_key; this way
10311 des_check_key behaves as it always did, but applications and
10312 the library itself, which was buggy for des_check_key == 1,
10313 have a cleaner way to pick the version they need.
10316 *) New function PKCS12_newpass() which changes the password of a
10320 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10321 dynamic mix. In both cases the ids can be used as an index into the
10322 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10323 functions so they accept a list of the field values and the
10324 application doesn't need to directly manipulate the X509_TRUST
10328 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10332 *) Modify the way the V3 extension code looks up extensions. This now
10333 works in a similar way to the object code: we have some "standard"
10334 extensions in a static table which is searched with OBJ_bsearch()
10335 and the application can add dynamic ones if needed. The file
10336 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10337 updated whenever a new extension is added to the core code and kept
10338 in ext_nid order. There is a simple program 'tabtest.c' which checks
10339 this. New extensions are not added too often so this file can readily
10340 be maintained manually.
10342 There are two big advantages in doing things this way. The extensions
10343 can be looked up immediately and no longer need to be "added" using
10344 X509V3_add_standard_extensions(): this function now does nothing.
10345 [Side note: I get *lots* of email saying the extension code doesn't
10346 work because people forget to call this function]
10347 Also no dynamic allocation is done unless new extensions are added:
10348 so if we don't add custom extensions there is no need to call
10349 X509V3_EXT_cleanup().
10352 *) Modify enc utility's salting as follows: make salting the default. Add a
10353 magic header, so unsalted files fail gracefully instead of just decrypting
10354 to garbage. This is because not salting is a big security hole, so people
10355 should be discouraged from doing it.
10358 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10359 digest to be passed on the command line but it only used this
10360 parameter when signing a certificate. Modified so all relevant
10361 operations are affected by the digest parameter including the
10362 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10363 DSA key was used because it didn't fix the digest.
10366 *) Initial certificate chain verify code. Currently tests the untrusted
10367 certificates for consistency with the verify purpose (which is set
10368 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10370 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10371 this is because it will reject chains with invalid extensions whereas
10372 every previous version of OpenSSL and SSLeay made no checks at all.
10374 Trust code: checks the root CA for the relevant trust settings. Trust
10375 settings have an initial value consistent with the verify purpose: e.g.
10376 if the verify purpose is for SSL client use it expects the CA to be
10377 trusted for SSL client use. However the default value can be changed to
10378 permit custom trust settings: one example of this would be to only trust
10379 certificates from a specific "secure" set of CAs.
10381 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10382 which should be used for version portability: especially since the
10383 verify structure is likely to change more often now.
10385 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10386 to set them. If not set then assume SSL clients will verify SSL servers
10389 Two new options to the verify program: -untrusted allows a set of
10390 untrusted certificates to be passed in and -purpose which sets the
10391 intended purpose of the certificate. If a purpose is set then the
10392 new chain verify code is used to check extension consistency.
10395 *) Support for the authority information access extension.
10398 *) Modify RSA and DSA PEM read routines to transparently handle
10399 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10400 public keys in a format compatible with certificate
10401 SubjectPublicKeyInfo structures. Unfortunately there were already
10402 functions called *_PublicKey_* which used various odd formats so
10403 these are retained for compatibility: however the DSA variants were
10404 never in a public release so they have been deleted. Changed dsa/rsa
10405 utilities to handle the new format: note no releases ever handled public
10406 keys so we should be OK.
10408 The primary motivation for this change is to avoid the same fiasco
10409 that dogs private keys: there are several incompatible private key
10410 formats some of which are standard and some OpenSSL specific and
10411 require various evil hacks to allow partial transparent handling and
10412 even then it doesn't work with DER formats. Given the option anything
10413 other than PKCS#8 should be dumped: but the other formats have to
10414 stay in the name of compatibility.
10416 With public keys and the benefit of hindsight one standard format
10417 is used which works with EVP_PKEY, RSA or DSA structures: though
10418 it clearly returns an error if you try to read the wrong kind of key.
10420 Added a -pubkey option to the 'x509' utility to output the public key.
10421 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10422 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10423 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10424 that do the same as the EVP_PKEY_assign_*() except they up the
10425 reference count of the added key (they don't "swallow" the
10429 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10430 CRLs would fail if the file contained no certificates or no CRLs:
10431 added a new function to read in both types and return the number
10432 read: this means that if none are read it will be an error. The
10433 DER versions of the certificate and CRL reader would always fail
10434 because it isn't possible to mix certificates and CRLs in DER format
10435 without choking one or the other routine. Changed this to just read
10436 a certificate: this is the best we can do. Also modified the code
10437 in apps/verify.c to take notice of return codes: it was previously
10438 attempting to read in certificates from NULL pointers and ignoring
10439 any errors: this is one reason why the cert and CRL reader seemed
10440 to work. It doesn't check return codes from the default certificate
10441 routines: these may well fail if the certificates aren't installed.
10444 *) Code to support otherName option in GeneralName.
10447 *) First update to verify code. Change the verify utility
10448 so it warns if it is passed a self signed certificate:
10449 for consistency with the normal behaviour. X509_verify
10450 has been modified to it will now verify a self signed
10451 certificate if *exactly* the same certificate appears
10452 in the store: it was previously impossible to trust a
10453 single self signed certificate. This means that:
10454 openssl verify ss.pem
10455 now gives a warning about a self signed certificate but
10456 openssl verify -CAfile ss.pem ss.pem
10460 *) For servers, store verify_result in SSL_SESSION data structure
10461 (and add it to external session representation).
10462 This is needed when client certificate verifications fails,
10463 but an application-provided verification callback (set by
10464 SSL_CTX_set_cert_verify_callback) allows accepting the session
10465 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10466 but returns 1): When the session is reused, we have to set
10467 ssl->verify_result to the appropriate error code to avoid
10469 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10471 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10472 case in PKCS7_dataInit() where the signed PKCS7 structure
10473 didn't contain any existing data because it was being created.
10474 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10476 *) Add a salt to the key derivation routines in enc.c. This
10477 forms the first 8 bytes of the encrypted file. Also add a
10478 -S option to allow a salt to be input on the command line.
10481 *) New function X509_cmp(). Oddly enough there wasn't a function
10482 to compare two certificates. We do this by working out the SHA1
10483 hash and comparing that. X509_cmp() will be needed by the trust
10487 *) SSL_get1_session() is like SSL_get_session(), but increments
10488 the reference count in the SSL_SESSION returned.
10489 [Geoff Thorpe <geoff@eu.c2.net>]
10491 *) Fix for 'req': it was adding a null to request attributes.
10492 Also change the X509_LOOKUP and X509_INFO code to handle
10493 certificate auxiliary information.
10496 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10500 *) Add the possibility to add extra information to the memory leak
10501 detecting output, to form tracebacks, showing from where each
10502 allocation was originated: CRYPTO_push_info("constant string") adds
10503 the string plus current file name and line number to a per-thread
10504 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10505 is like calling CYRPTO_pop_info() until the stack is empty.
10506 Also updated memory leak detection code to be multi-thread-safe.
10509 *) Add options -text and -noout to pkcs7 utility and delete the
10510 encryption options which never did anything. Update docs.
10513 *) Add options to some of the utilities to allow the pass phrase
10514 to be included on either the command line (not recommended on
10515 OSes like Unix) or read from the environment. Update the
10516 manpages and fix a few bugs.
10519 *) Add a few manpages for some of the openssl commands.
10522 *) Fix the -revoke option in ca. It was freeing up memory twice,
10523 leaking and not finding already revoked certificates.
10526 *) Extensive changes to support certificate auxiliary information.
10527 This involves the use of X509_CERT_AUX structure and X509_AUX
10528 functions. An X509_AUX function such as PEM_read_X509_AUX()
10529 can still read in a certificate file in the usual way but it
10530 will also read in any additional "auxiliary information". By
10531 doing things this way a fair degree of compatibility can be
10532 retained: existing certificates can have this information added
10533 using the new 'x509' options.
10535 Current auxiliary information includes an "alias" and some trust
10536 settings. The trust settings will ultimately be used in enhanced
10537 certificate chain verification routines: currently a certificate
10538 can only be trusted if it is self signed and then it is trusted
10542 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10543 The problem was that one of the replacement routines had not been working
10544 since SSLeay releases. For now the offending routine has been replaced
10545 with non-optimised assembler. Even so, this now gives around 95%
10546 performance improvement for 1024 bit RSA signs.
10549 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10550 handling. Most clients have the effective key size in bits equal to
10551 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10552 A few however don't do this and instead use the size of the decrypted key
10553 to determine the RC2 key length and the AlgorithmIdentifier to determine
10554 the effective key length. In this case the effective key length can still
10555 be 40 bits but the key length can be 168 bits for example. This is fixed
10556 by manually forcing an RC2 key into the EVP_PKEY structure because the
10557 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10558 the key length and effective key length are equal.
10561 *) Add a bunch of functions that should simplify the creation of
10562 X509_NAME structures. Now you should be able to do:
10563 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10564 and have it automatically work out the correct field type and fill in
10565 the structures. The more adventurous can try:
10566 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10567 and it will (hopefully) work out the correct multibyte encoding.
10570 *) Change the 'req' utility to use the new field handling and multibyte
10571 copy routines. Before the DN field creation was handled in an ad hoc
10572 way in req, ca, and x509 which was rather broken and didn't support
10573 BMPStrings or UTF8Strings. Since some software doesn't implement
10574 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10575 using the dirstring_type option. See the new comment in the default
10576 openssl.cnf for more info.
10579 *) Make crypto/rand/md_rand.c more robust:
10580 - Assure unique random numbers after fork().
10581 - Make sure that concurrent threads access the global counter and
10582 md serializably so that we never lose entropy in them
10583 or use exactly the same state in multiple threads.
10584 Access to the large state is not always serializable because
10585 the additional locking could be a performance killer, and
10586 md should be large enough anyway.
10589 *) New file apps/app_rand.c with commonly needed functionality
10590 for handling the random seed file.
10592 Use the random seed file in some applications that previously did not:
10594 dsaparam -genkey (which also ignored its '-rand' option),
10597 x509 (when signing).
10598 Except on systems with /dev/urandom, it is crucial to have a random
10599 seed file at least for key creation, DSA signing, and for DH exchanges;
10600 for RSA signatures we could do without one.
10602 gendh and gendsa (unlike genrsa) used to read only the first byte
10603 of each file listed in the '-rand' option. The function as previously
10604 found in genrsa is now in app_rand.c and is used by all programs
10605 that support '-rand'.
10608 *) In RAND_write_file, use mode 0600 for creating files;
10609 don't just chmod when it may be too late.
10612 *) Report an error from X509_STORE_load_locations
10613 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10616 *) New function ASN1_mbstring_copy() this copies a string in either
10617 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10618 into an ASN1_STRING type. A mask of permissible types is passed
10619 and it chooses the "minimal" type to use or an error if not type
10623 *) Add function equivalents to the various macros in asn1.h. The old
10624 macros are retained with an M_ prefix. Code inside the library can
10625 use the M_ macros. External code (including the openssl utility)
10626 should *NOT* in order to be "shared library friendly".
10629 *) Add various functions that can check a certificate's extensions
10630 to see if it usable for various purposes such as SSL client,
10631 server or S/MIME and CAs of these types. This is currently
10632 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10633 verification. Also added a -purpose flag to x509 utility to
10634 print out all the purposes.
10637 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10641 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10642 for, obtain and decode and extension and obtain its critical flag.
10643 This allows all the necessary extension code to be handled in a
10644 single function call.
10647 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10648 platforms. See crypto/rc4/rc4_enc.c for further details.
10651 *) New -noout option to asn1parse. This causes no output to be produced
10652 its main use is when combined with -strparse and -out to extract data
10653 from a file (which may not be in ASN.1 format).
10656 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10657 when producing the local key id.
10658 [Richard Levitte <levitte@stacken.kth.se>]
10660 *) New option -dhparam in s_server. This allows a DH parameter file to be
10661 stated explicitly. If it is not stated then it tries the first server
10662 certificate file. The previous behaviour hard coded the filename
10666 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10667 a public key to be input or output. For example:
10668 openssl rsa -in key.pem -pubout -out pubkey.pem
10669 Also added necessary DSA public key functions to handle this.
10672 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10673 in the message. This was handled by allowing
10674 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10675 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10677 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10678 to the end of the strings whereas this didn't. This would cause problems
10679 if strings read with d2i_ASN1_bytes() were later modified.
10680 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10682 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10683 data and it contains EOF it will end up returning an error. This is
10684 caused by input 46 bytes long. The cause is due to the way base64
10685 BIOs find the start of base64 encoded data. They do this by trying a
10686 trial decode on each line until they find one that works. When they
10687 do a flag is set and it starts again knowing it can pass all the
10688 data directly through the decoder. Unfortunately it doesn't reset
10689 the context it uses. This means that if EOF is reached an attempt
10690 is made to pass two EOFs through the context and this causes the
10691 resulting error. This can also cause other problems as well. As is
10692 usual with these problems it takes *ages* to find and the fix is
10693 trivial: move one line.
10694 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10696 *) Ugly workaround to get s_client and s_server working under Windows. The
10697 old code wouldn't work because it needed to select() on sockets and the
10698 tty (for keypresses and to see if data could be written). Win32 only
10699 supports select() on sockets so we select() with a 1s timeout on the
10700 sockets and then see if any characters are waiting to be read, if none
10701 are present then we retry, we also assume we can always write data to
10702 the tty. This isn't nice because the code then blocks until we've
10703 received a complete line of data and it is effectively polling the
10704 keyboard at 1s intervals: however it's quite a bit better than not
10705 working at all :-) A dedicated Windows application might handle this
10706 with an event loop for example.
10709 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10710 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10711 will be called when RSA_sign() and RSA_verify() are used. This is useful
10712 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10713 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10714 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10715 This necessitated the support of an extra signature type NID_md5_sha1
10716 for SSL signatures and modifications to the SSL library to use it instead
10717 of calling RSA_public_decrypt() and RSA_private_encrypt().
10720 *) Add new -verify -CAfile and -CApath options to the crl program, these
10721 will lookup a CRL issuers certificate and verify the signature in a
10722 similar way to the verify program. Tidy up the crl program so it
10723 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10724 less strict. It will now permit CRL extensions even if it is not
10725 a V2 CRL: this will allow it to tolerate some broken CRLs.
10728 *) Initialize all non-automatic variables each time one of the openssl
10729 sub-programs is started (this is necessary as they may be started
10730 multiple times from the "OpenSSL>" prompt).
10731 [Lennart Bang, Bodo Moeller]
10733 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10734 removing all other RSA functionality (this is what NO_RSA does). This
10735 is so (for example) those in the US can disable those operations covered
10736 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10740 *) Non-copying interface to BIO pairs.
10741 (still largely untested)
10744 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10745 ASCII string. This was handled independently in various places before.
10748 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10749 UTF8 strings a character at a time.
10752 *) Use client_version from client hello to select the protocol
10753 (s23_srvr.c) and for RSA client key exchange verification
10754 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10757 *) Add various utility functions to handle SPKACs, these were previously
10758 handled by poking round in the structure internals. Added new function
10759 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10760 print, verify and generate SPKACs. Based on an original idea from
10761 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10764 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10767 *) Allow the config file extension section to be overwritten on the
10768 command line. Based on an original idea from Massimiliano Pala
10769 <madwolf@comune.modena.it>. The new option is called -extensions
10770 and can be applied to ca, req and x509. Also -reqexts to override
10771 the request extensions in req and -crlexts to override the crl extensions
10775 *) Add new feature to the SPKAC handling in ca. Now you can include
10776 the same field multiple times by preceding it by "XXXX." for example:
10779 this is the same syntax as used in the req config file.
10782 *) Allow certificate extensions to be added to certificate requests. These
10783 are specified in a 'req_extensions' option of the req section of the
10784 config file. They can be printed out with the -text option to req but
10785 are otherwise ignored at present.
10788 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10789 data read consists of only the final block it would not decrypted because
10790 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10791 A misplaced 'break' also meant the decrypted final block might not be
10792 copied until the next read.
10795 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10796 a few extra parameters to the DH structure: these will be useful if
10797 for example we want the value of 'q' or implement X9.42 DH.
10800 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10801 provides hooks that allow the default DSA functions or functions on a
10802 "per key" basis to be replaced. This allows hardware acceleration and
10803 hardware key storage to be handled without major modification to the
10804 library. Also added low level modexp hooks and CRYPTO_EX structure and
10805 associated functions.
10808 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10809 as "read only": it can't be written to and the buffer it points to will
10810 not be freed. Reading from a read only BIO is much more efficient than
10811 a normal memory BIO. This was added because there are several times when
10812 an area of memory needs to be read from a BIO. The previous method was
10813 to create a memory BIO and write the data to it, this results in two
10814 copies of the data and an O(n^2) reading algorithm. There is a new
10815 function BIO_new_mem_buf() which creates a read only memory BIO from
10816 an area of memory. Also modified the PKCS#7 routines to use read only
10820 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10821 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10822 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10823 but a retry condition occurred while trying to read the rest.
10826 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10827 NID_pkcs7_encrypted by default: this was wrong since this should almost
10828 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10829 the encrypted data type: this is a more sensible place to put it and it
10830 allows the PKCS#12 code to be tidied up that duplicated this
10834 *) Changed obj_dat.pl script so it takes its input and output files on
10835 the command line. This should avoid shell escape redirection problems
10839 *) Initial support for certificate extension requests, these are included
10840 in things like Xenroll certificate requests. Included functions to allow
10841 extensions to be obtained and added.
10844 *) -crlf option to s_client and s_server for sending newlines as
10845 CRLF (as required by many protocols).
10848 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10850 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10851 [Ralf S. Engelschall]
10853 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10854 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10856 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10860 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10861 DH parameters/keys (q is lost during that conversion, but the resulting
10862 DH parameters contain its length).
10864 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10865 much faster than DH_generate_parameters (which creates parameters
10866 where p = 2*q + 1), and also the smaller q makes DH computations
10867 much more efficient (160-bit exponentiation instead of 1024-bit
10868 exponentiation); so this provides a convenient way to support DHE
10869 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10870 utter importance to use
10871 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10873 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10874 when such DH parameters are used, because otherwise small subgroup
10875 attacks may become possible!
10878 *) Avoid memory leak in i2d_DHparams.
10881 *) Allow the -k option to be used more than once in the enc program:
10882 this allows the same encrypted message to be read by multiple recipients.
10885 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10886 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10887 it will always use the numerical form of the OID, even if it has a short
10891 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10892 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10893 otherwise bn_mod_exp was called. In the case of hardware keys for example
10894 no private key components need be present and it might store extra data
10895 in the RSA structure, which cannot be accessed from bn_mod_exp.
10896 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10897 private key operations.
10900 *) Added support for SPARC Linux.
10903 *) pem_password_cb function type incompatibly changed from
10904 typedef int pem_password_cb(char *buf, int size, int rwflag);
10906 ....(char *buf, int size, int rwflag, void *userdata);
10907 so that applications can pass data to their callbacks:
10908 The PEM[_ASN1]_{read,write}... functions and macros now take an
10909 additional void * argument, which is just handed through whenever
10910 the password callback is called.
10911 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10913 New function SSL_CTX_set_default_passwd_cb_userdata.
10915 Compatibility note: As many C implementations push function arguments
10916 onto the stack in reverse order, the new library version is likely to
10917 interoperate with programs that have been compiled with the old
10918 pem_password_cb definition (PEM_whatever takes some data that
10919 happens to be on the stack as its last argument, and the callback
10920 just ignores this garbage); but there is no guarantee whatsoever that
10923 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10924 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10925 problems not only on Windows, but also on some Unix platforms.
10926 To avoid problematic command lines, these definitions are now in an
10927 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10928 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10931 *) MIPS III/IV assembler module is reimplemented.
10934 *) More DES library cleanups: remove references to srand/rand and
10935 delete an unused file.
10938 *) Add support for the the free Netwide assembler (NASM) under Win32,
10939 since not many people have MASM (ml) and it can be hard to obtain.
10940 This is currently experimental but it seems to work OK and pass all
10941 the tests. Check out INSTALL.W32 for info.
10944 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10945 without temporary keys kept an extra copy of the server key,
10946 and connections with temporary keys did not free everything in case
10950 *) New function RSA_check_key and new openssl rsa option -check
10951 for verifying the consistency of RSA keys.
10952 [Ulf Moeller, Bodo Moeller]
10954 *) Various changes to make Win32 compile work:
10955 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10956 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10957 comparison" warnings.
10958 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10961 *) Add a debugging option to PKCS#5 v2 key generation function: when
10962 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10963 derived keys are printed to stderr.
10966 *) Copy the flags in ASN1_STRING_dup().
10967 [Roman E. Pavlov <pre@mo.msk.ru>]
10969 *) The x509 application mishandled signing requests containing DSA
10970 keys when the signing key was also DSA and the parameters didn't match.
10972 It was supposed to omit the parameters when they matched the signing key:
10973 the verifying software was then supposed to automatically use the CA's
10974 parameters if they were absent from the end user certificate.
10976 Omitting parameters is no longer recommended. The test was also
10977 the wrong way round! This was probably due to unusual behaviour in
10978 EVP_cmp_parameters() which returns 1 if the parameters match.
10979 This meant that parameters were omitted when they *didn't* match and
10980 the certificate was useless. Certificates signed with 'ca' didn't have
10982 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10984 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10985 The interface is as follows:
10986 Applications can use
10987 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10988 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10989 "off" is now the default.
10990 The library internally uses
10991 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10992 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10993 to disable memory-checking temporarily.
10995 Some inconsistent states that previously were possible (and were
10996 even the default) are now avoided.
10998 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10999 with each memory chunk allocated; this is occasionally more helpful
11000 than just having a counter.
11002 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11004 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11008 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11009 which largely parallels "options", but is for changing API behaviour,
11010 whereas "options" are about protocol behaviour.
11011 Initial "mode" flags are:
11013 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11014 a single record has been written.
11015 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11016 retries use the same buffer location.
11017 (But all of the contents must be
11021 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11024 *) Fix problems with no-hmac etc.
11025 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11027 *) New functions RSA_get_default_method(), RSA_set_method() and
11028 RSA_get_method(). These allows replacement of RSA_METHODs without having
11029 to mess around with the internals of an RSA structure.
11032 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11033 Also really enable memory leak checks in openssl.c and in some
11035 [Chad C. Mulligan, Bodo Moeller]
11037 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11038 up the length of negative integers. This has now been simplified to just
11039 store the length when it is first determined and use it later, rather
11040 than trying to keep track of where data is copied and updating it to
11042 [Steve Henson, reported by Brien Wheeler
11043 <bwheeler@authentica-security.com>]
11045 *) Add a new function PKCS7_signatureVerify. This allows the verification
11046 of a PKCS#7 signature but with the signing certificate passed to the
11047 function itself. This contrasts with PKCS7_dataVerify which assumes the
11048 certificate is present in the PKCS#7 structure. This isn't always the
11049 case: certificates can be omitted from a PKCS#7 structure and be
11050 distributed by "out of band" means (such as a certificate database).
11053 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11054 function prototypes in pem.h, also change util/mkdef.pl to add the
11055 necessary function names.
11058 *) mk1mf.pl (used by Windows builds) did not properly read the
11059 options set by Configure in the top level Makefile, and Configure
11060 was not even able to write more than one option correctly.
11061 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11064 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11065 file to be loaded from a BIO or FILE pointer. The BIO version will
11066 for example allow memory BIOs to contain config info.
11069 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11070 Whoever hopes to achieve shared-library compatibility across versions
11071 must use this, not the compile-time macro.
11072 (Exercise 0.9.4: Which is the minimum library version required by
11074 Note: All this applies only to multi-threaded programs, others don't
11078 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11079 through a BIO pair triggered the default case, i.e.
11080 SSLerr(...,SSL_R_UNKNOWN_STATE).
11083 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11084 can use the SSL library even if none of the specific BIOs is
11088 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11089 for the encoded length.
11090 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11092 *) Add initial documentation of the X509V3 functions.
11095 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11096 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11097 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11098 secure PKCS#8 private key format with a high iteration count.
11101 *) Fix determination of Perl interpreter: A perl or perl5
11102 _directory_ in $PATH was also accepted as the interpreter.
11103 [Ralf S. Engelschall]
11105 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11106 wrong with it but it was very old and did things like calling
11107 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11108 unusual formatting.
11111 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11112 to use the new extension code.
11115 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11116 with macros. This should make it easier to change their form, add extra
11117 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11121 *) Add to configuration table a new entry that can specify an alternative
11122 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11123 according to Mark Crispin <MRC@Panda.COM>.
11127 *) DES CBC did not update the IV. Weird.
11130 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11131 Changing the behaviour of the former might break existing programs --
11132 where IV updating is needed, des_ncbc_encrypt can be used.
11135 *) When bntest is run from "make test" it drives bc to check its
11136 calculations, as well as internally checking them. If an internal check
11137 fails, it needs to cause bc to give a non-zero result or make test carries
11138 on without noticing the failure. Fixed.
11141 *) DES library cleanups.
11144 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11145 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11146 ciphers. NOTE: although the key derivation function has been verified
11147 against some published test vectors it has not been extensively tested
11148 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11152 *) Instead of "mkdir -p", which is not fully portable, use new
11153 Perl script "util/mkdir-p.pl".
11156 *) Rewrite the way password based encryption (PBE) is handled. It used to
11157 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11158 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11159 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11160 the 'parameter' field of the AlgorithmIdentifier is passed to the
11161 underlying key generation function so it must do its own ASN1 parsing.
11162 This has also changed the EVP_PBE_CipherInit() function which now has a
11163 'parameter' argument instead of literal salt and iteration count values
11164 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11167 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11168 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11169 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11170 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11171 value was just used as a "magic string" and not used directly its
11172 value doesn't matter.
11175 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11179 *) "linux-sparc64" configuration (ultrapenguin).
11180 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11181 "linux-sparc" configuration.
11182 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11184 *) config now generates no-xxx options for missing ciphers.
11187 *) Support the EBCDIC character set (work in progress).
11188 File ebcdic.c not yet included because it has a different license.
11189 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11191 *) Support BS2000/OSD-POSIX.
11192 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11194 *) Make callbacks for key generation use void * instead of char *.
11197 *) Make S/MIME samples compile (not yet tested).
11200 *) Additional typesafe stacks.
11203 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11207 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11209 *) New configuration variant "sco5-gcc".
11211 *) Updated some demos.
11212 [Sean O Riordain, Wade Scholine]
11214 *) Add missing BIO_free at exit of pkcs12 application.
11217 *) Fix memory leak in conf.c.
11220 *) Updates for Win32 to assembler version of MD5.
11223 *) Set #! path to perl in apps/der_chop to where we found it
11224 instead of using a fixed path.
11227 *) SHA library changes for irix64-mips4-cc.
11230 *) Improvements for VMS support.
11234 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11236 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11237 This also avoids the problems with SC4.2 and unpatched SC5.
11238 [Andy Polyakov <appro@fy.chalmers.se>]
11240 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11241 These are required because of the typesafe stack would otherwise break
11242 existing code. If old code used a structure member which used to be STACK
11243 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11244 sk_num or sk_value it would produce an error because the num, data members
11245 are not present in STACK_OF. Now it just produces a warning. sk_set
11246 replaces the old method of assigning a value to sk_value
11247 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11248 that does this will no longer work (and should use sk_set instead) but
11249 this could be regarded as a "questionable" behaviour anyway.
11252 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11253 correctly handle encrypted S/MIME data.
11256 *) Change type of various DES function arguments from des_cblock
11257 (which means, in function argument declarations, pointer to char)
11258 to des_cblock * (meaning pointer to array with 8 char elements),
11259 which allows the compiler to do more typechecking; it was like
11260 that back in SSLeay, but with lots of ugly casts.
11262 Introduce new type const_des_cblock.
11265 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11266 problems: find RecipientInfo structure that matches recipient certificate
11267 and initialise the ASN1 structures properly based on passed cipher.
11270 *) Belatedly make the BN tests actually check the results.
11273 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11274 to and from BNs: it was completely broken. New compilation option
11275 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11276 key elements as negative integers.
11279 *) Reorganize and speed up MD5.
11280 [Andy Polyakov <appro@fy.chalmers.se>]
11283 [Richard Levitte <richard@levitte.org>]
11285 *) New option -out to asn1parse to allow the parsed structure to be
11286 output to a file. This is most useful when combined with the -strparse
11287 option to examine the output of things like OCTET STRINGS.
11290 *) Make SSL library a little more fool-proof by not requiring any longer
11291 that SSL_set_{accept,connect}_state be called before
11292 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11293 in many applications because usually everything *appeared* to work as
11294 intended anyway -- now it really works as intended).
11297 *) Move openssl.cnf out of lib/.
11300 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11301 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11302 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11303 [Ralf S. Engelschall]
11305 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11306 handle PKCS#7 enveloped data properly.
11307 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11309 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11310 copying pointers. The cert_st handling is changed by this in
11311 various ways (and thus what used to be known as ctx->default_cert
11312 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11313 any longer when s->cert does not give us what we need).
11314 ssl_cert_instantiate becomes obsolete by this change.
11315 As soon as we've got the new code right (possibly it already is?),
11316 we have solved a couple of bugs of the earlier code where s->cert
11317 was used as if it could not have been shared with other SSL structures.
11319 Note that using the SSL API in certain dirty ways now will result
11320 in different behaviour than observed with earlier library versions:
11321 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11322 does not influence s as it used to.
11324 In order to clean up things more thoroughly, inside SSL_SESSION
11325 we don't use CERT any longer, but a new structure SESS_CERT
11326 that holds per-session data (if available); currently, this is
11327 the peer's certificate chain and, for clients, the server's certificate
11328 and temporary key. CERT holds only those values that can have
11329 meaningful defaults in an SSL_CTX.
11332 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11333 from the internal representation. Various PKCS#7 fixes: remove some
11334 evil casts and set the enc_dig_alg field properly based on the signing
11338 *) Allow PKCS#12 password to be set from the command line or the
11339 environment. Let 'ca' get its config file name from the environment
11340 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11344 *) Allow certificate policies extension to use an IA5STRING for the
11345 organization field. This is contrary to the PKIX definition but
11346 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11350 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11351 without disallowing inline assembler and the like for non-pedantic builds.
11354 *) Support Borland C++ builder.
11355 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11357 *) Support Mingw32.
11360 *) SHA-1 cleanups and performance enhancements.
11361 [Andy Polyakov <appro@fy.chalmers.se>]
11363 *) Sparc v8plus assembler for the bignum library.
11364 [Andy Polyakov <appro@fy.chalmers.se>]
11366 *) Accept any -xxx and +xxx compiler options in Configure.
11369 *) Update HPUX configuration.
11372 *) Add missing sk_<type>_unshift() function to safestack.h
11373 [Ralf S. Engelschall]
11375 *) New function SSL_CTX_use_certificate_chain_file that sets the
11376 "extra_cert"s in addition to the certificate. (This makes sense
11377 only for "PEM" format files, as chains as a whole are not
11381 *) Support verify_depth from the SSL API.
11382 x509_vfy.c had what can be considered an off-by-one-error:
11383 Its depth (which was not part of the external interface)
11384 was actually counting the number of certificates in a chain;
11385 now it really counts the depth.
11388 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11389 instead of X509err, which often resulted in confusing error
11390 messages since the error codes are not globally unique
11391 (e.g. an alleged error in ssl3_accept when a certificate
11392 didn't match the private key).
11394 *) New function SSL_CTX_set_session_id_context that allows to set a default
11395 value (so that you don't need SSL_set_session_id_context for each
11396 connection using the SSL_CTX).
11399 *) OAEP decoding bug fix.
11402 *) Support INSTALL_PREFIX for package builders, as proposed by
11406 *) New Configure options "threads" and "no-threads". For systems
11407 where the proper compiler options are known (currently Solaris
11408 and Linux), "threads" is the default.
11411 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11414 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11415 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11416 such as /usr/local/bin.
11419 *) "make linux-shared" to build shared libraries.
11420 [Niels Poppe <niels@netbox.org>]
11422 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11425 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11426 extension adding in x509 utility.
11429 *) Remove NOPROTO sections and error code comments.
11432 *) Partial rewrite of the DEF file generator to now parse the ANSI
11436 *) New Configure options --prefix=DIR and --openssldir=DIR.
11439 *) Complete rewrite of the error code script(s). It is all now handled
11440 by one script at the top level which handles error code gathering,
11441 header rewriting and C source file generation. It should be much better
11442 than the old method: it now uses a modified version of Ulf's parser to
11443 read the ANSI prototypes in all header files (thus the old K&R definitions
11444 aren't needed for error creation any more) and do a better job of
11445 translating function codes into names. The old 'ASN1 error code imbedded
11446 in a comment' is no longer necessary and it doesn't use .err files which
11447 have now been deleted. Also the error code call doesn't have to appear all
11448 on one line (which resulted in some large lines...).
11451 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11454 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11455 0 (which usually indicates a closed connection), but continue reading.
11458 *) Fix some race conditions.
11461 *) Add support for CRL distribution points extension. Add Certificate
11462 Policies and CRL distribution points documentation.
11465 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11468 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11469 8 of keying material. Merlin has also confirmed interop with this fix
11470 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11471 [Merlin Hughes <merlin@baltimore.ie>]
11473 *) Fix lots of warnings.
11474 [Richard Levitte <levitte@stacken.kth.se>]
11476 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11477 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11478 [Richard Levitte <levitte@stacken.kth.se>]
11480 *) Fix problems with sizeof(long) == 8.
11481 [Andy Polyakov <appro@fy.chalmers.se>]
11483 *) Change functions to ANSI C.
11486 *) Fix typos in error codes.
11487 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11489 *) Remove defunct assembler files from Configure.
11492 *) SPARC v8 assembler BIGNUM implementation.
11493 [Andy Polyakov <appro@fy.chalmers.se>]
11495 *) Support for Certificate Policies extension: both print and set.
11496 Various additions to support the r2i method this uses.
11499 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11500 return a const string when you are expecting an allocated buffer.
11503 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11504 types DirectoryString and DisplayText.
11507 *) Add code to allow r2i extensions to access the configuration database,
11508 add an LHASH database driver and add several ctx helper functions.
11511 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11512 fail when they extended the size of a BIGNUM.
11515 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11516 support typesafe stack.
11519 *) Fix typo in SSL_[gs]et_options().
11520 [Nils Frostberg <nils@medcom.se>]
11522 *) Delete various functions and files that belonged to the (now obsolete)
11523 old X509V3 handling code.
11526 *) New Configure option "rsaref".
11529 *) Don't auto-generate pem.h.
11532 *) Introduce type-safe ASN.1 SETs.
11535 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11536 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11538 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11539 that links with OpenSSL (well at least cause lots of warnings), but fear
11540 not: the conversion is trivial, and it eliminates loads of evil casts. A
11541 few STACKed things have been converted already. Feel free to convert more.
11542 In the fullness of time, I'll do away with the STACK type altogether.
11545 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11546 specified in <certfile> by updating the entry in the index.txt file.
11547 This way one no longer has to edit the index.txt file manually for
11548 revoking a certificate. The -revoke option does the gory details now.
11549 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11551 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11552 `-text' option at all and this way the `-noout -text' combination was
11553 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11554 [Ralf S. Engelschall]
11556 *) Make sure a corresponding plain text error message exists for the
11557 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11558 verify callback function determined that a certificate was revoked.
11559 [Ralf S. Engelschall]
11561 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11562 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11563 all available cipers including rc5, which was forgotten until now.
11564 In order to let the testing shell script know which algorithms
11565 are available, a new (up to now undocumented) command
11566 "openssl list-cipher-commands" is used.
11569 *) Bugfix: s_client occasionally would sleep in select() when
11570 it should have checked SSL_pending() first.
11573 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11574 the raw DSA values prior to ASN.1 encoding.
11577 *) Tweaks to Configure
11578 [Niels Poppe <niels@netbox.org>]
11580 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11584 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11587 *) New config option to avoid instructions that are illegal on the 80386.
11588 The default code is faster, but requires at least a 486.
11591 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11592 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11593 same as SSL2_VERSION anyway.
11596 *) New "-showcerts" option for s_client.
11599 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11600 application. Various cleanups and fixes.
11603 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11604 modify error routines to work internally. Add error codes and PBE init
11605 to library startup routines.
11608 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11609 packing functions to asn1 and evp. Changed function names and error
11610 codes along the way.
11613 *) PKCS12 integration: and so it begins... First of several patches to
11614 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11615 objects to objects.h
11618 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11619 and display support for Thawte strong extranet extension.
11622 *) Add LinuxPPC support.
11623 [Jeff Dubrule <igor@pobox.org>]
11625 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11626 bn_div_words in alpha.s.
11627 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11629 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11630 OAEP isn't supported when OpenSSL is built with RSAref.
11631 [Ulf Moeller <ulf@fitug.de>]
11633 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11634 so they no longer are missing under -DNOPROTO.
11635 [Soren S. Jorvang <soren@t.dk>]
11638 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11640 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11641 doesn't work when the session is reused. Coming soon!
11644 *) Fix a security hole, that allows sessions to be reused in the wrong
11645 context thus bypassing client cert protection! All software that uses
11646 client certs and session caches in multiple contexts NEEDS PATCHING to
11647 allow session reuse! A fuller solution is in the works.
11648 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11650 *) Some more source tree cleanups (removed obsolete files
11651 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11652 permission on "config" script to be executable) and a fix for the INSTALL
11654 [Ulf Moeller <ulf@fitug.de>]
11656 *) Remove some legacy and erroneous uses of malloc, free instead of
11658 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11660 *) Make rsa_oaep_test return non-zero on error.
11661 [Ulf Moeller <ulf@fitug.de>]
11663 *) Add support for native Solaris shared libraries. Configure
11664 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11665 if someone would make that last step automatic.
11666 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11668 *) ctx_size was not built with the right compiler during "make links". Fixed.
11671 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11672 except NULL ciphers". This means the default cipher list will no longer
11673 enable NULL ciphers. They need to be specifically enabled e.g. with
11674 the string "DEFAULT:eNULL".
11677 *) Fix to RSA private encryption routines: if p < q then it would
11678 occasionally produce an invalid result. This will only happen with
11679 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11682 *) Be less restrictive and allow also `perl util/perlpath.pl
11683 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11684 because this way one can also use an interpreter named `perl5' (which is
11685 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11686 installed as `perl').
11687 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11689 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11690 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11692 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11693 advapi32.lib to Win32 build and change the pem test comparison
11694 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11695 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11696 and crypto/des/ede_cbcm_enc.c.
11699 *) DES quad checksum was broken on big-endian architectures. Fixed.
11702 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11703 Win32 test batch file so it (might) work again. The Win32 test batch file
11704 is horrible: I feel ill....
11707 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11708 in e_os.h. Audit of header files to check ANSI and non ANSI
11709 sections: 10 functions were absent from non ANSI section and not exported
11710 from Windows DLLs. Fixed up libeay.num for new functions.
11713 *) Make `openssl version' output lines consistent.
11714 [Ralf S. Engelschall]
11716 *) Fix Win32 symbol export lists for BIO functions: Added
11717 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11718 to ms/libeay{16,32}.def.
11719 [Ralf S. Engelschall]
11721 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11722 fine under Unix and passes some trivial tests I've now added. But the
11723 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11724 added to make sure no one expects that this stuff really works in the
11725 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11726 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11728 [Ralf S. Engelschall]
11730 *) Fix the generation of two part addresses in perl.
11731 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11733 *) Add config entry for Linux on MIPS.
11734 [John Tobey <jtobey@channel1.com>]
11736 *) Make links whenever Configure is run, unless we are on Windoze.
11739 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11740 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11744 *) Add a useful kludge to allow package maintainers to specify compiler and
11745 other platforms details on the command line without having to patch the
11746 Configure script everytime: One now can use ``perl Configure
11747 <id>:<details>'', i.e. platform ids are allowed to have details appended
11748 to them (separated by colons). This is treated as there would be a static
11749 pre-configured entry in Configure's %table under key <id> with value
11750 <details> and ``perl Configure <id>'' is called. So, when you want to
11751 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11752 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11753 now, which overrides the FreeBSD-elf entry on-the-fly.
11754 [Ralf S. Engelschall]
11756 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11759 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11760 on the `perl Configure ...' command line. This way one can compile
11761 OpenSSL libraries with Position Independent Code (PIC) which is needed
11762 for linking it into DSOs.
11763 [Ralf S. Engelschall]
11765 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11769 *) Cleaned up the LICENSE document: The official contact for any license
11770 questions now is the OpenSSL core team under openssl-core@openssl.org.
11771 And add a paragraph about the dual-license situation to make sure people
11772 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11773 to the OpenSSL toolkit.
11774 [Ralf S. Engelschall]
11776 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11777 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11778 Additionally cleaned up the `make links' target: Remove unnecessary
11779 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11780 to speed processing and no longer clutter the display with confusing
11781 stuff. Instead only the actually done links are displayed.
11782 [Ralf S. Engelschall]
11784 *) Permit null encryption ciphersuites, used for authentication only. It used
11785 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11786 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11790 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11791 signed attributes when verifying signatures (this would break them),
11792 the detached data encoding was wrong and public keys obtained using
11793 X509_get_pubkey() weren't freed.
11796 *) Add text documentation for the BUFFER functions. Also added a work around
11797 to a Win95 console bug. This was triggered by the password read stuff: the
11798 last character typed gets carried over to the next fread(). If you were
11799 generating a new cert request using 'req' for example then the last
11800 character of the passphrase would be CR which would then enter the first
11804 *) Added the new `Includes OpenSSL Cryptography Software' button as
11805 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11806 button and can be used by applications based on OpenSSL to show the
11807 relationship to the OpenSSL project.
11808 [Ralf S. Engelschall]
11810 *) Remove confusing variables in function signatures in files
11811 ssl/ssl_lib.c and ssl/ssl.h.
11812 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11814 *) Don't install bss_file.c under PREFIX/include/
11815 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11817 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11818 functions that return function pointers and has support for NT specific
11819 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11820 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11821 unsigned to signed types: this was killing the Win32 compile.
11824 *) Add new certificate file to stack functions,
11825 SSL_add_dir_cert_subjects_to_stack() and
11826 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11827 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11828 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11829 This means that Apache-SSL and similar packages don't have to mess around
11830 to add as many CAs as they want to the preferred list.
11833 *) Experiment with doxygen documentation. Currently only partially applied to
11835 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11836 openssl.doxy as the configuration file.
11839 *) Get rid of remaining C++-style comments which strict C compilers hate.
11840 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11842 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11843 compiled in by default: it has problems with large keys.
11846 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11847 DH private keys and/or callback functions which directly correspond to
11848 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11849 is needed for applications which have to configure certificates on a
11850 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11852 For the RSA certificate situation is makes no difference, but
11853 for the DSA certificate situation this fixes the "no shared cipher"
11854 problem where the OpenSSL cipher selection procedure failed because the
11855 temporary keys were not overtaken from the context and the API provided
11856 no way to reconfigure them.
11857 The new functions now let applications reconfigure the stuff and they
11858 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11859 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11860 non-public-API function ssl_cert_instantiate() is used as a helper
11861 function and also to reduce code redundancy inside ssl_rsa.c.
11862 [Ralf S. Engelschall]
11864 *) Move s_server -dcert and -dkey options out of the undocumented feature
11865 area because they are useful for the DSA situation and should be
11866 recognized by the users.
11867 [Ralf S. Engelschall]
11869 *) Fix the cipher decision scheme for export ciphers: the export bits are
11870 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11871 SSL_EXP_MASK. So, the original variable has to be used instead of the
11872 already masked variable.
11873 [Richard Levitte <levitte@stacken.kth.se>]
11875 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11876 [Richard Levitte <levitte@stacken.kth.se>]
11878 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11879 from `int' to `unsigned int' because it's a length and initialized by
11880 EVP_DigestFinal() which expects an `unsigned int *'.
11881 [Richard Levitte <levitte@stacken.kth.se>]
11883 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11884 script. Instead use the usual Shell->Perl transition trick.
11885 [Ralf S. Engelschall]
11887 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11888 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11889 -noout -modulus' as it's already the case for `openssl rsa -noout
11890 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11891 currently the public key is printed (a decision which was already done by
11892 `openssl dsa -modulus' in the past) which serves a similar purpose.
11893 Additionally the NO_RSA no longer completely removes the whole -modulus
11894 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11896 [Ralf S. Engelschall]
11898 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11899 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11900 [Arne Ansper <arne@ats.cyber.ee>]
11902 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11903 to be added. Now both 'req' and 'ca' can use new objects defined in the
11907 *) Add cool BIO that does syslog (or event log on NT).
11908 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11910 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11911 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11912 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11913 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11916 *) Add preliminary config info for new extension code.
11919 *) Make RSA_NO_PADDING really use no padding.
11920 [Ulf Moeller <ulf@fitug.de>]
11922 *) Generate errors when private/public key check is done.
11925 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11926 for some CRL extensions and new objects added.
11929 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11930 key usage extension and fuller support for authority key id.
11933 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11934 padding method for RSA, which is recommended for new applications in PKCS
11935 #1 v2.0 (RFC 2437, October 1998).
11936 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11937 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11938 against Bleichbacher's attack on RSA.
11939 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11942 *) Updates to the new SSL compression code
11943 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11945 *) Fix so that the version number in the master secret, when passed
11946 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11947 (because the server will not accept higher), that the version number
11948 is 0x03,0x01, not 0x03,0x00
11949 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11951 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11952 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11953 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11956 *) Support for RAW extensions where an arbitrary extension can be
11957 created by including its DER encoding. See apps/openssl.cnf for
11961 *) Make sure latest Perl versions don't interpret some generated C array
11962 code as Perl array code in the crypto/err/err_genc.pl script.
11963 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11965 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11966 not many people have the assembler. Various Win32 compilation fixes and
11967 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11968 build instructions.
11971 *) Modify configure script 'Configure' to automatically create crypto/date.h
11972 file under Win32 and also build pem.h from pem.org. New script
11973 util/mkfiles.pl to create the MINFO file on environments that can't do a
11974 'make files': perl util/mkfiles.pl >MINFO should work.
11977 *) Major rework of DES function declarations, in the pursuit of correctness
11978 and purity. As a result, many evil casts evaporated, and some weirdness,
11979 too. You may find this causes warnings in your code. Zapping your evil
11980 casts will probably fix them. Mostly.
11983 *) Fix for a typo in asn1.h. Bug fix to object creation script
11984 obj_dat.pl. It considered a zero in an object definition to mean
11985 "end of object": none of the objects in objects.h have any zeros
11986 so it wasn't spotted.
11987 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11989 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11990 Masking (CBCM). In the absence of test vectors, the best I have been able
11991 to do is check that the decrypt undoes the encrypt, so far. Send me test
11992 vectors if you have them.
11995 *) Correct calculation of key length for export ciphers (too much space was
11996 allocated for null ciphers). This has not been tested!
11999 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12000 message is now correct (it understands "crypto" and "ssl" on its
12001 command line). There is also now an "update" option. This will update
12002 the util/ssleay.num and util/libeay.num files with any new functions.
12004 perl util/mkdef.pl crypto ssl update
12005 it will update them.
12008 *) Overhauled the Perl interface (perl/*):
12009 - ported BN stuff to OpenSSL's different BN library
12010 - made the perl/ source tree CVS-aware
12011 - renamed the package from SSLeay to OpenSSL (the files still contain
12012 their history because I've copied them in the repository)
12013 - removed obsolete files (the test scripts will be replaced
12014 by better Test::Harness variants in the future)
12015 [Ralf S. Engelschall]
12017 *) First cut for a very conservative source tree cleanup:
12018 1. merge various obsolete readme texts into doc/ssleay.txt
12019 where we collect the old documents and readme texts.
12020 2. remove the first part of files where I'm already sure that we no
12021 longer need them because of three reasons: either they are just temporary
12022 files which were left by Eric or they are preserved original files where
12023 I've verified that the diff is also available in the CVS via "cvs diff
12024 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12025 the crypto/md/ stuff).
12026 [Ralf S. Engelschall]
12028 *) More extension code. Incomplete support for subject and issuer alt
12029 name, issuer and authority key id. Change the i2v function parameters
12030 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12031 what that's for :-) Fix to ASN1 macro which messed up
12032 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12035 *) Preliminary support for ENUMERATED type. This is largely copied from the
12039 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12040 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12042 *) Make sure `make rehash' target really finds the `openssl' program.
12043 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12045 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12046 like to hear about it if this slows down other processors.
12049 *) Add CygWin32 platform information to Configure script.
12050 [Alan Batie <batie@aahz.jf.intel.com>]
12052 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12053 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12055 *) New program nseq to manipulate netscape certificate sequences
12058 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12062 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12063 but the BN code had some problems that would cause failures when
12064 doing certificate verification and some other functions.
12065 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12067 *) Add ASN1 and PEM code to support netscape certificate sequences.
12070 *) Add ASN1 and PEM code to support netscape certificate sequences.
12073 *) Add several PKIX and private extended key usage OIDs.
12076 *) Modify the 'ca' program to handle the new extension code. Modify
12077 openssl.cnf for new extension format, add comments.
12080 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12081 and add a sample to openssl.cnf so req -x509 now adds appropriate
12085 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12086 error code, add initial support to X509_print() and x509 application.
12089 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12090 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12091 stuff is currently isolated and isn't even compiled yet.
12094 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12095 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12096 Removed the versions check from X509 routines when loading extensions:
12097 this allows certain broken certificates that don't set the version
12098 properly to be processed.
12101 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12102 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12103 can still be regenerated with "make depend".
12106 *) Spelling mistake in C version of CAST-128.
12107 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12109 *) Changes to the error generation code. The perl script err-code.pl
12110 now reads in the old error codes and retains the old numbers, only
12111 adding new ones if necessary. It also only changes the .err files if new
12112 codes are added. The makefiles have been modified to only insert errors
12113 when needed (to avoid needlessly modifying header files). This is done
12114 by only inserting errors if the .err file is newer than the auto generated
12115 C file. To rebuild all the error codes from scratch (the old behaviour)
12116 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12117 or delete all the .err files.
12120 *) CAST-128 was incorrectly implemented for short keys. The C version has
12121 been fixed, but is untested. The assembler versions are also fixed, but
12122 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12123 to regenerate it if needed.
12124 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12125 Hagino <itojun@kame.net>]
12127 *) File was opened incorrectly in randfile.c.
12128 [Ulf Möller <ulf@fitug.de>]
12130 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12131 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12132 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12133 al: it's just almost always a UTCTime. Note this patch adds new error
12134 codes so do a "make errors" if there are problems.
12137 *) Correct Linux 1 recognition in config.
12138 [Ulf Möller <ulf@fitug.de>]
12140 *) Remove pointless MD5 hash when using DSA keys in ca.
12141 [Anonymous <nobody@replay.com>]
12143 *) Generate an error if given an empty string as a cert directory. Also
12144 generate an error if handed NULL (previously returned 0 to indicate an
12145 error, but didn't set one).
12146 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12148 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12151 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12152 parameters. This was causing a warning which killed off the Win32 compile.
12155 *) Remove C++ style comments from crypto/bn/bn_local.h.
12156 [Neil Costigan <neil.costigan@celocom.com>]
12158 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12159 based on a text string, looking up short and long names and finally
12160 "dot" format. The "dot" format stuff didn't work. Added new function
12161 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12162 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12163 OID is not part of the table.
12166 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12167 X509_LOOKUP_by_alias().
12170 *) Sort openssl functions by name.
12173 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12174 encryption from sample DSA keys (in case anyone is interested the password
12178 *) Make _all_ *_free functions accept a NULL pointer.
12179 [Frans Heymans <fheymans@isaserver.be>]
12181 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12183 [Anonymous <nobody@replay.com>]
12185 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12186 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12188 *) Don't blow it for numeric -newkey arguments to apps/req.
12189 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12191 *) Temp key "for export" tests were wrong in s3_srvr.c.
12192 [Anonymous <nobody@replay.com>]
12194 *) Add prototype for temp key callback functions
12195 SSL_CTX_set_tmp_{rsa,dh}_callback().
12198 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12199 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12202 *) X509_name_add_entry() freed the wrong thing after an error.
12203 [Arne Ansper <arne@ats.cyber.ee>]
12205 *) rsa_eay.c would attempt to free a NULL context.
12206 [Arne Ansper <arne@ats.cyber.ee>]
12208 *) BIO_s_socket() had a broken should_retry() on Windoze.
12209 [Arne Ansper <arne@ats.cyber.ee>]
12211 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12212 [Arne Ansper <arne@ats.cyber.ee>]
12214 *) Make sure the already existing X509_STORE->depth variable is initialized
12215 in X509_STORE_new(), but document the fact that this variable is still
12216 unused in the certificate verification process.
12217 [Ralf S. Engelschall]
12219 *) Fix the various library and apps files to free up pkeys obtained from
12220 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12223 *) Fix reference counting in X509_PUBKEY_get(). This makes
12224 demos/maurice/example2.c work, amongst others, probably.
12225 [Steve Henson and Ben Laurie]
12227 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12228 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12229 are no longer created. This way we have a single and consistent command
12230 line interface `openssl <command>', similar to `cvs <command>'.
12231 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12233 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12234 BIT STRING wrapper always have zero unused bits.
12237 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12240 *) Make the top-level INSTALL documentation easier to understand.
12243 *) Makefiles updated to exit if an error occurs in a sub-directory
12244 make (including if user presses ^C) [Paul Sutton]
12246 *) Make Montgomery context stuff explicit in RSA data structure.
12249 *) Fix build order of pem and err to allow for generated pem.h.
12252 *) Fix renumbering bug in X509_NAME_delete_entry().
12255 *) Enhanced the err-ins.pl script so it makes the error library number
12256 global and can add a library name. This is needed for external ASN1 and
12257 other error libraries.
12260 *) Fixed sk_insert which never worked properly.
12263 *) Fix ASN1 macros so they can handle indefinite length constructed
12264 EXPLICIT tags. Some non standard certificates use these: they can now
12268 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12269 into a single doc/ssleay.txt bundle. This way the information is still
12270 preserved but no longer messes up this directory. Now it's new room for
12271 the new set of documentation files.
12272 [Ralf S. Engelschall]
12274 *) SETs were incorrectly DER encoded. This was a major pain, because they
12275 shared code with SEQUENCEs, which aren't coded the same. This means that
12276 almost everything to do with SETs or SEQUENCEs has either changed name or
12277 number of arguments.
12278 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12280 *) Fix test data to work with the above.
12283 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12284 was already fixed by Eric for 0.9.1 it seems.
12285 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12287 *) Autodetect FreeBSD3.
12290 *) Fix various bugs in Configure. This affects the following platforms:
12294 unixware-2.0-pentium
12298 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12299 before they are needed.
12302 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12306 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12308 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12309 changed SSLeay to OpenSSL in version strings.
12310 [Ralf S. Engelschall]
12312 *) Some fixups to the top-level documents.
12315 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12316 because the symlink to include/ was missing.
12317 [Ralf S. Engelschall]
12319 *) Incorporated the popular no-RSA/DSA-only patches
12320 which allow to compile a RSA-free SSLeay.
12321 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12323 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12324 when "ssleay" is still not found.
12325 [Ralf S. Engelschall]
12327 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12328 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12330 *) Updated the README file.
12331 [Ralf S. Engelschall]
12333 *) Added various .cvsignore files in the CVS repository subdirs
12334 to make a "cvs update" really silent.
12335 [Ralf S. Engelschall]
12337 *) Recompiled the error-definition header files and added
12338 missing symbols to the Win32 linker tables.
12339 [Ralf S. Engelschall]
12341 *) Cleaned up the top-level documents;
12342 o new files: CHANGES and LICENSE
12343 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12344 o merged COPYRIGHT into LICENSE
12345 o removed obsolete TODO file
12346 o renamed MICROSOFT to INSTALL.W32
12347 [Ralf S. Engelschall]
12349 *) Removed dummy files from the 0.9.1b source tree:
12350 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12351 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12352 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12353 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12354 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12355 [Ralf S. Engelschall]
12357 *) Added various platform portability fixes.
12360 *) The Genesis of the OpenSSL rpject:
12361 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12362 Young and Tim J. Hudson created while they were working for C2Net until
12364 [The OpenSSL Project]
12367 Changes between 0.9.0b and 0.9.1b [not released]
12369 *) Updated a few CA certificates under certs/
12372 *) Changed some BIGNUM api stuff.
12375 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12376 DGUX x86, Linux Alpha, etc.
12379 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12380 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12384 *) Add -strparse option to asn1pars program which parses nested
12386 [Dr Stephen Henson <shenson@bigfoot.com>]
12388 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12391 *) DSA fix for "ca" program.
12394 *) Added "-genkey" option to "dsaparam" program.
12397 *) Added RIPE MD160 (rmd160) message digest.
12400 *) Added -a (all) option to "ssleay version" command.
12403 *) Added PLATFORM define which is the id given to Configure.
12406 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12409 *) Extended the ASN.1 parser routines.
12412 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12415 *) Added a BN_CTX to the BN library.
12418 *) Fixed the weak key values in DES library
12421 *) Changed API in EVP library for cipher aliases.
12424 *) Added support for RC2/64bit cipher.
12427 *) Converted the lhash library to the crypto/mem.c functions.
12430 *) Added more recognized ASN.1 object ids.
12433 *) Added more RSA padding checks for SSL/TLS.
12436 *) Added BIO proxy/filter functionality.
12439 *) Added extra_certs to SSL_CTX which can be used
12440 send extra CA certificates to the client in the CA cert chain sending
12441 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12444 *) Now Fortezza is denied in the authentication phase because
12445 this is key exchange mechanism is not supported by SSLeay at all.
12448 *) Additional PKCS1 checks.
12451 *) Support the string "TLSv1" for all TLS v1 ciphers.
12454 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12455 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12458 *) Fixed a few memory leaks.
12461 *) Fixed various code and comment typos.
12464 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12465 bytes sent in the client random.
12466 [Edward Bishop <ebishop@spyglass.com>]