5 Changes between 1.1.0a and 1.1.1 [xx XXX xxxx]
9 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
10 using the algorithm defined in
11 https://www.akkadia.org/drepper/SHA-crypt.txt
14 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
16 *) OCSP Status Request extension unbounded memory growth
18 A malicious client can send an excessively large OCSP Status Request
19 extension. If that client continually requests renegotiation, sending a
20 large OCSP Status Request extension each time, then there will be unbounded
21 memory growth on the server. This will eventually lead to a Denial Of
22 Service attack through memory exhaustion. Servers with a default
23 configuration are vulnerable even if they do not support OCSP. Builds using
24 the "no-ocsp" build time option are not affected.
26 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
30 *) SSL_peek() hang on empty record
32 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
33 sends an empty record. This could be exploited by a malicious peer in a
34 Denial Of Service attack.
36 This issue was reported to OpenSSL by Alex Gaynor.
40 *) Excessive allocation of memory in tls_get_message_header() and
41 dtls1_preprocess_fragment()
43 A (D)TLS message includes 3 bytes for its length in the header for the
44 message. This would allow for messages up to 16Mb in length. Messages of
45 this length are excessive and OpenSSL includes a check to ensure that a
46 peer is sending reasonably sized messages in order to avoid too much memory
47 being consumed to service a connection. A flaw in the logic of version
48 1.1.0 means that memory for the message is allocated too early, prior to
49 the excessive message length check. Due to way memory is allocated in
50 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
51 to service a connection. This could lead to a Denial of Service through
52 memory exhaustion. However, the excessive message length check still takes
53 place, and this would cause the connection to immediately fail. Assuming
54 that the application calls SSL_free() on the failed conneciton in a timely
55 manner then the 21Mb of allocated memory will then be immediately freed
56 again. Therefore the excessive memory allocation will be transitory in
57 nature. This then means that there is only a security impact if:
59 1) The application does not call SSL_free() in a timely manner in the event
60 that the connection fails
62 2) The application is working in a constrained environment where there is
63 very little free memory
65 3) The attacker initiates multiple connection attempts such that there are
66 multiple connections in a state where memory has been allocated for the
67 connection; SSL_free() has not yet been called; and there is insufficient
68 memory to service the multiple requests.
70 Except in the instance of (1) above any Denial Of Service is likely to be
71 transitory because as soon as the connection fails the memory is
72 subsequently freed again in the SSL_free() call. However there is an
73 increased risk during this period of application crashes due to the lack of
74 memory - which would then mean a more serious Denial of Service.
76 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
77 (CVE-2016-6307 and CVE-2016-6308)
80 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
81 had to be removed. Primary reason is that vendor assembler can't
82 assemble our modules with -KPIC flag. As result it, assembly
83 support, was not even available as option. But its lack means
84 lack of side-channel resistant code, which is incompatible with
85 security by todays standards. Fortunately gcc is readily available
86 prepackaged option, which we firmly point at...
89 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
91 *) Windows command-line tool supports UTF-8 opt-in option for arguments
92 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
93 (to any value) allows Windows user to access PKCS#12 file generated
94 with Windows CryptoAPI and protected with non-ASCII password, as well
95 as files generated under UTF-8 locale on Linux also protected with
99 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
100 have been disabled by default and removed from DEFAULT, just like RC4.
101 See the RC4 item below to re-enable both.
104 *) The method for finding the storage location for the Windows RAND seed file
105 has changed. First we check %RANDFILE%. If that is not set then we check
106 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
107 all else fails we fall back to C:\.
110 *) The EVP_EncryptUpdate() function has had its return type changed from void
111 to int. A return of 0 indicates and error while a return of 1 indicates
115 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
116 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
117 off the constant time implementation for RSA, DSA and DH have been made
118 no-ops and deprecated.
121 *) Windows RAND implementation was simplified to only get entropy by
122 calling CryptGenRandom(). Various other RAND-related tickets
124 [Joseph Wylie Yandle, Rich Salz]
126 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
127 and OPENSSL_LH_, respectively. The old names are available
128 with API compatibility. They new names are now completely documented.
131 *) Unify TYPE_up_ref(obj) methods signature.
132 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
133 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
134 int (instead of void) like all others TYPE_up_ref() methods.
135 So now these methods also check the return value of CRYPTO_atomic_add(),
136 and the validity of object reference counter.
137 [fdasilvayy@gmail.com]
139 *) With Windows Visual Studio builds, the .pdb files are installed
140 alongside the installed libraries and executables. For a static
141 library installation, ossl_static.pdb is the associate compiler
142 generated .pdb file to be used when linking programs.
145 *) Remove openssl.spec. Packaging files belong with the packagers.
148 *) Automatic Darwin/OSX configuration has had a refresh, it will now
149 recognise x86_64 architectures automatically. You can still decide
150 to build for a different bitness with the environment variable
151 KERNEL_BITS (can be 32 or 64), for example:
153 KERNEL_BITS=32 ./config
157 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
158 256 bit AES and HMAC with SHA256.
161 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
164 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
167 *) To enable users to have their own config files and build file templates,
168 Configure looks in the directory indicated by the environment variable
169 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
170 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
171 name and is used as is.
174 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
175 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
176 X509_CERT_FILE_CTX was removed.
179 *) "shared" builds are now the default. To create only static libraries use
180 the "no-shared" Configure option.
183 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
184 All of these option have not worked for some while and are fundamental
188 *) Make various cleanup routines no-ops and mark them as deprecated. Most
189 global cleanup functions are no longer required because they are handled
190 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
191 Explicitly de-initing can cause problems (e.g. where a library that uses
192 OpenSSL de-inits, but an application is still using it). The affected
193 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
194 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
195 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
199 *) --strict-warnings no longer enables runtime debugging options
200 such as REF_DEBUG. Instead, debug options are automatically
201 enabled with '--debug' builds.
202 [Andy Polyakov, Emilia Käsper]
204 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
205 have been moved out of the public header files. New functions for managing
206 these have been added.
209 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
210 objects have been moved out of the public header files. New
211 functions for managing these have been added.
214 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
215 have been moved out of the public header files. New functions for managing
216 these have been added.
219 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
220 moved out of the public header files. New functions for managing these
224 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
227 *) Removed the mk1mf build scripts.
230 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
231 it is always safe to #include a header now.
234 *) Removed the aged BC-32 config and all its supporting scripts
237 *) Removed support for Ultrix, Netware, and OS/2.
240 *) Add support for HKDF.
243 *) Add support for blake2b and blake2s
246 *) Added support for "pipelining". Ciphers that have the
247 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
248 encryptions/decryptions simultaneously. There are currently no built-in
249 ciphers with this property but the expectation is that engines will be able
250 to offer it to significantly improve throughput. Support has been extended
251 into libssl so that multiple records for a single connection can be
252 processed in one go (for >=TLS 1.1).
255 *) Added the AFALG engine. This is an async capable engine which is able to
256 offload work to the Linux kernel. In this initial version it only supports
257 AES128-CBC. The kernel must be version 4.1.0 or greater.
260 *) OpenSSL now uses a new threading API. It is no longer necessary to
261 set locking callbacks to use OpenSSL in a multi-threaded environment. There
262 are two supported threading models: pthreads and windows threads. It is
263 also possible to configure OpenSSL at compile time for "no-threads". The
264 old threading API should no longer be used. The functions have been
265 replaced with "no-op" compatibility macros.
266 [Alessandro Ghedini, Matt Caswell]
268 *) Modify behavior of ALPN to invoke callback after SNI/servername
269 callback, such that updates to the SSL_CTX affect ALPN.
272 *) Add SSL_CIPHER queries for authentication and key-exchange.
275 *) Changes to the DEFAULT cipherlist:
276 - Prefer (EC)DHE handshakes over plain RSA.
277 - Prefer AEAD ciphers over legacy ciphers.
278 - Prefer ECDSA over RSA when both certificates are available.
279 - Prefer TLSv1.2 ciphers/PRF.
280 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
284 *) Change the ECC default curve list to be this, in order: x25519,
285 secp256r1, secp521r1, secp384r1.
288 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
289 disabled by default. They can be re-enabled using the
290 enable-weak-ssl-ciphers option to Configure.
293 *) If the server has ALPN configured, but supports no protocols that the
294 client advertises, send a fatal "no_application_protocol" alert.
295 This behaviour is SHALL in RFC 7301, though it isn't universally
296 implemented by other servers.
299 *) Add X25519 support.
300 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
301 for public and private key encoding using the format documented in
302 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
303 key generation and key derivation.
305 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
309 *) Deprecate SRP_VBASE_get_by_user.
310 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
311 In order to fix an unavoidable memory leak (CVE-2016-0798),
312 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
313 seed, even if the seed is configured.
315 Users should use SRP_VBASE_get1_by_user instead. Note that in
316 SRP_VBASE_get1_by_user, caller must free the returned value. Note
317 also that even though configuring the SRP seed attempts to hide
318 invalid usernames by continuing the handshake with fake
319 credentials, this behaviour is not constant time and no strong
320 guarantees are made that the handshake is indistinguishable from
321 that of a valid user.
324 *) Configuration change; it's now possible to build dynamic engines
325 without having to build shared libraries and vice versa. This
326 only applies to the engines in engines/, those in crypto/engine/
327 will always be built into libcrypto (i.e. "static").
329 Building dynamic engines is enabled by default; to disable, use
330 the configuration option "disable-dynamic-engine".
332 The only requirements for building dynamic engines are the
333 presence of the DSO module and building with position independent
334 code, so they will also automatically be disabled if configuring
335 with "disable-dso" or "disable-pic".
337 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
338 are also taken away from openssl/opensslconf.h, as they are
342 *) Configuration change; if there is a known flag to compile
343 position independent code, it will always be applied on the
344 libcrypto and libssl object files, and never on the application
345 object files. This means other libraries that use routines from
346 libcrypto / libssl can be made into shared libraries regardless
347 of how OpenSSL was configured.
349 If this isn't desirable, the configuration options "disable-pic"
350 or "no-pic" can be used to disable the use of PIC. This will
351 also disable building shared libraries and dynamic engines.
354 *) Removed JPAKE code. It was experimental and has no wide use.
357 *) The INSTALL_PREFIX Makefile variable has been renamed to
358 DESTDIR. That makes for less confusion on what this variable
359 is for. Also, the configuration option --install_prefix is
363 *) Heartbeat for TLS has been removed and is disabled by default
364 for DTLS; configure with enable-heartbeats. Code that uses the
365 old #define's might need to be updated.
366 [Emilia Käsper, Rich Salz]
368 *) Rename REF_CHECK to REF_DEBUG.
371 *) New "unified" build system
373 The "unified" build system is aimed to be a common system for all
374 platforms we support. With it comes new support for VMS.
376 This system builds supports building in a different directory tree
377 than the source tree. It produces one Makefile (for unix family
378 or lookalikes), or one descrip.mms (for VMS).
380 The source of information to make the Makefile / descrip.mms is
381 small files called 'build.info', holding the necessary
382 information for each directory with source to compile, and a
383 template in Configurations, like unix-Makefile.tmpl or
386 We rely heavily on the perl module Text::Template.
389 *) Added support for auto-initialisation and de-initialisation of the library.
390 OpenSSL no longer requires explicit init or deinit routines to be called,
391 except in certain circumstances. See the OPENSSL_init_crypto() and
392 OPENSSL_init_ssl() man pages for further information.
395 *) The arguments to the DTLSv1_listen function have changed. Specifically the
396 "peer" argument is now expected to be a BIO_ADDR object.
398 *) Rewrite of BIO networking library. The BIO library lacked consistent
399 support of IPv6, and adding it required some more extensive
400 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
401 which hold all types of addresses and chains of address information.
402 It also introduces a new API, with functions like BIO_socket,
403 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
404 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
405 have been adapted accordingly.
408 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
412 *) CRIME protection: disable compression by default, even if OpenSSL is
413 compiled with zlib enabled. Applications can still enable compression
414 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
415 using the SSL_CONF library to configure compression.
418 *) The signature of the session callback configured with
419 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
420 was explicitly marked as 'const unsigned char*' instead of
424 *) Always DPURIFY. Remove the use of uninitialized memory in the
425 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
428 *) Removed many obsolete configuration items, including
429 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
430 MD2_CHAR, MD2_INT, MD2_LONG
432 IDEA_SHORT, IDEA_LONG
433 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
434 [Rich Salz, with advice from Andy Polyakov]
436 *) Many BN internals have been moved to an internal header file.
437 [Rich Salz with help from Andy Polyakov]
439 *) Configuration and writing out the results from it has changed.
440 Files such as Makefile include/openssl/opensslconf.h and are now
441 produced through general templates, such as Makefile.in and
442 crypto/opensslconf.h.in and some help from the perl module
445 Also, the center of configuration information is no longer
446 Makefile. Instead, Configure produces a perl module in
447 configdata.pm which holds most of the config data (in the hash
448 table %config), the target data that comes from the target
449 configuration in one of the Configurations/*.conf files (in
453 *) To clarify their intended purposes, the Configure options
454 --prefix and --openssldir change their semantics, and become more
455 straightforward and less interdependent.
457 --prefix shall be used exclusively to give the location INSTALLTOP
458 where programs, scripts, libraries, include files and manuals are
459 going to be installed. The default is now /usr/local.
461 --openssldir shall be used exclusively to give the default
462 location OPENSSLDIR where certificates, private keys, CRLs are
463 managed. This is also where the default openssl.cnf gets
465 If the directory given with this option is a relative path, the
466 values of both the --prefix value and the --openssldir value will
467 be combined to become OPENSSLDIR.
468 The default for --openssldir is INSTALLTOP/ssl.
470 Anyone who uses --openssldir to specify where OpenSSL is to be
471 installed MUST change to use --prefix instead.
474 *) The GOST engine was out of date and therefore it has been removed. An up
475 to date GOST engine is now being maintained in an external repository.
476 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
477 support for GOST ciphersuites (these are only activated if a GOST engine
481 *) EGD is no longer supported by default; use enable-egd when
483 [Ben Kaduk and Rich Salz]
485 *) The distribution now has Makefile.in files, which are used to
486 create Makefile's when Configure is run. *Configure must be run
487 before trying to build now.*
490 *) The return value for SSL_CIPHER_description() for error conditions
494 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
496 Obtaining and performing DNSSEC validation of TLSA records is
497 the application's responsibility. The application provides
498 the TLSA records of its choice to OpenSSL, and these are then
499 used to authenticate the peer.
501 The TLSA records need not even come from DNS. They can, for
502 example, be used to implement local end-entity certificate or
503 trust-anchor "pinning", where the "pin" data takes the form
504 of TLSA records, which can augment or replace verification
505 based on the usual WebPKI public certification authorities.
508 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
509 continues to support deprecated interfaces in default builds.
510 However, applications are strongly advised to compile their
511 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
512 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
513 or the 1.1.0 releases.
515 In environments in which all applications have been ported to
516 not use any deprecated interfaces OpenSSL's Configure script
517 should be used with the --api=1.1.0 option to entirely remove
518 support for the deprecated features from the library and
519 unconditionally disable them in the installed headers.
520 Essentially the same effect can be achieved with the "no-deprecated"
521 argument to Configure, except that this will always restrict
522 the build to just the latest API, rather than a fixed API
525 As applications are ported to future revisions of the API,
526 they should update their compile-time OPENSSL_API_COMPAT define
527 accordingly, but in most cases should be able to continue to
528 compile with later releases.
530 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
531 0x10000000L and 0x00908000L, respectively. However those
532 versions did not support the OPENSSL_API_COMPAT feature, and
533 so applications are not typically tested for explicit support
534 of just the undeprecated features of either release.
537 *) Add support for setting the minimum and maximum supported protocol.
538 It can bet set via the SSL_set_min_proto_version() and
539 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
540 MaxProtcol. It's recommended to use the new APIs to disable
541 protocols instead of disabling individual protocols using
542 SSL_set_options() or SSL_CONF's Protocol. This change also
543 removes support for disabling TLS 1.2 in the OpenSSL TLS
544 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
547 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
550 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
551 and integrates ECDSA and ECDH functionality into EC. Implementations can
552 now redirect key generation and no longer need to convert to or from
555 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
556 include the ec.h header file instead.
559 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
560 ciphers who are no longer supported and drops support the ephemeral RSA key
561 exchange. The LOW ciphers currently doesn't have any ciphers in it.
564 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
565 opaque. For HMAC_CTX, the following constructors and destructors
568 HMAC_CTX *HMAC_CTX_new(void);
569 void HMAC_CTX_free(HMAC_CTX *ctx);
571 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
572 destroy such methods has been added. See EVP_MD_meth_new(3) and
573 EVP_CIPHER_meth_new(3) for documentation.
576 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
577 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
578 EVP_MD_CTX_reset() should be called instead to reinitialise
579 an already created structure.
580 2) For consistency with the majority of our object creators and
581 destructors, EVP_MD_CTX_(create|destroy) were renamed to
582 EVP_MD_CTX_(new|free). The old names are retained as macros
583 for deprecated builds.
586 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
587 cryptographic operations to be performed asynchronously as long as an
588 asynchronous capable engine is used. See the ASYNC_start_job() man page for
589 further details. Libssl has also had this capability integrated with the
590 introduction of the new mode SSL_MODE_ASYNC and associated error
591 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
592 pages. This work was developed in partnership with Intel Corp.
595 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
596 always enabled now. If you want to disable the support you should
597 exclude it using the list of supported ciphers. This also means that the
598 "-no_ecdhe" option has been removed from s_server.
601 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
602 SSL_{CTX_}set1_curves() which can set a list.
605 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
606 curve you want to support using SSL_{CTX_}set1_curves().
609 *) State machine rewrite. The state machine code has been significantly
610 refactored in order to remove much duplication of code and solve issues
611 with the old code (see ssl/statem/README for further details). This change
612 does have some associated API changes. Notably the SSL_state() function
613 has been removed and replaced by SSL_get_state which now returns an
614 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
615 altogether. The previous handshake states defined in ssl.h and ssl3.h have
619 *) All instances of the string "ssleay" in the public API were replaced
620 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
621 Some error codes related to internal RSA_eay API's were renamed.
624 *) The demo files in crypto/threads were moved to demo/threads.
627 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
629 [Matt Caswell, Rich Salz]
631 *) New ASN.1 embed macro.
633 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
634 structure is not allocated: it is part of the parent. That is instead of
642 This reduces memory fragmentation and make it impossible to accidentally
643 set a mandatory field to NULL.
645 This currently only works for some fields specifically a SEQUENCE, CHOICE,
646 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
647 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
651 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
654 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
655 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
656 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
657 DES and RC4 ciphersuites.
660 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
661 This changes the decoding behaviour for some invalid messages,
662 though the change is mostly in the more lenient direction, and
663 legacy behaviour is preserved as much as possible.
666 *) Fix no-stdio build.
667 [ David Woodhouse <David.Woodhouse@intel.com> and also
668 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
670 *) New testing framework
671 The testing framework has been largely rewritten and is now using
672 perl and the perl modules Test::Harness and an extended variant of
673 Test::More called OpenSSL::Test to do its work. All test scripts in
674 test/ have been rewritten into test recipes, and all direct calls to
675 executables in test/Makefile have become individual recipes using the
676 simplified testing OpenSSL::Test::Simple.
678 For documentation on our testing modules, do:
680 perldoc test/testlib/OpenSSL/Test/Simple.pm
681 perldoc test/testlib/OpenSSL/Test.pm
685 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
686 are used; the latter aborts on memory leaks (usually checked on exit).
687 Some undocumented "set malloc, etc., hooks" functions were removed
688 and others were changed. All are now documented.
691 *) In DSA_generate_parameters_ex, if the provided seed is too short,
693 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
695 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
696 from RFC4279, RFC4785, RFC5487, RFC5489.
698 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
699 original RSA_PSK patch.
702 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
703 era flag was never set throughout the codebase (only read). Also removed
704 SSL3_FLAGS_POP_BUFFER which was only used if
705 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
708 *) Changed the default name options in the "ca", "crl", "req" and "x509"
709 to be "oneline" instead of "compat".
712 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
713 not aware of clients that still exhibit this bug, and the workaround
714 hasn't been working properly for a while.
717 *) The return type of BIO_number_read() and BIO_number_written() as well as
718 the corresponding num_read and num_write members in the BIO structure has
719 changed from unsigned long to uint64_t. On platforms where an unsigned
720 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
724 *) Given the pervasive nature of TLS extensions it is inadvisable to run
725 OpenSSL without support for them. It also means that maintaining
726 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
727 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
730 *) Removed support for the two export grade static DH ciphersuites
731 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
732 were newly added (along with a number of other static DH ciphersuites) to
733 1.0.2. However the two export ones have *never* worked since they were
734 introduced. It seems strange in any case to be adding new export
735 ciphersuites, and given "logjam" it also does not seem correct to fix them.
738 *) Version negotiation has been rewritten. In particular SSLv23_method(),
739 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
740 and turned into macros which simply call the new preferred function names
741 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
742 should use the new names instead. Also as part of this change the ssl23.h
743 header file has been removed.
746 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
747 code and the associated standard is no longer considered fit-for-purpose.
750 *) RT2547 was closed. When generating a private key, try to make the
751 output file readable only by the owner. This behavior change might
752 be noticeable when interacting with other software.
754 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
758 *) Added HTTP GET support to the ocsp command.
761 *) Changed default digest for the dgst and enc commands from MD5 to
765 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
768 *) Added support for TLS extended master secret from
769 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
770 initial patch which was a great help during development.
773 *) All libssl internal structures have been removed from the public header
774 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
775 now redundant). Users should not attempt to access internal structures
776 directly. Instead they should use the provided API functions.
779 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
780 Access to deprecated functions can be re-enabled by running config with
781 "enable-deprecated". In addition applications wishing to use deprecated
782 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
783 will, by default, disable some transitive includes that previously existed
784 in the header files (e.g. ec.h will no longer, by default, include bn.h)
787 *) Added support for OCB mode. OpenSSL has been granted a patent license
788 compatible with the OpenSSL license for use of OCB. Details are available
789 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
790 for OCB can be removed by calling config with no-ocb.
793 *) SSLv2 support has been removed. It still supports receiving a SSLv2
794 compatible client hello.
797 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
798 done while fixing the error code for the key-too-small case.
799 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
801 *) CA.sh has been removmed; use CA.pl instead.
804 *) Removed old DES API.
807 *) Remove various unsupported platforms:
813 Sinix/ReliantUNIX RM400
818 16-bit platforms such as WIN16
821 *) Clean up OPENSSL_NO_xxx #define's
822 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
823 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
824 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
825 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
826 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
827 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
828 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
829 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
830 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
831 Remove MS_STATIC; it's a relic from platforms <32 bits.
834 *) Cleaned up dead code
835 Remove all but one '#ifdef undef' which is to be looked at.
838 *) Clean up calling of xxx_free routines.
839 Just like free(), fix most of the xxx_free routines to accept
840 NULL. Remove the non-null checks from callers. Save much code.
843 *) Add secure heap for storage of private keys (when possible).
844 Add BIO_s_secmem(), CBIGNUM, etc.
845 Contributed by Akamai Technologies under our Corporate CLA.
848 *) Experimental support for a new, fast, unbiased prime candidate generator,
849 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
850 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
852 *) New output format NSS in the sess_id command line tool. This allows
853 exporting the session id and the master key in NSS keylog format.
854 [Martin Kaiser <martin@kaiser.cx>]
856 *) Harmonize version and its documentation. -f flag is used to display
858 [mancha <mancha1@zoho.com>]
860 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
861 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
862 [mancha <mancha1@zoho.com>]
864 *) Fix some double frees. These are not thought to be exploitable.
865 [mancha <mancha1@zoho.com>]
867 *) A missing bounds check in the handling of the TLS heartbeat extension
868 can be used to reveal up to 64k of memory to a connected client or
871 Thanks for Neel Mehta of Google Security for discovering this bug and to
872 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
873 preparing the fix (CVE-2014-0160)
874 [Adam Langley, Bodo Moeller]
876 *) Fix for the attack described in the paper "Recovering OpenSSL
877 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
878 by Yuval Yarom and Naomi Benger. Details can be obtained from:
879 http://eprint.iacr.org/2014/140
881 Thanks to Yuval Yarom and Naomi Benger for discovering this
882 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
883 [Yuval Yarom and Naomi Benger]
885 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
886 this fixes a limitation in previous versions of OpenSSL.
889 *) Experimental encrypt-then-mac support.
891 Experimental support for encrypt then mac from
892 draft-gutmann-tls-encrypt-then-mac-02.txt
894 To enable it set the appropriate extension number (0x42 for the test
895 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
897 For non-compliant peers (i.e. just about everything) this should have no
900 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
904 *) Add EVP support for key wrapping algorithms, to avoid problems with
905 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
906 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
907 algorithms and include tests cases.
910 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
914 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
915 MGF1 digest and OAEP label.
918 *) Make openssl verify return errors.
919 [Chris Palmer <palmer@google.com> and Ben Laurie]
921 *) New function ASN1_TIME_diff to calculate the difference between two
922 ASN1_TIME structures or one structure and the current time.
925 *) Update fips_test_suite to support multiple command line options. New
926 test to induce all self test errors in sequence and check expected
930 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
931 sign or verify all in one operation.
934 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
935 test programs and fips_test_suite. Includes functionality to parse
936 the minimal script output of fipsalgest.pl directly.
939 *) Add authorisation parameter to FIPS_module_mode_set().
942 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
945 *) Use separate DRBG fields for internal and external flags. New function
946 FIPS_drbg_health_check() to perform on demand health checking. Add
947 generation tests to fips_test_suite with reduced health check interval to
948 demonstrate periodic health checking. Add "nodh" option to
949 fips_test_suite to skip very slow DH test.
952 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
956 *) More extensive health check for DRBG checking many more failure modes.
957 New function FIPS_selftest_drbg_all() to handle every possible DRBG
958 combination: call this in fips_test_suite.
961 *) Add support for canonical generation of DSA parameter 'g'. See
964 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
965 POST to handle HMAC cases.
968 *) Add functions FIPS_module_version() and FIPS_module_version_text()
969 to return numerical and string versions of the FIPS module number.
972 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
973 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
974 outside the validated module in the FIPS capable OpenSSL.
977 *) Minor change to DRBG entropy callback semantics. In some cases
978 there is no multiple of the block length between min_len and
979 max_len. Allow the callback to return more than max_len bytes
980 of entropy but discard any extra: it is the callback's responsibility
981 to ensure that the extra data discarded does not impact the
982 requested amount of entropy.
985 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
986 information in FIPS186-3, SP800-57 and SP800-131A.
989 *) CCM support via EVP. Interface is very similar to GCM case except we
990 must supply all data in one chunk (i.e. no update, final) and the
991 message length must be supplied if AAD is used. Add algorithm test
995 *) Initial version of POST overhaul. Add POST callback to allow the status
996 of POST to be monitored and/or failures induced. Modify fips_test_suite
997 to use callback. Always run all selftests even if one fails.
1000 *) XTS support including algorithm test driver in the fips_gcmtest program.
1001 Note: this does increase the maximum key length from 32 to 64 bytes but
1002 there should be no binary compatibility issues as existing applications
1003 will never use XTS mode.
1006 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1007 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1008 performs algorithm blocking for unapproved PRNG types. Also do not
1009 set PRNG type in FIPS_mode_set(): leave this to the application.
1010 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1011 the standard OpenSSL PRNG: set additional data to a date time vector.
1014 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1015 This shouldn't present any incompatibility problems because applications
1016 shouldn't be using these directly and any that are will need to rethink
1017 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1020 *) Extensive self tests and health checking required by SP800-90 DRBG.
1021 Remove strength parameter from FIPS_drbg_instantiate and always
1022 instantiate at maximum supported strength.
1025 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1028 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1031 *) New function DH_compute_key_padded() to compute a DH key and pad with
1032 leading zeroes if needed: this complies with SP800-56A et al.
1035 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1036 anything, incomplete, subject to change and largely untested at present.
1039 *) Modify fipscanisteronly build option to only build the necessary object
1040 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1043 *) Add experimental option FIPSSYMS to give all symbols in
1044 fipscanister.o and FIPS or fips prefix. This will avoid
1045 conflicts with future versions of OpenSSL. Add perl script
1046 util/fipsas.pl to preprocess assembly language source files
1047 and rename any affected symbols.
1050 *) Add selftest checks and algorithm block of non-fips algorithms in
1051 FIPS mode. Remove DES2 from selftests.
1054 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1055 return internal method without any ENGINE dependencies. Add new
1056 tiny fips sign and verify functions.
1059 *) New build option no-ec2m to disable characteristic 2 code.
1062 *) New build option "fipscanisteronly". This only builds fipscanister.o
1063 and (currently) associated fips utilities. Uses the file Makefile.fips
1064 instead of Makefile.org as the prototype.
1067 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1068 Update fips_gcmtest to use IV generator.
1071 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1072 setting output buffer to NULL. The *Final function must be
1073 called although it will not retrieve any additional data. The tag
1074 can be set or retrieved with a ctrl. The IV length is by default 12
1075 bytes (96 bits) but can be set to an alternative value. If the IV
1076 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1080 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1081 underlying do_cipher function handles all cipher semantics itself
1082 including padding and finalisation. This is useful if (for example)
1083 an ENGINE cipher handles block padding itself. The behaviour of
1084 do_cipher is subtly changed if this flag is set: the return value
1085 is the number of characters written to the output buffer (zero is
1086 no longer an error code) or a negative error code. Also if the
1087 input buffer is NULL and length 0 finalisation should be performed.
1090 *) If a candidate issuer certificate is already part of the constructed
1091 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1094 *) Improve forward-security support: add functions
1096 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1097 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1099 for use by SSL/TLS servers; the callback function will be called whenever a
1100 new session is created, and gets to decide whether the session may be
1101 cached to make it resumable (return 0) or not (return 1). (As by the
1102 SSL/TLS protocol specifications, the session_id sent by the server will be
1103 empty to indicate that the session is not resumable; also, the server will
1104 not generate RFC 4507 (RFC 5077) session tickets.)
1106 A simple reasonable callback implementation is to return is_forward_secure.
1107 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1108 by the SSL/TLS server library, indicating whether it can provide forward
1110 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1112 *) New -verify_name option in command line utilities to set verification
1116 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1117 Add CMAC pkey methods.
1120 *) Experimental renegotiation in s_server -www mode. If the client
1121 browses /reneg connection is renegotiated. If /renegcert it is
1122 renegotiated requesting a certificate.
1125 *) Add an "external" session cache for debugging purposes to s_server. This
1126 should help trace issues which normally are only apparent in deployed
1127 multi-process servers.
1130 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1131 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1132 BIO_set_cipher() and some obscure PEM functions were changed so they
1133 can now return an error. The RAND changes required a change to the
1134 RAND_METHOD structure.
1137 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1138 a gcc attribute to warn if the result of a function is ignored. This
1139 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1140 whose return value is often ignored.
1143 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1144 These allow SCTs (signed certificate timestamps) to be requested and
1145 validated when establishing a connection.
1146 [Rob Percival <robpercival@google.com>]
1148 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1150 *) Prevent padding oracle in AES-NI CBC MAC check
1152 A MITM attacker can use a padding oracle attack to decrypt traffic
1153 when the connection uses an AES CBC cipher and the server support
1156 This issue was introduced as part of the fix for Lucky 13 padding
1157 attack (CVE-2013-0169). The padding check was rewritten to be in
1158 constant time by making sure that always the same bytes are read and
1159 compared against either the MAC or padding bytes. But it no longer
1160 checked that there was enough data to have both the MAC and padding
1163 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1167 *) Fix EVP_EncodeUpdate overflow
1169 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1170 Base64 encoding of binary data. If an attacker is able to supply very large
1171 amounts of input data then a length check can overflow resulting in a heap
1174 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1175 the PEM_write_bio* family of functions. These are mainly used within the
1176 OpenSSL command line applications, so any application which processes data
1177 from an untrusted source and outputs it as a PEM file should be considered
1178 vulnerable to this issue. User applications that call these APIs directly
1179 with large amounts of untrusted data may also be vulnerable.
1181 This issue was reported by Guido Vranken.
1185 *) Fix EVP_EncryptUpdate overflow
1187 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1188 is able to supply very large amounts of input data after a previous call to
1189 EVP_EncryptUpdate() with a partial block then a length check can overflow
1190 resulting in a heap corruption. Following an analysis of all OpenSSL
1191 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1192 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1193 the first called function after an EVP_EncryptInit(), and therefore that
1194 specific call must be safe. The second form is where the length passed to
1195 EVP_EncryptUpdate() can be seen from the code to be some small value and
1196 therefore there is no possibility of an overflow. Since all instances are
1197 one of these two forms, it is believed that there can be no overflows in
1198 internal code due to this problem. It should be noted that
1199 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1200 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1201 of these calls have also been analysed too and it is believed there are no
1202 instances in internal usage where an overflow could occur.
1204 This issue was reported by Guido Vranken.
1208 *) Prevent ASN.1 BIO excessive memory allocation
1210 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1211 a short invalid encoding can cause allocation of large amounts of memory
1212 potentially consuming excessive resources or exhausting memory.
1214 Any application parsing untrusted data through d2i BIO functions is
1215 affected. The memory based functions such as d2i_X509() are *not* affected.
1216 Since the memory based functions are used by the TLS library, TLS
1217 applications are not affected.
1219 This issue was reported by Brian Carpenter.
1225 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1226 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1227 in arbitrary stack data being returned in the buffer.
1229 This issue was reported by Guido Vranken.
1233 *) Modify behavior of ALPN to invoke callback after SNI/servername
1234 callback, such that updates to the SSL_CTX affect ALPN.
1237 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1241 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1242 methods are enabled and ssl2 is disabled the methods return NULL.
1245 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1247 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1248 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1249 provide any "EXPORT" or "LOW" strength ciphers.
1252 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1253 is by default disabled at build-time. Builds that are not configured with
1254 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1255 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1256 will need to explicitly call either of:
1258 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1260 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1262 as appropriate. Even if either of those is used, or the application
1263 explicitly uses the version-specific SSLv2_method() or its client and
1264 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1265 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1266 ciphers, and SSLv2 56-bit DES are no longer available.
1270 *) Fix a double-free in DSA code
1272 A double free bug was discovered when OpenSSL parses malformed DSA private
1273 keys and could lead to a DoS attack or memory corruption for applications
1274 that receive DSA private keys from untrusted sources. This scenario is
1277 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1282 *) Disable SRP fake user seed to address a server memory leak.
1284 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1286 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1287 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1288 was changed to ignore the "fake user" SRP seed, even if the seed
1291 Users should use SRP_VBASE_get1_by_user instead. Note that in
1292 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1293 also that even though configuring the SRP seed attempts to hide
1294 invalid usernames by continuing the handshake with fake
1295 credentials, this behaviour is not constant time and no strong
1296 guarantees are made that the handshake is indistinguishable from
1297 that of a valid user.
1301 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1303 In the BN_hex2bn function the number of hex digits is calculated using an
1304 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1305 large values of |i| this can result in |bn_expand| not allocating any
1306 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1307 field as NULL leading to a subsequent NULL ptr deref. For very large values
1308 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1309 In this case memory is allocated to the internal BIGNUM data field, but it
1310 is insufficiently sized leading to heap corruption. A similar issue exists
1311 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1312 is ever called by user applications with very large untrusted hex/dec data.
1313 This is anticipated to be a rare occurrence.
1315 All OpenSSL internal usage of these functions use data that is not expected
1316 to be untrusted, e.g. config file data or application command line
1317 arguments. If user developed applications generate config file data based
1318 on untrusted data then it is possible that this could also lead to security
1319 consequences. This is also anticipated to be rare.
1321 This issue was reported to OpenSSL by Guido Vranken.
1325 *) Fix memory issues in BIO_*printf functions
1327 The internal |fmtstr| function used in processing a "%s" format string in
1328 the BIO_*printf functions could overflow while calculating the length of a
1329 string and cause an OOB read when printing very long strings.
1331 Additionally the internal |doapr_outch| function can attempt to write to an
1332 OOB memory location (at an offset from the NULL pointer) in the event of a
1333 memory allocation failure. In 1.0.2 and below this could be caused where
1334 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1335 could be in processing a very long "%s" format string. Memory leaks can
1338 The first issue may mask the second issue dependent on compiler behaviour.
1339 These problems could enable attacks where large amounts of untrusted data
1340 is passed to the BIO_*printf functions. If applications use these functions
1341 in this way then they could be vulnerable. OpenSSL itself uses these
1342 functions when printing out human-readable dumps of ASN.1 data. Therefore
1343 applications that print this data could be vulnerable if the data is from
1344 untrusted sources. OpenSSL command line applications could also be
1345 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1346 as command line arguments.
1348 Libssl is not considered directly vulnerable. Additionally certificates etc
1349 received via remote connections via libssl are also unlikely to be able to
1350 trigger these issues because of message size limits enforced within libssl.
1352 This issue was reported to OpenSSL Guido Vranken.
1356 *) Side channel attack on modular exponentiation
1358 A side-channel attack was found which makes use of cache-bank conflicts on
1359 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1360 of RSA keys. The ability to exploit this issue is limited as it relies on
1361 an attacker who has control of code in a thread running on the same
1362 hyper-threaded core as the victim thread which is performing decryptions.
1364 This issue was reported to OpenSSL by Yuval Yarom, The University of
1365 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1366 Nadia Heninger, University of Pennsylvania with more information at
1367 http://cachebleed.info.
1371 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1372 if no keysize is specified with default_bits. This fixes an
1373 omission in an earlier change that changed all RSA/DSA key generation
1374 apps to use 2048 bits by default.
1377 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1378 *) DH small subgroups
1380 Historically OpenSSL only ever generated DH parameters based on "safe"
1381 primes. More recently (in version 1.0.2) support was provided for
1382 generating X9.42 style parameter files such as those required for RFC 5114
1383 support. The primes used in such files may not be "safe". Where an
1384 application is using DH configured with parameters based on primes that are
1385 not "safe" then an attacker could use this fact to find a peer's private
1386 DH exponent. This attack requires that the attacker complete multiple
1387 handshakes in which the peer uses the same private DH exponent. For example
1388 this could be used to discover a TLS server's private DH exponent if it's
1389 reusing the private DH exponent or it's using a static DH ciphersuite.
1391 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1392 TLS. It is not on by default. If the option is not set then the server
1393 reuses the same private DH exponent for the life of the server process and
1394 would be vulnerable to this attack. It is believed that many popular
1395 applications do set this option and would therefore not be at risk.
1397 The fix for this issue adds an additional check where a "q" parameter is
1398 available (as is the case in X9.42 based parameters). This detects the
1399 only known attack, and is the only possible defense for static DH
1400 ciphersuites. This could have some performance impact.
1402 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1403 default and cannot be disabled. This could have some performance impact.
1405 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1409 *) SSLv2 doesn't block disabled ciphers
1411 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1412 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1413 been disabled, provided that the SSLv2 protocol was not also disabled via
1416 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1417 and Sebastian Schinzel.
1421 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1423 *) BN_mod_exp may produce incorrect results on x86_64
1425 There is a carry propagating bug in the x86_64 Montgomery squaring
1426 procedure. No EC algorithms are affected. Analysis suggests that attacks
1427 against RSA and DSA as a result of this defect would be very difficult to
1428 perform and are not believed likely. Attacks against DH are considered just
1429 feasible (although very difficult) because most of the work necessary to
1430 deduce information about a private key may be performed offline. The amount
1431 of resources required for such an attack would be very significant and
1432 likely only accessible to a limited number of attackers. An attacker would
1433 additionally need online access to an unpatched system using the target
1434 private key in a scenario with persistent DH parameters and a private
1435 key that is shared between multiple clients. For example this can occur by
1436 default in OpenSSL DHE based SSL/TLS ciphersuites.
1438 This issue was reported to OpenSSL by Hanno Böck.
1442 *) Certificate verify crash with missing PSS parameter
1444 The signature verification routines will crash with a NULL pointer
1445 dereference if presented with an ASN.1 signature using the RSA PSS
1446 algorithm and absent mask generation function parameter. Since these
1447 routines are used to verify certificate signature algorithms this can be
1448 used to crash any certificate verification operation and exploited in a
1449 DoS attack. Any application which performs certificate verification is
1450 vulnerable including OpenSSL clients and servers which enable client
1453 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1457 *) X509_ATTRIBUTE memory leak
1459 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1460 memory. This structure is used by the PKCS#7 and CMS routines so any
1461 application which reads PKCS#7 or CMS data from untrusted sources is
1462 affected. SSL/TLS is not affected.
1464 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1469 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1470 This changes the decoding behaviour for some invalid messages,
1471 though the change is mostly in the more lenient direction, and
1472 legacy behaviour is preserved as much as possible.
1475 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1477 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1479 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1481 *) Alternate chains certificate forgery
1483 During certificate verification, OpenSSL will attempt to find an
1484 alternative certificate chain if the first attempt to build such a chain
1485 fails. An error in the implementation of this logic can mean that an
1486 attacker could cause certain checks on untrusted certificates to be
1487 bypassed, such as the CA flag, enabling them to use a valid leaf
1488 certificate to act as a CA and "issue" an invalid certificate.
1490 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1494 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1496 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1497 incompatibility in the handling of HMAC. The previous ABI has now been
1501 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1503 *) Malformed ECParameters causes infinite loop
1505 When processing an ECParameters structure OpenSSL enters an infinite loop
1506 if the curve specified is over a specially malformed binary polynomial
1509 This can be used to perform denial of service against any
1510 system which processes public keys, certificate requests or
1511 certificates. This includes TLS clients and TLS servers with
1512 client authentication enabled.
1514 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1518 *) Exploitable out-of-bounds read in X509_cmp_time
1520 X509_cmp_time does not properly check the length of the ASN1_TIME
1521 string and can read a few bytes out of bounds. In addition,
1522 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1525 An attacker can use this to craft malformed certificates and CRLs of
1526 various sizes and potentially cause a segmentation fault, resulting in
1527 a DoS on applications that verify certificates or CRLs. TLS clients
1528 that verify CRLs are affected. TLS clients and servers with client
1529 authentication enabled may be affected if they use custom verification
1532 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1533 independently by Hanno Böck.
1537 *) PKCS7 crash with missing EnvelopedContent
1539 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1540 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1541 with missing content and trigger a NULL pointer dereference on parsing.
1543 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1544 structures from untrusted sources are affected. OpenSSL clients and
1545 servers are not affected.
1547 This issue was reported to OpenSSL by Michal Zalewski (Google).
1551 *) CMS verify infinite loop with unknown hash function
1553 When verifying a signedData message the CMS code can enter an infinite loop
1554 if presented with an unknown hash function OID. This can be used to perform
1555 denial of service against any system which verifies signedData messages using
1557 This issue was reported to OpenSSL by Johannes Bauer.
1561 *) Race condition handling NewSessionTicket
1563 If a NewSessionTicket is received by a multi-threaded client when attempting to
1564 reuse a previous ticket then a race condition can occur potentially leading to
1565 a double free of the ticket data.
1569 *) Only support 256-bit or stronger elliptic curves with the
1570 'ecdh_auto' setting (server) or by default (client). Of supported
1571 curves, prefer P-256 (both).
1574 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1576 *) ClientHello sigalgs DoS fix
1578 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1579 invalid signature algorithms extension a NULL pointer dereference will
1580 occur. This can be exploited in a DoS attack against the server.
1582 This issue was was reported to OpenSSL by David Ramos of Stanford
1585 [Stephen Henson and Matt Caswell]
1587 *) Multiblock corrupted pointer fix
1589 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1590 feature only applies on 64 bit x86 architecture platforms that support AES
1591 NI instructions. A defect in the implementation of "multiblock" can cause
1592 OpenSSL's internal write buffer to become incorrectly set to NULL when
1593 using non-blocking IO. Typically, when the user application is using a
1594 socket BIO for writing, this will only result in a failed connection.
1595 However if some other BIO is used then it is likely that a segmentation
1596 fault will be triggered, thus enabling a potential DoS attack.
1598 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1602 *) Segmentation fault in DTLSv1_listen fix
1604 The DTLSv1_listen function is intended to be stateless and processes the
1605 initial ClientHello from many peers. It is common for user code to loop
1606 over the call to DTLSv1_listen until a valid ClientHello is received with
1607 an associated cookie. A defect in the implementation of DTLSv1_listen means
1608 that state is preserved in the SSL object from one invocation to the next
1609 that can lead to a segmentation fault. Errors processing the initial
1610 ClientHello can trigger this scenario. An example of such an error could be
1611 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1614 This issue was reported to OpenSSL by Per Allansson.
1618 *) Segmentation fault in ASN1_TYPE_cmp fix
1620 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1621 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1622 certificate signature algorithm consistency this can be used to crash any
1623 certificate verification operation and exploited in a DoS attack. Any
1624 application which performs certificate verification is vulnerable including
1625 OpenSSL clients and servers which enable client authentication.
1629 *) Segmentation fault for invalid PSS parameters fix
1631 The signature verification routines will crash with a NULL pointer
1632 dereference if presented with an ASN.1 signature using the RSA PSS
1633 algorithm and invalid parameters. Since these routines are used to verify
1634 certificate signature algorithms this can be used to crash any
1635 certificate verification operation and exploited in a DoS attack. Any
1636 application which performs certificate verification is vulnerable including
1637 OpenSSL clients and servers which enable client authentication.
1639 This issue was was reported to OpenSSL by Brian Carpenter.
1643 *) ASN.1 structure reuse memory corruption fix
1645 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1646 memory corruption via an invalid write. Such reuse is and has been
1647 strongly discouraged and is believed to be rare.
1649 Applications that parse structures containing CHOICE or ANY DEFINED BY
1650 components may be affected. Certificate parsing (d2i_X509 and related
1651 functions) are however not affected. OpenSSL clients and servers are
1656 *) PKCS7 NULL pointer dereferences fix
1658 The PKCS#7 parsing code does not handle missing outer ContentInfo
1659 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1660 missing content and trigger a NULL pointer dereference on parsing.
1662 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1663 otherwise parse PKCS#7 structures from untrusted sources are
1664 affected. OpenSSL clients and servers are not affected.
1666 This issue was reported to OpenSSL by Michal Zalewski (Google).
1670 *) DoS via reachable assert in SSLv2 servers fix
1672 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1673 servers that both support SSLv2 and enable export cipher suites by sending
1674 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1676 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1677 (OpenSSL development team).
1681 *) Empty CKE with client auth and DHE fix
1683 If client auth is used then a server can seg fault in the event of a DHE
1684 ciphersuite being selected and a zero length ClientKeyExchange message
1685 being sent by the client. This could be exploited in a DoS attack.
1689 *) Handshake with unseeded PRNG fix
1691 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1692 with an unseeded PRNG. The conditions are:
1693 - The client is on a platform where the PRNG has not been seeded
1694 automatically, and the user has not seeded manually
1695 - A protocol specific client method version has been used (i.e. not
1696 SSL_client_methodv23)
1697 - A ciphersuite is used that does not require additional random data from
1698 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1700 If the handshake succeeds then the client random that has been used will
1701 have been generated from a PRNG with insufficient entropy and therefore the
1702 output may be predictable.
1704 For example using the following command with an unseeded openssl will
1705 succeed on an unpatched platform:
1707 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1711 *) Use After Free following d2i_ECPrivatekey error fix
1713 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1714 could cause a use after free condition. This, in turn, could cause a double
1715 free in several private key parsing functions (such as d2i_PrivateKey
1716 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1717 for applications that receive EC private keys from untrusted
1718 sources. This scenario is considered rare.
1720 This issue was discovered by the BoringSSL project and fixed in their
1725 *) X509_to_X509_REQ NULL pointer deref fix
1727 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1728 the certificate key is invalid. This function is rarely used in practice.
1730 This issue was discovered by Brian Carpenter.
1734 *) Removed the export ciphers from the DEFAULT ciphers
1737 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1739 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1740 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1741 So far those who have to target multiple platforms would compromise
1742 and argue that binary targeting say ARMv5 would still execute on
1743 ARMv8. "Universal" build resolves this compromise by providing
1744 near-optimal performance even on newer platforms.
1747 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1748 (other platforms pending).
1749 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1751 *) Add support for the SignedCertificateTimestampList certificate and
1752 OCSP response extensions from RFC6962.
1755 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1756 for corner cases. (Certain input points at infinity could lead to
1757 bogus results, with non-infinity inputs mapped to infinity too.)
1760 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1761 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1762 common cases are optimized and there still is room for further
1763 improvements. Vector Permutation AES for Altivec is also added.
1766 *) Add support for little-endian ppc64 Linux target.
1767 [Marcelo Cerri (IBM)]
1769 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1770 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1771 are optimized and there still is room for further improvements.
1772 Both 32- and 64-bit modes are supported.
1773 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1775 *) Improved ARMv7 NEON support.
1778 *) Support for SPARC Architecture 2011 crypto extensions, first
1779 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1780 SHA256/512, MD5, GHASH and modular exponentiation.
1781 [Andy Polyakov, David Miller]
1783 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1785 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1787 *) Support for new and upcoming Intel processors, including AVX2,
1788 BMI and SHA ISA extensions. This includes additional "stitched"
1789 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1792 This work was sponsored by Intel Corp.
1795 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1796 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1797 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1800 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1801 this fixes a limitation in previous versions of OpenSSL.
1804 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1805 MGF1 digest and OAEP label.
1808 *) Add EVP support for key wrapping algorithms, to avoid problems with
1809 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1810 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1811 algorithms and include tests cases.
1814 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1816 [Douglas E. Engert, Steve Henson]
1818 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1819 difference in days and seconds between two tm or ASN1_TIME structures.
1822 *) Add -rev test option to s_server to just reverse order of characters
1823 received by client and send back to server. Also prints an abbreviated
1824 summary of the connection parameters.
1827 *) New option -brief for s_client and s_server to print out a brief summary
1828 of connection parameters.
1831 *) Add callbacks for arbitrary TLS extensions.
1832 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1834 *) New option -crl_download in several openssl utilities to download CRLs
1835 from CRLDP extension in certificates.
1838 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1841 *) New function X509_CRL_diff to generate a delta CRL from the difference
1842 of two full CRLs. Add support to "crl" utility.
1845 *) New functions to set lookup_crls function and to retrieve
1846 X509_STORE from X509_STORE_CTX.
1849 *) Print out deprecated issuer and subject unique ID fields in
1853 *) Extend OCSP I/O functions so they can be used for simple general purpose
1854 HTTP as well as OCSP. New wrapper function which can be used to download
1855 CRLs using the OCSP API.
1858 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1861 *) SSL_CONF* functions. These provide a common framework for application
1862 configuration using configuration files or command lines.
1865 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1866 message callback and prints the results. Needs compile time option
1867 "enable-ssl-trace". New options to s_client and s_server to enable
1871 *) New ctrl and macro to retrieve supported points extensions.
1872 Print out extension in s_server and s_client.
1875 *) New functions to retrieve certificate signature and signature
1879 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1883 *) New Suite B modes for TLS code. These use and enforce the requirements
1884 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1885 only use Suite B curves. The Suite B modes can be set by using the
1886 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1889 *) New chain verification flags for Suite B levels of security. Check
1890 algorithms are acceptable when flags are set in X509_verify_cert.
1893 *) Make tls1_check_chain return a set of flags indicating checks passed
1894 by a certificate chain. Add additional tests to handle client
1895 certificates: checks for matching certificate type and issuer name
1899 *) If an attempt is made to use a signature algorithm not in the peer
1900 preference list abort the handshake. If client has no suitable
1901 signature algorithms in response to a certificate request do not
1902 use the certificate.
1905 *) If server EC tmp key is not in client preference list abort handshake.
1908 *) Add support for certificate stores in CERT structure. This makes it
1909 possible to have different stores per SSL structure or one store in
1910 the parent SSL_CTX. Include distinct stores for certificate chain
1911 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1912 to build and store a certificate chain in CERT structure: returning
1913 an error if the chain cannot be built: this will allow applications
1914 to test if a chain is correctly configured.
1916 Note: if the CERT based stores are not set then the parent SSL_CTX
1917 store is used to retain compatibility with existing behaviour.
1921 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1922 mask based on the current session, check mask when sending client
1923 hello and checking the requested ciphersuite.
1926 *) New ctrls to retrieve and set certificate types in a certificate
1927 request message. Print out received values in s_client. If certificate
1928 types is not set with custom values set sensible values based on
1929 supported signature algorithms.
1932 *) Support for distinct client and server supported signature algorithms.
1935 *) Add certificate callback. If set this is called whenever a certificate
1936 is required by client or server. An application can decide which
1937 certificate chain to present based on arbitrary criteria: for example
1938 supported signature algorithms. Add very simple example to s_server.
1939 This fixes many of the problems and restrictions of the existing client
1940 certificate callback: for example you can now clear an existing
1941 certificate and specify the whole chain.
1944 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1945 the certificate can be used for (if anything). Set valid_flags field
1946 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1947 to have similar checks in it.
1949 Add new "cert_flags" field to CERT structure and include a "strict mode".
1950 This enforces some TLS certificate requirements (such as only permitting
1951 certificate signature algorithms contained in the supported algorithms
1952 extension) which some implementations ignore: this option should be used
1953 with caution as it could cause interoperability issues.
1956 *) Update and tidy signature algorithm extension processing. Work out
1957 shared signature algorithms based on preferences and peer algorithms
1958 and print them out in s_client and s_server. Abort handshake if no
1959 shared signature algorithms.
1962 *) Add new functions to allow customised supported signature algorithms
1963 for SSL and SSL_CTX structures. Add options to s_client and s_server
1967 *) New function SSL_certs_clear() to delete all references to certificates
1968 from an SSL structure. Before this once a certificate had been added
1969 it couldn't be removed.
1972 *) Integrate hostname, email address and IP address checking with certificate
1973 verification. New verify options supporting checking in openssl utility.
1976 *) Fixes and wildcard matching support to hostname and email checking
1977 functions. Add manual page.
1978 [Florian Weimer (Red Hat Product Security Team)]
1980 *) New functions to check a hostname email or IP address against a
1981 certificate. Add options x509 utility to print results of checks against
1985 *) Fix OCSP checking.
1986 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1988 *) Initial experimental support for explicitly trusted non-root CAs.
1989 OpenSSL still tries to build a complete chain to a root but if an
1990 intermediate CA has a trust setting included that is used. The first
1991 setting is used: whether to trust (e.g., -addtrust option to the x509
1995 *) Add -trusted_first option which attempts to find certificates in the
1996 trusted store even if an untrusted chain is also supplied.
1999 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2000 platform support for Linux and Android.
2003 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2006 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2007 When in FIPS mode the approved implementations are used as normal,
2008 when not in FIPS mode the internal unapproved versions are used instead.
2009 This means that the FIPS capable OpenSSL isn't forced to use the
2010 (often lower performance) FIPS implementations outside FIPS mode.
2013 *) Transparently support X9.42 DH parameters when calling
2014 PEM_read_bio_DHparameters. This means existing applications can handle
2015 the new parameter format automatically.
2018 *) Initial experimental support for X9.42 DH parameter format: mainly
2019 to support use of 'q' parameter for RFC5114 parameters.
2022 *) Add DH parameters from RFC5114 including test data to dhtest.
2025 *) Support for automatic EC temporary key parameter selection. If enabled
2026 the most preferred EC parameters are automatically used instead of
2027 hardcoded fixed parameters. Now a server just has to call:
2028 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2029 support ECDH and use the most appropriate parameters.
2032 *) Enhance and tidy EC curve and point format TLS extension code. Use
2033 static structures instead of allocation if default values are used.
2034 New ctrls to set curves we wish to support and to retrieve shared curves.
2035 Print out shared curves in s_server. New options to s_server and s_client
2036 to set list of supported curves.
2039 *) New ctrls to retrieve supported signature algorithms and
2040 supported curve values as an array of NIDs. Extend openssl utility
2041 to print out received values.
2044 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2045 between NIDs and the more common NIST names such as "P-256". Enhance
2046 ecparam utility and ECC method to recognise the NIST names for curves.
2049 *) Enhance SSL/TLS certificate chain handling to support different
2050 chains for each certificate instead of one chain in the parent SSL_CTX.
2053 *) Support for fixed DH ciphersuite client authentication: where both
2054 server and client use DH certificates with common parameters.
2057 *) Support for fixed DH ciphersuites: those requiring DH server
2061 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2063 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2064 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2065 X509_CINF_get_signature were reverted post internal team review.
2067 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2069 *) Build fixes for the Windows and OpenVMS platforms
2070 [Matt Caswell and Richard Levitte]
2072 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2074 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2075 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2076 dereference. This could lead to a Denial Of Service attack. Thanks to
2077 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2081 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2082 dtls1_buffer_record function under certain conditions. In particular this
2083 could occur if an attacker sent repeated DTLS records with the same
2084 sequence number but for the next epoch. The memory leak could be exploited
2085 by an attacker in a Denial of Service attack through memory exhaustion.
2086 Thanks to Chris Mueller for reporting this issue.
2090 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2091 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2092 method would be set to NULL which could later result in a NULL pointer
2093 dereference. Thanks to Frank Schmirler for reporting this issue.
2097 *) Abort handshake if server key exchange message is omitted for ephemeral
2100 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2101 reporting this issue.
2105 *) Remove non-export ephemeral RSA code on client and server. This code
2106 violated the TLS standard by allowing the use of temporary RSA keys in
2107 non-export ciphersuites and could be used by a server to effectively
2108 downgrade the RSA key length used to a value smaller than the server
2109 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2110 INRIA or reporting this issue.
2114 *) Fixed issue where DH client certificates are accepted without verification.
2115 An OpenSSL server will accept a DH certificate for client authentication
2116 without the certificate verify message. This effectively allows a client to
2117 authenticate without the use of a private key. This only affects servers
2118 which trust a client certificate authority which issues certificates
2119 containing DH keys: these are extremely rare and hardly ever encountered.
2120 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2125 *) Ensure that the session ID context of an SSL is updated when its
2126 SSL_CTX is updated via SSL_set_SSL_CTX.
2128 The session ID context is typically set from the parent SSL_CTX,
2129 and can vary with the CTX.
2132 *) Fix various certificate fingerprint issues.
2134 By using non-DER or invalid encodings outside the signed portion of a
2135 certificate the fingerprint can be changed without breaking the signature.
2136 Although no details of the signed portion of the certificate can be changed
2137 this can cause problems with some applications: e.g. those using the
2138 certificate fingerprint for blacklists.
2140 1. Reject signatures with non zero unused bits.
2142 If the BIT STRING containing the signature has non zero unused bits reject
2143 the signature. All current signature algorithms require zero unused bits.
2145 2. Check certificate algorithm consistency.
2147 Check the AlgorithmIdentifier inside TBS matches the one in the
2148 certificate signature. NB: this will result in signature failure
2149 errors for some broken certificates.
2151 Thanks to Konrad Kraszewski from Google for reporting this issue.
2153 3. Check DSA/ECDSA signatures use DER.
2155 Re-encode DSA/ECDSA signatures and compare with the original received
2156 signature. Return an error if there is a mismatch.
2158 This will reject various cases including garbage after signature
2159 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2160 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2161 (negative or with leading zeroes).
2163 Further analysis was conducted and fixes were developed by Stephen Henson
2164 of the OpenSSL core team.
2169 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2170 results on some platforms, including x86_64. This bug occurs at random
2171 with a very low probability, and is not known to be exploitable in any
2172 way, though its exact impact is difficult to determine. Thanks to Pieter
2173 Wuille (Blockstream) who reported this issue and also suggested an initial
2174 fix. Further analysis was conducted by the OpenSSL development team and
2175 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2176 the OpenSSL core team.
2180 *) Do not resume sessions on the server if the negotiated protocol
2181 version does not match the session's version. Resuming with a different
2182 version, while not strictly forbidden by the RFC, is of questionable
2183 sanity and breaks all known clients.
2184 [David Benjamin, Emilia Käsper]
2186 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2187 early CCS messages during renegotiation. (Note that because
2188 renegotiation is encrypted, this early CCS was not exploitable.)
2191 *) Tighten client-side session ticket handling during renegotiation:
2192 ensure that the client only accepts a session ticket if the server sends
2193 the extension anew in the ServerHello. Previously, a TLS client would
2194 reuse the old extension state and thus accept a session ticket if one was
2195 announced in the initial ServerHello.
2197 Similarly, ensure that the client requires a session ticket if one
2198 was advertised in the ServerHello. Previously, a TLS client would
2199 ignore a missing NewSessionTicket message.
2202 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2204 *) SRTP Memory Leak.
2206 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2207 sends a carefully crafted handshake message, to cause OpenSSL to fail
2208 to free up to 64k of memory causing a memory leak. This could be
2209 exploited in a Denial Of Service attack. This issue affects OpenSSL
2210 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2211 whether SRTP is used or configured. Implementations of OpenSSL that
2212 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2214 The fix was developed by the OpenSSL team.
2218 *) Session Ticket Memory Leak.
2220 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2221 integrity of that ticket is first verified. In the event of a session
2222 ticket integrity check failing, OpenSSL will fail to free memory
2223 causing a memory leak. By sending a large number of invalid session
2224 tickets an attacker could exploit this issue in a Denial Of Service
2229 *) Build option no-ssl3 is incomplete.
2231 When OpenSSL is configured with "no-ssl3" as a build option, servers
2232 could accept and complete a SSL 3.0 handshake, and clients could be
2233 configured to send them.
2235 [Akamai and the OpenSSL team]
2237 *) Add support for TLS_FALLBACK_SCSV.
2238 Client applications doing fallback retries should call
2239 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2241 [Adam Langley, Bodo Moeller]
2243 *) Add additional DigestInfo checks.
2245 Re-encode DigestInto in DER and check against the original when
2246 verifying RSA signature: this will reject any improperly encoded
2247 DigestInfo structures.
2249 Note: this is a precautionary measure and no attacks are currently known.
2253 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2255 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2256 SRP code can be overrun an internal buffer. Add sanity check that
2257 g, A, B < N to SRP code.
2259 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2260 Group for discovering this issue.
2264 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2265 TLS 1.0 instead of higher protocol versions when the ClientHello message
2266 is badly fragmented. This allows a man-in-the-middle attacker to force a
2267 downgrade to TLS 1.0 even if both the server and the client support a
2268 higher protocol version, by modifying the client's TLS records.
2270 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2271 researching this issue.
2275 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2276 to a denial of service attack. A malicious server can crash the client
2277 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2278 ciphersuite and sending carefully crafted handshake messages.
2280 Thanks to Felix Gröbert (Google) for discovering and researching this
2285 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2286 to leak memory. This can be exploited through a Denial of Service attack.
2287 Thanks to Adam Langley for discovering and researching this issue.
2291 *) An attacker can force openssl to consume large amounts of memory whilst
2292 processing DTLS handshake messages. This can be exploited through a
2293 Denial of Service attack.
2294 Thanks to Adam Langley for discovering and researching this issue.
2298 *) An attacker can force an error condition which causes openssl to crash
2299 whilst processing DTLS packets due to memory being freed twice. This
2300 can be exploited through a Denial of Service attack.
2301 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2306 *) If a multithreaded client connects to a malicious server using a resumed
2307 session and the server sends an ec point format extension it could write
2308 up to 255 bytes to freed memory.
2310 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2315 *) A malicious server can crash an OpenSSL client with a null pointer
2316 dereference (read) by specifying an SRP ciphersuite even though it was not
2317 properly negotiated with the client. This can be exploited through a
2318 Denial of Service attack.
2320 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2321 discovering and researching this issue.
2325 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2326 X509_name_oneline, X509_name_print_ex et al. to leak some information
2327 from the stack. Applications may be affected if they echo pretty printing
2328 output to the attacker.
2330 Thanks to Ivan Fratric (Google) for discovering this issue.
2332 [Emilia Käsper, and Steve Henson]
2334 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2335 for corner cases. (Certain input points at infinity could lead to
2336 bogus results, with non-infinity inputs mapped to infinity too.)
2339 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2341 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2342 handshake can force the use of weak keying material in OpenSSL
2343 SSL/TLS clients and servers.
2345 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2346 researching this issue. (CVE-2014-0224)
2347 [KIKUCHI Masashi, Steve Henson]
2349 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2350 OpenSSL DTLS client the code can be made to recurse eventually crashing
2353 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2355 [Imre Rad, Steve Henson]
2357 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2358 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2359 client or server. This is potentially exploitable to run arbitrary
2360 code on a vulnerable client or server.
2362 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2363 [Jüri Aedla, Steve Henson]
2365 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2366 are subject to a denial of service attack.
2368 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2369 this issue. (CVE-2014-3470)
2370 [Felix Gröbert, Ivan Fratric, Steve Henson]
2372 *) Harmonize version and its documentation. -f flag is used to display
2374 [mancha <mancha1@zoho.com>]
2376 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2377 in i2d_ECPrivateKey.
2378 [mancha <mancha1@zoho.com>]
2380 *) Fix some double frees. These are not thought to be exploitable.
2381 [mancha <mancha1@zoho.com>]
2383 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2385 *) A missing bounds check in the handling of the TLS heartbeat extension
2386 can be used to reveal up to 64k of memory to a connected client or
2389 Thanks for Neel Mehta of Google Security for discovering this bug and to
2390 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2391 preparing the fix (CVE-2014-0160)
2392 [Adam Langley, Bodo Moeller]
2394 *) Fix for the attack described in the paper "Recovering OpenSSL
2395 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2396 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2397 http://eprint.iacr.org/2014/140
2399 Thanks to Yuval Yarom and Naomi Benger for discovering this
2400 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2401 [Yuval Yarom and Naomi Benger]
2403 *) TLS pad extension: draft-agl-tls-padding-03
2405 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2406 TLS client Hello record length value would otherwise be > 255 and
2407 less that 512 pad with a dummy extension containing zeroes so it
2408 is at least 512 bytes long.
2410 [Adam Langley, Steve Henson]
2412 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2414 *) Fix for TLS record tampering bug. A carefully crafted invalid
2415 handshake could crash OpenSSL with a NULL pointer exception.
2416 Thanks to Anton Johansson for reporting this issues.
2419 *) Keep original DTLS digest and encryption contexts in retransmission
2420 structures so we can use the previous session parameters if they need
2421 to be resent. (CVE-2013-6450)
2424 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2425 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2426 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2427 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2428 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2429 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2430 [Rob Stradling, Adam Langley]
2432 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2434 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2435 supporting platforms or when small records were transferred.
2436 [Andy Polyakov, Steve Henson]
2438 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2440 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2442 This addresses the flaw in CBC record processing discovered by
2443 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2444 at: http://www.isg.rhul.ac.uk/tls/
2446 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2447 Security Group at Royal Holloway, University of London
2448 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2449 Emilia Käsper for the initial patch.
2451 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2453 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2454 ciphersuites which can be exploited in a denial of service attack.
2455 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2456 and detecting this bug and to Wolfgang Ettlinger
2457 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2461 *) Return an error when checking OCSP signatures when key is NULL.
2462 This fixes a DoS attack. (CVE-2013-0166)
2465 *) Make openssl verify return errors.
2466 [Chris Palmer <palmer@google.com> and Ben Laurie]
2468 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2469 the right response is stapled. Also change SSL_get_certificate()
2470 so it returns the certificate actually sent.
2471 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2472 [Rob Stradling <rob.stradling@comodo.com>]
2474 *) Fix possible deadlock when decoding public keys.
2477 *) Don't use TLS 1.0 record version number in initial client hello
2481 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2483 *) Sanity check record length before skipping explicit IV in TLS
2484 1.2, 1.1 and DTLS to fix DoS attack.
2486 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2487 fuzzing as a service testing platform.
2491 *) Initialise tkeylen properly when encrypting CMS messages.
2492 Thanks to Solar Designer of Openwall for reporting this issue.
2495 *) In FIPS mode don't try to use composite ciphers as they are not
2499 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2501 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2502 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2503 mean any application compiled against OpenSSL 1.0.0 headers setting
2504 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2505 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2506 0x10000000L Any application which was previously compiled against
2507 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2508 will need to be recompiled as a result. Letting be results in
2509 inability to disable specifically TLS 1.1 and in client context,
2510 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2513 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2514 disable just protocol X, but all protocols above X *if* there are
2515 protocols *below* X still enabled. In more practical terms it means
2516 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2517 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2518 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2522 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2524 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2525 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2526 in CRYPTO_realloc_clean.
2528 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2529 issue and to Adam Langley <agl@chromium.org> for fixing it.
2531 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2533 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2536 *) Workarounds for some broken servers that "hang" if a client hello
2537 record length exceeds 255 bytes.
2539 1. Do not use record version number > TLS 1.0 in initial client
2540 hello: some (but not all) hanging servers will now work.
2541 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2542 the number of ciphers sent in the client hello. This should be
2543 set to an even number, such as 50, for example by passing:
2544 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2545 Most broken servers should now work.
2546 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2547 TLS 1.2 client support entirely.
2550 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2553 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2555 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2556 STRING form instead of a DigestInfo.
2559 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2560 and the RSA_sign/RSA_verify functions. This was made more apparent when
2561 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2562 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2563 the correct format in RSA_verify so both forms transparently work.
2566 *) Some servers which support TLS 1.0 can choke if we initially indicate
2567 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2568 encrypted premaster secret. As a workaround use the maximum permitted
2569 client version in client hello, this should keep such servers happy
2570 and still work with previous versions of OpenSSL.
2573 *) Add support for TLS/DTLS heartbeats.
2574 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2576 *) Add support for SCTP.
2577 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2579 *) Improved PRNG seeding for VOS.
2580 [Paul Green <Paul.Green@stratus.com>]
2582 *) Extensive assembler packs updates, most notably:
2584 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2585 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2586 - x86_64: bit-sliced AES implementation;
2587 - ARM: NEON support, contemporary platforms optimizations;
2588 - s390x: z196 support;
2589 - *: GHASH and GF(2^m) multiplication implementations;
2593 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2594 (removal of unnecessary code)
2595 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2597 *) Add TLS key material exporter from RFC 5705.
2600 *) Add DTLS-SRTP negotiation from RFC 5764.
2603 *) Add Next Protocol Negotiation,
2604 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2605 disabled with a no-npn flag to config or Configure. Code donated
2607 [Adam Langley <agl@google.com> and Ben Laurie]
2609 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2610 NIST-P256, NIST-P521, with constant-time single point multiplication on
2611 typical inputs. Compiler support for the nonstandard type __uint128_t is
2612 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2613 Code made available under Apache License version 2.0.
2615 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2616 line to include this in your build of OpenSSL, and run "make depend" (or
2617 "make update"). This enables the following EC_METHODs:
2619 EC_GFp_nistp224_method()
2620 EC_GFp_nistp256_method()
2621 EC_GFp_nistp521_method()
2623 EC_GROUP_new_by_curve_name() will automatically use these (while
2624 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2626 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2628 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2629 all platforms. Move ssize_t definition from e_os.h to the public
2630 header file e_os2.h as it now appears in public header file cms.h
2633 *) New -sigopt option to the ca, req and x509 utilities. Additional
2634 signature parameters can be passed using this option and in
2638 *) Add RSA PSS signing function. This will generate and set the
2639 appropriate AlgorithmIdentifiers for PSS based on those in the
2640 corresponding EVP_MD_CTX structure. No application support yet.
2643 *) Support for companion algorithm specific ASN1 signing routines.
2644 New function ASN1_item_sign_ctx() signs a pre-initialised
2645 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2646 the appropriate parameters.
2649 *) Add new algorithm specific ASN1 verification initialisation function
2650 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2651 handling will be the same no matter what EVP_PKEY_METHOD is used.
2652 Add a PSS handler to support verification of PSS signatures: checked
2653 against a number of sample certificates.
2656 *) Add signature printing for PSS. Add PSS OIDs.
2657 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2659 *) Add algorithm specific signature printing. An individual ASN1 method
2660 can now print out signatures instead of the standard hex dump.
2662 More complex signatures (e.g. PSS) can print out more meaningful
2663 information. Include DSA version that prints out the signature
2667 *) Password based recipient info support for CMS library: implementing
2671 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2672 neatly separates the code into cipher and PBE sections and is required
2673 for some algorithms that split PBES2 into separate pieces (such as
2674 password based CMS).
2677 *) Session-handling fixes:
2678 - Fix handling of connections that are resuming with a session ID,
2679 but also support Session Tickets.
2680 - Fix a bug that suppressed issuing of a new ticket if the client
2681 presented a ticket with an expired session.
2682 - Try to set the ticket lifetime hint to something reasonable.
2683 - Make tickets shorter by excluding irrelevant information.
2684 - On the client side, don't ignore renewed tickets.
2685 [Adam Langley, Bodo Moeller (Google)]
2687 *) Fix PSK session representation.
2690 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2692 This work was sponsored by Intel.
2695 *) Add GCM support to TLS library. Some custom code is needed to split
2696 the IV between the fixed (from PRF) and explicit (from TLS record)
2697 portions. This adds all GCM ciphersuites supported by RFC5288 and
2698 RFC5289. Generalise some AES* cipherstrings to include GCM and
2699 add a special AESGCM string for GCM only.
2702 *) Expand range of ctrls for AES GCM. Permit setting invocation
2703 field on decrypt and retrieval of invocation field only on encrypt.
2706 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2707 As required by RFC5289 these ciphersuites cannot be used if for
2708 versions of TLS earlier than 1.2.
2711 *) For FIPS capable OpenSSL interpret a NULL default public key method
2712 as unset and return the appropriate default but do *not* set the default.
2713 This means we can return the appropriate method in applications that
2714 switch between FIPS and non-FIPS modes.
2717 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2718 ENGINE is used then we cannot handle that in the FIPS module so we
2719 keep original code iff non-FIPS operations are allowed.
2722 *) Add -attime option to openssl utilities.
2723 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2725 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2728 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2729 FIPS EC methods unconditionally for now.
2732 *) New build option no-ec2m to disable characteristic 2 code.
2735 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2736 all cases can be covered as some introduce binary incompatibilities.
2739 *) Redirect RSA operations to FIPS module including keygen,
2740 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2743 *) Add similar low level API blocking to ciphers.
2746 *) Low level digest APIs are not approved in FIPS mode: any attempt
2747 to use these will cause a fatal error. Applications that *really* want
2748 to use them can use the private_* version instead.
2751 *) Redirect cipher operations to FIPS module for FIPS builds.
2754 *) Redirect digest operations to FIPS module for FIPS builds.
2757 *) Update build system to add "fips" flag which will link in fipscanister.o
2758 for static and shared library builds embedding a signature if needed.
2761 *) Output TLS supported curves in preference order instead of numerical
2762 order. This is currently hardcoded for the highest order curves first.
2763 This should be configurable so applications can judge speed vs strength.
2766 *) Add TLS v1.2 server support for client authentication.
2769 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2773 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2774 FIPS modules versions.
2777 *) Add TLS v1.2 client side support for client authentication. Keep cache
2778 of handshake records longer as we don't know the hash algorithm to use
2779 until after the certificate request message is received.
2782 *) Initial TLS v1.2 client support. Add a default signature algorithms
2783 extension including all the algorithms we support. Parse new signature
2784 format in client key exchange. Relax some ECC signing restrictions for
2785 TLS v1.2 as indicated in RFC5246.
2788 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2789 to new signature format when needed using client digest preference.
2790 All server ciphersuites should now work correctly in TLS v1.2. No client
2791 support yet and no support for client certificates.
2794 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2795 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2796 ciphersuites. At present only RSA key exchange ciphersuites work with
2797 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2798 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2799 and version checking.
2802 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2803 with this defined it will not be affected by any changes to ssl internal
2804 structures. Add several utility functions to allow openssl application
2805 to work with OPENSSL_NO_SSL_INTERN defined.
2809 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2811 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2814 *) Permit abbreviated handshakes when renegotiating using the function
2815 SSL_renegotiate_abbreviated().
2816 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2818 *) Add call to ENGINE_register_all_complete() to
2819 ENGINE_load_builtin_engines(), so some implementations get used
2820 automatically instead of needing explicit application support.
2823 *) Add support for TLS key exporter as described in RFC5705.
2824 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2826 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2827 a few changes are required:
2829 Add SSL_OP_NO_TLSv1_1 flag.
2830 Add TLSv1_1 methods.
2831 Update version checking logic to handle version 1.1.
2832 Add explicit IV handling (ported from DTLS code).
2833 Add command line options to s_client/s_server.
2836 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2838 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2839 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2840 content decryption and always return the same error. Note: this attack
2841 needs on average 2^20 messages so it only affects automated senders. The
2842 old behaviour can be re-enabled in the CMS code by setting the
2843 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2844 an MMA defence is not necessary.
2845 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2846 this issue. (CVE-2012-0884)
2849 *) Fix CVE-2011-4619: make sure we really are receiving a
2850 client hello before rejecting multiple SGC restarts. Thanks to
2851 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2854 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2856 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2857 Thanks to Antonio Martin, Enterprise Secure Access Research and
2858 Development, Cisco Systems, Inc. for discovering this bug and
2859 preparing a fix. (CVE-2012-0050)
2862 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2864 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2865 of the Vaudenay padding oracle attack on CBC mode encryption
2866 which enables an efficient plaintext recovery attack against
2867 the OpenSSL implementation of DTLS. Their attack exploits timing
2868 differences arising during decryption processing. A research
2869 paper describing this attack can be found at:
2870 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2871 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2872 Security Group at Royal Holloway, University of London
2873 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2874 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2875 for preparing the fix. (CVE-2011-4108)
2876 [Robin Seggelmann, Michael Tuexen]
2878 *) Clear bytes used for block padding of SSL 3.0 records.
2880 [Adam Langley (Google)]
2882 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2883 Kadianakis <desnacked@gmail.com> for discovering this issue and
2884 Adam Langley for preparing the fix. (CVE-2011-4619)
2885 [Adam Langley (Google)]
2887 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2888 [Andrey Kulikov <amdeich@gmail.com>]
2890 *) Prevent malformed RFC3779 data triggering an assertion failure.
2891 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2892 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2893 [Rob Austein <sra@hactrn.net>]
2895 *) Improved PRNG seeding for VOS.
2896 [Paul Green <Paul.Green@stratus.com>]
2898 *) Fix ssl_ciph.c set-up race.
2899 [Adam Langley (Google)]
2901 *) Fix spurious failures in ecdsatest.c.
2902 [Emilia Käsper (Google)]
2904 *) Fix the BIO_f_buffer() implementation (which was mixing different
2905 interpretations of the '..._len' fields).
2906 [Adam Langley (Google)]
2908 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2909 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2910 threads won't reuse the same blinding coefficients.
2912 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2913 lock to call BN_BLINDING_invert_ex, and avoids one use of
2914 BN_BLINDING_update for each BN_BLINDING structure (previously,
2915 the last update always remained unused).
2916 [Emilia Käsper (Google)]
2918 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2919 [Bob Buckholz (Google)]
2921 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2923 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2924 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2925 [Kaspar Brand <ossl@velox.ch>]
2927 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2928 for multi-threaded use of ECDH. (CVE-2011-3210)
2929 [Adam Langley (Google)]
2931 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2934 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2935 signature public key algorithm by using OID xref utilities instead.
2936 Before this you could only use some ECC ciphersuites with SHA1 only.
2939 *) Add protection against ECDSA timing attacks as mentioned in the paper
2940 by Billy Bob Brumley and Nicola Tuveri, see:
2942 http://eprint.iacr.org/2011/232.pdf
2944 [Billy Bob Brumley and Nicola Tuveri]
2946 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2948 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2949 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2951 *) Fix bug in string printing code: if *any* escaping is enabled we must
2952 escape the escape character (backslash) or the resulting string is
2956 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2958 *) Disable code workaround for ancient and obsolete Netscape browsers
2959 and servers: an attacker can use it in a ciphersuite downgrade attack.
2960 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2963 *) Fixed J-PAKE implementation error, originally discovered by
2964 Sebastien Martini, further info and confirmation from Stefan
2965 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2968 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2970 *) Fix extension code to avoid race conditions which can result in a buffer
2971 overrun vulnerability: resumed sessions must not be modified as they can
2972 be shared by multiple threads. CVE-2010-3864
2975 *) Fix WIN32 build system to correctly link an ENGINE directory into
2979 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2981 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2983 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2985 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2987 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2988 context. The operation can be customised via the ctrl mechanism in
2989 case ENGINEs want to include additional functionality.
2992 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2995 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2996 output hashes compatible with older versions of OpenSSL.
2997 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2999 *) Fix compression algorithm handling: if resuming a session use the
3000 compression algorithm of the resumed session instead of determining
3001 it from client hello again. Don't allow server to change algorithm.
3004 *) Add load_crls() function to apps tidying load_certs() too. Add option
3005 to verify utility to allow additional CRLs to be included.
3008 *) Update OCSP request code to permit adding custom headers to the request:
3009 some responders need this.
3012 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3014 [Julia Lawall <julia@diku.dk>]
3016 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3017 needlessly dereferenced structures, used obsolete functions and
3018 didn't handle all updated verify codes correctly.
3021 *) Disable MD2 in the default configuration.
3024 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3025 indicate the initial BIO being pushed or popped. This makes it possible
3026 to determine whether the BIO is the one explicitly called or as a result
3027 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3028 it handles reference counts correctly and doesn't zero out the I/O bio
3029 when it is not being explicitly popped. WARNING: applications which
3030 included workarounds for the old buggy behaviour will need to be modified
3031 or they could free up already freed BIOs.
3034 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3035 renaming to all platforms (within the 0.9.8 branch, this was
3036 done conditionally on Netware platforms to avoid a name clash).
3037 [Guenter <lists@gknw.net>]
3039 *) Add ECDHE and PSK support to DTLS.
3040 [Michael Tuexen <tuexen@fh-muenster.de>]
3042 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3046 *) Add "missing" function EVP_MD_flags() (without this the only way to
3047 retrieve a digest flags is by accessing the structure directly. Update
3048 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3049 or cipher is registered as in the "from" argument. Print out all
3050 registered digests in the dgst usage message instead of manually
3051 attempting to work them out.
3054 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3055 this allows the use of compression and extensions. Change default cipher
3056 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3057 by default unless an application cipher string requests it.
3060 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3061 key ids to find matching certificates and keys but some PKCS#12 files
3062 don't follow the (somewhat unwritten) rules and this strategy fails.
3063 Now just gather all certificates together and the first private key
3064 then look for the first certificate that matches the key.
3067 *) Support use of registered digest and cipher names for dgst and cipher
3068 commands instead of having to add each one as a special case. So now
3075 openssl dgst -sha256 foo
3077 and this works for ENGINE based algorithms too.
3081 *) Update Gost ENGINE to support parameter files.
3082 [Victor B. Wagner <vitus@cryptocom.ru>]
3084 *) Support GeneralizedTime in ca utility.
3085 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3087 *) Enhance the hash format used for certificate directory links. The new
3088 form uses the canonical encoding (meaning equivalent names will work
3089 even if they aren't identical) and uses SHA1 instead of MD5. This form
3090 is incompatible with the older format and as a result c_rehash should
3091 be used to rebuild symbolic links.
3094 *) Make PKCS#8 the default write format for private keys, replacing the
3095 traditional format. This form is standardised, more secure and doesn't
3096 include an implicit MD5 dependency.
3099 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3100 committed to OpenSSL should pass this lot as a minimum.
3103 *) Add session ticket override functionality for use by EAP-FAST.
3104 [Jouni Malinen <j@w1.fi>]
3106 *) Modify HMAC functions to return a value. Since these can be implemented
3107 in an ENGINE errors can occur.
3110 *) Type-checked OBJ_bsearch_ex.
3113 *) Type-checked OBJ_bsearch. Also some constification necessitated
3114 by type-checking. Still to come: TXT_DB, bsearch(?),
3115 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3119 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3120 seconds to a tm structure directly, instead of going through OS
3121 specific date routines. This avoids any issues with OS routines such
3122 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3123 and X509_time_adj_ex() to cover the extended range. The existing
3124 X509_time_adj() is still usable and will no longer have any date issues.
3127 *) Delta CRL support. New use deltas option which will attempt to locate
3128 and search any appropriate delta CRLs available.
3130 This work was sponsored by Google.
3133 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3134 code and add additional score elements. Validate alternate CRL paths
3135 as part of the CRL checking and indicate a new error "CRL path validation
3136 error" in this case. Applications wanting additional details can use
3137 the verify callback and check the new "parent" field. If this is not
3138 NULL CRL path validation is taking place. Existing applications won't
3139 see this because it requires extended CRL support which is off by
3142 This work was sponsored by Google.
3145 *) Support for freshest CRL extension.
3147 This work was sponsored by Google.
3150 *) Initial indirect CRL support. Currently only supported in the CRLs
3151 passed directly and not via lookup. Process certificate issuer
3152 CRL entry extension and lookup CRL entries by bother issuer name
3153 and serial number. Check and process CRL issuer entry in IDP extension.
3155 This work was sponsored by Google.
3158 *) Add support for distinct certificate and CRL paths. The CRL issuer
3159 certificate is validated separately in this case. Only enabled if
3160 an extended CRL support flag is set: this flag will enable additional
3161 CRL functionality in future.
3163 This work was sponsored by Google.
3166 *) Add support for policy mappings extension.
3168 This work was sponsored by Google.
3171 *) Fixes to pathlength constraint, self issued certificate handling,
3172 policy processing to align with RFC3280 and PKITS tests.
3174 This work was sponsored by Google.
3177 *) Support for name constraints certificate extension. DN, email, DNS
3178 and URI types are currently supported.
3180 This work was sponsored by Google.
3183 *) To cater for systems that provide a pointer-based thread ID rather
3184 than numeric, deprecate the current numeric thread ID mechanism and
3185 replace it with a structure and associated callback type. This
3186 mechanism allows a numeric "hash" to be extracted from a thread ID in
3187 either case, and on platforms where pointers are larger than 'long',
3188 mixing is done to help ensure the numeric 'hash' is usable even if it
3189 can't be guaranteed unique. The default mechanism is to use "&errno"
3190 as a pointer-based thread ID to distinguish between threads.
3192 Applications that want to provide their own thread IDs should now use
3193 CRYPTO_THREADID_set_callback() to register a callback that will call
3194 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3196 Note that ERR_remove_state() is now deprecated, because it is tied
3197 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3198 to free the current thread's error state should be replaced by
3199 ERR_remove_thread_state(NULL).
3201 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3202 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3203 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3204 application was previously providing a numeric thread callback that
3205 was inappropriate for distinguishing threads, then uniqueness might
3206 have been obtained with &errno that happened immediately in the
3207 intermediate development versions of OpenSSL; this is no longer the
3208 case, the numeric thread callback will now override the automatic use
3210 [Geoff Thorpe, with help from Bodo Moeller]
3212 *) Initial support for different CRL issuing certificates. This covers a
3213 simple case where the self issued certificates in the chain exist and
3214 the real CRL issuer is higher in the existing chain.
3216 This work was sponsored by Google.
3219 *) Removed effectively defunct crypto/store from the build.
3222 *) Revamp of STACK to provide stronger type-checking. Still to come:
3223 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3224 ASN1_STRING, CONF_VALUE.
3227 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3228 RAM on SSL connections. This option can save about 34k per idle SSL.
3231 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3232 STACK, TXT_DB, bsearch, qsort.
3235 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3236 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3237 support for data, signedData, compressedData, digestedData and
3238 encryptedData, envelopedData types included. Scripts to check against
3239 RFC4134 examples draft and interop and consistency checks of many
3240 content types and variants.
3243 *) Add options to enc utility to support use of zlib compression BIO.
3246 *) Extend mk1mf to support importing of options and assembly language
3247 files from Configure script, currently only included in VC-WIN32.
3248 The assembly language rules can now optionally generate the source
3249 files from the associated perl scripts.
3252 *) Implement remaining functionality needed to support GOST ciphersuites.
3253 Interop testing has been performed using CryptoPro implementations.
3254 [Victor B. Wagner <vitus@cryptocom.ru>]
3256 *) s390x assembler pack.
3259 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3263 *) Implement Opaque PRF Input TLS extension as specified in
3264 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3265 official specification yet and no extension type assignment by
3266 IANA exists, this extension (for now) will have to be explicitly
3267 enabled when building OpenSSL by providing the extension number
3268 to use. For example, specify an option
3270 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3272 to the "config" or "Configure" script to enable the extension,
3273 assuming extension number 0x9527 (which is a completely arbitrary
3274 and unofficial assignment based on the MD5 hash of the Internet
3275 Draft). Note that by doing so, you potentially lose
3276 interoperability with other TLS implementations since these might
3277 be using the same extension number for other purposes.
3279 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3280 opaque PRF input value to use in the handshake. This will create
3281 an interal copy of the length-'len' string at 'src', and will
3282 return non-zero for success.
3284 To get more control and flexibility, provide a callback function
3287 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3288 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3292 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3295 Callback function 'cb' will be called in handshakes, and is
3296 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3297 Argument 'arg' is for application purposes (the value as given to
3298 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3299 be provided to the callback function). The callback function
3300 has to return non-zero to report success: usually 1 to use opaque
3301 PRF input just if possible, or 2 to enforce use of the opaque PRF
3302 input. In the latter case, the library will abort the handshake
3303 if opaque PRF input is not successfully negotiated.
3305 Arguments 'peerinput' and 'len' given to the callback function
3306 will always be NULL and 0 in the case of a client. A server will
3307 see the client's opaque PRF input through these variables if
3308 available (NULL and 0 otherwise). Note that if the server
3309 provides an opaque PRF input, the length must be the same as the
3310 length of the client's opaque PRF input.
3312 Note that the callback function will only be called when creating
3313 a new session (session resumption can resume whatever was
3314 previously negotiated), and will not be called in SSL 2.0
3315 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3316 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3317 for applications that need to enforce opaque PRF input.
3321 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3324 [Victor B. Wagner <vitus@cryptocom.ru>]
3326 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3327 RFC4507bis. The encrypted ticket format is an encrypted encoded
3328 SSL_SESSION structure, that way new session features are automatically
3331 If a client application caches session in an SSL_SESSION structure
3332 support is transparent because tickets are now stored in the encoded
3335 The SSL_CTX structure automatically generates keys for ticket
3336 protection in servers so again support should be possible
3337 with no application modification.
3339 If a client or server wishes to disable RFC4507 support then the option
3340 SSL_OP_NO_TICKET can be set.
3342 Add a TLS extension debugging callback to allow the contents of any client
3343 or server extensions to be examined.
3345 This work was sponsored by Google.
3348 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3349 OpenSSL should now compile cleanly on gcc 4.2
3350 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3352 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3353 support including streaming MAC support: this is required for GOST
3354 ciphersuite support.
3355 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3357 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3358 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3359 to output in BER and PEM format.
3362 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3363 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3364 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3365 ENGINE support for HMAC keys which are unextractable. New -mac and
3366 -macopt options to dgst utility.
3369 *) New option -sigopt to dgst utility. Update dgst to use
3370 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3371 alternative signing parameters such as X9.31 or PSS in the dgst
3375 *) Change ssl_cipher_apply_rule(), the internal function that does
3376 the work each time a ciphersuite string requests enabling
3377 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3378 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3379 the order of disabled ciphersuites such that those ciphersuites
3380 that most recently went from enabled to disabled not only stay
3381 in order with respect to each other, but also have higher priority
3382 than other disabled ciphersuites the next time ciphersuites are
3385 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3386 the same ciphersuites as with "HIGH" alone, but in a specific
3387 order where the PSK ciphersuites come first (since they are the
3388 most recently disabled ciphersuites when "HIGH" is parsed).
3390 Also, change ssl_create_cipher_list() (using this new
3391 funcionality) such that between otherwise identical
3392 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3396 *) Change ssl_create_cipher_list() so that it automatically
3397 arranges the ciphersuites in reasonable order before starting
3398 to process the rule string. Thus, the definition for "DEFAULT"
3399 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3400 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3401 This makes it much easier to arrive at a reasonable default order
3402 in applications for which anonymous ciphers are OK (meaning
3403 that you can't actually use DEFAULT).
3404 [Bodo Moeller; suggested by Victor Duchovni]
3406 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3407 processing) into multiple integers instead of setting
3408 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3409 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3410 (These masks as well as the individual bit definitions are hidden
3411 away into the non-exported interface ssl/ssl_locl.h, so this
3412 change to the definition of the SSL_CIPHER structure shouldn't
3413 affect applications.) This give us more bits for each of these
3414 categories, so there is no longer a need to coagulate AES128 and
3415 AES256 into a single algorithm bit, and to coagulate Camellia128
3416 and Camellia256 into a single algorithm bit, which has led to all
3419 Thus, among other things, the kludge introduced in 0.9.7m and
3420 0.9.8e for masking out AES256 independently of AES128 or masking
3421 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3423 With the change, we also introduce new ciphersuite aliases that
3424 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3428 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3429 Use the leftmost N bytes of the signature input if the input is
3430 larger than the prime q (with N being the size in bytes of q).
3433 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3434 it yet and it is largely untested.
3437 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3440 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3441 some compilers (gcc 4.2 and later) reject their use. Safestack is
3442 reimplemented. Update ASN1 to avoid use of legacy functions.
3445 *) Win32/64 targets are linked with Winsock2.
3448 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3449 to external functions. This can be used to increase CRL handling
3450 efficiency especially when CRLs are very large by (for example) storing
3451 the CRL revoked certificates in a database.
3454 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3455 new CRLs added to a directory can be used. New command line option
3456 -verify_return_error to s_client and s_server. This causes real errors
3457 to be returned by the verify callback instead of carrying on no matter
3458 what. This reflects the way a "real world" verify callback would behave.
3461 *) GOST engine, supporting several GOST algorithms and public key formats.
3462 Kindly donated by Cryptocom.
3465 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3466 partitioned by DP are handled but no indirect CRL or reason partitioning
3467 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3468 selected via a scoring technique which handles IDP and AKID in CRLs.
3471 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3472 will ultimately be used for all verify operations: this will remove the
3473 X509_STORE dependency on certificate verification and allow alternative
3474 lookup methods. X509_STORE based implementations of these two callbacks.
3477 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3478 Modify get_crl() to find a valid (unexpired) CRL if possible.
3481 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3482 this would be called X509_CRL_cmp() but that name is already used by
3483 a function that just compares CRL issuer names. Cache several CRL
3484 extensions in X509_CRL structure and cache CRLDP in X509.
3487 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3488 this maps equivalent X509_NAME structures into a consistent structure.
3489 Name comparison can then be performed rapidly using memcmp().
3492 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3496 *) Allow digests to supply their own micalg string for S/MIME type using
3497 the ctrl EVP_MD_CTRL_MICALG.
3500 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3501 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3502 ctrl. It can then customise the structure before and/or after signing
3506 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3507 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3508 to free up any added signature OIDs.
3511 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3512 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3513 digest and cipher tables. New options added to openssl utility:
3514 list-message-digest-algorithms and list-cipher-algorithms.
3517 *) Change the array representation of binary polynomials: the list
3518 of degrees of non-zero coefficients is now terminated with -1.
3519 Previously it was terminated with 0, which was also part of the
3520 value; thus, the array representation was not applicable to
3521 polynomials where t^0 has coefficient zero. This change makes
3522 the array representation useful in a more general context.
3525 *) Various modifications and fixes to SSL/TLS cipher string
3526 handling. For ECC, the code now distinguishes between fixed ECDH
3527 with RSA certificates on the one hand and with ECDSA certificates
3528 on the other hand, since these are separate ciphersuites. The
3529 unused code for Fortezza ciphersuites has been removed.
3531 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3532 (not "ECDHE"). For consistency with the code for DH
3533 certificates, use of ECDH certificates is now considered ECDH
3534 authentication, not RSA or ECDSA authentication (the latter is
3535 merely the CA's signing algorithm and not actively used in the
3538 The temporary ciphersuite alias "ECCdraft" is no longer
3539 available, and ECC ciphersuites are no longer excluded from "ALL"
3540 and "DEFAULT". The following aliases now exist for RFC 4492
3541 ciphersuites, most of these by analogy with the DH case:
3543 kECDHr - ECDH cert, signed with RSA
3544 kECDHe - ECDH cert, signed with ECDSA
3545 kECDH - ECDH cert (signed with either RSA or ECDSA)
3546 kEECDH - ephemeral ECDH
3547 ECDH - ECDH cert or ephemeral ECDH
3553 AECDH - anonymous ECDH
3554 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3558 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3559 Use correct micalg parameters depending on digest(s) in signed message.
3562 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3563 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3566 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3567 an engine to register a method. Add ENGINE lookups for methods and
3568 functional reference processing.
3571 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3572 EVP_{Sign,Verify}* which allow an application to customise the signature
3576 *) New -resign option to smime utility. This adds one or more signers
3577 to an existing PKCS#7 signedData structure. Also -md option to use an
3578 alternative message digest algorithm for signing.
3581 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3582 create PKCS7 structures containing multiple signers. Update smime
3583 application to support multiple signers.
3586 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3590 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3591 Reorganize PBE internals to lookup from a static table using NIDs,
3592 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3593 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3594 PRF which will be automatically used with PBES2.
3597 *) Replace the algorithm specific calls to generate keys in "req" with the
3601 *) Update PKCS#7 enveloped data routines to use new API. This is now
3602 supported by any public key method supporting the encrypt operation. A
3603 ctrl is added to allow the public key algorithm to examine or modify
3604 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3608 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3609 a default digest type to use. In most cases this will be SHA1 but some
3610 algorithms (such as GOST) need to specify an alternative digest. The
3611 return value indicates how strong the preference is 1 means optional and
3612 2 is mandatory (that is it is the only supported type). Modify
3613 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3614 use the default md. Update openssl utilities to use the default digest
3615 type for signing if it is not explicitly indicated.
3618 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3619 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3620 signing method from the key type. This effectively removes the link
3621 between digests and public key types.
3624 *) Add an OID cross reference table and utility functions. Its purpose is to
3625 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3626 rsaEncryption. This will allow some of the algorithm specific hackery
3627 needed to use the correct OID to be removed.
3630 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3631 structures for PKCS7_sign(). They are now set up by the relevant public
3635 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3638 *) Add support for key derivation (agreement) in the API, DH method and
3642 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3643 public and private key formats. As a side effect these add additional
3644 command line functionality not previously available: DSA signatures can be
3645 generated and verified using pkeyutl and DH key support and generation in
3650 [Oliver Tappe <zooey@hirschkaefer.de>]
3652 *) New make target "install_html_docs" installs HTML renditions of the
3654 [Oliver Tappe <zooey@hirschkaefer.de>]
3656 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3657 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3658 support key and parameter generation and add initial key generation
3659 functionality for RSA.
3662 *) Add functions for main EVP_PKEY_method operations. The undocumented
3663 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3664 EVP_PKEY_{encrypt,decrypt}_old.
3667 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3668 key API, doesn't do much yet.
3671 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3672 public key algorithms. New option to openssl utility:
3673 "list-public-key-algorithms" to print out info.
3676 *) Implement the Supported Elliptic Curves Extension for
3677 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3680 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3681 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3684 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3685 utilities such as rsa, dsa, dsaparam etc except they process any key
3689 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3690 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3691 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3695 *) Initial support for pluggable public key ASN1.
3696 De-spaghettify the public key ASN1 handling. Move public and private
3697 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3698 algorithm specific handling to a single module within the relevant
3699 algorithm directory. Add functions to allow (near) opaque processing
3700 of public and private key structures.
3703 *) Implement the Supported Point Formats Extension for
3704 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3707 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3708 for the psk identity [hint] and the psk callback functions to the
3709 SSL_SESSION, SSL and SSL_CTX structure.
3712 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3716 SSL_CTX_use_psk_identity_hint
3717 SSL_get_psk_identity_hint
3718 SSL_get_psk_identity
3719 SSL_use_psk_identity_hint
3721 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3723 *) Add RFC 3161 compliant time stamp request creation, response generation
3724 and response verification functionality.
3725 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3727 *) Add initial support for TLS extensions, specifically for the server_name
3728 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3729 have new members for a host name. The SSL data structure has an
3730 additional member SSL_CTX *initial_ctx so that new sessions can be
3731 stored in that context to allow for session resumption, even after the
3732 SSL has been switched to a new SSL_CTX in reaction to a client's
3733 server_name extension.
3735 New functions (subject to change):
3737 SSL_get_servername()
3738 SSL_get_servername_type()
3741 New CTRL codes and macros (subject to change):
3743 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3744 - SSL_CTX_set_tlsext_servername_callback()
3745 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3746 - SSL_CTX_set_tlsext_servername_arg()
3747 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3749 openssl s_client has a new '-servername ...' option.
3751 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3752 '-key2 ...', '-servername_fatal' (subject to change). This allows
3753 testing the HostName extension for a specific single host name ('-cert'
3754 and '-key' remain fallbacks for handshakes without HostName
3755 negotiation). If the unrecognized_name alert has to be sent, this by
3756 default is a warning; it becomes fatal with the '-servername_fatal'
3759 [Peter Sylvester, Remy Allais, Christophe Renou]
3761 *) Whirlpool hash implementation is added.
3764 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3765 bn(64,32). Because of instruction set limitations it doesn't have
3766 any negative impact on performance. This was done mostly in order
3767 to make it possible to share assembler modules, such as bn_mul_mont
3768 implementations, between 32- and 64-bit builds without hassle.
3771 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3772 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3776 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3777 dedicated Montgomery multiplication procedure, is introduced.
3778 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3779 "64-bit" performance on certain 32-bit targets.
3782 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3783 in SSL structures. New SSL ctrl to set maximum send fragment size.
3784 Save memory by seeting the I/O buffer sizes dynamically instead of
3785 using the maximum available value.
3788 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3789 in addition to the text details.
3792 *) Very, very preliminary EXPERIMENTAL support for printing of general
3793 ASN1 structures. This currently produces rather ugly output and doesn't
3794 handle several customised structures at all.
3797 *) Integrated support for PVK file format and some related formats such
3798 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3799 these in the 'rsa' and 'dsa' utilities.
3802 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3805 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3806 place for the (very old) "NETSCAPE" format certificates which are now
3807 handled using new ASN1 code equivalents.
3810 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3811 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3812 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3815 *) Modify CRL distribution points extension code to print out previously
3816 unsupported fields. Enhance extension setting code to allow setting of
3820 *) Add print and set support for Issuing Distribution Point CRL extension.
3823 *) Change 'Configure' script to enable Camellia by default.
3826 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3828 *) When rejecting SSL/TLS records due to an incorrect version number, never
3829 update s->server with a new major version number. As of
3830 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3831 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3832 the previous behavior could result in a read attempt at NULL when
3833 receiving specific incorrect SSL/TLS records once record payload
3834 protection is active. (CVE-2010-0740)
3835 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3837 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3838 could be crashed if the relevant tables were not present (e.g. chrooted).
3839 [Tomas Hoger <thoger@redhat.com>]
3841 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3843 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3844 [Martin Olsson, Neel Mehta]
3846 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3847 accommodate for stack sorting, always a write lock!).
3850 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3851 excessive delays in the RAND_poll(): over a minute. As a workaround
3852 include a time check in the inner Heap32Next loop too.
3855 *) The code that handled flushing of data in SSL/TLS originally used the
3856 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3857 the problem outlined in PR#1949. The fix suggested there however can
3858 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3859 of Apache). So instead simplify the code to flush unconditionally.
3860 This should be fine since flushing with no data to flush is a no op.
3863 *) Handle TLS versions 2.0 and later properly and correctly use the
3864 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3865 off ancient servers have a habit of sticking around for a while...
3868 *) Modify compression code so it frees up structures without using the
3869 ex_data callbacks. This works around a problem where some applications
3870 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3871 restarting) then use compression (e.g. SSL with compression) later.
3872 This results in significant per-connection memory leaks and
3873 has caused some security issues including CVE-2008-1678 and
3877 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3878 change when encrypting or decrypting.
3881 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3882 connect and renegotiate with servers which do not support RI.
3883 Until RI is more widely deployed this option is enabled by default.
3886 *) Add "missing" ssl ctrls to clear options and mode.
3889 *) If client attempts to renegotiate and doesn't support RI respond with
3890 a no_renegotiation alert as required by RFC5746. Some renegotiating
3891 TLS clients will continue a connection gracefully when they receive
3892 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3893 waiting for a server hello which it will never receive. Now we treat a
3894 received no_renegotiation alert as a fatal error. This is because
3895 applications requesting a renegotiation might well expect it to succeed
3896 and would have no code in place to handle the server denying it so the
3897 only safe thing to do is to terminate the connection.
3900 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3901 peer supports secure renegotiation and 0 otherwise. Print out peer
3902 renegotiation support in s_client/s_server.
3905 *) Replace the highly broken and deprecated SPKAC certification method with
3906 the updated NID creation version. This should correctly handle UTF8.
3909 *) Implement RFC5746. Re-enable renegotiation but require the extension
3910 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3911 turns out to be a bad idea. It has been replaced by
3912 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3913 SSL_CTX_set_options(). This is really not recommended unless you
3914 know what you are doing.
3915 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3917 *) Fixes to stateless session resumption handling. Use initial_ctx when
3918 issuing and attempting to decrypt tickets in case it has changed during
3919 servername handling. Use a non-zero length session ID when attempting
3920 stateless session resumption: this makes it possible to determine if
3921 a resumption has occurred immediately after receiving server hello
3922 (several places in OpenSSL subtly assume this) instead of later in
3926 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3927 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3928 fixes for a few places where the return code is not checked
3930 [Julia Lawall <julia@diku.dk>]
3932 *) Add --strict-warnings option to Configure script to include devteam
3933 warnings in other configurations.
3936 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3937 makes it possible to install openssl libraries in locations which
3938 have names other than "lib", for example "/usr/lib64" which some
3940 [Steve Henson, based on patch from Jeremy Utley]
3942 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3943 X690 8.9.12 and can produce some misleading textual output of OIDs.
3944 [Steve Henson, reported by Dan Kaminsky]
3946 *) Delete MD2 from algorithm tables. This follows the recommendation in
3947 several standards that it is not used in new applications due to
3948 several cryptographic weaknesses. For binary compatibility reasons
3949 the MD2 API is still compiled in by default.
3952 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3956 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3957 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3959 [Guenter <lists@gknw.net>]
3961 *) Fix the server certificate chain building code to use X509_verify_cert(),
3962 it used to have an ad-hoc builder which was unable to cope with anything
3963 other than a simple chain.
3964 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3966 *) Don't check self signed certificate signatures in X509_verify_cert()
3967 by default (a flag can override this): it just wastes time without
3968 adding any security. As a useful side effect self signed root CAs
3969 with non-FIPS digests are now usable in FIPS mode.
3972 *) In dtls1_process_out_of_seq_message() the check if the current message
3973 is already buffered was missing. For every new message was memory
3974 allocated, allowing an attacker to perform an denial of service attack
3975 with sending out of seq handshake messages until there is no memory
3976 left. Additionally every future messege was buffered, even if the
3977 sequence number made no sense and would be part of another handshake.
3978 So only messages with sequence numbers less than 10 in advance will be
3979 buffered. (CVE-2009-1378)
3980 [Robin Seggelmann, discovered by Daniel Mentz]
3982 *) Records are buffered if they arrive with a future epoch to be
3983 processed after finishing the corresponding handshake. There is
3984 currently no limitation to this buffer allowing an attacker to perform
3985 a DOS attack with sending records with future epochs until there is no
3986 memory left. This patch adds the pqueue_size() function to determine
3987 the size of a buffer and limits the record buffer to 100 entries.
3989 [Robin Seggelmann, discovered by Daniel Mentz]
3991 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3992 parent structure is freed. (CVE-2009-1379)
3995 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3996 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3999 [Ilya O. <vrghost@gmail.com>]
4001 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4003 *) Disable renegotiation completely - this fixes a severe security
4004 problem (CVE-2009-3555) at the cost of breaking all
4005 renegotiation. Renegotiation can be re-enabled by setting
4006 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4007 run-time. This is really not recommended unless you know what
4011 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4013 *) Don't set val to NULL when freeing up structures, it is freed up by
4014 underlying code. If sizeof(void *) > sizeof(long) this can result in
4015 zeroing past the valid field. (CVE-2009-0789)
4016 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4018 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4019 checked correctly. This would allow some invalid signed attributes to
4020 appear to verify correctly. (CVE-2009-0591)
4021 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4023 *) Reject UniversalString and BMPString types with invalid lengths. This
4024 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4025 a legal length. (CVE-2009-0590)
4028 *) Set S/MIME signing as the default purpose rather than setting it
4029 unconditionally. This allows applications to override it at the store
4033 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4034 to handle some structures.
4037 *) Improve efficiency of mem_gets: don't search whole buffer each time
4039 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4041 *) New -hex option for openssl rand.
4044 *) Print out UTF8String and NumericString when parsing ASN1.
4047 *) Support NumericString type for name components.
4050 *) Allow CC in the environment to override the automatically chosen
4051 compiler. Note that nothing is done to ensure flags work with the
4055 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4057 *) Properly check EVP_VerifyFinal() and similar return values
4059 [Ben Laurie, Bodo Moeller, Google Security Team]
4061 *) Enable TLS extensions by default.
4064 *) Allow the CHIL engine to be loaded, whether the application is
4065 multithreaded or not. (This does not release the developer from the
4066 obligation to set up the dynamic locking callbacks.)
4067 [Sander Temme <sander@temme.net>]
4069 *) Use correct exit code if there is an error in dgst command.
4070 [Steve Henson; problem pointed out by Roland Dirlewanger]
4072 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4073 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4076 *) Add experimental JPAKE support, including demo authentication in
4077 s_client and s_server.
4080 *) Set the comparison function in v3_addr_canonize().
4081 [Rob Austein <sra@hactrn.net>]
4083 *) Add support for XMPP STARTTLS in s_client.
4084 [Philip Paeps <philip@freebsd.org>]
4086 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4087 to ensure that even with this option, only ciphersuites in the
4088 server's preference list will be accepted. (Note that the option
4089 applies only when resuming a session, so the earlier behavior was
4090 just about the algorithm choice for symmetric cryptography.)
4093 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4095 *) Fix NULL pointer dereference if a DTLS server received
4096 ChangeCipherSpec as first record (CVE-2009-1386).
4099 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4100 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4103 *) The fix in 0.9.8c that supposedly got rid of unsafe
4104 double-checked locking was incomplete for RSA blinding,
4105 addressing just one layer of what turns out to have been
4106 doubly unsafe triple-checked locking.
4108 So now fix this for real by retiring the MONT_HELPER macro
4109 in crypto/rsa/rsa_eay.c.
4111 [Bodo Moeller; problem pointed out by Marius Schilder]
4113 *) Various precautionary measures:
4115 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4117 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4118 (NB: This would require knowledge of the secret session ticket key
4119 to exploit, in which case you'd be SOL either way.)
4121 - Change bn_nist.c so that it will properly handle input BIGNUMs
4122 outside the expected range.
4124 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4127 [Neel Mehta, Bodo Moeller]
4129 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4130 the load fails. Useful for distros.
4131 [Ben Laurie and the FreeBSD team]
4133 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4136 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4139 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4141 This work was sponsored by Logica.
4144 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4145 keystores. Support for SSL/TLS client authentication too.
4146 Not compiled unless enable-capieng specified to Configure.
4148 This work was sponsored by Logica.
4151 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4152 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4153 attribute creation routines such as certificate requests and PKCS#12
4157 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4159 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4160 handshake which could lead to a cilent crash as found using the
4161 Codenomicon TLS test suite (CVE-2008-1672)
4162 [Steve Henson, Mark Cox]
4164 *) Fix double free in TLS server name extensions which could lead to
4165 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4168 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4170 Clear the error queue to ensure that error entries left from
4171 older function calls do not interfere with the correct operation.
4172 [Lutz Jaenicke, Erik de Castro Lopo]
4174 *) Remove root CA certificates of commercial CAs:
4176 The OpenSSL project does not recommend any specific CA and does not
4177 have any policy with respect to including or excluding any CA.
4178 Therefore it does not make any sense to ship an arbitrary selection
4179 of root CA certificates with the OpenSSL software.
4182 *) RSA OAEP patches to fix two separate invalid memory reads.
4183 The first one involves inputs when 'lzero' is greater than
4184 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4185 before the beginning of from). The second one involves inputs where
4186 the 'db' section contains nothing but zeroes (there is a one-byte
4187 invalid read after the end of 'db').
4188 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4190 *) Partial backport from 0.9.9-dev:
4192 Introduce bn_mul_mont (dedicated Montgomery multiplication
4193 procedure) as a candidate for BIGNUM assembler implementation.
4194 While 0.9.9-dev uses assembler for various architectures, only
4195 x86_64 is available by default here in the 0.9.8 branch, and
4196 32-bit x86 is available through a compile-time setting.
4198 To try the 32-bit x86 assembler implementation, use Configure
4199 option "enable-montasm" (which exists only for this backport).
4201 As "enable-montasm" for 32-bit x86 disclaims code stability
4202 anyway, in this constellation we activate additional code
4203 backported from 0.9.9-dev for further performance improvements,
4204 namely BN_from_montgomery_word. (To enable this otherwise,
4205 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4207 [Andy Polyakov (backport partially by Bodo Moeller)]
4209 *) Add TLS session ticket callback. This allows an application to set
4210 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4211 values. This is useful for key rollover for example where several key
4212 sets may exist with different names.
4215 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4216 This was broken until now in 0.9.8 releases, such that the only way
4217 a registered ENGINE could be used (assuming it initialises
4218 successfully on the host) was to explicitly set it as the default
4219 for the relevant algorithms. This is in contradiction with 0.9.7
4220 behaviour and the documentation. With this fix, when an ENGINE is
4221 registered into a given algorithm's table of implementations, the
4222 'uptodate' flag is reset so that auto-discovery will be used next
4223 time a new context for that algorithm attempts to select an
4225 [Ian Lister (tweaked by Geoff Thorpe)]
4227 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4228 implementation in the following ways:
4230 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4233 Lack of BER streaming support means one pass streaming processing is
4234 only supported if data is detached: setting the streaming flag is
4235 ignored for embedded content.
4237 CMS support is disabled by default and must be explicitly enabled
4238 with the enable-cms configuration option.
4241 *) Update the GMP engine glue to do direct copies between BIGNUM and
4242 mpz_t when openssl and GMP use the same limb size. Otherwise the
4243 existing "conversion via a text string export" trick is still used.
4244 [Paul Sheer <paulsheer@gmail.com>]
4246 *) Zlib compression BIO. This is a filter BIO which compressed and
4247 uncompresses any data passed through it.
4250 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4251 RFC3394 compatible AES key wrapping.
4254 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4255 sets string data without copying. X509_ALGOR_set0() and
4256 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4257 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4258 from an X509_ATTRIBUTE structure optionally checking it occurs only
4259 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4263 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4264 to get the expected BN_FLG_CONSTTIME behavior.
4265 [Bodo Moeller (Google)]
4269 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4270 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4271 - added some more tests to do_tests.pl
4272 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4273 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4274 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4275 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4276 - various changes to netware.pl to enable gcc-cross builds on Win32
4278 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4279 - various changes to fix missing prototype warnings
4280 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4281 - added AES, WHIRLPOOL and CPUID assembler code to build files
4282 - added missing AES assembler make rules to mk1mf.pl
4283 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4284 [Guenter Knauf <eflash@gmx.net>]
4286 *) Implement certificate status request TLS extension defined in RFC3546.
4287 A client can set the appropriate parameters and receive the encoded
4288 OCSP response via a callback. A server can query the supplied parameters
4289 and set the encoded OCSP response in the callback. Add simplified examples
4290 to s_client and s_server.
4293 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4295 *) Fix various bugs:
4296 + Binary incompatibility of ssl_ctx_st structure
4297 + DTLS interoperation with non-compliant servers
4298 + Don't call get_session_cb() without proposed session
4299 + Fix ia64 assembler code
4300 [Andy Polyakov, Steve Henson]
4302 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4304 *) DTLS Handshake overhaul. There were longstanding issues with
4305 OpenSSL DTLS implementation, which were making it impossible for
4306 RFC 4347 compliant client to communicate with OpenSSL server.
4307 Unfortunately just fixing these incompatibilities would "cut off"
4308 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4309 server keeps tolerating non RFC compliant syntax. The opposite is
4310 not true, 0.9.8f client can not communicate with earlier server.
4311 This update even addresses CVE-2007-4995.
4314 *) Changes to avoid need for function casts in OpenSSL: some compilers
4315 (gcc 4.2 and later) reject their use.
4316 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4319 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4320 RFC4507bis. The encrypted ticket format is an encrypted encoded
4321 SSL_SESSION structure, that way new session features are automatically
4324 If a client application caches session in an SSL_SESSION structure
4325 support is transparent because tickets are now stored in the encoded
4328 The SSL_CTX structure automatically generates keys for ticket
4329 protection in servers so again support should be possible
4330 with no application modification.
4332 If a client or server wishes to disable RFC4507 support then the option
4333 SSL_OP_NO_TICKET can be set.
4335 Add a TLS extension debugging callback to allow the contents of any client
4336 or server extensions to be examined.
4338 This work was sponsored by Google.
4341 *) Add initial support for TLS extensions, specifically for the server_name
4342 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4343 have new members for a host name. The SSL data structure has an
4344 additional member SSL_CTX *initial_ctx so that new sessions can be
4345 stored in that context to allow for session resumption, even after the
4346 SSL has been switched to a new SSL_CTX in reaction to a client's
4347 server_name extension.
4349 New functions (subject to change):
4351 SSL_get_servername()
4352 SSL_get_servername_type()
4355 New CTRL codes and macros (subject to change):
4357 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4358 - SSL_CTX_set_tlsext_servername_callback()
4359 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4360 - SSL_CTX_set_tlsext_servername_arg()
4361 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4363 openssl s_client has a new '-servername ...' option.
4365 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4366 '-key2 ...', '-servername_fatal' (subject to change). This allows
4367 testing the HostName extension for a specific single host name ('-cert'
4368 and '-key' remain fallbacks for handshakes without HostName
4369 negotiation). If the unrecognized_name alert has to be sent, this by
4370 default is a warning; it becomes fatal with the '-servername_fatal'
4373 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4375 *) Add AES and SSE2 assembly language support to VC++ build.
4378 *) Mitigate attack on final subtraction in Montgomery reduction.
4381 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4382 (which previously caused an internal error).
4385 *) Squeeze another 10% out of IGE mode when in != out.
4388 *) AES IGE mode speedup.
4389 [Dean Gaudet (Google)]
4391 *) Add the Korean symmetric 128-bit cipher SEED (see
4392 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4393 add SEED ciphersuites from RFC 4162:
4395 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4396 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4397 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4398 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4400 To minimize changes between patchlevels in the OpenSSL 0.9.8
4401 series, SEED remains excluded from compilation unless OpenSSL
4402 is configured with 'enable-seed'.
4403 [KISA, Bodo Moeller]
4405 *) Mitigate branch prediction attacks, which can be practical if a
4406 single processor is shared, allowing a spy process to extract
4407 information. For detailed background information, see
4408 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4409 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4410 and Necessary Software Countermeasures"). The core of the change
4411 are new versions BN_div_no_branch() and
4412 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4413 respectively, which are slower, but avoid the security-relevant
4414 conditional branches. These are automatically called by BN_div()
4415 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4416 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4417 remove a conditional branch.
4419 BN_FLG_CONSTTIME is the new name for the previous
4420 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4421 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4422 in the exponent causes BN_mod_exp_mont() to use the alternative
4423 implementation in BN_mod_exp_mont_consttime().) The old name
4424 remains as a deprecated alias.
4426 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4427 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4428 constant-time implementations for more than just exponentiation.
4429 Here too the old name is kept as a deprecated alias.
4431 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4432 the BN_BLINDING structure gets an independent copy of the
4433 modulus. This means that the previous "BIGNUM *m" argument to
4434 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4435 essentially becomes "const BIGNUM *m", although we can't actually
4436 change this in the header file before 0.9.9. It allows
4437 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4438 enable BN_FLG_CONSTTIME.
4440 [Matthew D Wood (Intel Corp)]
4442 *) In the SSL/TLS server implementation, be strict about session ID
4443 context matching (which matters if an application uses a single
4444 external cache for different purposes). Previously,
4445 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4446 set. This did ensure strict client verification, but meant that,
4447 with applications using a single external cache for quite
4448 different requirements, clients could circumvent ciphersuite
4449 restrictions for a given session ID context by starting a session
4450 in a different context.
4453 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4454 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4455 authentication-only ciphersuites.
4458 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4459 not complete and could lead to a possible single byte overflow
4460 (CVE-2007-5135) [Ben Laurie]
4462 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4464 *) Since AES128 and AES256 (and similarly Camellia128 and
4465 Camellia256) share a single mask bit in the logic of
4466 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4467 kludge to work properly if AES128 is available and AES256 isn't
4468 (or if Camellia128 is available and Camellia256 isn't).
4471 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4472 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4473 When a point or a seed is encoded in a BIT STRING, we need to
4474 prevent the removal of trailing zero bits to get the proper DER
4475 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4476 of a NamedBitList, for which trailing 0 bits need to be removed.)
4479 *) Have SSL/TLS server implementation tolerate "mismatched" record
4480 protocol version while receiving ClientHello even if the
4481 ClientHello is fragmented. (The server can't insist on the
4482 particular protocol version it has chosen before the ServerHello
4483 message has informed the client about his choice.)
4486 *) Add RFC 3779 support.
4487 [Rob Austein for ARIN, Ben Laurie]
4489 *) Load error codes if they are not already present instead of using a
4490 static variable. This allows them to be cleanly unloaded and reloaded.
4491 Improve header file function name parsing.
4494 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4495 or CAPABILITY handshake as required by RFCs.
4498 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4500 *) Introduce limits to prevent malicious keys being able to
4501 cause a denial of service. (CVE-2006-2940)
4502 [Steve Henson, Bodo Moeller]
4504 *) Fix ASN.1 parsing of certain invalid structures that can result
4505 in a denial of service. (CVE-2006-2937) [Steve Henson]
4507 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4508 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4510 *) Fix SSL client code which could crash if connecting to a
4511 malicious SSLv2 server. (CVE-2006-4343)
4512 [Tavis Ormandy and Will Drewry, Google Security Team]
4514 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4515 match only those. Before that, "AES256-SHA" would be interpreted
4516 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4517 the same strength classification in 0.9.7h) as we currently only
4518 have a single AES bit in the ciphersuite description bitmap.
4519 That change, however, also applied to ciphersuite strings such as
4520 "RC4-MD5" that intentionally matched multiple ciphersuites --
4521 namely, SSL 2.0 ciphersuites in addition to the more common ones
4522 from SSL 3.0/TLS 1.0.
4524 So we change the selection algorithm again: Naming an explicit
4525 ciphersuite selects this one ciphersuite, and any other similar
4526 ciphersuite (same bitmap) from *other* protocol versions.
4527 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4528 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4530 Since SSL 2.0 does not have any ciphersuites for which the
4531 128/256 bit distinction would be relevant, this works for now.
4532 The proper fix will be to use different bits for AES128 and
4533 AES256, which would have avoided the problems from the beginning;
4534 however, bits are scarce, so we can only do this in a new release
4535 (not just a patchlevel) when we can change the SSL_CIPHER
4536 definition to split the single 'unsigned long mask' bitmap into
4537 multiple values to extend the available space.
4541 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4543 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4544 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4546 *) Add AES IGE and biIGE modes.
4549 *) Change the Unix randomness entropy gathering to use poll() when
4550 possible instead of select(), since the latter has some
4551 undesirable limitations.
4552 [Darryl Miles via Richard Levitte and Bodo Moeller]
4554 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4555 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4556 cannot be implicitly activated as part of, e.g., the "AES" alias.
4557 However, please upgrade to OpenSSL 0.9.9[-dev] for
4558 non-experimental use of the ECC ciphersuites to get TLS extension
4559 support, which is required for curve and point format negotiation
4560 to avoid potential handshake problems.
4563 *) Disable rogue ciphersuites:
4565 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4566 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4567 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4569 The latter two were purportedly from
4570 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4573 Also deactivate the remaining ciphersuites from
4574 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4575 unofficial, and the ID has long expired.
4578 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4579 dual-core machines) and other potential thread-safety issues.
4582 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4583 versions), which is now available for royalty-free use
4584 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4585 Also, add Camellia TLS ciphersuites from RFC 4132.
4587 To minimize changes between patchlevels in the OpenSSL 0.9.8
4588 series, Camellia remains excluded from compilation unless OpenSSL
4589 is configured with 'enable-camellia'.
4592 *) Disable the padding bug check when compression is in use. The padding
4593 bug check assumes the first packet is of even length, this is not
4594 necessarily true if compresssion is enabled and can result in false
4595 positives causing handshake failure. The actual bug test is ancient
4596 code so it is hoped that implementations will either have fixed it by
4597 now or any which still have the bug do not support compression.
4600 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4602 *) When applying a cipher rule check to see if string match is an explicit
4603 cipher suite and only match that one cipher suite if it is.
4606 *) Link in manifests for VC++ if needed.
4607 [Austin Ziegler <halostatue@gmail.com>]
4609 *) Update support for ECC-based TLS ciphersuites according to
4610 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4611 TLS extensions, which are supported starting with the 0.9.9
4612 branch, not in the OpenSSL 0.9.8 branch).
4615 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4616 opaque EVP_CIPHER_CTX handling.
4619 *) Fixes and enhancements to zlib compression code. We now only use
4620 "zlib1.dll" and use the default __cdecl calling convention on Win32
4621 to conform with the standards mentioned here:
4622 http://www.zlib.net/DLL_FAQ.txt
4623 Static zlib linking now works on Windows and the new --with-zlib-include
4624 --with-zlib-lib options to Configure can be used to supply the location
4625 of the headers and library. Gracefully handle case where zlib library
4629 *) Several fixes and enhancements to the OID generation code. The old code
4630 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4631 handle numbers larger than ULONG_MAX, truncated printing and had a
4632 non standard OBJ_obj2txt() behaviour.
4635 *) Add support for building of engines under engine/ as shared libraries
4636 under VC++ build system.
4639 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4640 Hopefully, we will not see any false combination of paths any more.
4643 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4645 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4646 (part of SSL_OP_ALL). This option used to disable the
4647 countermeasure against man-in-the-middle protocol-version
4648 rollback in the SSL 2.0 server implementation, which is a bad
4649 idea. (CVE-2005-2969)
4651 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4652 for Information Security, National Institute of Advanced Industrial
4653 Science and Technology [AIST], Japan)]
4655 *) Add two function to clear and return the verify parameter flags.
4658 *) Keep cipherlists sorted in the source instead of sorting them at
4659 runtime, thus removing the need for a lock.
4662 *) Avoid some small subgroup attacks in Diffie-Hellman.
4663 [Nick Mathewson and Ben Laurie]
4665 *) Add functions for well-known primes.
4668 *) Extended Windows CE support.
4669 [Satoshi Nakamura and Andy Polyakov]
4671 *) Initialize SSL_METHOD structures at compile time instead of during
4672 runtime, thus removing the need for a lock.
4675 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4676 attempting to decrypt each encrypted key in turn. Add support to
4680 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4682 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4685 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4688 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4689 key into the same file any more.
4692 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4695 *) Add -utf8 command line and config file option to 'ca'.
4696 [Stefan <stf@udoma.org]
4698 *) Removed the macro des_crypt(), as it seems to conflict with some
4699 libraries. Use DES_crypt().
4702 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4703 involves renaming the source and generated shared-libs for
4704 both. The engines will accept the corrected or legacy ids
4705 ('ncipher' and '4758_cca' respectively) when binding. NB,
4706 this only applies when building 'shared'.
4707 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4709 *) Add attribute functions to EVP_PKEY structure. Modify
4710 PKCS12_create() to recognize a CSP name attribute and
4711 use it. Make -CSP option work again in pkcs12 utility.
4714 *) Add new functionality to the bn blinding code:
4715 - automatic re-creation of the BN_BLINDING parameters after
4716 a fixed number of uses (currently 32)
4717 - add new function for parameter creation
4718 - introduce flags to control the update behaviour of the
4719 BN_BLINDING parameters
4720 - hide BN_BLINDING structure
4721 Add a second BN_BLINDING slot to the RSA structure to improve
4722 performance when a single RSA object is shared among several
4726 *) Add support for DTLS.
4727 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4729 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4730 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4733 *) Remove buggy and incomplete DH cert support from
4734 ssl/ssl_rsa.c and ssl/s3_both.c
4737 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4738 the apps/openssl applications.
4741 *) Compile clean with "-Wall -Wmissing-prototypes
4742 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4743 DEBUG_SAFESTACK must also be set.
4746 *) Change ./Configure so that certain algorithms can be disabled by default.
4747 The new counterpiece to "no-xxx" is "enable-xxx".
4749 The patented RC5 and MDC2 algorithms will now be disabled unless
4750 "enable-rc5" and "enable-mdc2", respectively, are specified.
4752 (IDEA remains enabled despite being patented. This is because IDEA
4753 is frequently required for interoperability, and there is no license
4754 fee for non-commercial use. As before, "no-idea" can be used to
4755 avoid this algorithm.)
4759 *) Add processing of proxy certificates (see RFC 3820). This work was
4760 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4761 EGEE (Enabling Grids for E-science in Europe).
4764 *) RC4 performance overhaul on modern architectures/implementations, such
4765 as Intel P4, IA-64 and AMD64.
4768 *) New utility extract-section.pl. This can be used specify an alternative
4769 section number in a pod file instead of having to treat each file as
4770 a separate case in Makefile. This can be done by adding two lines to the
4773 =for comment openssl_section:XXX
4775 The blank line is mandatory.
4779 *) New arguments -certform, -keyform and -pass for s_client and s_server
4780 to allow alternative format key and certificate files and passphrase
4784 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4785 update associated structures and add various utility functions.
4787 Add new policy related verify parameters, include policy checking in
4788 standard verify code. Enhance 'smime' application with extra parameters
4789 to support policy checking and print out.
4792 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4793 Nehemiah processors. These extensions support AES encryption in hardware
4794 as well as RNG (though RNG support is currently disabled).
4795 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4797 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4800 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4801 [Andy Polyakov and a number of other people]
4803 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4804 implementation contributed by IBM.
4805 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4807 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4808 exponent rather than 'unsigned long'. There is a corresponding change to
4809 the new 'rsa_keygen' element of the RSA_METHOD structure.
4810 [Jelte Jansen, Geoff Thorpe]
4812 *) Functionality for creating the initial serial number file is now
4813 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4815 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4816 number file to 1, which is bound to cause problems. To avoid
4817 the problems while respecting compatibility between different 0.9.7
4818 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4819 CA.pl for serial number initialization. With the new release 0.9.8,
4820 we can fix the problem directly in the 'ca' utility.)
4823 *) Reduced header interdepencies by declaring more opaque objects in
4824 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4825 give fewer recursive includes, which could break lazy source code - so
4826 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4827 developers should define this symbol when building and using openssl to
4828 ensure they track the recommended behaviour, interfaces, [etc], but
4829 backwards-compatible behaviour prevails when this isn't defined.
4832 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4835 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4836 This will generate a random key of the appropriate length based on the
4837 cipher context. The EVP_CIPHER can provide its own random key generation
4838 routine to support keys of a specific form. This is used in the des and
4839 3des routines to generate a key of the correct parity. Update S/MIME
4840 code to use new functions and hence generate correct parity DES keys.
4841 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4842 valid (weak or incorrect parity).
4845 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4846 as looking them up. This is useful when the verified structure may contain
4847 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4848 present unless the new PKCS7_NO_CRL flag is asserted.
4851 *) Extend ASN1 oid configuration module. It now additionally accepts the
4854 shortName = some long name, 1.2.3.4
4857 *) Reimplemented the BN_CTX implementation. There is now no more static
4858 limitation on the number of variables it can handle nor the depth of the
4859 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4860 information can now expand as required, and rather than having a single
4861 static array of bignums, BN_CTX now uses a linked-list of such arrays
4862 allowing it to expand on demand whilst maintaining the usefulness of
4863 BN_CTX's "bundling".
4866 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4867 to allow all RSA operations to function using a single BN_CTX.
4870 *) Preliminary support for certificate policy evaluation and checking. This
4871 is initially intended to pass the tests outlined in "Conformance Testing
4872 of Relying Party Client Certificate Path Processing Logic" v1.07.
4875 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4876 remained unused and not that useful. A variety of other little bignum
4877 tweaks and fixes have also been made continuing on from the audit (see
4881 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4882 associated ASN1, EVP and SSL functions and old ASN1 macros.
4885 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4886 and this should never fail. So the return value from the use of
4887 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4888 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4891 *) BN_CTX_get() should return zero-valued bignums, providing the same
4892 initialised value as BN_new().
4893 [Geoff Thorpe, suggested by Ulf Möller]
4895 *) Support for inhibitAnyPolicy certificate extension.
4898 *) An audit of the BIGNUM code is underway, for which debugging code is
4899 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4900 is considered valid when processing BIGNUMs, and causes execution to
4901 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4902 further steps are taken to deliberately pollute unused data in BIGNUM
4903 structures to try and expose faulty code further on. For now, openssl will
4904 (in its default mode of operation) continue to tolerate the inconsistent
4905 forms that it has tolerated in the past, but authors and packagers should
4906 consider trying openssl and their own applications when compiled with
4907 these debugging symbols defined. It will help highlight potential bugs in
4908 their own code, and will improve the test coverage for OpenSSL itself. At
4909 some point, these tighter rules will become openssl's default to improve
4910 maintainability, though the assert()s and other overheads will remain only
4911 in debugging configurations. See bn.h for more details.
4912 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4914 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4915 that can only be obtained through BN_CTX_new() (which implicitly
4916 initialises it). The presence of this function only made it possible
4917 to overwrite an existing structure (and cause memory leaks).
4920 *) Because of the callback-based approach for implementing LHASH as a
4921 template type, lh_insert() adds opaque objects to hash-tables and
4922 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4923 to clean up those corresponding objects before destroying the hash table
4924 (and losing the object pointers). So some over-zealous constifications in
4925 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4926 objects as "const" and the lh_doall[_arg] callback wrappers are not
4927 prototyped to have "const" restrictions on the object pointers they are
4928 given (and so aren't required to cast them away any more).
4931 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4932 (speed) prefers to use its own implementation. The two implementations
4933 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4934 its object type properly exposed (MS_TM) instead of casting to/from "char
4935 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4936 aren't necessarily the greatest nomenclatures - but this is what was used
4937 internally to the implementation so I've used that for now.
4940 *) Ensure that deprecated functions do not get compiled when
4941 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4942 the self-tests were still using deprecated key-generation functions so
4943 these have been updated also.
4946 *) Reorganise PKCS#7 code to separate the digest location functionality
4947 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4948 New function PKCS7_set_digest() to set the digest type for PKCS#7
4949 digestedData type. Add additional code to correctly generate the
4950 digestedData type and add support for this type in PKCS7 initialization
4954 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4955 structure of type "other".
4958 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4959 sure the loop does correctly stop and breaking ("division by zero")
4960 modulus operations are not performed. The (pre-generated) prime
4961 table crypto/bn/bn_prime.h was already correct, but it could not be
4962 re-generated on some platforms because of the "division by zero"
4963 situation in the script.
4964 [Ralf S. Engelschall]
4966 *) Update support for ECC-based TLS ciphersuites according to
4967 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4968 SHA-1 now is only used for "small" curves (where the
4969 representation of a field element takes up to 24 bytes); for
4970 larger curves, the field element resulting from ECDH is directly
4971 used as premaster secret.
4972 [Douglas Stebila (Sun Microsystems Laboratories)]
4974 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
4975 curve secp160r1 to the tests.
4976 [Douglas Stebila (Sun Microsystems Laboratories)]
4978 *) Add the possibility to load symbols globally with DSO.
4979 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
4981 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
4982 control of the error stack.
4985 *) Add support for STORE in ENGINE.
4988 *) Add the STORE type. The intention is to provide a common interface
4989 to certificate and key stores, be they simple file-based stores, or
4990 HSM-type store, or LDAP stores, or...
4991 NOTE: The code is currently UNTESTED and isn't really used anywhere.
4994 *) Add a generic structure called OPENSSL_ITEM. This can be used to
4995 pass a list of arguments to any function as well as provide a way
4996 for a function to pass data back to the caller.
4999 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5000 works like BUF_strdup() but can be used to duplicate a portion of
5001 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5005 *) Add the function sk_find_ex() which works like sk_find(), but will
5006 return an index to an element even if an exact match couldn't be
5007 found. The index is guaranteed to point at the element where the
5008 searched-for key would be inserted to preserve sorting order.
5011 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5012 takes an extra flags argument for optional functionality. Currently,
5013 the following flags are defined:
5015 OBJ_BSEARCH_VALUE_ON_NOMATCH
5016 This one gets OBJ_bsearch_ex() to return a pointer to the first
5017 element where the comparing function returns a negative or zero
5020 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5021 This one gets OBJ_bsearch_ex() to return a pointer to the first
5022 element where the comparing function returns zero. This is useful
5023 if there are more than one element where the comparing function
5027 *) Make it possible to create self-signed certificates with 'openssl ca'
5028 in such a way that the self-signed certificate becomes part of the
5029 CA database and uses the same mechanisms for serial number generation
5030 as all other certificate signing. The new flag '-selfsign' enables
5031 this functionality. Adapt CA.sh and CA.pl.in.
5034 *) Add functionality to check the public key of a certificate request
5035 against a given private. This is useful to check that a certificate
5036 request can be signed by that key (self-signing).
5039 *) Make it possible to have multiple active certificates with the same
5040 subject in the CA index file. This is done only if the keyword
5041 'unique_subject' is set to 'no' in the main CA section (default
5042 if 'CA_default') of the configuration file. The value is saved
5043 with the database itself in a separate index attribute file,
5044 named like the index file with '.attr' appended to the name.
5047 *) Generate muti valued AVAs using '+' notation in config files for
5051 *) Support for nameConstraints certificate extension.
5054 *) Support for policyConstraints certificate extension.
5057 *) Support for policyMappings certificate extension.
5060 *) Make sure the default DSA_METHOD implementation only uses its
5061 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5062 and change its own handlers to be NULL so as to remove unnecessary
5063 indirection. This lets alternative implementations fallback to the
5064 default implementation more easily.
5067 *) Support for directoryName in GeneralName related extensions
5071 *) Make it possible to link applications using Makefile.shared.
5072 Make that possible even when linking against static libraries!
5075 *) Support for single pass processing for S/MIME signing. This now
5076 means that S/MIME signing can be done from a pipe, in addition
5077 cleartext signing (multipart/signed type) is effectively streaming
5078 and the signed data does not need to be all held in memory.
5080 This is done with a new flag PKCS7_STREAM. When this flag is set
5081 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5082 is done after the data is output (and digests calculated) in
5083 SMIME_write_PKCS7().
5086 *) Add full support for -rpath/-R, both in shared libraries and
5087 applications, at least on the platforms where it's known how
5091 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5092 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5093 will now compute a table of multiples of the generator that
5094 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5095 faster (notably in the case of a single point multiplication,
5096 scalar * generator).
5097 [Nils Larsch, Bodo Moeller]
5099 *) IPv6 support for certificate extensions. The various extensions
5100 which use the IP:a.b.c.d can now take IPv6 addresses using the
5101 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5105 *) Added an ENGINE that implements RSA by performing private key
5106 exponentiations with the GMP library. The conversions to and from
5107 GMP's mpz_t format aren't optimised nor are any montgomery forms
5108 cached, and on x86 it appears OpenSSL's own performance has caught up.
5109 However there are likely to be other architectures where GMP could
5110 provide a boost. This ENGINE is not built in by default, but it can be
5111 specified at Configure time and should be accompanied by the necessary
5112 linker additions, eg;
5113 ./config -DOPENSSL_USE_GMP -lgmp
5116 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5117 testing availability of engines with "-t" - the old behaviour is
5118 produced by increasing the feature's verbosity with "-tt".
5121 *) ECDSA routines: under certain error conditions uninitialized BN objects
5122 could be freed. Solution: make sure initialization is performed early
5123 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5127 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5128 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5129 software implementations. For DSA and DH, parameter generation can
5130 also be overridden by providing the appropriate method callbacks.
5133 *) Change the "progress" mechanism used in key-generation and
5134 primality testing to functions that take a new BN_GENCB pointer in
5135 place of callback/argument pairs. The new API functions have "_ex"
5136 postfixes and the older functions are reimplemented as wrappers for
5137 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5138 declarations of the old functions to help (graceful) attempts to
5139 migrate to the new functions. Also, the new key-generation API
5140 functions operate on a caller-supplied key-structure and return
5141 success/failure rather than returning a key or NULL - this is to
5142 help make "keygen" another member function of RSA_METHOD etc.
5144 Example for using the new callback interface:
5146 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5150 BN_GENCB_set(&my_cb, my_callback, my_arg);
5152 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5153 /* For the meaning of a, b in calls to my_callback(), see the
5154 * documentation of the function that calls the callback.
5155 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5156 * my_callback should return 1 if it wants BN_is_prime_ex()
5157 * to continue, or 0 to stop.
5162 *) Change the ZLIB compression method to be stateful, and make it
5163 available to TLS with the number defined in
5164 draft-ietf-tls-compression-04.txt.
5167 *) Add the ASN.1 structures and functions for CertificatePair, which
5168 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5170 CertificatePair ::= SEQUENCE {
5171 forward [0] Certificate OPTIONAL,
5172 reverse [1] Certificate OPTIONAL,
5173 -- at least one of the pair shall be present -- }
5175 Also implement the PEM functions to read and write certificate
5176 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5178 This needed to be defined, mostly for the sake of the LDAP
5179 attribute crossCertificatePair, but may prove useful elsewhere as
5183 *) Make it possible to inhibit symlinking of shared libraries in
5184 Makefile.shared, for Cygwin's sake.
5187 *) Extend the BIGNUM API by creating a function
5188 void BN_set_negative(BIGNUM *a, int neg);
5189 and a macro that behave like
5190 int BN_is_negative(const BIGNUM *a);
5192 to avoid the need to access 'a->neg' directly in applications.
5195 *) Implement fast modular reduction for pseudo-Mersenne primes
5196 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5197 EC_GROUP_new_curve_GFp() will now automatically use this
5199 [Nils Larsch <nla@trustcenter.de>]
5201 *) Add new lock type (CRYPTO_LOCK_BN).
5204 *) Change the ENGINE framework to automatically load engines
5205 dynamically from specific directories unless they could be
5206 found to already be built in or loaded. Move all the
5207 current engines except for the cryptodev one to a new
5209 The engines in engines/ are built as shared libraries if
5210 the "shared" options was given to ./Configure or ./config.
5211 Otherwise, they are inserted in libcrypto.a.
5212 /usr/local/ssl/engines is the default directory for dynamic
5213 engines, but that can be overridden at configure time through
5214 the usual use of --prefix and/or --openssldir, and at run
5215 time with the environment variable OPENSSL_ENGINES.
5216 [Geoff Thorpe and Richard Levitte]
5218 *) Add Makefile.shared, a helper makefile to build shared
5219 libraries. Adapt Makefile.org.
5222 *) Add version info to Win32 DLLs.
5223 [Peter 'Luna' Runestig" <peter@runestig.com>]
5225 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5226 can be added using this API to created arbitrary PKCS#12
5227 files while avoiding the low level API.
5229 New options to PKCS12_create(), key or cert can be NULL and
5230 will then be omitted from the output file. The encryption
5231 algorithm NIDs can be set to -1 for no encryption, the mac
5232 iteration count can be set to 0 to omit the mac.
5234 Enhance pkcs12 utility by making the -nokeys and -nocerts
5235 options work when creating a PKCS#12 file. New option -nomac
5236 to omit the mac, NONE can be set for an encryption algorithm.
5237 New code is modified to use the enhanced PKCS12_create()
5238 instead of the low level API.
5241 *) Extend ASN1 encoder to support indefinite length constructed
5242 encoding. This can output sequences tags and octet strings in
5243 this form. Modify pk7_asn1.c to support indefinite length
5244 encoding. This is experimental and needs additional code to
5245 be useful, such as an ASN1 bio and some enhanced streaming
5248 Extend template encode functionality so that tagging is passed
5249 down to the template encoder.
5252 *) Let 'openssl req' fail if an argument to '-newkey' is not
5253 recognized instead of using RSA as a default.
5256 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5257 As these are not official, they are not included in "ALL";
5258 the "ECCdraft" ciphersuite group alias can be used to select them.
5259 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5261 *) Add ECDH engine support.
5262 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5264 *) Add ECDH in new directory crypto/ecdh/.
5265 [Douglas Stebila (Sun Microsystems Laboratories)]
5267 *) Let BN_rand_range() abort with an error after 100 iterations
5268 without success (which indicates a broken PRNG).
5271 *) Change BN_mod_sqrt() so that it verifies that the input value
5272 is really the square of the return value. (Previously,
5273 BN_mod_sqrt would show GIGO behaviour.)
5276 *) Add named elliptic curves over binary fields from X9.62, SECG,
5277 and WAP/WTLS; add OIDs that were still missing.
5279 [Sheueling Chang Shantz and Douglas Stebila
5280 (Sun Microsystems Laboratories)]
5282 *) Extend the EC library for elliptic curves over binary fields
5283 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5286 EC_GF2m_simple_method
5290 EC_GROUP_new_curve_GF2m
5291 EC_GROUP_set_curve_GF2m
5292 EC_GROUP_get_curve_GF2m
5293 EC_POINT_set_affine_coordinates_GF2m
5294 EC_POINT_get_affine_coordinates_GF2m
5295 EC_POINT_set_compressed_coordinates_GF2m
5297 Point compression for binary fields is disabled by default for
5298 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5301 As binary polynomials are represented as BIGNUMs, various members
5302 of the EC_GROUP and EC_POINT data structures can be shared
5303 between the implementations for prime fields and binary fields;
5304 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5305 are essentially identical to their ..._GFp counterparts.
5306 (For simplicity, the '..._GFp' prefix has been dropped from
5307 various internal method names.)
5309 An internal 'field_div' method (similar to 'field_mul' and
5310 'field_sqr') has been added; this is used only for binary fields.
5312 [Sheueling Chang Shantz and Douglas Stebila
5313 (Sun Microsystems Laboratories)]
5315 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5316 through methods ('mul', 'precompute_mult').
5318 The generic implementations (now internally called 'ec_wNAF_mul'
5319 and 'ec_wNAF_precomputed_mult') remain the default if these
5320 methods are undefined.
5322 [Sheueling Chang Shantz and Douglas Stebila
5323 (Sun Microsystems Laboratories)]
5325 *) New function EC_GROUP_get_degree, which is defined through
5326 EC_METHOD. For curves over prime fields, this returns the bit
5327 length of the modulus.
5329 [Sheueling Chang Shantz and Douglas Stebila
5330 (Sun Microsystems Laboratories)]
5332 *) New functions EC_GROUP_dup, EC_POINT_dup.
5333 (These simply call ..._new and ..._copy).
5335 [Sheueling Chang Shantz and Douglas Stebila
5336 (Sun Microsystems Laboratories)]
5338 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5339 Polynomials are represented as BIGNUMs (where the sign bit is not
5340 used) in the following functions [macros]:
5343 BN_GF2m_sub [= BN_GF2m_add]
5344 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5345 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5346 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5348 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5349 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5350 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5351 BN_GF2m_cmp [= BN_ucmp]
5353 (Note that only the 'mod' functions are actually for fields GF(2^m).
5354 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5356 For some functions, an the irreducible polynomial defining a
5357 field can be given as an 'unsigned int[]' with strictly
5358 decreasing elements giving the indices of those bits that are set;
5359 i.e., p[] represents the polynomial
5360 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5362 p[0] > p[1] > ... > p[k] = 0.
5363 This applies to the following functions:
5368 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5369 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5371 BN_GF2m_mod_sqrt_arr
5372 BN_GF2m_mod_solve_quad_arr
5376 Conversion can be performed by the following functions:
5381 bntest.c has additional tests for binary polynomial arithmetic.
5383 Two implementations for BN_GF2m_mod_div() are available.
5384 The default algorithm simply uses BN_GF2m_mod_inv() and
5385 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5386 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5387 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5389 [Sheueling Chang Shantz and Douglas Stebila
5390 (Sun Microsystems Laboratories)]
5392 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5393 functionality is disabled at compile-time.
5394 [Douglas Stebila <douglas.stebila@sun.com>]
5396 *) Change default behaviour of 'openssl asn1parse' so that more
5397 information is visible when viewing, e.g., a certificate:
5399 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5400 mode the content of non-printable OCTET STRINGs is output in a
5401 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5402 avoid the appearance of a printable string.
5403 [Nils Larsch <nla@trustcenter.de>]
5405 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5407 EC_GROUP_set_asn1_flag()
5408 EC_GROUP_get_asn1_flag()
5409 EC_GROUP_set_point_conversion_form()
5410 EC_GROUP_get_point_conversion_form()
5411 These control ASN1 encoding details:
5412 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5413 has been set to OPENSSL_EC_NAMED_CURVE.
5414 - Points are encoded in uncompressed form by default; options for
5415 asn1_for are as for point2oct, namely
5416 POINT_CONVERSION_COMPRESSED
5417 POINT_CONVERSION_UNCOMPRESSED
5418 POINT_CONVERSION_HYBRID
5420 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5423 EC_GROUP_get0_seed()
5424 EC_GROUP_get_seed_len()
5425 This is used only for ASN1 purposes (so far).
5426 [Nils Larsch <nla@trustcenter.de>]
5428 *) Add 'field_type' member to EC_METHOD, which holds the NID
5429 of the appropriate field type OID. The new function
5430 EC_METHOD_get_field_type() returns this value.
5431 [Nils Larsch <nla@trustcenter.de>]
5436 EC_POINT_point2hex()
5437 EC_POINT_hex2point()
5438 providing useful interfaces to EC_POINT_point2oct() and
5439 EC_POINT_oct2point().
5440 [Nils Larsch <nla@trustcenter.de>]
5442 *) Change internals of the EC library so that the functions
5443 EC_GROUP_set_generator()
5444 EC_GROUP_get_generator()
5445 EC_GROUP_get_order()
5446 EC_GROUP_get_cofactor()
5447 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5448 to methods, which would lead to unnecessary code duplication when
5449 adding different types of curves.
5450 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5452 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5453 arithmetic, and such that modified wNAFs are generated
5454 (which avoid length expansion in many cases).
5457 *) Add a function EC_GROUP_check_discriminant() (defined via
5458 EC_METHOD) that verifies that the curve discriminant is non-zero.
5460 Add a function EC_GROUP_check() that makes some sanity tests
5461 on a EC_GROUP, its generator and order. This includes
5462 EC_GROUP_check_discriminant().
5463 [Nils Larsch <nla@trustcenter.de>]
5465 *) Add ECDSA in new directory crypto/ecdsa/.
5467 Add applications 'openssl ecparam' and 'openssl ecdsa'
5468 (these are based on 'openssl dsaparam' and 'openssl dsa').
5470 ECDSA support is also included in various other files across the
5471 library. Most notably,
5472 - 'openssl req' now has a '-newkey ecdsa:file' option;
5473 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5474 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5475 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5476 them suitable for ECDSA where domain parameters must be
5477 extracted before the specific public key;
5478 - ECDSA engine support has been added.
5479 [Nils Larsch <nla@trustcenter.de>]
5481 *) Include some named elliptic curves, and add OIDs from X9.62,
5482 SECG, and WAP/WTLS. Each curve can be obtained from the new
5484 EC_GROUP_new_by_curve_name(),
5485 and the list of available named curves can be obtained with
5486 EC_get_builtin_curves().
5487 Also add a 'curve_name' member to EC_GROUP objects, which can be
5489 EC_GROUP_set_curve_name()
5490 EC_GROUP_get_curve_name()
5491 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5493 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5494 was actually never needed) and in BN_mul(). The removal in BN_mul()
5495 required a small change in bn_mul_part_recursive() and the addition
5496 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5497 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5498 bn_sub_words() and bn_add_words() except they take arrays with
5502 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5504 *) Cleanse PEM buffers before freeing them since they may contain
5506 [Benjamin Bennett <ben@psc.edu>]
5508 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5509 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5510 authentication-only ciphersuites.
5513 *) Since AES128 and AES256 share a single mask bit in the logic of
5514 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5515 kludge to work properly if AES128 is available and AES256 isn't.
5518 *) Expand security boundary to match 1.1.1 module.
5521 *) Remove redundant features: hash file source, editing of test vectors
5522 modify fipsld to use external fips_premain.c signature.
5525 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5526 run algorithm test programs.
5529 *) Make algorithm test programs more tolerant of whitespace.
5532 *) Have SSL/TLS server implementation tolerate "mismatched" record
5533 protocol version while receiving ClientHello even if the
5534 ClientHello is fragmented. (The server can't insist on the
5535 particular protocol version it has chosen before the ServerHello
5536 message has informed the client about his choice.)
5539 *) Load error codes if they are not already present instead of using a
5540 static variable. This allows them to be cleanly unloaded and reloaded.
5543 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5545 *) Introduce limits to prevent malicious keys being able to
5546 cause a denial of service. (CVE-2006-2940)
5547 [Steve Henson, Bodo Moeller]
5549 *) Fix ASN.1 parsing of certain invalid structures that can result
5550 in a denial of service. (CVE-2006-2937) [Steve Henson]
5552 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5553 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5555 *) Fix SSL client code which could crash if connecting to a
5556 malicious SSLv2 server. (CVE-2006-4343)
5557 [Tavis Ormandy and Will Drewry, Google Security Team]
5559 *) Change ciphersuite string processing so that an explicit
5560 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5561 will no longer include "AES128-SHA"), and any other similar
5562 ciphersuite (same bitmap) from *other* protocol versions (so that
5563 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5564 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5565 changes from 0.9.8b and 0.9.8d.
5568 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5570 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5571 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5573 *) Change the Unix randomness entropy gathering to use poll() when
5574 possible instead of select(), since the latter has some
5575 undesirable limitations.
5576 [Darryl Miles via Richard Levitte and Bodo Moeller]
5578 *) Disable rogue ciphersuites:
5580 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5581 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5582 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5584 The latter two were purportedly from
5585 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5588 Also deactive the remaining ciphersuites from
5589 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5590 unofficial, and the ID has long expired.
5593 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5594 dual-core machines) and other potential thread-safety issues.
5597 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5599 *) Adapt fipsld and the build system to link against the validated FIPS
5600 module in FIPS mode.
5603 *) Fixes for VC++ 2005 build under Windows.
5606 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5607 from a Windows bash shell such as MSYS. It is autodetected from the
5608 "config" script when run from a VC++ environment. Modify standard VC++
5609 build to use fipscanister.o from the GNU make build.
5612 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5614 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5615 The value now differs depending on if you build for FIPS or not.
5616 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5617 safely run with a non-FIPSed libcrypto, as it may crash because of
5618 the difference induced by this change.
5621 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5623 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5624 (part of SSL_OP_ALL). This option used to disable the
5625 countermeasure against man-in-the-middle protocol-version
5626 rollback in the SSL 2.0 server implementation, which is a bad
5627 idea. (CVE-2005-2969)
5629 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5630 for Information Security, National Institute of Advanced Industrial
5631 Science and Technology [AIST], Japan)]
5633 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5634 mainly for FIPS compliance and not fully integrated at this stage.
5637 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5638 the exponentiation using a fixed-length exponent. (Otherwise,
5639 the information leaked through timing could expose the secret key
5640 after many signatures; cf. Bleichenbacher's attack on DSA with
5644 *) Make a new fixed-window mod_exp implementation the default for
5645 RSA, DSA, and DH private-key operations so that the sequence of
5646 squares and multiplies and the memory access pattern are
5647 independent of the particular secret key. This will mitigate
5648 cache-timing and potential related attacks.
5650 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5651 and this is automatically used by BN_mod_exp_mont() if the new flag
5652 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5653 will use this BN flag for private exponents unless the flag
5654 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5655 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5657 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5659 *) Change the client implementation for SSLv23_method() and
5660 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5661 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5662 (Previously, the SSL 2.0 backwards compatible Client Hello
5663 message format would be used even with SSL_OP_NO_SSLv2.)
5666 *) Add support for smime-type MIME parameter in S/MIME messages which some
5670 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5671 a threadsafe manner. Modify rsa code to use new function and add calls
5672 to dsa and dh code (which had race conditions before).
5675 *) Include the fixed error library code in the C error file definitions
5676 instead of fixing them up at runtime. This keeps the error code
5677 structures constant.
5680 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5682 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5685 *) Fixes for newer kerberos headers. NB: the casts are needed because
5686 the 'length' field is signed on one version and unsigned on another
5687 with no (?) obvious way to tell the difference, without these VC++
5688 complains. Also the "definition" of FAR (blank) is no longer included
5689 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5690 some needed definitions.
5693 *) Undo Cygwin change.
5696 *) Added support for proxy certificates according to RFC 3820.
5697 Because they may be a security thread to unaware applications,
5698 they must be explicitly allowed in run-time. See
5699 docs/HOWTO/proxy_certificates.txt for further information.
5702 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5704 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5705 server and client random values. Previously
5706 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5707 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5709 This change has negligible security impact because:
5711 1. Server and client random values still have 24 bytes of pseudo random
5714 2. Server and client random values are sent in the clear in the initial
5717 3. The master secret is derived using the premaster secret (48 bytes in
5718 size for static RSA ciphersuites) as well as client server and random
5721 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5724 [Stephen Henson, reported by UK NISCC]
5726 *) Use Windows randomness collection on Cygwin.
5729 *) Fix hang in EGD/PRNGD query when communication socket is closed
5730 prematurely by EGD/PRNGD.
5731 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5733 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5736 *) Back-port of selected performance improvements from development
5737 branch, as well as improved support for PowerPC platforms.
5740 *) Add lots of checks for memory allocation failure, error codes to indicate
5741 failure and freeing up memory if a failure occurs.
5742 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5744 *) Add new -passin argument to dgst.
5747 *) Perform some character comparisons of different types in X509_NAME_cmp:
5748 this is needed for some certificates that re-encode DNs into UTF8Strings
5749 (in violation of RFC3280) and can't or won't issue name rollover
5753 *) Make an explicit check during certificate validation to see that
5754 the CA setting in each certificate on the chain is correct. As a
5755 side effect always do the following basic checks on extensions,
5756 not just when there's an associated purpose to the check:
5758 - if there is an unhandled critical extension (unless the user
5759 has chosen to ignore this fault)
5760 - if the path length has been exceeded (if one is set at all)
5761 - that certain extensions fit the associated purpose (if one has
5765 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5767 *) Avoid a race condition when CRLs are checked in a multi threaded
5768 environment. This would happen due to the reordering of the revoked
5769 entries during signature checking and serial number lookup. Now the
5770 encoding is cached and the serial number sort performed under a lock.
5771 Add new STACK function sk_is_sorted().
5774 *) Add Delta CRL to the extension code.
5777 *) Various fixes to s3_pkt.c so alerts are sent properly.
5778 [David Holmes <d.holmes@f5.com>]
5780 *) Reduce the chances of duplicate issuer name and serial numbers (in
5781 violation of RFC3280) using the OpenSSL certificate creation utilities.
5782 This is done by creating a random 64 bit value for the initial serial
5783 number when a serial number file is created or when a self signed
5784 certificate is created using 'openssl req -x509'. The initial serial
5785 number file is created using 'openssl x509 -next_serial' in CA.pl
5786 rather than being initialized to 1.
5789 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5791 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5792 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5793 [Joe Orton, Steve Henson]
5795 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5797 [Joe Orton, Steve Henson]
5799 *) Make it possible to have multiple active certificates with the same
5800 subject in the CA index file. This is done only if the keyword
5801 'unique_subject' is set to 'no' in the main CA section (default
5802 if 'CA_default') of the configuration file. The value is saved
5803 with the database itself in a separate index attribute file,
5804 named like the index file with '.attr' appended to the name.
5807 *) X509 verify fixes. Disable broken certificate workarounds when
5808 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5809 keyUsage extension present. Don't accept CRLs with unhandled critical
5810 extensions: since verify currently doesn't process CRL extensions this
5811 rejects a CRL with *any* critical extensions. Add new verify error codes
5815 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5816 A clarification of RFC2560 will require the use of OCTET STRINGs and
5817 some implementations cannot handle the current raw format. Since OpenSSL
5818 copies and compares OCSP nonces as opaque blobs without any attempt at
5819 parsing them this should not create any compatibility issues.
5822 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5823 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5824 this HMAC (and other) operations are several times slower than OpenSSL
5828 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5829 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5831 *) Use the correct content when signing type "other".
5834 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5836 *) Fix various bugs revealed by running the NISCC test suite:
5838 Stop out of bounds reads in the ASN1 code when presented with
5839 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5841 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5843 If verify callback ignores invalid public key errors don't try to check
5844 certificate signature with the NULL public key.
5848 *) New -ignore_err option in ocsp application to stop the server
5849 exiting on the first error in a request.
5852 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5853 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5857 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5858 extra data after the compression methods not only for TLS 1.0
5859 but also for SSL 3.0 (as required by the specification).
5860 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5862 *) Change X509_certificate_type() to mark the key as exported/exportable
5863 when it's 512 *bits* long, not 512 bytes.
5866 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5867 blocks during encryption.
5870 *) Various fixes to base64 BIO and non blocking I/O. On write
5871 flushes were not handled properly if the BIO retried. On read
5872 data was not being buffered properly and had various logic bugs.
5873 This also affects blocking I/O when the data being decoded is a
5877 *) Various S/MIME bugfixes and compatibility changes:
5878 output correct application/pkcs7 MIME type if
5879 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5880 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5881 of files as .eml work). Correctly handle very long lines in MIME
5885 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5887 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5888 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5889 a protocol version number mismatch like a decryption error
5890 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5893 *) Turn on RSA blinding by default in the default implementation
5894 to avoid a timing attack. Applications that don't want it can call
5895 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5896 They would be ill-advised to do so in most cases.
5897 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5899 *) Change RSA blinding code so that it works when the PRNG is not
5900 seeded (in this case, the secret RSA exponent is abused as
5901 an unpredictable seed -- if it is not unpredictable, there
5902 is no point in blinding anyway). Make RSA blinding thread-safe
5903 by remembering the creator's thread ID in rsa->blinding and
5904 having all other threads use local one-time blinding factors
5905 (this requires more computation than sharing rsa->blinding, but
5906 avoids excessive locking; and if an RSA object is not shared
5907 between threads, blinding will still be very fast).
5910 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5911 ENGINE as defaults for all supported algorithms irrespective of
5912 the 'flags' parameter. 'flags' is now honoured, so applications
5913 should make sure they are passing it correctly.
5916 *) Target "mingw" now allows native Windows code to be generated in
5917 the Cygwin environment as well as with the MinGW compiler.
5920 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5922 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5923 via timing by performing a MAC computation even if incorrect
5924 block cipher padding has been found. This is a countermeasure
5925 against active attacks where the attacker has to distinguish
5926 between bad padding and a MAC verification error. (CVE-2003-0078)
5928 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5929 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5930 Martin Vuagnoux (EPFL, Ilion)]
5932 *) Make the no-err option work as intended. The intention with no-err
5933 is not to have the whole error stack handling routines removed from
5934 libcrypto, it's only intended to remove all the function name and
5935 reason texts, thereby removing some of the footprint that may not
5936 be interesting if those errors aren't displayed anyway.
5938 NOTE: it's still possible for any application or module to have it's
5939 own set of error texts inserted. The routines are there, just not
5940 used by default when no-err is given.
5943 *) Add support for FreeBSD on IA64.
5944 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5946 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5947 Kerberos function mit_des_cbc_cksum(). Before this change,
5948 the value returned by DES_cbc_cksum() was like the one from
5949 mit_des_cbc_cksum(), except the bytes were swapped.
5950 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5952 *) Allow an application to disable the automatic SSL chain building.
5953 Before this a rather primitive chain build was always performed in
5954 ssl3_output_cert_chain(): an application had no way to send the
5955 correct chain if the automatic operation produced an incorrect result.
5957 Now the chain builder is disabled if either:
5959 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5961 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5963 The reasoning behind this is that an application would not want the
5964 auto chain building to take place if extra chain certificates are
5965 present and it might also want a means of sending no additional
5966 certificates (for example the chain has two certificates and the
5970 *) Add the possibility to build without the ENGINE framework.
5971 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5973 *) Under Win32 gmtime() can return NULL: check return value in
5974 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
5977 *) DSA routines: under certain error conditions uninitialized BN objects
5978 could be freed. Solution: make sure initialization is performed early
5979 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
5980 Nils Larsch <nla@trustcenter.de> via PR#459)
5983 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
5984 checked on reconnect on the client side, therefore session resumption
5985 could still fail with a "ssl session id is different" error. This
5986 behaviour is masked when SSL_OP_ALL is used due to
5987 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
5988 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
5989 followup to PR #377.
5992 *) IA-32 assembler support enhancements: unified ELF targets, support
5993 for SCO/Caldera platforms, fix for Cygwin shared build.
5996 *) Add support for FreeBSD on sparc64. As a consequence, support for
5997 FreeBSD on non-x86 processors is separate from x86 processors on
5998 the config script, much like the NetBSD support.
5999 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6001 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6003 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6006 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6007 code (06) was taken as the first octet of the session ID and the last
6008 octet was ignored consequently. As a result SSLv2 client side session
6009 caching could not have worked due to the session ID mismatch between
6011 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6015 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6016 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6020 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6021 seems that in spite of existing for more than a year, many application
6022 author have done nothing to provide the necessary callbacks, which
6023 means that this particular engine will not work properly anywhere.
6024 This is a very unfortunate situation which forces us, in the name
6025 of usability, to give the hw_ncipher.c a static lock, which is part
6027 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6028 appear in 0.9.8 or later. We EXPECT application authors to have
6029 dealt properly with this when 0.9.8 is released (unless we actually
6030 make such changes in the libcrypto locking code that changes will
6031 have to be made anyway).
6034 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6035 octets have been read, EOF or an error occurs. Without this change
6036 some truncated ASN1 structures will not produce an error.
6039 *) Disable Heimdal support, since it hasn't been fully implemented.
6040 Still give the possibility to force the use of Heimdal, but with
6041 warnings and a request that patches get sent to openssl-dev.
6044 *) Add the VC-CE target, introduce the WINCE sysname, and add
6045 INSTALL.WCE and appropriate conditionals to make it build.
6046 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6048 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6049 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6050 edit numbers of the version.
6051 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6053 *) Introduce safe string copy and catenation functions
6054 (BUF_strlcpy() and BUF_strlcat()).
6055 [Ben Laurie (CHATS) and Richard Levitte]
6057 *) Avoid using fixed-size buffers for one-line DNs.
6058 [Ben Laurie (CHATS)]
6060 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6061 resizing buffers containing secrets, and use where appropriate.
6062 [Ben Laurie (CHATS)]
6064 *) Avoid using fixed size buffers for configuration file location.
6065 [Ben Laurie (CHATS)]
6067 *) Avoid filename truncation for various CA files.
6068 [Ben Laurie (CHATS)]
6070 *) Use sizeof in preference to magic numbers.
6071 [Ben Laurie (CHATS)]
6073 *) Avoid filename truncation in cert requests.
6074 [Ben Laurie (CHATS)]
6076 *) Add assertions to check for (supposedly impossible) buffer
6078 [Ben Laurie (CHATS)]
6080 *) Don't cache truncated DNS entries in the local cache (this could
6081 potentially lead to a spoofing attack).
6082 [Ben Laurie (CHATS)]
6084 *) Fix various buffers to be large enough for hex/decimal
6085 representations in a platform independent manner.
6086 [Ben Laurie (CHATS)]
6088 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6089 resizing buffers containing secrets, and use where appropriate.
6090 [Ben Laurie (CHATS)]
6092 *) Add BIO_indent() to avoid much slightly worrying code to do
6094 [Ben Laurie (CHATS)]
6096 *) Convert sprintf()/BIO_puts() to BIO_printf().
6097 [Ben Laurie (CHATS)]
6099 *) buffer_gets() could terminate with the buffer only half
6101 [Ben Laurie (CHATS)]
6103 *) Add assertions to prevent user-supplied crypto functions from
6104 overflowing internal buffers by having large block sizes, etc.
6105 [Ben Laurie (CHATS)]
6107 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6109 [Ben Laurie (CHATS)]
6111 *) Eliminate unused copy of key in RC4.
6112 [Ben Laurie (CHATS)]
6114 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6115 [Ben Laurie (CHATS)]
6117 *) Fix off-by-one error in EGD path.
6118 [Ben Laurie (CHATS)]
6120 *) If RANDFILE path is too long, ignore instead of truncating.
6121 [Ben Laurie (CHATS)]
6123 *) Eliminate unused and incorrectly sized X.509 structure
6125 [Ben Laurie (CHATS)]
6127 *) Eliminate unused and dangerous function knumber().
6128 [Ben Laurie (CHATS)]
6130 *) Eliminate unused and dangerous structure, KSSL_ERR.
6131 [Ben Laurie (CHATS)]
6133 *) Protect against overlong session ID context length in an encoded
6134 session object. Since these are local, this does not appear to be
6136 [Ben Laurie (CHATS)]
6138 *) Change from security patch (see 0.9.6e below) that did not affect
6139 the 0.9.6 release series:
6141 Remote buffer overflow in SSL3 protocol - an attacker could
6142 supply an oversized master key in Kerberos-enabled versions.
6144 [Ben Laurie (CHATS)]
6146 *) Change the SSL kerb5 codes to match RFC 2712.
6149 *) Make -nameopt work fully for req and add -reqopt switch.
6150 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6152 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6153 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6155 *) Make sure tests can be performed even if the corresponding algorithms
6156 have been removed entirely. This was also the last step to make
6157 OpenSSL compilable with DJGPP under all reasonable conditions.
6158 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6160 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6161 to allow version independent disabling of normally unselected ciphers,
6162 which may be activated as a side-effect of selecting a single cipher.
6164 (E.g., cipher list string "RSA" enables ciphersuites that are left
6165 out of "ALL" because they do not provide symmetric encryption.
6166 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6167 [Lutz Jaenicke, Bodo Moeller]
6169 *) Add appropriate support for separate platform-dependent build
6170 directories. The recommended way to make a platform-dependent
6171 build directory is the following (tested on Linux), maybe with
6174 # Place yourself outside of the OpenSSL source tree. In
6175 # this example, the environment variable OPENSSL_SOURCE
6176 # is assumed to contain the absolute OpenSSL source directory.
6177 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6178 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6179 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6180 mkdir -p `dirname $F`
6181 ln -s $OPENSSL_SOURCE/$F $F
6184 To be absolutely sure not to disturb the source tree, a "make clean"
6185 is a good thing. If it isn't successful, don't worry about it,
6186 it probably means the source directory is very clean.
6189 *) Make sure any ENGINE control commands make local copies of string
6190 pointers passed to them whenever necessary. Otherwise it is possible
6191 the caller may have overwritten (or deallocated) the original string
6192 data when a later ENGINE operation tries to use the stored values.
6193 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6195 *) Improve diagnostics in file reading and command-line digests.
6196 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6198 *) Add AES modes CFB and OFB to the object database. Correct an
6199 error in AES-CFB decryption.
6202 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6203 allows existing EVP_CIPHER_CTX structures to be reused after
6204 calling EVP_*Final(). This behaviour is used by encryption
6205 BIOs and some applications. This has the side effect that
6206 applications must explicitly clean up cipher contexts with
6207 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6210 *) Check the values of dna and dnb in bn_mul_recursive before calling
6211 bn_mul_comba (a non zero value means the a or b arrays do not contain
6212 n2 elements) and fallback to bn_mul_normal if either is not zero.
6215 *) Fix escaping of non-ASCII characters when using the -subj option
6216 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6219 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6220 form for "surname", serialNumber has no short form.
6221 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6222 therefore remove "mail" short name for "internet 7".
6223 The OID for unique identifiers in X509 certificates is
6224 x500UniqueIdentifier, not uniqueIdentifier.
6225 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6228 *) Add an "init" command to the ENGINE config module and auto initialize
6229 ENGINEs. Without any "init" command the ENGINE will be initialized
6230 after all ctrl commands have been executed on it. If init=1 the
6231 ENGINE is initailized at that point (ctrls before that point are run
6232 on the uninitialized ENGINE and after on the initialized one). If
6233 init=0 then the ENGINE will not be iniatialized at all.
6236 *) Fix the 'app_verify_callback' interface so that the user-defined
6237 argument is actually passed to the callback: In the
6238 SSL_CTX_set_cert_verify_callback() prototype, the callback
6239 declaration has been changed from
6242 int (*cb)(X509_STORE_CTX *,void *);
6243 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6244 i=s->ctx->app_verify_callback(&ctx)
6245 has been changed into
6246 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6248 To update applications using SSL_CTX_set_cert_verify_callback(),
6249 a dummy argument can be added to their callback functions.
6250 [D. K. Smetters <smetters@parc.xerox.com>]
6252 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6253 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6255 *) Add and OPENSSL_LOAD_CONF define which will cause
6256 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6257 This allows older applications to transparently support certain
6258 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6259 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6260 load the config file and OPENSSL_add_all_algorithms_conf() which will
6261 always load it have also been added.
6264 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6265 Adjust NIDs and EVP layer.
6266 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6268 *) Config modules support in openssl utility.
6270 Most commands now load modules from the config file,
6271 though in a few (such as version) this isn't done
6272 because it couldn't be used for anything.
6274 In the case of ca and req the config file used is
6275 the same as the utility itself: that is the -config
6276 command line option can be used to specify an
6280 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6281 use "openssl_conf" if filename is NULL use default openssl config file.
6284 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6285 config section name. Add a new flag to tolerate a missing config file
6286 and move code to CONF_modules_load_file().
6289 *) Support for crypto accelerator cards from Accelerated Encryption
6290 Processing, www.aep.ie. (Use engine 'aep')
6291 The support was copied from 0.9.6c [engine] and adapted/corrected
6292 to work with the new engine framework.
6293 [AEP Inc. and Richard Levitte]
6295 *) Support for SureWare crypto accelerator cards from Baltimore
6296 Technologies. (Use engine 'sureware')
6297 The support was copied from 0.9.6c [engine] and adapted
6298 to work with the new engine framework.
6301 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6302 make the newer ENGINE framework commands for the CHIL engine work.
6303 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6305 *) Make it possible to produce shared libraries on ReliantUNIX.
6306 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6308 *) Add the configuration target debug-linux-ppro.
6309 Make 'openssl rsa' use the general key loading routines
6310 implemented in apps.c, and make those routines able to
6311 handle the key format FORMAT_NETSCAPE and the variant
6313 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6315 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6316 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6318 *) Add -keyform to rsautl, and document -engine.
6319 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6321 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6322 BIO_R_NO_SUCH_FILE error code rather than the generic
6323 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6326 *) Add new functions
6328 ERR_peek_last_error_line
6329 ERR_peek_last_error_line_data.
6330 These are similar to
6333 ERR_peek_error_line_data,
6334 but report on the latest error recorded rather than the first one
6335 still in the error queue.
6336 [Ben Laurie, Bodo Moeller]
6338 *) default_algorithms option in ENGINE config module. This allows things
6340 default_algorithms = ALL
6341 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6344 *) Preliminary ENGINE config module.
6347 *) New experimental application configuration code.
6350 *) Change the AES code to follow the same name structure as all other
6351 symmetric ciphers, and behave the same way. Move everything to
6352 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6353 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6355 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6356 [Ben Laurie and Theo de Raadt]
6358 *) Add option to output public keys in req command.
6359 [Massimiliano Pala madwolf@openca.org]
6361 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6362 (up to about 10% better than before for P-192 and P-224).
6365 *) New functions/macros
6367 SSL_CTX_set_msg_callback(ctx, cb)
6368 SSL_CTX_set_msg_callback_arg(ctx, arg)
6369 SSL_set_msg_callback(ssl, cb)
6370 SSL_set_msg_callback_arg(ssl, arg)
6372 to request calling a callback function
6374 void cb(int write_p, int version, int content_type,
6375 const void *buf, size_t len, SSL *ssl, void *arg)
6377 whenever a protocol message has been completely received
6378 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6379 protocol version according to which the SSL library interprets
6380 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6381 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6382 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6383 specification (change_cipher_spec(20), alert(21), handshake(22)).
6384 'buf' and 'len' point to the actual message, 'ssl' to the
6385 SSL object, and 'arg' is the application-defined value set by
6386 SSL[_CTX]_set_msg_callback_arg().
6388 'openssl s_client' and 'openssl s_server' have new '-msg' options
6389 to enable a callback that displays all protocol messages.
6392 *) Change the shared library support so shared libraries are built as
6393 soon as the corresponding static library is finished, and thereby get
6394 openssl and the test programs linked against the shared library.
6395 This still only happens when the keyword "shard" has been given to
6396 the configuration scripts.
6398 NOTE: shared library support is still an experimental thing, and
6399 backward binary compatibility is still not guaranteed.
6400 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6402 *) Add support for Subject Information Access extension.
6403 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6405 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6406 additional bytes when new memory had to be allocated, not just
6407 when reusing an existing buffer.
6410 *) New command line and configuration option 'utf8' for the req command.
6411 This allows field values to be specified as UTF8 strings.
6414 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6415 runs for the former and machine-readable output for the latter.
6418 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6419 of the e-mail address in the DN (i.e., it will go into a certificate
6420 extension only). The new configuration file option 'email_in_dn = no'
6421 has the same effect.
6422 [Massimiliano Pala madwolf@openca.org]
6424 *) Change all functions with names starting with des_ to be starting
6425 with DES_ instead. Add wrappers that are compatible with libdes,
6426 but are named _ossl_old_des_*. Finally, add macros that map the
6427 des_* symbols to the corresponding _ossl_old_des_* if libdes
6428 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6429 desired, the des_* symbols will be mapped to DES_*, with one
6432 Since we provide two compatibility mappings, the user needs to
6433 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6434 compatibility is desired. The default (i.e., when that macro
6435 isn't defined) is OpenSSL 0.9.6c compatibility.
6437 There are also macros that enable and disable the support of old
6438 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6439 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6440 are defined, the default will apply: to support the old des routines.
6442 In either case, one must include openssl/des.h to get the correct
6443 definitions. Do not try to just include openssl/des_old.h, that
6446 NOTE: This is a major break of an old API into a new one. Software
6447 authors are encouraged to switch to the DES_ style functions. Some
6448 time in the future, des_old.h and the libdes compatibility functions
6449 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6450 default), and then completely removed.
6453 *) Test for certificates which contain unsupported critical extensions.
6454 If such a certificate is found during a verify operation it is
6455 rejected by default: this behaviour can be overridden by either
6456 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6457 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6458 X509_supported_extension() has also been added which returns 1 if a
6459 particular extension is supported.
6462 *) Modify the behaviour of EVP cipher functions in similar way to digests
6463 to retain compatibility with existing code.
6466 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6467 compatibility with existing code. In particular the 'ctx' parameter does
6468 not have to be to be initialized before the call to EVP_DigestInit() and
6469 it is tidied up after a call to EVP_DigestFinal(). New function
6470 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6471 EVP_MD_CTX_copy() changed to not require the destination to be
6472 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6473 requires the destination to be valid.
6475 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6476 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6479 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6480 so that complete 'Handshake' protocol structures are kept in memory
6481 instead of overwriting 'msg_type' and 'length' with 'body' data.
6484 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6485 [Massimo Santin via Richard Levitte]
6487 *) Major restructuring to the underlying ENGINE code. This includes
6488 reduction of linker bloat, separation of pure "ENGINE" manipulation
6489 (initialisation, etc) from functionality dealing with implementations
6490 of specific crypto iterfaces. This change also introduces integrated
6491 support for symmetric ciphers and digest implementations - so ENGINEs
6492 can now accelerate these by providing EVP_CIPHER and EVP_MD
6493 implementations of their own. This is detailed in crypto/engine/README
6494 as it couldn't be adequately described here. However, there are a few
6495 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6496 were changed in the original introduction of ENGINE code have now
6497 reverted back - the hooking from this code to ENGINE is now a good
6498 deal more passive and at run-time, operations deal directly with
6499 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6500 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6501 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6502 they were not being used by the framework as there is no concept of a
6503 BIGNUM_METHOD and they could not be generalised to the new
6504 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6505 ENGINE_cpy() has been removed as it cannot be consistently defined in
6509 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6512 *) Change mkdef.pl to sort symbols that get the same entry number,
6513 and make sure the automatically generated functions ERR_load_*
6514 become part of libeay.num as well.
6517 *) New function SSL_renegotiate_pending(). This returns true once
6518 renegotiation has been requested (either SSL_renegotiate() call
6519 or HelloRequest/ClientHello received from the peer) and becomes
6520 false once a handshake has been completed.
6521 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6522 sends a HelloRequest, but does not ensure that a handshake takes
6523 place. SSL_renegotiate_pending() is useful for checking if the
6524 client has followed the request.)
6527 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6528 By default, clients may request session resumption even during
6529 renegotiation (if session ID contexts permit); with this option,
6530 session resumption is possible only in the first handshake.
6532 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6533 more bits available for options that should not be part of
6534 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6537 *) Add some demos for certificate and certificate request creation.
6540 *) Make maximum certificate chain size accepted from the peer application
6541 settable (SSL*_get/set_max_cert_list()), as proposed by
6542 "Douglas E. Engert" <deengert@anl.gov>.
6545 *) Add support for shared libraries for Unixware-7
6546 (Boyd Lynn Gerber <gerberb@zenez.com>).
6549 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6550 be done prior to destruction. Use this to unload error strings from
6551 ENGINEs that load their own error strings. NB: This adds two new API
6552 functions to "get" and "set" this destroy handler in an ENGINE.
6555 *) Alter all existing ENGINE implementations (except "openssl" and
6556 "openbsd") to dynamically instantiate their own error strings. This
6557 makes them more flexible to be built both as statically-linked ENGINEs
6558 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6559 Also, add stub code to each that makes building them as self-contained
6560 shared-libraries easier (see README.ENGINE).
6563 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6564 implementations into applications that are completely implemented in
6565 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6566 commands that can be used to configure what shared-library to load and
6567 to control aspects of the way it is handled. Also, made an update to
6568 the README.ENGINE file that brings its information up-to-date and
6569 provides some information and instructions on the "dynamic" ENGINE
6570 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6573 *) Make it possible to unload ranges of ERR strings with a new
6574 "ERR_unload_strings" function.
6577 *) Add a copy() function to EVP_MD.
6580 *) Make EVP_MD routines take a context pointer instead of just the
6581 md_data void pointer.
6584 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6585 that the digest can only process a single chunk of data
6586 (typically because it is provided by a piece of
6587 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6588 is only going to provide a single chunk of data, and hence the
6589 framework needn't accumulate the data for oneshot drivers.
6592 *) As with "ERR", make it possible to replace the underlying "ex_data"
6593 functions. This change also alters the storage and management of global
6594 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6595 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6596 index counters. The API functions that use this state have been changed
6597 to take a "class_index" rather than pointers to the class's local STACK
6598 and counter, and there is now an API function to dynamically create new
6599 classes. This centralisation allows us to (a) plug a lot of the
6600 thread-safety problems that existed, and (b) makes it possible to clean
6601 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6602 such data would previously have always leaked in application code and
6603 workarounds were in place to make the memory debugging turn a blind eye
6604 to it. Application code that doesn't use this new function will still
6605 leak as before, but their memory debugging output will announce it now
6606 rather than letting it slide.
6608 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6609 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6610 has a return value to indicate success or failure.
6613 *) Make it possible to replace the underlying "ERR" functions such that the
6614 global state (2 LHASH tables and 2 locks) is only used by the "default"
6615 implementation. This change also adds two functions to "get" and "set"
6616 the implementation prior to it being automatically set the first time
6617 any other ERR function takes place. Ie. an application can call "get",
6618 pass the return value to a module it has just loaded, and that module
6619 can call its own "set" function using that value. This means the
6620 module's "ERR" operations will use (and modify) the error state in the
6621 application and not in its own statically linked copy of OpenSSL code.
6624 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6625 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6626 the operation, and provides a more encapsulated way for external code
6627 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6628 to use these functions rather than manually incrementing the counts.
6630 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6633 *) Add EVP test program.
6636 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6639 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6640 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6641 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6642 These allow a CRL to be built without having to access X509_CRL fields
6643 directly. Modify 'ca' application to use new functions.
6646 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6647 bug workarounds. Rollback attack detection is a security feature.
6648 The problem will only arise on OpenSSL servers when TLSv1 is not
6649 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6650 Software authors not wanting to support TLSv1 will have special reasons
6651 for their choice and can explicitly enable this option.
6652 [Bodo Moeller, Lutz Jaenicke]
6654 *) Rationalise EVP so it can be extended: don't include a union of
6655 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6656 (similar to those existing for EVP_CIPHER_CTX).
6661 EVP_MD_CTX_init(&md); /* new function call */
6662 EVP_DigestInit(&md, EVP_sha1());
6663 EVP_DigestUpdate(&md, in, len);
6664 EVP_DigestFinal(&md, out, NULL);
6665 EVP_MD_CTX_cleanup(&md); /* new function call */
6669 *) Make DES key schedule conform to the usual scheme, as well as
6670 correcting its structure. This means that calls to DES functions
6671 now have to pass a pointer to a des_key_schedule instead of a
6672 plain des_key_schedule (which was actually always a pointer
6675 des_key_schedule ks;
6677 des_set_key_checked(..., &ks);
6678 des_ncbc_encrypt(..., &ks, ...);
6680 (Note that a later change renames 'des_...' into 'DES_...'.)
6683 *) Initial reduction of linker bloat: the use of some functions, such as
6684 PEM causes large amounts of unused functions to be linked in due to
6685 poor organisation. For example pem_all.c contains every PEM function
6686 which has a knock on effect of linking in large amounts of (unused)
6687 ASN1 code. Grouping together similar functions and splitting unrelated
6688 functions prevents this.
6691 *) Cleanup of EVP macros.
6694 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6695 correct _ecb suffix.
6698 *) Add initial OCSP responder support to ocsp application. The
6699 revocation information is handled using the text based index
6700 use by the ca application. The responder can either handle
6701 requests generated internally, supplied in files (for example
6702 via a CGI script) or using an internal minimal server.
6705 *) Add configuration choices to get zlib compression for TLS.
6708 *) Changes to Kerberos SSL for RFC 2712 compliance:
6709 1. Implemented real KerberosWrapper, instead of just using
6710 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6711 2. Implemented optional authenticator field of KerberosWrapper.
6713 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6714 and authenticator structs; see crypto/krb5/.
6716 Generalized Kerberos calls to support multiple Kerberos libraries.
6717 [Vern Staats <staatsvr@asc.hpc.mil>,
6718 Jeffrey Altman <jaltman@columbia.edu>
6719 via Richard Levitte]
6721 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6722 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6723 values for each of the key sizes rather than having just
6724 parameters (and 'speed' generating keys each time).
6727 *) Speed up EVP routines.
6730 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6731 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6732 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6733 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6735 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6736 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6737 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6740 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6742 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6745 *) Added the OS2-EMX target.
6746 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6748 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6749 to support NCONF routines in extension code. New function CONF_set_nconf()
6750 to allow functions which take an NCONF to also handle the old LHASH
6751 structure: this means that the old CONF compatible routines can be
6752 retained (in particular wrt extensions) without having to duplicate the
6753 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6756 *) Enhance the general user interface with mechanisms for inner control
6757 and with possibilities to have yes/no kind of prompts.
6760 *) Change all calls to low level digest routines in the library and
6761 applications to use EVP. Add missing calls to HMAC_cleanup() and
6762 don't assume HMAC_CTX can be copied using memcpy().
6763 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6765 *) Add the possibility to control engines through control names but with
6766 arbitrary arguments instead of just a string.
6767 Change the key loaders to take a UI_METHOD instead of a callback
6768 function pointer. NOTE: this breaks binary compatibility with earlier
6769 versions of OpenSSL [engine].
6770 Adapt the nCipher code for these new conditions and add a card insertion
6774 *) Enhance the general user interface with mechanisms to better support
6775 dialog box interfaces, application-defined prompts, the possibility
6776 to use defaults (for example default passwords from somewhere else)
6777 and interrupts/cancellations.
6780 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6781 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6784 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6785 tidy up some unnecessarily weird code in 'sk_new()').
6786 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6788 *) Change the key loading routines for ENGINEs to use the same kind
6789 callback (pem_password_cb) as all other routines that need this
6793 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6794 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6795 than this minimum value is recommended.
6798 *) New random seeder for OpenVMS, using the system process statistics
6799 that are easily reachable.
6802 *) Windows apparently can't transparently handle global
6803 variables defined in DLLs. Initialisations such as:
6805 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6807 won't compile. This is used by the any applications that need to
6808 declare their own ASN1 modules. This was fixed by adding the option
6809 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6810 needed for static libraries under Win32.
6813 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6814 setting of purpose and trust fields. New X509_STORE trust and
6815 purpose functions and tidy up setting in other SSL functions.
6818 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6819 structure. These are inherited by X509_STORE_CTX when it is
6820 initialised. This allows various defaults to be set in the
6821 X509_STORE structure (such as flags for CRL checking and custom
6822 purpose or trust settings) for functions which only use X509_STORE_CTX
6823 internally such as S/MIME.
6825 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6826 trust settings if they are not set in X509_STORE. This allows X509_STORE
6827 purposes and trust (in S/MIME for example) to override any set by default.
6829 Add command line options for CRL checking to smime, s_client and s_server
6833 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6834 are set then the CRL is looked up in the X509_STORE structure and
6835 its validity and signature checked, then if the certificate is found
6836 in the CRL the verify fails with a revoked error.
6838 Various new CRL related callbacks added to X509_STORE_CTX structure.
6840 Command line options added to 'verify' application to support this.
6842 This needs some additional work, such as being able to handle multiple
6843 CRLs with different times, extension based lookup (rather than just
6844 by subject name) and ultimately more complete V2 CRL extension
6848 *) Add a general user interface API (crypto/ui/). This is designed
6849 to replace things like des_read_password and friends (backward
6850 compatibility functions using this new API are provided).
6851 The purpose is to remove prompting functions from the DES code
6852 section as well as provide for prompting through dialog boxes in
6853 a window system and the like.
6856 *) Add "ex_data" support to ENGINE so implementations can add state at a
6857 per-structure level rather than having to store it globally.
6860 *) Make it possible for ENGINE structures to be copied when retrieved by
6861 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6862 This causes the "original" ENGINE structure to act like a template,
6863 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6864 operational state can be localised to each ENGINE structure, despite the
6865 fact they all share the same "methods". New ENGINE structures returned in
6866 this case have no functional references and the return value is the single
6867 structural reference. This matches the single structural reference returned
6868 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6872 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6873 needs to match any other type at all we need to manually clear the
6877 *) Changes to the "openssl engine" utility to include;
6878 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6879 about an ENGINE's available control commands.
6880 - executing control commands from command line arguments using the
6881 '-pre' and '-post' switches. '-post' is only used if '-t' is
6882 specified and the ENGINE is successfully initialised. The syntax for
6883 the individual commands are colon-separated, for example;
6884 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6887 *) New dynamic control command support for ENGINEs. ENGINEs can now
6888 declare their own commands (numbers), names (strings), descriptions,
6889 and input types for run-time discovery by calling applications. A
6890 subset of these commands are implicitly classed as "executable"
6891 depending on their input type, and only these can be invoked through
6892 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6893 can be based on user input, config files, etc). The distinction is
6894 that "executable" commands cannot return anything other than a boolean
6895 result and can only support numeric or string input, whereas some
6896 discoverable commands may only be for direct use through
6897 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6898 pointers, or other custom uses. The "executable" commands are to
6899 support parameterisations of ENGINE behaviour that can be
6900 unambiguously defined by ENGINEs and used consistently across any
6901 OpenSSL-based application. Commands have been added to all the
6902 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6903 control over shared-library paths without source code alterations.
6906 *) Changed all ENGINE implementations to dynamically allocate their
6907 ENGINEs rather than declaring them statically. Apart from this being
6908 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6909 this also allows the implementations to compile without using the
6910 internal engine_int.h header.
6913 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6914 'const' value. Any code that should be able to modify a RAND_METHOD
6915 should already have non-const pointers to it (ie. they should only
6916 modify their own ones).
6919 *) Made a variety of little tweaks to the ENGINE code.
6920 - "atalla" and "ubsec" string definitions were moved from header files
6921 to C code. "nuron" string definitions were placed in variables
6922 rather than hard-coded - allowing parameterisation of these values
6923 later on via ctrl() commands.
6924 - Removed unused "#if 0"'d code.
6925 - Fixed engine list iteration code so it uses ENGINE_free() to release
6926 structural references.
6927 - Constified the RAND_METHOD element of ENGINE structures.
6928 - Constified various get/set functions as appropriate and added
6929 missing functions (including a catch-all ENGINE_cpy that duplicates
6930 all ENGINE values onto a new ENGINE except reference counts/state).
6931 - Removed NULL parameter checks in get/set functions. Setting a method
6932 or function to NULL is a way of cancelling out a previously set
6933 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6934 and doesn't justify the extra error symbols and code.
6935 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6936 flags from engine_int.h to engine.h.
6937 - Changed prototypes for ENGINE handler functions (init(), finish(),
6938 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6941 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6942 to the algorithm using long division. The binary algorithm can be
6943 used only if the modulus is odd. On 32-bit systems, it is faster
6944 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6945 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6946 up to 450 bits. In 64-bit environments, the binary algorithm
6947 appears to be advantageous for much longer moduli; here we use it
6948 for moduli up to 2048 bits.
6951 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6952 could not support the combine flag in choice fields.
6955 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6956 extensions from a certificate request to the certificate.
6959 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6960 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6961 file: this allows the display of the certificate about to be
6962 signed to be customised, to allow certain fields to be included
6963 or excluded and extension details. The old system didn't display
6964 multicharacter strings properly, omitted fields not in the policy
6965 and couldn't display additional details such as extensions.
6968 *) Function EC_POINTs_mul for multiple scalar multiplication
6969 of an arbitrary number of elliptic curve points
6970 \sum scalars[i]*points[i],
6971 optionally including the generator defined for the EC_GROUP:
6972 scalar*generator + \sum scalars[i]*points[i].
6974 EC_POINT_mul is a simple wrapper function for the typical case
6975 that the point list has just one item (besides the optional
6979 *) First EC_METHODs for curves over GF(p):
6981 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
6982 operations and provides various method functions that can also
6983 operate with faster implementations of modular arithmetic.
6985 EC_GFp_mont_method() reuses most functions that are part of
6986 EC_GFp_simple_method, but uses Montgomery arithmetic.
6988 [Bodo Moeller; point addition and point doubling
6989 implementation directly derived from source code provided by
6990 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
6992 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
6993 crypto/ec/ec_lib.c):
6995 Curves are EC_GROUP objects (with an optional group generator)
6996 based on EC_METHODs that are built into the library.
6998 Points are EC_POINT objects based on EC_GROUP objects.
7000 Most of the framework would be able to handle curves over arbitrary
7001 finite fields, but as there are no obvious types for fields other
7002 than GF(p), some functions are limited to that for now.
7005 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7006 that the file contains a complete HTTP response.
7009 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7010 change the def and num file printf format specifier from "%-40sXXX"
7011 to "%-39s XXX". The latter will always guarantee a space after the
7012 field while the former will cause them to run together if the field
7013 is 40 of more characters long.
7016 *) Constify the cipher and digest 'method' functions and structures
7017 and modify related functions to take constant EVP_MD and EVP_CIPHER
7021 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7022 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7025 *) Modify EVP_Digest*() routines so they now return values. Although the
7026 internal software routines can never fail additional hardware versions
7030 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7032 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7033 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7036 ERR_R_NESTED_ASN1_ERROR
7038 ERR_R_MISSING_ASN1_EOS
7039 were 4 .. 9, conflicting with
7040 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7042 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7043 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7045 Add new error code 'ERR_R_INTERNAL_ERROR'.
7048 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7052 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7053 sets the subject name for a new request or supersedes the
7054 subject name in a given request. Formats that can be parsed are
7055 'CN=Some Name, OU=myOU, C=IT'
7057 'CN=Some Name/OU=myOU/C=IT'.
7059 Add options '-batch' and '-verbose' to 'openssl req'.
7060 [Massimiliano Pala <madwolf@hackmasters.net>]
7062 *) Introduce the possibility to access global variables through
7063 functions on platform were that's the best way to handle exporting
7064 global variables in shared libraries. To enable this functionality,
7065 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7066 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7067 is normally done by Configure or something similar).
7069 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7070 in the source file (foo.c) like this:
7072 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7073 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7075 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7076 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7078 OPENSSL_DECLARE_GLOBAL(int,foo);
7079 #define foo OPENSSL_GLOBAL_REF(foo)
7080 OPENSSL_DECLARE_GLOBAL(double,bar);
7081 #define bar OPENSSL_GLOBAL_REF(bar)
7083 The #defines are very important, and therefore so is including the
7084 header file everywhere where the defined globals are used.
7086 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7087 of ASN.1 items, but that structure is a bit different.
7089 The largest change is in util/mkdef.pl which has been enhanced with
7090 better and easier to understand logic to choose which symbols should
7091 go into the Windows .def files as well as a number of fixes and code
7092 cleanup (among others, algorithm keywords are now sorted
7093 lexicographically to avoid constant rewrites).
7096 *) In BN_div() keep a copy of the sign of 'num' before writing the
7097 result to 'rm' because if rm==num the value will be overwritten
7098 and produce the wrong result if 'num' is negative: this caused
7099 problems with BN_mod() and BN_nnmod().
7102 *) Function OCSP_request_verify(). This checks the signature on an
7103 OCSP request and verifies the signer certificate. The signer
7104 certificate is just checked for a generic purpose and OCSP request
7108 *) Add OCSP_check_validity() function to check the validity of OCSP
7109 responses. OCSP responses are prepared in real time and may only
7110 be a few seconds old. Simply checking that the current time lies
7111 between thisUpdate and nextUpdate max reject otherwise valid responses
7112 caused by either OCSP responder or client clock inaccuracy. Instead
7113 we allow thisUpdate and nextUpdate to fall within a certain period of
7114 the current time. The age of the response can also optionally be
7115 checked. Two new options -validity_period and -status_age added to
7119 *) If signature or public key algorithm is unrecognized print out its
7120 OID rather that just UNKNOWN.
7123 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7124 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7125 ID to be generated from the issuer certificate alone which can then be
7126 passed to OCSP_id_issuer_cmp().
7129 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7130 ASN1 modules to export functions returning ASN1_ITEM pointers
7131 instead of the ASN1_ITEM structures themselves. This adds several
7132 new macros which allow the underlying ASN1 function/structure to
7133 be accessed transparently. As a result code should not use ASN1_ITEM
7134 references directly (such as &X509_it) but instead use the relevant
7135 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7136 use of the new ASN1 code on platforms where exporting structures
7137 is problematical (for example in shared libraries) but exporting
7138 functions returning pointers to structures is not.
7141 *) Add support for overriding the generation of SSL/TLS session IDs.
7142 These callbacks can be registered either in an SSL_CTX or per SSL.
7143 The purpose of this is to allow applications to control, if they wish,
7144 the arbitrary values chosen for use as session IDs, particularly as it
7145 can be useful for session caching in multiple-server environments. A
7146 command-line switch for testing this (and any client code that wishes
7147 to use such a feature) has been added to "s_server".
7148 [Geoff Thorpe, Lutz Jaenicke]
7150 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7151 of the form '#if defined(...) || defined(...) || ...' and
7152 '#if !defined(...) && !defined(...) && ...'. This also avoids
7153 the growing number of special cases it was previously handling.
7156 *) Make all configuration macros available for application by making
7157 sure they are available in opensslconf.h, by giving them names starting
7158 with "OPENSSL_" to avoid conflicts with other packages and by making
7159 sure e_os2.h will cover all platform-specific cases together with
7161 Additionally, it is now possible to define configuration/platform-
7162 specific names (called "system identities"). In the C code, these
7163 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7164 macro with the name beginning with "OPENSSL_SYS_", which is determined
7165 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7169 *) New option -set_serial to 'req' and 'x509' this allows the serial
7170 number to use to be specified on the command line. Previously self
7171 signed certificates were hard coded with serial number 0 and the
7172 CA options of 'x509' had to use a serial number in a file which was
7176 *) New options to 'ca' utility to support V2 CRL entry extensions.
7177 Currently CRL reason, invalidity date and hold instruction are
7178 supported. Add new CRL extensions to V3 code and some new objects.
7181 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7182 disable standard block padding (aka PKCS#5 padding) in the EVP
7183 API, which was previously mandatory. This means that the data is
7184 not padded in any way and so the total length much be a multiple
7185 of the block size, otherwise an error occurs.
7188 *) Initial (incomplete) OCSP SSL support.
7191 *) New function OCSP_parse_url(). This splits up a URL into its host,
7192 port and path components: primarily to parse OCSP URLs. New -url
7193 option to ocsp utility.
7196 *) New nonce behavior. The return value of OCSP_check_nonce() now
7197 reflects the various checks performed. Applications can decide
7198 whether to tolerate certain situations such as an absent nonce
7199 in a response when one was present in a request: the ocsp application
7200 just prints out a warning. New function OCSP_add1_basic_nonce()
7201 this is to allow responders to include a nonce in a response even if
7202 the request is nonce-less.
7205 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7206 skipped when using openssl x509 multiple times on a single input file,
7207 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7210 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7211 set string type: to handle setting ASN1_TIME structures. Fix ca
7212 utility to correctly initialize revocation date of CRLs.
7215 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7216 the clients preferred ciphersuites and rather use its own preferences.
7217 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7218 Internet Explorer by ensuring unchanged hash method during stepup.
7219 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7222 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7223 to aes and add a new 'exist' option to print out symbols that don't
7227 *) Additional options to ocsp utility to allow flags to be set and
7228 additional certificates supplied.
7231 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7232 OCSP client a number of certificate to only verify the response
7236 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7237 handle the new API. Currently only ECB, CBC modes supported. Add new
7240 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7241 Encryption Standard (AES) Ciphersuites for Transport Layer
7242 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7243 not enabled by default and were not part of the "ALL" ciphersuite
7244 alias because they were not yet official; they could be
7245 explicitly requested by specifying the "AESdraft" ciphersuite
7246 group alias. In the final release of OpenSSL 0.9.7, the group
7247 alias is called "AES" and is part of "ALL".)
7248 [Ben Laurie, Steve Henson, Bodo Moeller]
7250 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7251 request to response.
7254 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7255 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7256 extract information from a certificate request. OCSP_response_create()
7257 creates a response and optionally adds a basic response structure.
7258 OCSP_basic_add1_status() adds a complete single response to a basic
7259 response and returns the OCSP_SINGLERESP structure just added (to allow
7260 extensions to be included for example). OCSP_basic_add1_cert() adds a
7261 certificate to a basic response and OCSP_basic_sign() signs a basic
7262 response with various flags. New helper functions ASN1_TIME_check()
7263 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7264 (converts ASN1_TIME to GeneralizedTime).
7267 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7268 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7269 structure from a certificate. X509_pubkey_digest() digests the public_key
7270 contents: this is used in various key identifiers.
7273 *) Make sk_sort() tolerate a NULL argument.
7274 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7276 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7277 passed by the function are trusted implicitly. If any of them signed the
7278 response then it is assumed to be valid and is not verified.
7281 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7282 to data. This was previously part of the PKCS7 ASN1 code. This
7283 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7284 [Steve Henson, reported by Kenneth R. Robinette
7285 <support@securenetterm.com>]
7287 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7288 routines: without these tracing memory leaks is very painful.
7289 Fix leaks in PKCS12 and PKCS7 routines.
7292 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7293 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7294 effectively meant GeneralizedTime would never be used. Now it
7295 is initialised to -1 but X509_time_adj() now has to check the value
7296 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7297 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7298 [Steve Henson, reported by Kenneth R. Robinette
7299 <support@securenetterm.com>]
7301 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7302 result in a zero length in the ASN1_INTEGER structure which was
7303 not consistent with the structure when d2i_ASN1_INTEGER() was used
7304 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7305 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7306 where it did not print out a minus for negative ASN1_INTEGER.
7309 *) Add summary printout to ocsp utility. The various functions which
7310 convert status values to strings have been renamed to:
7311 OCSP_response_status_str(), OCSP_cert_status_str() and
7312 OCSP_crl_reason_str() and are no longer static. New options
7313 to verify nonce values and to disable verification. OCSP response
7314 printout format cleaned up.
7317 *) Add additional OCSP certificate checks. These are those specified
7318 in RFC2560. This consists of two separate checks: the CA of the
7319 certificate being checked must either be the OCSP signer certificate
7320 or the issuer of the OCSP signer certificate. In the latter case the
7321 OCSP signer certificate must contain the OCSP signing extended key
7322 usage. This check is performed by attempting to match the OCSP
7323 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7324 in the OCSP_CERTID structures of the response.
7327 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7328 and related routines. This uses the standard OpenSSL certificate
7329 verify routines to perform initial checks (just CA validity) and
7330 to obtain the certificate chain. Then additional checks will be
7331 performed on the chain. Currently the root CA is checked to see
7332 if it is explicitly trusted for OCSP signing. This is used to set
7333 a root CA as a global signing root: that is any certificate that
7334 chains to that CA is an acceptable OCSP signing certificate.
7337 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7338 extensions from a separate configuration file.
7339 As when reading extensions from the main configuration file,
7340 the '-extensions ...' option may be used for specifying the
7342 [Massimiliano Pala <madwolf@comune.modena.it>]
7344 *) New OCSP utility. Allows OCSP requests to be generated or
7345 read. The request can be sent to a responder and the output
7346 parsed, outputed or printed in text form. Not complete yet:
7347 still needs to check the OCSP response validity.
7350 *) New subcommands for 'openssl ca':
7351 'openssl ca -status <serial>' prints the status of the cert with
7352 the given serial number (according to the index file).
7353 'openssl ca -updatedb' updates the expiry status of certificates
7355 [Massimiliano Pala <madwolf@comune.modena.it>]
7357 *) New '-newreq-nodes' command option to CA.pl. This is like
7358 '-newreq', but calls 'openssl req' with the '-nodes' option
7359 so that the resulting key is not encrypted.
7360 [Damien Miller <djm@mindrot.org>]
7362 *) New configuration for the GNU Hurd.
7363 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7365 *) Initial code to implement OCSP basic response verify. This
7366 is currently incomplete. Currently just finds the signer's
7367 certificate and verifies the signature on the response.
7370 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7371 value of OPENSSLDIR. This is available via the new '-d' option
7372 to 'openssl version', and is also included in 'openssl version -a'.
7375 *) Allowing defining memory allocation callbacks that will be given
7376 file name and line number information in additional arguments
7377 (a const char* and an int). The basic functionality remains, as
7378 well as the original possibility to just replace malloc(),
7379 realloc() and free() by functions that do not know about these
7380 additional arguments. To register and find out the current
7381 settings for extended allocation functions, the following
7382 functions are provided:
7384 CRYPTO_set_mem_ex_functions
7385 CRYPTO_set_locked_mem_ex_functions
7386 CRYPTO_get_mem_ex_functions
7387 CRYPTO_get_locked_mem_ex_functions
7389 These work the same way as CRYPTO_set_mem_functions and friends.
7390 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7391 extended allocation function is enabled.
7392 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7393 a conventional allocation function is enabled.
7394 [Richard Levitte, Bodo Moeller]
7396 *) Finish off removing the remaining LHASH function pointer casts.
7397 There should no longer be any prototype-casting required when using
7398 the LHASH abstraction, and any casts that remain are "bugs". See
7399 the callback types and macros at the head of lhash.h for details
7400 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7403 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7404 If /dev/[u]random devices are not available or do not return enough
7405 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7407 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7408 /etc/entropy will be queried once each in this sequence, querying stops
7409 when enough entropy was collected without querying more sockets.
7412 *) Change the Unix RAND_poll() variant to be able to poll several
7413 random devices, as specified by DEVRANDOM, until a sufficient amount
7414 of data has been collected. We spend at most 10 ms on each file
7415 (select timeout) and read in non-blocking mode. DEVRANDOM now
7416 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7417 (previously it was just the string "/dev/urandom"), so on typical
7418 platforms the 10 ms delay will never occur.
7419 Also separate out the Unix variant to its own file, rand_unix.c.
7420 For VMS, there's a currently-empty rand_vms.c.
7423 *) Move OCSP client related routines to ocsp_cl.c. These
7424 provide utility functions which an application needing
7425 to issue a request to an OCSP responder and analyse the
7426 response will typically need: as opposed to those which an
7427 OCSP responder itself would need which will be added later.
7429 OCSP_request_sign() signs an OCSP request with an API similar
7430 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7431 response. OCSP_response_get1_basic() extracts basic response
7432 from response. OCSP_resp_find_status(): finds and extracts status
7433 information from an OCSP_CERTID structure (which will be created
7434 when the request structure is built). These are built from lower
7435 level functions which work on OCSP_SINGLERESP structures but
7436 won't normally be used unless the application wishes to examine
7437 extensions in the OCSP response for example.
7439 Replace nonce routines with a pair of functions.
7440 OCSP_request_add1_nonce() adds a nonce value and optionally
7441 generates a random value. OCSP_check_nonce() checks the
7442 validity of the nonce in an OCSP response.
7445 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7446 This doesn't copy the supplied OCSP_CERTID and avoids the
7447 need to free up the newly created id. Change return type
7448 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7449 This can then be used to add extensions to the request.
7450 Deleted OCSP_request_new(), since most of its functionality
7451 is now in OCSP_REQUEST_new() (and the case insensitive name
7452 clash) apart from the ability to set the request name which
7453 will be added elsewhere.
7456 *) Update OCSP API. Remove obsolete extensions argument from
7457 various functions. Extensions are now handled using the new
7458 OCSP extension code. New simple OCSP HTTP function which
7459 can be used to send requests and parse the response.
7462 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7463 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7464 uses the special reorder version of SET OF to sort the attributes
7465 and reorder them to match the encoded order. This resolves a long
7466 standing problem: a verify on a PKCS7 structure just after signing
7467 it used to fail because the attribute order did not match the
7468 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7469 it uses the received order. This is necessary to tolerate some broken
7470 software that does not order SET OF. This is handled by encoding
7471 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7472 to produce the required SET OF.
7475 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7476 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7477 files to get correct declarations of the ASN.1 item variables.
7480 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7481 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7482 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7483 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7484 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7485 ASN1_ITEM and no wrapper functions.
7488 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7489 replace the old function pointer based I/O routines. Change most of
7490 the *_d2i_bio() and *_d2i_fp() functions to use these.
7493 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7494 lines, recognice more "algorithms" that can be deselected, and make
7495 it complain about algorithm deselection that isn't recognised.
7498 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7499 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7500 to use new functions. Add NO_ASN1_OLD which can be set to remove
7501 some old style ASN1 functions: this can be used to determine if old
7502 code will still work when these eventually go away.
7505 *) New extension functions for OCSP structures, these follow the
7506 same conventions as certificates and CRLs.
7509 *) New function X509V3_add1_i2d(). This automatically encodes and
7510 adds an extension. Its behaviour can be customised with various
7511 flags to append, replace or delete. Various wrappers added for
7512 certificates and CRLs.
7515 *) Fix to avoid calling the underlying ASN1 print routine when
7516 an extension cannot be parsed. Correct a typo in the
7517 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7520 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7521 entries for variables.
7524 *) Add functionality to apps/openssl.c for detecting locking
7525 problems: As the program is single-threaded, all we have
7526 to do is register a locking callback using an array for
7527 storing which locks are currently held by the program.
7530 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7531 SSL_get_ex_data_X509_STORE_idx(), which is used in
7532 ssl_verify_cert_chain() and thus can be called at any time
7533 during TLS/SSL handshakes so that thread-safety is essential.
7534 Unfortunately, the ex_data design is not at all suited
7535 for multi-threaded use, so it probably should be abolished.
7538 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7539 [Broadcom, tweaked and integrated by Geoff Thorpe]
7541 *) Move common extension printing code to new function
7542 X509V3_print_extensions(). Reorganise OCSP print routines and
7543 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7546 *) New function X509_signature_print() to remove duplication in some
7550 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7551 set (this was treated exactly the same as SET OF previously). This
7552 is used to reorder the STACK representing the structure to match the
7553 encoding. This will be used to get round a problem where a PKCS7
7554 structure which was signed could not be verified because the STACK
7555 order did not reflect the encoded order.
7558 *) Reimplement the OCSP ASN1 module using the new code.
7561 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7562 for its ASN1 operations. The old style function pointers still exist
7563 for now but they will eventually go away.
7566 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7567 completely replaces the old ASN1 functionality with a table driven
7568 encoder and decoder which interprets an ASN1_ITEM structure describing
7569 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7570 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7571 has also been converted to the new form.
7574 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7575 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7576 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7577 for negative moduli.
7580 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7581 of not touching the result's sign bit.
7584 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7588 *) Changed the LHASH code to use prototypes for callbacks, and created
7589 macros to declare and implement thin (optionally static) functions
7590 that provide type-safety and avoid function pointer casting for the
7591 type-specific callbacks.
7594 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7596 [Veers Staats <staatsvr@asc.hpc.mil>,
7597 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7599 *) Reformat the FAQ so the different questions and answers can be divided
7600 in sections depending on the subject.
7603 *) Have the zlib compression code load ZLIB.DLL dynamically under
7607 *) New function BN_mod_sqrt for computing square roots modulo a prime
7608 (using the probabilistic Tonelli-Shanks algorithm unless
7609 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7610 be handled deterministically).
7611 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7613 *) Make BN_mod_inverse faster by explicitly handling small quotients
7614 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7615 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7618 *) New function BN_kronecker.
7621 *) Fix BN_gcd so that it works on negative inputs; the result is
7622 positive unless both parameters are zero.
7623 Previously something reasonably close to an infinite loop was
7624 possible because numbers could be growing instead of shrinking
7625 in the implementation of Euclid's algorithm.
7628 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7629 sign of the number in question.
7631 Fix BN_is_word(a,w) to work correctly for w == 0.
7633 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7634 because its test if the absolute value of 'a' equals 'w'.
7635 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7636 it exists mostly for use in the implementations of BN_is_zero(),
7637 BN_is_one(), and BN_is_word().
7640 *) New function BN_swap.
7643 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7644 the exponentiation functions are more likely to produce reasonable
7645 results on negative inputs.
7648 *) Change BN_mod_mul so that the result is always non-negative.
7649 Previously, it could be negative if one of the factors was negative;
7650 I don't think anyone really wanted that behaviour.
7653 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7654 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7655 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7656 and add new functions:
7665 BN_mod_lshift1_quick
7669 These functions always generate non-negative results.
7671 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7672 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7674 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7675 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7676 be reduced modulo m.
7677 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7680 The following entry accidentally appeared in the CHANGES file
7681 distributed with OpenSSL 0.9.7. The modifications described in
7682 it do *not* apply to OpenSSL 0.9.7.
7684 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7685 was actually never needed) and in BN_mul(). The removal in BN_mul()
7686 required a small change in bn_mul_part_recursive() and the addition
7687 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7688 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7689 bn_sub_words() and bn_add_words() except they take arrays with
7694 *) In 'openssl passwd', verify passwords read from the terminal
7695 unless the '-salt' option is used (which usually means that
7696 verification would just waste user's time since the resulting
7697 hash is going to be compared with some given password hash)
7698 or the new '-noverify' option is used.
7700 This is an incompatible change, but it does not affect
7701 non-interactive use of 'openssl passwd' (passwords on the command
7702 line, '-stdin' option, '-in ...' option) and thus should not
7706 *) Remove all references to RSAref, since there's no more need for it.
7709 *) Make DSO load along a path given through an environment variable
7710 (SHLIB_PATH) with shl_load().
7713 *) Constify the ENGINE code as a result of BIGNUM constification.
7714 Also constify the RSA code and most things related to it. In a
7715 few places, most notable in the depth of the ASN.1 code, ugly
7716 casts back to non-const were required (to be solved at a later
7720 *) Make it so the openssl application has all engines loaded by default.
7723 *) Constify the BIGNUM routines a little more.
7726 *) Add the following functions:
7728 ENGINE_load_cswift()
7730 ENGINE_load_atalla()
7732 ENGINE_load_builtin_engines()
7734 That way, an application can itself choose if external engines that
7735 are built-in in OpenSSL shall ever be used or not. The benefit is
7736 that applications won't have to be linked with libdl or other dso
7737 libraries unless it's really needed.
7739 Changed 'openssl engine' to load all engines on demand.
7740 Changed the engine header files to avoid the duplication of some
7741 declarations (they differed!).
7744 *) 'openssl engine' can now list capabilities.
7747 *) Better error reporting in 'openssl engine'.
7750 *) Never call load_dh_param(NULL) in s_server.
7753 *) Add engine application. It can currently list engines by name and
7754 identity, and test if they are actually available.
7757 *) Improve RPM specification file by forcing symbolic linking and making
7758 sure the installed documentation is also owned by root.root.
7759 [Damien Miller <djm@mindrot.org>]
7761 *) Give the OpenSSL applications more possibilities to make use of
7762 keys (public as well as private) handled by engines.
7765 *) Add OCSP code that comes from CertCo.
7768 *) Add VMS support for the Rijndael code.
7771 *) Added untested support for Nuron crypto accelerator.
7774 *) Add support for external cryptographic devices. This code was
7775 previously distributed separately as the "engine" branch.
7776 [Geoff Thorpe, Richard Levitte]
7778 *) Rework the filename-translation in the DSO code. It is now possible to
7779 have far greater control over how a "name" is turned into a filename
7780 depending on the operating environment and any oddities about the
7781 different shared library filenames on each system.
7784 *) Support threads on FreeBSD-elf in Configure.
7787 *) Fix for SHA1 assembly problem with MASM: it produces
7788 warnings about corrupt line number information when assembling
7789 with debugging information. This is caused by the overlapping
7791 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7794 NCONF_get_number() has no error checking at all. As a replacement,
7795 NCONF_get_number_e() is defined (_e for "error checking") and is
7796 promoted strongly. The old NCONF_get_number is kept around for
7797 binary backward compatibility.
7798 Make it possible for methods to load from something other than a BIO,
7799 by providing a function pointer that is given a name instead of a BIO.
7800 For example, this could be used to load configuration data from an
7804 *) Fix for non blocking accept BIOs. Added new I/O special reason
7805 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7806 with non blocking I/O was not possible because no retry code was
7807 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7811 *) Added the beginnings of Rijndael support.
7814 *) Fix for bug in DirectoryString mask setting. Add support for
7815 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7816 to allow certificate printing to more controllable, additional
7817 'certopt' option to 'x509' to allow new printing options to be
7821 *) Clean old EAY MD5 hack from e_os.h.
7824 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7826 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7827 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7828 [Joe Orton, Steve Henson]
7830 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7832 *) Fix additional bug revealed by the NISCC test suite:
7834 Stop bug triggering large recursion when presented with
7835 certain ASN.1 tags (CVE-2003-0851)
7838 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7840 *) Fix various bugs revealed by running the NISCC test suite:
7842 Stop out of bounds reads in the ASN1 code when presented with
7843 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7845 If verify callback ignores invalid public key errors don't try to check
7846 certificate signature with the NULL public key.
7850 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7851 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7855 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7856 extra data after the compression methods not only for TLS 1.0
7857 but also for SSL 3.0 (as required by the specification).
7858 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7860 *) Change X509_certificate_type() to mark the key as exported/exportable
7861 when it's 512 *bits* long, not 512 bytes.
7864 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7866 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7867 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7868 a protocol version number mismatch like a decryption error
7869 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7872 *) Turn on RSA blinding by default in the default implementation
7873 to avoid a timing attack. Applications that don't want it can call
7874 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7875 They would be ill-advised to do so in most cases.
7876 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7878 *) Change RSA blinding code so that it works when the PRNG is not
7879 seeded (in this case, the secret RSA exponent is abused as
7880 an unpredictable seed -- if it is not unpredictable, there
7881 is no point in blinding anyway). Make RSA blinding thread-safe
7882 by remembering the creator's thread ID in rsa->blinding and
7883 having all other threads use local one-time blinding factors
7884 (this requires more computation than sharing rsa->blinding, but
7885 avoids excessive locking; and if an RSA object is not shared
7886 between threads, blinding will still be very fast).
7889 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7891 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7892 via timing by performing a MAC computation even if incorrrect
7893 block cipher padding has been found. This is a countermeasure
7894 against active attacks where the attacker has to distinguish
7895 between bad padding and a MAC verification error. (CVE-2003-0078)
7897 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7898 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7899 Martin Vuagnoux (EPFL, Ilion)]
7901 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7903 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7904 memory from it's contents. This is done with a counter that will
7905 place alternating values in each byte. This can be used to solve
7906 two issues: 1) the removal of calls to memset() by highly optimizing
7907 compilers, and 2) cleansing with other values than 0, since those can
7908 be read through on certain media, for example a swap space on disk.
7911 *) Bugfix: client side session caching did not work with external caching,
7912 because the session->cipher setting was not restored when reloading
7913 from the external cache. This problem was masked, when
7914 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7915 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7918 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7919 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7920 [Zeev Lieber <zeev-l@yahoo.com>]
7922 *) Undo an undocumented change introduced in 0.9.6e which caused
7923 repeated calls to OpenSSL_add_all_ciphers() and
7924 OpenSSL_add_all_digests() to be ignored, even after calling
7928 *) Change the default configuration reader to deal with last line not
7929 being properly terminated.
7932 *) Change X509_NAME_cmp() so it applies the special rules on handling
7933 DN values that are of type PrintableString, as well as RDNs of type
7934 emailAddress where the value has the type ia5String.
7935 [stefank@valicert.com via Richard Levitte]
7937 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7938 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7939 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7940 the bitwise-OR of the two for use by the majority of applications
7941 wanting this behaviour, and update the docs. The documented
7942 behaviour and actual behaviour were inconsistent and had been
7943 changing anyway, so this is more a bug-fix than a behavioural
7945 [Geoff Thorpe, diagnosed by Nadav Har'El]
7947 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7948 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7951 *) Fix initialization code race conditions in
7952 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7953 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7954 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7955 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7956 ssl2_get_cipher_by_char(),
7957 ssl3_get_cipher_by_char().
7958 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7960 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7961 the cached sessions are flushed, as the remove_cb() might use ex_data
7962 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7963 (see [openssl.org #212]).
7964 [Geoff Thorpe, Lutz Jaenicke]
7966 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7967 length, instead of the encoding length to d2i_ASN1_OBJECT.
7970 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7972 *) [In 0.9.6g-engine release:]
7973 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
7974 [Lynn Gazis <lgazis@rainbow.com>]
7976 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
7978 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
7979 and get fix the header length calculation.
7980 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
7981 Alon Kantor <alonk@checkpoint.com> (and others),
7984 *) Use proper error handling instead of 'assertions' in buffer
7985 overflow checks added in 0.9.6e. This prevents DoS (the
7986 assertions could call abort()).
7987 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
7989 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
7991 *) Add various sanity checks to asn1_get_length() to reject
7992 the ASN1 length bytes if they exceed sizeof(long), will appear
7993 negative or the content length exceeds the length of the
7995 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
7997 *) Fix cipher selection routines: ciphers without encryption had no flags
7998 for the cipher strength set and where therefore not handled correctly
7999 by the selection routines (PR #130).
8002 *) Fix EVP_dsa_sha macro.
8006 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8007 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8008 that was added in OpenSSL 0.9.6d.
8010 As the countermeasure turned out to be incompatible with some
8011 broken SSL implementations, the new option is part of SSL_OP_ALL.
8012 SSL_OP_ALL is usually employed when compatibility with weird SSL
8013 implementations is desired (e.g. '-bugs' option to 's_client' and
8014 's_server'), so the new option is automatically set in many
8018 *) Changes in security patch:
8020 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8021 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8022 Air Force Materiel Command, USAF, under agreement number
8025 *) Add various sanity checks to asn1_get_length() to reject
8026 the ASN1 length bytes if they exceed sizeof(long), will appear
8027 negative or the content length exceeds the length of the
8028 supplied buffer. (CVE-2002-0659)
8029 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8031 *) Assertions for various potential buffer overflows, not known to
8033 [Ben Laurie (CHATS)]
8035 *) Various temporary buffers to hold ASCII versions of integers were
8036 too small for 64 bit platforms. (CVE-2002-0655)
8037 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8039 *) Remote buffer overflow in SSL3 protocol - an attacker could
8040 supply an oversized session ID to a client. (CVE-2002-0656)
8041 [Ben Laurie (CHATS)]
8043 *) Remote buffer overflow in SSL2 protocol - an attacker could
8044 supply an oversized client master key. (CVE-2002-0656)
8045 [Ben Laurie (CHATS)]
8047 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8049 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8050 encoded as NULL) with id-dsa-with-sha1.
8051 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8053 *) Check various X509_...() return values in apps/req.c.
8054 [Nils Larsch <nla@trustcenter.de>]
8056 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8057 an end-of-file condition would erroneously be flagged, when the CRLF
8058 was just at the end of a processed block. The bug was discovered when
8059 processing data through a buffering memory BIO handing the data to a
8060 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8061 <ptsekov@syntrex.com> and Nedelcho Stanev.
8064 *) Implement a countermeasure against a vulnerability recently found
8065 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8066 before application data chunks to avoid the use of known IVs
8067 with data potentially chosen by the attacker.
8070 *) Fix length checks in ssl3_get_client_hello().
8073 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8074 to prevent ssl3_read_internal() from incorrectly assuming that
8075 ssl3_read_bytes() found application data while handshake
8076 processing was enabled when in fact s->s3->in_read_app_data was
8077 merely automatically cleared during the initial handshake.
8078 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8080 *) Fix object definitions for Private and Enterprise: they were not
8081 recognized in their shortname (=lowercase) representation. Extend
8082 obj_dat.pl to issue an error when using undefined keywords instead
8083 of silently ignoring the problem (Svenning Sorensen
8084 <sss@sss.dnsalias.net>).
8087 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8088 generators, i.e. generators other than 2 and 5. (Previously, the
8089 code did not properly initialise the 'add' and 'rem' values to
8090 BN_generate_prime().)
8092 In the new general case, we do not insist that 'generator' is
8093 actually a primitive root: This requirement is rather pointless;
8094 a generator of the order-q subgroup is just as good, if not
8098 *) Map new X509 verification errors to alerts. Discovered and submitted by
8099 Tom Wu <tom@arcot.com>.
8102 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8103 returning non-zero before the data has been completely received
8104 when using non-blocking I/O.
8105 [Bodo Moeller; problem pointed out by John Hughes]
8107 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8108 [Ben Laurie, Lutz Jaenicke]
8110 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8111 Yoram Zahavi <YoramZ@gilian.com>).
8114 *) Add information about CygWin 1.3 and on, and preserve proper
8115 configuration for the versions before that.
8116 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8118 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8119 check whether we deal with a copy of a session and do not delete from
8120 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8121 <izhar@checkpoint.com>.
8124 *) Do not store session data into the internal session cache, if it
8125 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8126 flag is set). Proposed by Aslam <aslam@funk.com>.
8129 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8133 *) [In 0.9.6d-engine release:]
8134 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8135 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8137 *) Add the configuration target linux-s390x.
8138 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8140 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8141 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8142 variable as an indication that a ClientHello message has been
8143 received. As the flag value will be lost between multiple
8144 invocations of ssl3_accept when using non-blocking I/O, the
8145 function may not be aware that a handshake has actually taken
8146 place, thus preventing a new session from being added to the
8149 To avoid this problem, we now set s->new_session to 2 instead of
8150 using a local variable.
8151 [Lutz Jaenicke, Bodo Moeller]
8153 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8154 if the SSL_R_LENGTH_MISMATCH error is detected.
8155 [Geoff Thorpe, Bodo Moeller]
8157 *) New 'shared_ldflag' column in Configure platform table.
8160 *) Fix EVP_CIPHER_mode macro.
8161 ["Dan S. Camper" <dan@bti.net>]
8163 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8164 type, we must throw them away by setting rr->length to 0.
8165 [D P Chang <dpc@qualys.com>]
8167 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8169 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8170 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8171 worked incorrectly for those cases where range = 10..._2 and
8172 3*range is two bits longer than range.)
8175 *) Only add signing time to PKCS7 structures if it is not already
8179 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8180 OBJ_ld_ce should be OBJ_id_ce.
8181 Also some ip-pda OIDs in crypto/objects/objects.txt were
8182 incorrect (cf. RFC 3039).
8183 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8185 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8186 returns early because it has nothing to do.
8187 [Andy Schneider <andy.schneider@bjss.co.uk>]
8189 *) [In 0.9.6c-engine release:]
8190 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8191 [Andy Schneider <andy.schneider@bjss.co.uk>]
8193 *) [In 0.9.6c-engine release:]
8194 Add support for Cryptographic Appliance's keyserver technology.
8195 (Use engine 'keyclient')
8196 [Cryptographic Appliances and Geoff Thorpe]
8198 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8199 is called via tools/c89.sh because arguments have to be
8200 rearranged (all '-L' options must appear before the first object
8202 [Richard Shapiro <rshapiro@abinitio.com>]
8204 *) [In 0.9.6c-engine release:]
8205 Add support for Broadcom crypto accelerator cards, backported
8207 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8209 *) [In 0.9.6c-engine release:]
8210 Add support for SureWare crypto accelerator cards from
8211 Baltimore Technologies. (Use engine 'sureware')
8212 [Baltimore Technologies and Mark Cox]
8214 *) [In 0.9.6c-engine release:]
8215 Add support for crypto accelerator cards from Accelerated
8216 Encryption Processing, www.aep.ie. (Use engine 'aep')
8217 [AEP Inc. and Mark Cox]
8219 *) Add a configuration entry for gcc on UnixWare.
8220 [Gary Benson <gbenson@redhat.com>]
8222 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8223 messages are stored in a single piece (fixed-length part and
8224 variable-length part combined) and fix various bugs found on the way.
8227 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8228 instead. BIO_gethostbyname() does not know what timeouts are
8229 appropriate, so entries would stay in cache even when they have
8231 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8233 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8234 faced with a pathologically small ClientHello fragment that does
8235 not contain client_version: Instead of aborting with an error,
8236 simply choose the highest available protocol version (i.e.,
8237 TLS 1.0 unless it is disabled). In practice, ClientHello
8238 messages are never sent like this, but this change gives us
8239 strictly correct behaviour at least for TLS.
8242 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8243 never resets s->method to s->ctx->method when called from within
8244 one of the SSL handshake functions.
8245 [Bodo Moeller; problem pointed out by Niko Baric]
8247 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8248 (sent using the client's version number) if client_version is
8249 smaller than the protocol version in use. Also change
8250 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8251 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8252 the client will at least see that alert.
8255 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8259 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8260 client receives HelloRequest while in a handshake.
8261 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8263 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8264 should end in 'break', not 'goto end' which circumvents various
8265 cleanups done in state SSL_ST_OK. But session related stuff
8266 must be disabled for SSL_ST_OK in the case that we just sent a
8269 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8270 before just sending a HelloRequest.
8271 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8273 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8274 reveal whether illegal block cipher padding was found or a MAC
8275 verification error occurred. (Neither SSLerr() codes nor alerts
8276 are directly visible to potential attackers, but the information
8277 may leak via logfiles.)
8279 Similar changes are not required for the SSL 2.0 implementation
8280 because the number of padding bytes is sent in clear for SSL 2.0,
8281 and the extra bytes are just ignored. However ssl/s2_pkt.c
8282 failed to verify that the purported number of padding bytes is in
8286 *) Add OpenUNIX-8 support including shared libraries
8287 (Boyd Lynn Gerber <gerberb@zenez.com>).
8290 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8291 'wristwatch attack' using huge encoding parameters (cf.
8292 James H. Manger's CRYPTO 2001 paper). Note that the
8293 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8294 encoding parameters and hence was not vulnerable.
8297 *) BN_sqr() bug fix.
8298 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8300 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8301 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8302 followed by modular reduction.
8303 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8305 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8306 equivalent based on BN_pseudo_rand() instead of BN_rand().
8309 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8310 This function was broken, as the check for a new client hello message
8311 to handle SGC did not allow these large messages.
8312 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8315 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8318 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8319 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8322 *) Rework the configuration and shared library support for Tru64 Unix.
8323 The configuration part makes use of modern compiler features and
8324 still retains old compiler behavior for those that run older versions
8325 of the OS. The shared library support part includes a variant that
8326 uses the RPATH feature, and is available through the special
8327 configuration target "alpha-cc-rpath", which will never be selected
8329 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8331 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8332 with the same message size as in ssl3_get_certificate_request().
8333 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8334 messages might inadvertently be reject as too long.
8335 [Petr Lampa <lampa@fee.vutbr.cz>]
8337 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8340 *) Modified SSL library such that the verify_callback that has been set
8341 specificly for an SSL object with SSL_set_verify() is actually being
8342 used. Before the change, a verify_callback set with this function was
8343 ignored and the verify_callback() set in the SSL_CTX at the time of
8344 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8345 to allow the necessary settings.
8348 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8349 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8350 done automatically (in contradiction to the requirements of the C
8351 standard). This made problems when used from OpenSSH.
8354 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8355 dh->length and always used
8357 BN_rand_range(priv_key, dh->p).
8359 BN_rand_range() is not necessary for Diffie-Hellman, and this
8360 specific range makes Diffie-Hellman unnecessarily inefficient if
8361 dh->length (recommended exponent length) is much smaller than the
8362 length of dh->p. We could use BN_rand_range() if the order of
8363 the subgroup was stored in the DH structure, but we only have
8368 BN_rand(priv_key, l, ...)
8370 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8376 RSA_eay_public_encrypt
8377 RSA_eay_private_decrypt
8378 RSA_eay_private_encrypt (signing)
8379 RSA_eay_public_decrypt (signature verification)
8381 (default implementations for RSA_public_encrypt,
8382 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8383 always reject numbers >= n.
8386 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8387 to synchronize access to 'locking_thread'. This is necessary on
8388 systems where access to 'locking_thread' (an 'unsigned long'
8389 variable) is not atomic.
8392 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8393 *before* setting the 'crypto_lock_rand' flag. The previous code had
8394 a race condition if 0 is a valid thread ID.
8395 [Travis Vitek <vitek@roguewave.com>]
8397 *) Add support for shared libraries under Irix.
8398 [Albert Chin-A-Young <china@thewrittenword.com>]
8400 *) Add configuration option to build on Linux on both big-endian and
8402 [Ralf Baechle <ralf@uni-koblenz.de>]
8404 *) Add the possibility to create shared libraries on HP-UX.
8407 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8409 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8410 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8411 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8412 PRNG state recovery was possible based on the output of
8413 one PRNG request appropriately sized to gain knowledge on
8414 'md' followed by enough consecutive 1-byte PRNG requests
8415 to traverse all of 'state'.
8417 1. When updating 'md_local' (the current thread's copy of 'md')
8418 during PRNG output generation, hash all of the previous
8419 'md_local' value, not just the half used for PRNG output.
8421 2. Make the number of bytes from 'state' included into the hash
8422 independent from the number of PRNG bytes requested.
8424 The first measure alone would be sufficient to avoid
8425 Markku-Juhani's attack. (Actually it had never occurred
8426 to me that the half of 'md_local' used for chaining was the
8427 half from which PRNG output bytes were taken -- I had always
8428 assumed that the secret half would be used.) The second
8429 measure makes sure that additional data from 'state' is never
8430 mixed into 'md_local' in small portions; this heuristically
8431 further strengthens the PRNG.
8434 *) Fix crypto/bn/asm/mips3.s.
8437 *) When only the key is given to "enc", the IV is undefined. Print out
8438 an error message in this case.
8441 *) Handle special case when X509_NAME is empty in X509 printing routines.
8444 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8445 positive and less than q.
8448 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8449 used: it isn't thread safe and the add_lock_callback should handle
8451 [Paul Rose <Paul.Rose@bridge.com>]
8453 *) Verify that incoming data obeys the block size in
8454 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8458 [Ulf Möller, Bodo Möller]
8460 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8461 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8462 when fixing the server behaviour for backwards-compatible 'client
8463 hello' messages. (Note that the attack is impractical against
8464 SSL 3.0 and TLS 1.0 anyway because length and version checking
8465 means that the probability of guessing a valid ciphertext is
8466 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8469 Before 0.9.5, the countermeasure (hide the error by generating a
8470 random 'decryption result') did not work properly because
8471 ERR_clear_error() was missing, meaning that SSL_get_error() would
8472 detect the supposedly ignored error.
8474 Both problems are now fixed.
8477 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8478 (previously it was 1024).
8481 *) Fix for compatibility mode trust settings: ignore trust settings
8482 unless some valid trust or reject settings are present.
8485 *) Fix for blowfish EVP: its a variable length cipher.
8488 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8489 parameters in DSA public key structures and return an error in the
8490 DSA routines if parameters are absent.
8493 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8494 in the current directory if neither $RANDFILE nor $HOME was set.
8495 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8496 caused some confusion to Windows users who haven't defined $HOME.
8497 Thus RAND_file_name() is changed again: e_os.h can define a
8498 DEFAULT_HOME, which will be used if $HOME is not set.
8499 For Windows, we use "C:"; on other platforms, we still require
8500 environment variables.
8502 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8503 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8504 having multiple threads call RAND_poll() concurrently.
8507 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8508 combination of a flag and a thread ID variable.
8509 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8510 flag), *other* threads can enter ssleay_add_bytes without obeying
8511 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8512 that they do not hold after the first thread unsets add_do_not_lock).
8515 *) Change bctest again: '-x' expressions are not available in all
8519 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8521 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8522 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8524 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8525 the default extension for executables, if any. Also, make the perl
8526 scripts that use symlink() to test if it really exists and use "cp"
8527 if it doesn't. All this made OpenSSL compilable and installable in
8531 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8532 If SEQUENCE is length is indefinite just set c->slen to the total
8533 amount of data available.
8534 [Steve Henson, reported by shige@FreeBSD.org]
8535 [This change does not apply to 0.9.7.]
8537 *) Change bctest to avoid here-documents inside command substitution
8538 (workaround for FreeBSD /bin/sh bug).
8539 For compatibility with Ultrix, avoid shell functions (introduced
8540 in the bctest version that searches along $PATH).
8543 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8544 with des_encrypt() defined on some operating systems, like Solaris
8548 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8549 On the Importance of Eliminating Errors in Cryptographic
8550 Computations, J. Cryptology 14 (2001) 2, 101-119,
8551 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8554 *) MIPS assembler BIGNUM division bug fix.
8557 *) Disabled incorrect Alpha assembler code.
8560 *) Fix PKCS#7 decode routines so they correctly update the length
8561 after reading an EOC for the EXPLICIT tag.
8563 [This change does not apply to 0.9.7.]
8565 *) Fix bug in PKCS#12 key generation routines. This was triggered
8566 if a 3DES key was generated with a 0 initial byte. Include
8567 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8568 (but broken) behaviour.
8571 *) Enhance bctest to search for a working bc along $PATH and print
8573 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8575 *) Fix memory leaks in err.c: free err_data string if necessary;
8576 don't write to the wrong index in ERR_set_error_data.
8579 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8583 *) Replace rdtsc with _emit statements for VC++ version 5.
8584 [Jeremy Cooper <jeremy@baymoo.org>]
8586 *) Make it possible to reuse SSLv2 sessions.
8589 *) In copy_email() check for >= 0 as a return value for
8590 X509_NAME_get_index_by_NID() since 0 is a valid index.
8591 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8593 *) Avoid coredump with unsupported or invalid public keys by checking if
8594 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8595 PKCS7_verify() fails with non detached data.
8598 *) Don't use getenv in library functions when run as setuid/setgid.
8599 New function OPENSSL_issetugid().
8602 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8603 due to incorrect handling of multi-threading:
8605 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8607 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8609 3. Count how many times MemCheck_off() has been called so that
8610 nested use can be treated correctly. This also avoids
8611 inband-signalling in the previous code (which relied on the
8612 assumption that thread ID 0 is impossible).
8615 *) Add "-rand" option also to s_client and s_server.
8618 *) Fix CPU detection on Irix 6.x.
8619 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8620 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8622 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8625 [This change does not apply to 0.9.7.]
8627 *) Use the cached encoding of an X509_NAME structure rather than
8628 copying it. This is apparently the reason for the libsafe "errors"
8629 but the code is actually correct.
8632 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8633 Bleichenbacher's DSA attack.
8634 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8635 to be set and top=0 forces the highest bit to be set; top=-1 is new
8636 and leaves the highest bit random.
8637 [Ulf Moeller, Bodo Moeller]
8639 *) In the NCONF_...-based implementations for CONF_... queries
8640 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8641 a temporary CONF structure with the data component set to NULL
8642 (which gives segmentation faults in lh_retrieve).
8643 Instead, use NULL for the CONF pointer in CONF_get_string and
8644 CONF_get_number (which may use environment variables) and directly
8645 return NULL from CONF_get_section.
8648 *) Fix potential buffer overrun for EBCDIC.
8651 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8652 keyUsage if basicConstraints absent for a CA.
8655 *) Make SMIME_write_PKCS7() write mail header values with a format that
8656 is more generally accepted (no spaces before the semicolon), since
8657 some programs can't parse those values properly otherwise. Also make
8658 sure BIO's that break lines after each write do not create invalid
8662 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8663 macros previously used would not encode an empty SEQUENCE OF
8664 and break the signature.
8666 [This change does not apply to 0.9.7.]
8668 *) Zero the premaster secret after deriving the master secret in
8672 *) Add some EVP_add_digest_alias registrations (as found in
8673 OpenSSL_add_all_digests()) to SSL_library_init()
8674 aka OpenSSL_add_ssl_algorithms(). This provides improved
8675 compatibility with peers using X.509 certificates
8676 with unconventional AlgorithmIdentifier OIDs.
8679 *) Fix for Irix with NO_ASM.
8680 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8682 *) ./config script fixes.
8683 [Ulf Moeller, Richard Levitte]
8685 *) Fix 'openssl passwd -1'.
8688 *) Change PKCS12_key_gen_asc() so it can cope with non null
8689 terminated strings whose length is passed in the passlen
8690 parameter, for example from PEM callbacks. This was done
8691 by adding an extra length parameter to asc2uni().
8692 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8694 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8695 call failed, free the DSA structure.
8698 *) Fix to uni2asc() to cope with zero length Unicode strings.
8699 These are present in some PKCS#12 files.
8702 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8703 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8704 when writing a 32767 byte record.
8705 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8707 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8708 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8710 (RSA objects have a reference count access to which is protected
8711 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8712 so they are meant to be shared between threads.)
8713 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8714 "Reddie, Steven" <Steven.Reddie@ca.com>]
8716 *) Fix a deadlock in CRYPTO_mem_leaks().
8719 *) Use better test patterns in bntest.
8722 *) rand_win.c fix for Borland C.
8725 *) BN_rshift bugfix for n == 0.
8728 *) Add a 'bctest' script that checks for some known 'bc' bugs
8729 so that 'make test' does not abort just because 'bc' is broken.
8732 *) Store verify_result within SSL_SESSION also for client side to
8733 avoid potential security hole. (Re-used sessions on the client side
8734 always resulted in verify_result==X509_V_OK, not using the original
8735 result of the server certificate verification.)
8738 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8739 SSL3_RT_APPLICATION_DATA, return 0.
8740 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8744 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8745 releases, have been re-implemented by renaming the previous
8746 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8747 and ssl3_read_internal, respectively, and adding 'peek' parameters
8748 to them. The new ssl[23]_{read,peek} functions are calls to
8749 ssl[23]_read_internal with the 'peek' flag set appropriately.
8750 A 'peek' parameter has also been added to ssl3_read_bytes, which
8751 does the actual work for ssl3_read_internal.
8754 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8755 the method-specific "init()" handler. Also clean up ex_data after
8756 calling the method-specific "finish()" handler. Previously, this was
8757 happening the other way round.
8760 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8761 The previous value, 12, was not always sufficient for BN_mod_exp().
8764 *) Make sure that shared libraries get the internal name engine with
8765 the full version number and not just 0. This should mark the
8766 shared libraries as not backward compatible. Of course, this should
8767 be changed again when we can guarantee backward binary compatibility.
8770 *) Fix typo in get_cert_by_subject() in by_dir.c
8771 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8773 *) Rework the system to generate shared libraries:
8775 - Make note of the expected extension for the shared libraries and
8776 if there is a need for symbolic links from for example libcrypto.so.0
8777 to libcrypto.so.0.9.7. There is extended info in Configure for
8780 - Make as few rebuilds of the shared libraries as possible.
8782 - Still avoid linking the OpenSSL programs with the shared libraries.
8784 - When installing, install the shared libraries separately from the
8788 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8790 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8791 and not in SSL_clear because the latter is also used by the
8792 accept/connect functions; previously, the settings made by
8793 SSL_set_read_ahead would be lost during the handshake.
8794 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8796 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8797 Previously, it would create entries for disabled algorithms no
8801 *) Added several new manual pages for SSL_* function.
8804 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8806 *) In ssl23_get_client_hello, generate an error message when faced
8807 with an initial SSL 3.0/TLS record that is too small to contain the
8808 first two bytes of the ClientHello message, i.e. client_version.
8809 (Note that this is a pathologic case that probably has never happened
8810 in real life.) The previous approach was to use the version number
8811 from the record header as a substitute; but our protocol choice
8812 should not depend on that one because it is not authenticated
8813 by the Finished messages.
8816 *) More robust randomness gathering functions for Windows.
8817 [Jeffrey Altman <jaltman@columbia.edu>]
8819 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8820 not set then we don't setup the error code for issuer check errors
8821 to avoid possibly overwriting other errors which the callback does
8822 handle. If an application does set the flag then we assume it knows
8823 what it is doing and can handle the new informational codes
8827 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8828 a general "ANY" type, as such it should be able to decode anything
8829 including tagged types. However it didn't check the class so it would
8830 wrongly interpret tagged types in the same way as their universal
8831 counterpart and unknown types were just rejected. Changed so that the
8832 tagged and unknown types are handled in the same way as a SEQUENCE:
8833 that is the encoding is stored intact. There is also a new type
8834 "V_ASN1_OTHER" which is used when the class is not universal, in this
8835 case we have no idea what the actual type is so we just lump them all
8839 *) On VMS, stdout may very well lead to a file that is written to
8840 in a record-oriented fashion. That means that every write() will
8841 write a separate record, which will be read separately by the
8842 programs trying to read from it. This can be very confusing.
8844 The solution is to put a BIO filter in the way that will buffer
8845 text until a linefeed is reached, and then write everything a
8846 line at a time, so every record written will be an actual line,
8847 not chunks of lines and not (usually doesn't happen, but I've
8848 seen it once) several lines in one record. BIO_f_linebuffer() is
8851 Currently, it's a VMS-only method, because that's where it has
8852 been tested well enough.
8855 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8856 it can return incorrect results.
8857 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8858 but it was in 0.9.6-beta[12].)
8861 *) Disable the check for content being present when verifying detached
8862 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8863 include zero length content when signing messages.
8866 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8867 BIO_ctrl (for BIO pairs).
8870 *) Add DSO method for VMS.
8873 *) Bug fix: Montgomery multiplication could produce results with the
8877 *) Add RPM specification openssl.spec and modify it to build three
8878 packages. The default package contains applications, application
8879 documentation and run-time libraries. The devel package contains
8880 include files, static libraries and function documentation. The
8881 doc package contains the contents of the doc directory. The original
8882 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8885 *) Add a large number of documentation files for many SSL routines.
8886 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8888 *) Add a configuration entry for Sony News 4.
8889 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8891 *) Don't set the two most significant bits to one when generating a
8892 random number < q in the DSA library.
8895 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8896 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8897 the underlying transport is blocking) if a handshake took place.
8898 (The default behaviour is needed by applications such as s_client
8899 and s_server that use select() to determine when to use SSL_read;
8900 but for applications that know in advance when to expect data, it
8901 just makes things more complicated.)
8904 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8908 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8909 work better on such systems.
8910 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8912 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8913 Update PKCS12_parse() so it copies the friendlyName and the
8914 keyid to the certificates aux info.
8917 *) Fix bug in PKCS7_verify() which caused an infinite loop
8918 if there was more than one signature.
8919 [Sven Uszpelkat <su@celocom.de>]
8921 *) Major change in util/mkdef.pl to include extra information
8922 about each symbol, as well as presenting variables as well
8923 as functions. This change means that there's n more need
8924 to rebuild the .num files when some algorithms are excluded.
8927 *) Allow the verify time to be set by an application,
8928 rather than always using the current time.
8931 *) Phase 2 verify code reorganisation. The certificate
8932 verify code now looks up an issuer certificate by a
8933 number of criteria: subject name, authority key id
8934 and key usage. It also verifies self signed certificates
8935 by the same criteria. The main comparison function is
8936 X509_check_issued() which performs these checks.
8938 Lot of changes were necessary in order to support this
8939 without completely rewriting the lookup code.
8941 Authority and subject key identifier are now cached.
8943 The LHASH 'certs' is X509_STORE has now been replaced
8944 by a STACK_OF(X509_OBJECT). This is mainly because an
8945 LHASH can't store or retrieve multiple objects with
8946 the same hash value.
8948 As a result various functions (which were all internal
8949 use only) have changed to handle the new X509_STORE
8950 structure. This will break anything that messed round
8951 with X509_STORE internally.
8953 The functions X509_STORE_add_cert() now checks for an
8954 exact match, rather than just subject name.
8956 The X509_STORE API doesn't directly support the retrieval
8957 of multiple certificates matching a given criteria, however
8958 this can be worked round by performing a lookup first
8959 (which will fill the cache with candidate certificates)
8960 and then examining the cache for matches. This is probably
8961 the best we can do without throwing out X509_LOOKUP
8962 entirely (maybe later...).
8964 The X509_VERIFY_CTX structure has been enhanced considerably.
8966 All certificate lookup operations now go via a get_issuer()
8967 callback. Although this currently uses an X509_STORE it
8968 can be replaced by custom lookups. This is a simple way
8969 to bypass the X509_STORE hackery necessary to make this
8970 work and makes it possible to use more efficient techniques
8971 in future. A very simple version which uses a simple
8972 STACK for its trusted certificate store is also provided
8973 using X509_STORE_CTX_trusted_stack().
8975 The verify_cb() and verify() callbacks now have equivalents
8976 in the X509_STORE_CTX structure.
8978 X509_STORE_CTX also has a 'flags' field which can be used
8979 to customise the verify behaviour.
8982 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
8983 excludes S/MIME capabilities.
8986 *) When a certificate request is read in keep a copy of the
8987 original encoding of the signed data and use it when outputting
8988 again. Signatures then use the original encoding rather than
8989 a decoded, encoded version which may cause problems if the
8990 request is improperly encoded.
8993 *) For consistency with other BIO_puts implementations, call
8994 buffer_write(b, ...) directly in buffer_puts instead of calling
8997 In BIO_puts, increment b->num_write as in BIO_write.
8998 [Peter.Sylvester@EdelWeb.fr]
9000 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9001 BN_zero, we may not return a BIGNUM with an array consisting of
9005 *) Avoid calling abort() from within the library when problems are
9006 detected, except if preprocessor symbols have been defined
9007 (such as REF_CHECK, BN_DEBUG etc.).
9010 *) New openssl application 'rsautl'. This utility can be
9011 used for low level RSA operations. DER public key
9012 BIO/fp routines also added.
9015 *) New Configure entry and patches for compiling on QNX 4.
9016 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9018 *) A demo state-machine implementation was sponsored by
9019 Nuron (http://www.nuron.com/) and is now available in
9020 demos/state_machine.
9023 *) New options added to the 'dgst' utility for signature
9024 generation and verification.
9027 *) Unrecognized PKCS#7 content types are now handled via a
9028 catch all ASN1_TYPE structure. This allows unsupported
9029 types to be stored as a "blob" and an application can
9030 encode and decode it manually.
9033 *) Fix various signed/unsigned issues to make a_strex.c
9035 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9037 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9038 length if passed a buffer. ASN1_INTEGER_to_BN failed
9039 if passed a NULL BN and its argument was negative.
9040 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9042 *) Modification to PKCS#7 encoding routines to output definite
9043 length encoding. Since currently the whole structures are in
9044 memory there's not real point in using indefinite length
9045 constructed encoding. However if OpenSSL is compiled with
9046 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9049 *) Added BIO_vprintf() and BIO_vsnprintf().
9052 *) Added more prefixes to parse for in the the strings written
9053 through a logging bio, to cover all the levels that are available
9054 through syslog. The prefixes are now:
9056 PANIC, EMERG, EMR => LOG_EMERG
9057 ALERT, ALR => LOG_ALERT
9058 CRIT, CRI => LOG_CRIT
9059 ERROR, ERR => LOG_ERR
9060 WARNING, WARN, WAR => LOG_WARNING
9061 NOTICE, NOTE, NOT => LOG_NOTICE
9062 INFO, INF => LOG_INFO
9063 DEBUG, DBG => LOG_DEBUG
9065 and as before, if none of those prefixes are present at the
9066 beginning of the string, LOG_ERR is chosen.
9068 On Win32, the LOG_* levels are mapped according to this:
9070 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9071 LOG_WARNING => EVENTLOG_WARNING_TYPE
9072 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9076 *) Made it possible to reconfigure with just the configuration
9077 argument "reconf" or "reconfigure". The command line arguments
9078 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9079 and are retrieved from there when reconfiguring.
9083 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9085 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9088 *) The obj_dat.pl script was messing up the sorting of object
9089 names. The reason was that it compared the quoted version
9090 of strings as a result "OCSP" > "OCSP Signing" because
9091 " > SPACE. Changed script to store unquoted versions of
9092 names and add quotes on output. It was also omitting some
9093 names from the lookup table if they were given a default
9094 value (that is if SN is missing it is given the same
9095 value as LN and vice versa), these are now added on the
9096 grounds that if an object has a name we should be able to
9097 look it up. Finally added warning output when duplicate
9098 short or long names are found.
9101 *) Changes needed for Tandem NSK.
9102 [Scott Uroff <scott@xypro.com>]
9104 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9105 RSA_padding_check_SSLv23(), special padding was never detected
9106 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9107 version rollback attacks was not effective.
9109 In s23_clnt.c, don't use special rollback-attack detection padding
9110 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9111 client; similarly, in s23_srvr.c, don't do the rollback check if
9112 SSL 2.0 is the only protocol enabled in the server.
9115 *) Make it possible to get hexdumps of unprintable data with 'openssl
9116 asn1parse'. By implication, the functions ASN1_parse_dump() and
9117 BIO_dump_indent() are added.
9120 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9121 these print out strings and name structures based on various
9122 flags including RFC2253 support and proper handling of
9123 multibyte characters. Added options to the 'x509' utility
9124 to allow the various flags to be set.
9127 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9128 Also change the functions X509_cmp_current_time() and
9129 X509_gmtime_adj() work with an ASN1_TIME structure,
9130 this will enable certificates using GeneralizedTime in validity
9131 dates to be checked.
9134 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9135 negative public key encodings) on by default,
9136 NO_NEG_PUBKEY_BUG can be set to disable it.
9139 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9140 content octets. An i2c_ASN1_OBJECT is unnecessary because
9141 the encoding can be trivially obtained from the structure.
9144 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9145 not read locks (CRYPTO_r_[un]lock).
9148 *) A first attempt at creating official support for shared
9149 libraries through configuration. I've kept it so the
9150 default is static libraries only, and the OpenSSL programs
9151 are always statically linked for now, but there are
9152 preparations for dynamic linking in place.
9153 This has been tested on Linux and Tru64.
9156 *) Randomness polling function for Win9x, as described in:
9157 Peter Gutmann, Software Generation of Practically Strong
9161 *) Fix so PRNG is seeded in req if using an already existing
9165 *) New options to smime application. -inform and -outform
9166 allow alternative formats for the S/MIME message including
9167 PEM and DER. The -content option allows the content to be
9168 specified separately. This should allow things like Netscape
9169 form signing output easier to verify.
9172 *) Fix the ASN1 encoding of tags using the 'long form'.
9175 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9176 STRING types. These convert content octets to and from the
9177 underlying type. The actual tag and length octets are
9178 already assumed to have been read in and checked. These
9179 are needed because all other string types have virtually
9180 identical handling apart from the tag. By having versions
9181 of the ASN1 functions that just operate on content octets
9182 IMPLICIT tagging can be handled properly. It also allows
9183 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9184 and ASN1_INTEGER are identical apart from the tag.
9187 *) Change the handling of OID objects as follows:
9189 - New object identifiers are inserted in objects.txt, following
9190 the syntax given in objects.README.
9191 - objects.pl is used to process obj_mac.num and create a new
9193 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9196 This is currently kind of a hack, and the perl code in objects.pl
9197 isn't very elegant, but it works as I intended. The simplest way
9198 to check that it worked correctly is to look in obj_dat.h and
9199 check the array nid_objs and make sure the objects haven't moved
9200 around (this is important!). Additions are OK, as well as
9201 consistent name changes.
9204 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9207 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9208 The given file adds to whatever has already been seeded into the
9209 random pool through the RANDFILE configuration file option or
9210 environment variable, or the default random state file.
9213 *) mkstack.pl now sorts each macro group into lexical order.
9214 Previously the output order depended on the order the files
9215 appeared in the directory, resulting in needless rewriting
9219 *) Patches to make OpenSSL compile under Win32 again. Mostly
9220 work arounds for the VC++ problem that it treats func() as
9221 func(void). Also stripped out the parts of mkdef.pl that
9222 added extra typesafe functions: these no longer exist.
9225 *) Reorganisation of the stack code. The macros are now all
9226 collected in safestack.h . Each macro is defined in terms of
9227 a "stack macro" of the form SKM_<name>(type, a, b). The
9228 DEBUG_SAFESTACK is now handled in terms of function casts,
9229 this has the advantage of retaining type safety without the
9230 use of additional functions. If DEBUG_SAFESTACK is not defined
9231 then the non typesafe macros are used instead. Also modified the
9232 mkstack.pl script to handle the new form. Needs testing to see
9233 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9234 the default if no major problems. Similar behaviour for ASN1_SET_OF
9235 and PKCS12_STACK_OF.
9238 *) When some versions of IIS use the 'NET' form of private key the
9239 key derivation algorithm is different. Normally MD5(password) is
9240 used as a 128 bit RC4 key. In the modified case
9241 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9242 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9243 as the old Netscape_RSA functions except they have an additional
9244 'sgckey' parameter which uses the modified algorithm. Also added
9245 an -sgckey command line option to the rsa utility. Thanks to
9246 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9247 algorithm to openssl-dev.
9250 *) The evp_local.h macros were using 'c.##kname' which resulted in
9251 invalid expansion on some systems (SCO 5.0.5 for example).
9252 Corrected to 'c.kname'.
9253 [Phillip Porch <root@theporch.com>]
9255 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9256 a STACK of email addresses from a certificate or request, these look
9257 in the subject name and the subject alternative name extensions and
9258 omit any duplicate addresses.
9261 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9262 This makes DSA verification about 2 % faster.
9265 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9266 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9267 plus overhead for 1024 bit moduli).
9268 This makes exponentiations about 0.5 % faster for 1024 bit
9269 exponents (as measured by "openssl speed rsa2048").
9272 *) Rename memory handling macros to avoid conflicts with other
9274 Malloc => OPENSSL_malloc
9275 Malloc_locked => OPENSSL_malloc_locked
9276 Realloc => OPENSSL_realloc
9277 Free => OPENSSL_free
9280 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9281 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9284 *) CygWin32 support.
9285 [John Jarvie <jjarvie@newsguy.com>]
9287 *) The type-safe stack code has been rejigged. It is now only compiled
9288 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9289 by default all type-specific stack functions are "#define"d back to
9290 standard stack functions. This results in more streamlined output
9291 but retains the type-safety checking possibilities of the original
9295 *) The STACK code has been cleaned up, and certain type declarations
9296 that didn't make a lot of sense have been brought in line. This has
9297 also involved a cleanup of sorts in safestack.h to more correctly
9298 map type-safe stack functions onto their plain stack counterparts.
9299 This work has also resulted in a variety of "const"ifications of
9300 lots of the code, especially "_cmp" operations which should normally
9301 be prototyped with "const" parameters anyway.
9304 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9305 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9306 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9307 where all of 'md' is used each time the PRNG is used, but 'state'
9308 is used only indexed by a cyclic counter. As entropy may not be
9309 well distributed from the beginning, 'md' is important as a
9310 chaining variable. However, the output function chains only half
9311 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9312 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9313 in all of 'state' being rewritten, with the new values depending
9314 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9317 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9318 the handshake is continued after ssl_verify_cert_chain();
9319 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9320 can lead to 'unexplainable' connection aborts later.
9321 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9323 *) Major EVP API cipher revision.
9324 Add hooks for extra EVP features. This allows various cipher
9325 parameters to be set in the EVP interface. Support added for variable
9326 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9327 setting of RC2 and RC5 parameters.
9329 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9332 Remove lots of duplicated code from the EVP library. For example *every*
9333 cipher init() function handles the 'iv' in the same way according to the
9334 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9335 for CFB and OFB modes they zero ctx->num.
9337 New functionality allows removal of S/MIME code RC2 hack.
9339 Most of the routines have the same form and so can be declared in terms
9342 By shifting this to the top level EVP_CipherInit() it can be removed from
9343 all individual ciphers. If the cipher wants to handle IVs or keys
9344 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9347 Change lots of functions like EVP_EncryptUpdate() to now return a
9348 value: although software versions of the algorithms cannot fail
9349 any installed hardware versions can.
9352 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9353 this option is set, tolerate broken clients that send the negotiated
9354 protocol version number instead of the requested protocol version
9358 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9359 i.e. non-zero for export ciphersuites, zero otherwise.
9360 Previous versions had this flag inverted, inconsistent with
9361 rsa_tmp_cb (..._TMP_RSA_CB).
9362 [Bodo Moeller; problem reported by Amit Chopra]
9364 *) Add missing DSA library text string. Work around for some IIS
9365 key files with invalid SEQUENCE encoding.
9368 *) Add a document (doc/standards.txt) that list all kinds of standards
9369 and so on that are implemented in OpenSSL.
9372 *) Enhance c_rehash script. Old version would mishandle certificates
9373 with the same subject name hash and wouldn't handle CRLs at all.
9374 Added -fingerprint option to crl utility, to support new c_rehash
9378 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9381 *) Fix for SSL server purpose checking. Server checking was
9382 rejecting certificates which had extended key usage present
9383 but no ssl client purpose.
9384 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9386 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9387 is a little unclear about how a blank password is handled.
9388 Since the password in encoded as a BMPString with terminating
9389 double NULL a zero length password would end up as just the
9390 double NULL. However no password at all is different and is
9391 handled differently in the PKCS#12 key generation code. NS
9392 treats a blank password as zero length. MSIE treats it as no
9393 password on export: but it will try both on import. We now do
9394 the same: PKCS12_parse() tries zero length and no password if
9395 the password is set to "" or NULL (NULL is now a valid password:
9396 it wasn't before) as does the pkcs12 application.
9399 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9400 perror when PEM_read_bio_X509_REQ fails, the error message must
9401 be obtained from the error queue.
9404 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9405 it in ERR_remove_state if appropriate, and change ERR_get_state
9406 accordingly to avoid race conditions (this is necessary because
9407 thread_hash is no longer constant once set).
9410 *) Bugfix for linux-elf makefile.one.
9413 *) RSA_get_default_method() will now cause a default
9414 RSA_METHOD to be chosen if one doesn't exist already.
9415 Previously this was only set during a call to RSA_new()
9416 or RSA_new_method(NULL) meaning it was possible for
9417 RSA_get_default_method() to return NULL.
9420 *) Added native name translation to the existing DSO code
9421 that will convert (if the flag to do so is set) filenames
9422 that are sufficiently small and have no path information
9423 into a canonical native form. Eg. "blah" converted to
9424 "libblah.so" or "blah.dll" etc.
9427 *) New function ERR_error_string_n(e, buf, len) which is like
9428 ERR_error_string(e, buf), but writes at most 'len' bytes
9429 including the 0 terminator. For ERR_error_string_n, 'buf'
9431 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9433 *) CONF library reworked to become more general. A new CONF
9434 configuration file reader "class" is implemented as well as a
9435 new functions (NCONF_*, for "New CONF") to handle it. The now
9436 old CONF_* functions are still there, but are reimplemented to
9437 work in terms of the new functions. Also, a set of functions
9438 to handle the internal storage of the configuration data is
9439 provided to make it easier to write new configuration file
9440 reader "classes" (I can definitely see something reading a
9441 configuration file in XML format, for example), called _CONF_*,
9442 or "the configuration storage API"...
9444 The new configuration file reading functions are:
9446 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9447 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9449 NCONF_default, NCONF_WIN32
9451 NCONF_dump_fp, NCONF_dump_bio
9453 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9454 NCONF_new creates a new CONF object. This works in the same way
9455 as other interfaces in OpenSSL, like the BIO interface.
9456 NCONF_dump_* dump the internal storage of the configuration file,
9457 which is useful for debugging. All other functions take the same
9458 arguments as the old CONF_* functions wth the exception of the
9459 first that must be a `CONF *' instead of a `LHASH *'.
9461 To make it easer to use the new classes with the old CONF_* functions,
9462 the function CONF_set_default_method is provided.
9465 *) Add '-tls1' option to 'openssl ciphers', which was already
9466 mentioned in the documentation but had not been implemented.
9467 (This option is not yet really useful because even the additional
9468 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9471 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9472 OpenSSL-based applications) load shared libraries and bind to
9473 them in a portable way.
9474 [Geoff Thorpe, with contributions from Richard Levitte]
9476 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9478 *) Make sure _lrotl and _lrotr are only used with MSVC.
9480 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9481 (the default implementation of RAND_status).
9483 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9484 to '-clrext' (= clear extensions), as intended and documented.
9485 [Bodo Moeller; inconsistency pointed out by Michael Attili
9488 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9489 was larger than the MD block size.
9490 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9492 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9493 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9494 using the passed key: if the passed key was a private key the result
9495 of X509_print(), for example, would be to print out all the private key
9499 *) des_quad_cksum() byte order bug fix.
9500 [Ulf Möller, using the problem description in krb4-0.9.7, where
9501 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9503 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9505 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9507 *) For easily testing in shell scripts whether some command
9508 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9509 returns with exit code 0 iff no command of the given name is available.
9510 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9511 the output goes to stdout and nothing is printed to stderr.
9512 Additional arguments are always ignored.
9514 Since for each cipher there is a command of the same name,
9515 the 'no-cipher' compilation switches can be tested this way.
9517 ('openssl no-XXX' is not able to detect pseudo-commands such
9518 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9521 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9524 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9525 is set; it will be thrown away anyway because each handshake creates
9527 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9528 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9529 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9530 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9533 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9534 'Q' and 'R' lose their special meanings (quit/renegotiate).
9535 This is part of what -quiet does; unlike -quiet, -ign_eof
9536 does not suppress any output.
9539 *) Add compatibility options to the purpose and trust code. The
9540 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9541 accepts a certificate or CA, this was the previous behaviour,
9542 with all the associated security issues.
9544 X509_TRUST_COMPAT is the old trust behaviour: only and
9545 automatically trust self signed roots in certificate store. A
9546 new trust setting X509_TRUST_DEFAULT is used to specify that
9547 a purpose has no associated trust setting and it should instead
9548 use the value in the default purpose.
9551 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9552 and fix a memory leak.
9555 *) In util/mkerr.pl (which implements 'make errors'), preserve
9556 reason strings from the previous version of the .c file, as
9557 the default to have only downcase letters (and digits) in
9558 automatically generated reasons codes is not always appropriate.
9561 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9562 using strerror. Previously, ERR_reason_error_string() returned
9563 library names as reason strings for SYSerr; but SYSerr is a special
9564 case where small numbers are errno values, not library numbers.
9567 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9568 converts DSA parameters into DH parameters. (When creating parameters,
9569 DSA_generate_parameters is used.)
9572 *) Include 'length' (recommended exponent length) in C code generated
9573 by 'openssl dhparam -C'.
9576 *) The second argument to set_label in perlasm was already being used
9577 so couldn't be used as a "file scope" flag. Moved to third argument
9581 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9582 instead of RAND_bytes for encryption IVs and salts.
9585 *) Include RAND_status() into RAND_METHOD instead of implementing
9586 it only for md_rand.c Otherwise replacing the PRNG by calling
9587 RAND_set_rand_method would be impossible.
9590 *) Don't let DSA_generate_key() enter an infinite loop if the random
9591 number generation fails.
9594 *) New 'rand' application for creating pseudo-random output.
9597 *) Added configuration support for Linux/IA64
9598 [Rolf Haberrecker <rolf@suse.de>]
9600 *) Assembler module support for Mingw32.
9603 *) Shared library support for HPUX (in shlib/).
9604 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9606 *) Shared library support for Solaris gcc.
9607 [Lutz Behnke <behnke@trustcenter.de>]
9609 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9611 *) PKCS7_encrypt() was adding text MIME headers twice because they
9612 were added manually and by SMIME_crlf_copy().
9615 *) In bntest.c don't call BN_rand with zero bits argument.
9616 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9618 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9619 case was implemented. This caused BN_div_recp() to fail occasionally.
9622 *) Add an optional second argument to the set_label() in the perl
9623 assembly language builder. If this argument exists and is set
9624 to 1 it signals that the assembler should use a symbol whose
9625 scope is the entire file, not just the current function. This
9626 is needed with MASM which uses the format label:: for this scope.
9627 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9629 *) Change the ASN1 types so they are typedefs by default. Before
9630 almost all types were #define'd to ASN1_STRING which was causing
9631 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9635 *) Change names of new functions to the new get1/get0 naming
9636 convention: After 'get1', the caller owns a reference count
9637 and has to call ..._free; 'get0' returns a pointer to some
9638 data structure without incrementing reference counters.
9639 (Some of the existing 'get' functions increment a reference
9640 counter, some don't.)
9641 Similarly, 'set1' and 'add1' functions increase reference
9642 counters or duplicate objects.
9645 *) Allow for the possibility of temp RSA key generation failure:
9646 the code used to assume it always worked and crashed on failure.
9649 *) Fix potential buffer overrun problem in BIO_printf().
9650 [Ulf Möller, using public domain code by Patrick Powell; problem
9651 pointed out by David Sacerdote <das33@cornell.edu>]
9653 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9654 RAND_egd() and RAND_status(). In the command line application,
9655 the EGD socket can be specified like a seed file using RANDFILE
9659 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9660 Some CAs (e.g. Verisign) distribute certificates in this form.
9663 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9664 list to exclude them. This means that no special compilation option
9665 is needed to use anonymous DH: it just needs to be included in the
9669 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9670 EVP_MD_type. The old functionality is available in a new macro called
9671 EVP_MD_md(). Change code that uses it and update docs.
9674 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9675 where the 'void *' argument is replaced by a function pointer argument.
9676 Previously 'void *' was abused to point to functions, which works on
9677 many platforms, but is not correct. As these functions are usually
9678 called by macros defined in OpenSSL header files, most source code
9679 should work without changes.
9682 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9683 sections with information on -D... compiler switches used for
9684 compiling the library so that applications can see them. To enable
9685 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9686 must be defined. E.g.,
9687 #define OPENSSL_ALGORITHM_DEFINES
9688 #include <openssl/opensslconf.h>
9689 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9690 [Richard Levitte, Ulf and Bodo Möller]
9692 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9696 *) Change the 'other' type in certificate aux info to a STACK_OF
9697 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9698 the required ASN1 format: arbitrary types determined by an OID.
9701 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9702 argument to 'req'. This is not because the function is newer or
9703 better than others it just uses the work 'NEW' in the certificate
9704 request header lines. Some software needs this.
9707 *) Reorganise password command line arguments: now passwords can be
9708 obtained from various sources. Delete the PEM_cb function and make
9709 it the default behaviour: i.e. if the callback is NULL and the
9710 usrdata argument is not NULL interpret it as a null terminated pass
9711 phrase. If usrdata and the callback are NULL then the pass phrase
9712 is prompted for as usual.
9715 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9716 the support is automatically enabled. The resulting binaries will
9717 autodetect the card and use it if present.
9718 [Ben Laurie and Compaq Inc.]
9720 *) Work around for Netscape hang bug. This sends certificate request
9721 and server done in one record. Since this is perfectly legal in the
9722 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9723 the bugs/SSLv3 entry for more info.
9726 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9729 *) Add -rand argument to smime and pkcs12 applications and read/write
9733 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9736 *) Add command line password options to the remaining applications.
9739 *) Bug fix for BN_div_recp() for numerators with an even number of
9743 *) More tests in bntest.c, and changed test_bn output.
9746 *) ./config recognizes MacOS X now.
9749 *) Bug fix for BN_div() when the first words of num and divsor are
9750 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9753 *) Add support for various broken PKCS#8 formats, and command line
9754 options to produce them.
9757 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9758 get temporary BIGNUMs from a BN_CTX.
9761 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9765 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9766 include a #define from the old name to the new. The original intent
9767 was that statically linked binaries could for example just call
9768 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9769 link with digests. This never worked because SSLeay_add_all_digests()
9770 and SSLeay_add_all_ciphers() were in the same source file so calling
9771 one would link with the other. They are now in separate source files.
9774 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9777 *) Use a less unusual form of the Miller-Rabin primality test (it used
9778 a binary algorithm for exponentiation integrated into the Miller-Rabin
9779 loop, our standard modexp algorithms are faster).
9782 *) Support for the EBCDIC character set completed.
9783 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9785 *) Source code cleanups: use const where appropriate, eliminate casts,
9786 use void * instead of char * in lhash.
9789 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9790 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9791 this the server could overwrite ephemeral keys that the client
9795 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9796 using 50 iterations of the Rabin-Miller test.
9798 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9799 iterations of the Rabin-Miller test as required by the appendix
9800 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9801 As BN_is_prime_fasttest includes trial division, DSA parameter
9802 generation becomes much faster.
9804 This implies a change for the callback functions in DSA_is_prime
9805 and DSA_generate_parameters: The callback function is called once
9806 for each positive witness in the Rabin-Miller test, not just
9807 occasionally in the inner loop; and the parameters to the
9808 callback function now provide an iteration count for the outer
9809 loop rather than for the current invocation of the inner loop.
9810 DSA_generate_parameters additionally can call the callback
9811 function with an 'iteration count' of -1, meaning that a
9812 candidate has passed the trial division test (when q is generated
9813 from an application-provided seed, trial division is skipped).
9816 *) New function BN_is_prime_fasttest that optionally does trial
9817 division before starting the Rabin-Miller test and has
9818 an additional BN_CTX * argument (whereas BN_is_prime always
9819 has to allocate at least one BN_CTX).
9820 'callback(1, -1, cb_arg)' is called when a number has passed the
9821 trial division stage.
9824 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9828 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9831 *) New function BN_pseudo_rand().
9834 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9835 bignum version of BN_from_montgomery() with the working code from
9836 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9840 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9841 made it impossible to use the same SSL_SESSION data structure in
9842 SSL2 clients in multiple threads.
9845 *) The return value of RAND_load_file() no longer counts bytes obtained
9846 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9847 to seed the PRNG (previously an explicit byte count was required).
9848 [Ulf Möller, Bodo Möller]
9850 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9851 used (char *) instead of (void *) and had casts all over the place.
9854 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9857 *) Retain source code compatibility for BN_prime_checks macro:
9858 BN_is_prime(..., BN_prime_checks, ...) now uses
9859 BN_prime_checks_for_size to determine the appropriate number of
9860 Rabin-Miller iterations.
9863 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9864 DH_CHECK_P_NOT_SAFE_PRIME.
9865 (Check if this is true? OpenPGP calls them "strong".)
9868 *) Merge the functionality of "dh" and "gendh" programs into a new program
9869 "dhparam". The old programs are retained for now but will handle DH keys
9870 (instead of parameters) in future.
9873 *) Make the ciphers, s_server and s_client programs check the return values
9874 when a new cipher list is set.
9877 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9878 ciphers. Before when the 56bit ciphers were enabled the sorting was
9881 The syntax for the cipher sorting has been extended to support sorting by
9882 cipher-strength (using the strength_bits hard coded in the tables).
9883 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9885 Fix a bug in the cipher-command parser: when supplying a cipher command
9886 string with an "undefined" symbol (neither command nor alphanumeric
9887 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9888 an error is flagged.
9890 Due to the strength-sorting extension, the code of the
9891 ssl_create_cipher_list() function was completely rearranged. I hope that
9892 the readability was also increased :-)
9893 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9895 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9896 for the first serial number and places 2 in the serial number file. This
9897 avoids problems when the root CA is created with serial number zero and
9898 the first user certificate has the same issuer name and serial number
9902 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9903 the new code. Add documentation for this stuff.
9906 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9907 X509_*() to X509at_*() on the grounds that they don't handle X509
9908 structures and behave in an analogous way to the X509v3 functions:
9909 they shouldn't be called directly but wrapper functions should be used
9912 So we also now have some wrapper functions that call the X509at functions
9913 when passed certificate requests. (TO DO: similar things can be done with
9914 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9915 things. Some of these need some d2i or i2d and print functionality
9916 because they handle more complex structures.)
9919 *) Add missing #ifndefs that caused missing symbols when building libssl
9920 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9921 NO_RSA in ssl/s2*.c.
9922 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9924 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9925 has a return value which indicates the quality of the random data
9926 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9927 error queue. New function RAND_pseudo_bytes() generates output that is
9928 guaranteed to be unique but not unpredictable. RAND_add is like
9929 RAND_seed, but takes an extra argument for an entropy estimate
9930 (RAND_seed always assumes full entropy).
9933 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9934 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9935 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9936 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9937 false-positive rate of at most 2^-80 for random input.
9940 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9943 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9944 in the 0.9.5 release), this returns the chain
9945 from an X509_CTX structure with a dup of the stack and all
9946 the X509 reference counts upped: so the stack will exist
9947 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9950 Also make SSL_SESSION_print() print out the verify return
9954 *) Add manpage for the pkcs12 command. Also change the default
9955 behaviour so MAC iteration counts are used unless the new
9956 -nomaciter option is used. This improves file security and
9957 only older versions of MSIE (4.0 for example) need it.
9960 *) Honor the no-xxx Configure options when creating .DEF files.
9963 *) Add PKCS#10 attributes to field table: challengePassword,
9964 unstructuredName and unstructuredAddress. These are taken from
9965 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9966 international characters are used.
9968 More changes to X509_ATTRIBUTE code: allow the setting of types
9969 based on strings. Remove the 'loc' parameter when adding
9970 attributes because these will be a SET OF encoding which is sorted
9974 *) Initial changes to the 'req' utility to allow request generation
9975 automation. This will allow an application to just generate a template
9976 file containing all the field values and have req construct the
9979 Initial support for X509_ATTRIBUTE handling. Stacks of these are
9980 used all over the place including certificate requests and PKCS#7
9981 structures. They are currently handled manually where necessary with
9982 some primitive wrappers for PKCS#7. The new functions behave in a
9983 manner analogous to the X509 extension functions: they allow
9984 attributes to be looked up by NID and added.
9986 Later something similar to the X509V3 code would be desirable to
9987 automatically handle the encoding, decoding and printing of the
9988 more complex types. The string types like challengePassword can
9989 be handled by the string table functions.
9991 Also modified the multi byte string table handling. Now there is
9992 a 'global mask' which masks out certain types. The table itself
9993 can use the flag STABLE_NO_MASK to ignore the mask setting: this
9994 is useful when for example there is only one permissible type
9995 (as in countryName) and using the mask might result in no valid
9999 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10000 SSL_get_peer_finished to allow applications to obtain the latest
10001 Finished messages sent to the peer or expected from the peer,
10002 respectively. (SSL_get_peer_finished is usually the Finished message
10003 actually received from the peer, otherwise the protocol will be aborted.)
10005 As the Finished message are message digests of the complete handshake
10006 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10007 be used for external authentication procedures when the authentication
10008 provided by SSL/TLS is not desired or is not enough.
10011 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10012 the host supports BWX extension and if Compaq C is present on the
10013 $PATH. Just exploiting of the BWX extension results in 20-30%
10014 performance kick for some algorithms, e.g. DES and RC4 to mention
10015 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10019 *) Add support for MS "fast SGC". This is arguably a violation of the
10020 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10021 weak crypto and after checking the certificate is SGC a second one
10022 with strong crypto. MS SGC stops the first handshake after receiving
10023 the server certificate message and sends a second client hello. Since
10024 a server will typically do all the time consuming operations before
10025 expecting any further messages from the client (server key exchange
10026 is the most expensive) there is little difference between the two.
10028 To get OpenSSL to support MS SGC we have to permit a second client
10029 hello message after we have sent server done. In addition we have to
10030 reset the MAC if we do get this second client hello.
10033 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10034 if a DER encoded private key is RSA or DSA traditional format. Changed
10035 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10036 format DER encoded private key. Newer code should use PKCS#8 format which
10037 has the key type encoded in the ASN1 structure. Added DER private key
10038 support to pkcs8 application.
10041 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10042 ciphersuites has been selected (as required by the SSL 3/TLS 1
10043 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10044 is set, we interpret this as a request to violate the specification
10045 (the worst that can happen is a handshake failure, and 'correct'
10046 behaviour would result in a handshake failure anyway).
10049 *) In SSL_CTX_add_session, take into account that there might be multiple
10050 SSL_SESSION structures with the same session ID (e.g. when two threads
10051 concurrently obtain them from an external cache).
10052 The internal cache can handle only one SSL_SESSION with a given ID,
10053 so if there's a conflict, we now throw out the old one to achieve
10057 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10058 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10059 some routines that use cipher OIDs: some ciphers do not have OIDs
10060 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10064 *) Simplify the trust setting structure and code. Now we just have
10065 two sequences of OIDs for trusted and rejected settings. These will
10066 typically have values the same as the extended key usage extension
10067 and any application specific purposes.
10069 The trust checking code now has a default behaviour: it will just
10070 check for an object with the same NID as the passed id. Functions can
10071 be provided to override either the default behaviour or the behaviour
10072 for a given id. SSL client, server and email already have functions
10073 in place for compatibility: they check the NID and also return "trusted"
10074 if the certificate is self signed.
10077 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10078 traditional format into an EVP_PKEY structure.
10081 *) Add a password callback function PEM_cb() which either prompts for
10082 a password if usr_data is NULL or otherwise assumes it is a null
10083 terminated password. Allow passwords to be passed on command line
10084 environment or config files in a few more utilities.
10087 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10088 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10089 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10090 Update documentation.
10093 *) Support for ASN1 "NULL" type. This could be handled before by using
10094 ASN1_TYPE but there wasn't any function that would try to read a NULL
10095 and produce an error if it couldn't. For compatibility we also have
10096 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10097 don't allocate anything because they don't need to.
10100 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10102 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10104 *) Rebuild of the memory allocation routines used by OpenSSL code and
10105 possibly others as well. The purpose is to make an interface that
10106 provide hooks so anyone can build a separate set of allocation and
10107 deallocation routines to be used by OpenSSL, for example memory
10108 pool implementations, or something else, which was previously hard
10109 since Malloc(), Realloc() and Free() were defined as macros having
10110 the values malloc, realloc and free, respectively (except for Win32
10111 compilations). The same is provided for memory debugging code.
10112 OpenSSL already comes with functionality to find memory leaks, but
10113 this gives people a chance to debug other memory problems.
10115 With these changes, a new set of functions and macros have appeared:
10117 CRYPTO_set_mem_debug_functions() [F]
10118 CRYPTO_get_mem_debug_functions() [F]
10119 CRYPTO_dbg_set_options() [F]
10120 CRYPTO_dbg_get_options() [F]
10121 CRYPTO_malloc_debug_init() [M]
10123 The memory debug functions are NULL by default, unless the library
10124 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10125 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10126 gives the standard debugging functions that come with OpenSSL) or
10127 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10128 provided by the library user) must be used. When the standard
10129 debugging functions are used, CRYPTO_dbg_set_options can be used to
10130 request additional information:
10131 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10132 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10134 Also, things like CRYPTO_set_mem_functions will always give the
10135 expected result (the new set of functions is used for allocation
10136 and deallocation) at all times, regardless of platform and compiler
10139 To finish it up, some functions that were never use in any other
10140 way than through macros have a new API and new semantic:
10142 CRYPTO_dbg_malloc()
10143 CRYPTO_dbg_realloc()
10146 All macros of value have retained their old syntax.
10147 [Richard Levitte and Bodo Moeller]
10149 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10150 ordering of SMIMECapabilities wasn't in "strength order" and there
10151 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10155 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10156 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10157 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10159 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10160 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10161 functionality to handle multipart/signed properly) and a utility
10162 called 'smime' to call all this stuff. This is based on code I
10163 originally wrote for Celo who have kindly allowed it to be
10164 included in OpenSSL.
10167 *) Add variants des_set_key_checked and des_set_key_unchecked of
10168 des_set_key (aka des_key_sched). Global variable des_check_key
10169 decides which of these is called by des_set_key; this way
10170 des_check_key behaves as it always did, but applications and
10171 the library itself, which was buggy for des_check_key == 1,
10172 have a cleaner way to pick the version they need.
10175 *) New function PKCS12_newpass() which changes the password of a
10179 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10180 dynamic mix. In both cases the ids can be used as an index into the
10181 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10182 functions so they accept a list of the field values and the
10183 application doesn't need to directly manipulate the X509_TRUST
10187 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10191 *) Modify the way the V3 extension code looks up extensions. This now
10192 works in a similar way to the object code: we have some "standard"
10193 extensions in a static table which is searched with OBJ_bsearch()
10194 and the application can add dynamic ones if needed. The file
10195 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10196 updated whenever a new extension is added to the core code and kept
10197 in ext_nid order. There is a simple program 'tabtest.c' which checks
10198 this. New extensions are not added too often so this file can readily
10199 be maintained manually.
10201 There are two big advantages in doing things this way. The extensions
10202 can be looked up immediately and no longer need to be "added" using
10203 X509V3_add_standard_extensions(): this function now does nothing.
10204 [Side note: I get *lots* of email saying the extension code doesn't
10205 work because people forget to call this function]
10206 Also no dynamic allocation is done unless new extensions are added:
10207 so if we don't add custom extensions there is no need to call
10208 X509V3_EXT_cleanup().
10211 *) Modify enc utility's salting as follows: make salting the default. Add a
10212 magic header, so unsalted files fail gracefully instead of just decrypting
10213 to garbage. This is because not salting is a big security hole, so people
10214 should be discouraged from doing it.
10217 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10218 digest to be passed on the command line but it only used this
10219 parameter when signing a certificate. Modified so all relevant
10220 operations are affected by the digest parameter including the
10221 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10222 DSA key was used because it didn't fix the digest.
10225 *) Initial certificate chain verify code. Currently tests the untrusted
10226 certificates for consistency with the verify purpose (which is set
10227 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10229 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10230 this is because it will reject chains with invalid extensions whereas
10231 every previous version of OpenSSL and SSLeay made no checks at all.
10233 Trust code: checks the root CA for the relevant trust settings. Trust
10234 settings have an initial value consistent with the verify purpose: e.g.
10235 if the verify purpose is for SSL client use it expects the CA to be
10236 trusted for SSL client use. However the default value can be changed to
10237 permit custom trust settings: one example of this would be to only trust
10238 certificates from a specific "secure" set of CAs.
10240 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10241 which should be used for version portability: especially since the
10242 verify structure is likely to change more often now.
10244 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10245 to set them. If not set then assume SSL clients will verify SSL servers
10248 Two new options to the verify program: -untrusted allows a set of
10249 untrusted certificates to be passed in and -purpose which sets the
10250 intended purpose of the certificate. If a purpose is set then the
10251 new chain verify code is used to check extension consistency.
10254 *) Support for the authority information access extension.
10257 *) Modify RSA and DSA PEM read routines to transparently handle
10258 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10259 public keys in a format compatible with certificate
10260 SubjectPublicKeyInfo structures. Unfortunately there were already
10261 functions called *_PublicKey_* which used various odd formats so
10262 these are retained for compatibility: however the DSA variants were
10263 never in a public release so they have been deleted. Changed dsa/rsa
10264 utilities to handle the new format: note no releases ever handled public
10265 keys so we should be OK.
10267 The primary motivation for this change is to avoid the same fiasco
10268 that dogs private keys: there are several incompatible private key
10269 formats some of which are standard and some OpenSSL specific and
10270 require various evil hacks to allow partial transparent handling and
10271 even then it doesn't work with DER formats. Given the option anything
10272 other than PKCS#8 should be dumped: but the other formats have to
10273 stay in the name of compatibility.
10275 With public keys and the benefit of hindsight one standard format
10276 is used which works with EVP_PKEY, RSA or DSA structures: though
10277 it clearly returns an error if you try to read the wrong kind of key.
10279 Added a -pubkey option to the 'x509' utility to output the public key.
10280 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10281 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10282 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10283 that do the same as the EVP_PKEY_assign_*() except they up the
10284 reference count of the added key (they don't "swallow" the
10288 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10289 CRLs would fail if the file contained no certificates or no CRLs:
10290 added a new function to read in both types and return the number
10291 read: this means that if none are read it will be an error. The
10292 DER versions of the certificate and CRL reader would always fail
10293 because it isn't possible to mix certificates and CRLs in DER format
10294 without choking one or the other routine. Changed this to just read
10295 a certificate: this is the best we can do. Also modified the code
10296 in apps/verify.c to take notice of return codes: it was previously
10297 attempting to read in certificates from NULL pointers and ignoring
10298 any errors: this is one reason why the cert and CRL reader seemed
10299 to work. It doesn't check return codes from the default certificate
10300 routines: these may well fail if the certificates aren't installed.
10303 *) Code to support otherName option in GeneralName.
10306 *) First update to verify code. Change the verify utility
10307 so it warns if it is passed a self signed certificate:
10308 for consistency with the normal behaviour. X509_verify
10309 has been modified to it will now verify a self signed
10310 certificate if *exactly* the same certificate appears
10311 in the store: it was previously impossible to trust a
10312 single self signed certificate. This means that:
10313 openssl verify ss.pem
10314 now gives a warning about a self signed certificate but
10315 openssl verify -CAfile ss.pem ss.pem
10319 *) For servers, store verify_result in SSL_SESSION data structure
10320 (and add it to external session representation).
10321 This is needed when client certificate verifications fails,
10322 but an application-provided verification callback (set by
10323 SSL_CTX_set_cert_verify_callback) allows accepting the session
10324 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10325 but returns 1): When the session is reused, we have to set
10326 ssl->verify_result to the appropriate error code to avoid
10328 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10330 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10331 case in PKCS7_dataInit() where the signed PKCS7 structure
10332 didn't contain any existing data because it was being created.
10333 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10335 *) Add a salt to the key derivation routines in enc.c. This
10336 forms the first 8 bytes of the encrypted file. Also add a
10337 -S option to allow a salt to be input on the command line.
10340 *) New function X509_cmp(). Oddly enough there wasn't a function
10341 to compare two certificates. We do this by working out the SHA1
10342 hash and comparing that. X509_cmp() will be needed by the trust
10346 *) SSL_get1_session() is like SSL_get_session(), but increments
10347 the reference count in the SSL_SESSION returned.
10348 [Geoff Thorpe <geoff@eu.c2.net>]
10350 *) Fix for 'req': it was adding a null to request attributes.
10351 Also change the X509_LOOKUP and X509_INFO code to handle
10352 certificate auxiliary information.
10355 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10359 *) Add the possibility to add extra information to the memory leak
10360 detecting output, to form tracebacks, showing from where each
10361 allocation was originated: CRYPTO_push_info("constant string") adds
10362 the string plus current file name and line number to a per-thread
10363 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10364 is like calling CYRPTO_pop_info() until the stack is empty.
10365 Also updated memory leak detection code to be multi-thread-safe.
10368 *) Add options -text and -noout to pkcs7 utility and delete the
10369 encryption options which never did anything. Update docs.
10372 *) Add options to some of the utilities to allow the pass phrase
10373 to be included on either the command line (not recommended on
10374 OSes like Unix) or read from the environment. Update the
10375 manpages and fix a few bugs.
10378 *) Add a few manpages for some of the openssl commands.
10381 *) Fix the -revoke option in ca. It was freeing up memory twice,
10382 leaking and not finding already revoked certificates.
10385 *) Extensive changes to support certificate auxiliary information.
10386 This involves the use of X509_CERT_AUX structure and X509_AUX
10387 functions. An X509_AUX function such as PEM_read_X509_AUX()
10388 can still read in a certificate file in the usual way but it
10389 will also read in any additional "auxiliary information". By
10390 doing things this way a fair degree of compatibility can be
10391 retained: existing certificates can have this information added
10392 using the new 'x509' options.
10394 Current auxiliary information includes an "alias" and some trust
10395 settings. The trust settings will ultimately be used in enhanced
10396 certificate chain verification routines: currently a certificate
10397 can only be trusted if it is self signed and then it is trusted
10401 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10402 The problem was that one of the replacement routines had not been working
10403 since SSLeay releases. For now the offending routine has been replaced
10404 with non-optimised assembler. Even so, this now gives around 95%
10405 performance improvement for 1024 bit RSA signs.
10408 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10409 handling. Most clients have the effective key size in bits equal to
10410 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10411 A few however don't do this and instead use the size of the decrypted key
10412 to determine the RC2 key length and the AlgorithmIdentifier to determine
10413 the effective key length. In this case the effective key length can still
10414 be 40 bits but the key length can be 168 bits for example. This is fixed
10415 by manually forcing an RC2 key into the EVP_PKEY structure because the
10416 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10417 the key length and effective key length are equal.
10420 *) Add a bunch of functions that should simplify the creation of
10421 X509_NAME structures. Now you should be able to do:
10422 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10423 and have it automatically work out the correct field type and fill in
10424 the structures. The more adventurous can try:
10425 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10426 and it will (hopefully) work out the correct multibyte encoding.
10429 *) Change the 'req' utility to use the new field handling and multibyte
10430 copy routines. Before the DN field creation was handled in an ad hoc
10431 way in req, ca, and x509 which was rather broken and didn't support
10432 BMPStrings or UTF8Strings. Since some software doesn't implement
10433 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10434 using the dirstring_type option. See the new comment in the default
10435 openssl.cnf for more info.
10438 *) Make crypto/rand/md_rand.c more robust:
10439 - Assure unique random numbers after fork().
10440 - Make sure that concurrent threads access the global counter and
10441 md serializably so that we never lose entropy in them
10442 or use exactly the same state in multiple threads.
10443 Access to the large state is not always serializable because
10444 the additional locking could be a performance killer, and
10445 md should be large enough anyway.
10448 *) New file apps/app_rand.c with commonly needed functionality
10449 for handling the random seed file.
10451 Use the random seed file in some applications that previously did not:
10453 dsaparam -genkey (which also ignored its '-rand' option),
10456 x509 (when signing).
10457 Except on systems with /dev/urandom, it is crucial to have a random
10458 seed file at least for key creation, DSA signing, and for DH exchanges;
10459 for RSA signatures we could do without one.
10461 gendh and gendsa (unlike genrsa) used to read only the first byte
10462 of each file listed in the '-rand' option. The function as previously
10463 found in genrsa is now in app_rand.c and is used by all programs
10464 that support '-rand'.
10467 *) In RAND_write_file, use mode 0600 for creating files;
10468 don't just chmod when it may be too late.
10471 *) Report an error from X509_STORE_load_locations
10472 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10475 *) New function ASN1_mbstring_copy() this copies a string in either
10476 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10477 into an ASN1_STRING type. A mask of permissible types is passed
10478 and it chooses the "minimal" type to use or an error if not type
10482 *) Add function equivalents to the various macros in asn1.h. The old
10483 macros are retained with an M_ prefix. Code inside the library can
10484 use the M_ macros. External code (including the openssl utility)
10485 should *NOT* in order to be "shared library friendly".
10488 *) Add various functions that can check a certificate's extensions
10489 to see if it usable for various purposes such as SSL client,
10490 server or S/MIME and CAs of these types. This is currently
10491 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10492 verification. Also added a -purpose flag to x509 utility to
10493 print out all the purposes.
10496 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10500 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10501 for, obtain and decode and extension and obtain its critical flag.
10502 This allows all the necessary extension code to be handled in a
10503 single function call.
10506 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10507 platforms. See crypto/rc4/rc4_enc.c for further details.
10510 *) New -noout option to asn1parse. This causes no output to be produced
10511 its main use is when combined with -strparse and -out to extract data
10512 from a file (which may not be in ASN.1 format).
10515 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10516 when producing the local key id.
10517 [Richard Levitte <levitte@stacken.kth.se>]
10519 *) New option -dhparam in s_server. This allows a DH parameter file to be
10520 stated explicitly. If it is not stated then it tries the first server
10521 certificate file. The previous behaviour hard coded the filename
10525 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10526 a public key to be input or output. For example:
10527 openssl rsa -in key.pem -pubout -out pubkey.pem
10528 Also added necessary DSA public key functions to handle this.
10531 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10532 in the message. This was handled by allowing
10533 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10534 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10536 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10537 to the end of the strings whereas this didn't. This would cause problems
10538 if strings read with d2i_ASN1_bytes() were later modified.
10539 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10541 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10542 data and it contains EOF it will end up returning an error. This is
10543 caused by input 46 bytes long. The cause is due to the way base64
10544 BIOs find the start of base64 encoded data. They do this by trying a
10545 trial decode on each line until they find one that works. When they
10546 do a flag is set and it starts again knowing it can pass all the
10547 data directly through the decoder. Unfortunately it doesn't reset
10548 the context it uses. This means that if EOF is reached an attempt
10549 is made to pass two EOFs through the context and this causes the
10550 resulting error. This can also cause other problems as well. As is
10551 usual with these problems it takes *ages* to find and the fix is
10552 trivial: move one line.
10553 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10555 *) Ugly workaround to get s_client and s_server working under Windows. The
10556 old code wouldn't work because it needed to select() on sockets and the
10557 tty (for keypresses and to see if data could be written). Win32 only
10558 supports select() on sockets so we select() with a 1s timeout on the
10559 sockets and then see if any characters are waiting to be read, if none
10560 are present then we retry, we also assume we can always write data to
10561 the tty. This isn't nice because the code then blocks until we've
10562 received a complete line of data and it is effectively polling the
10563 keyboard at 1s intervals: however it's quite a bit better than not
10564 working at all :-) A dedicated Windows application might handle this
10565 with an event loop for example.
10568 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10569 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10570 will be called when RSA_sign() and RSA_verify() are used. This is useful
10571 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10572 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10573 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10574 This necessitated the support of an extra signature type NID_md5_sha1
10575 for SSL signatures and modifications to the SSL library to use it instead
10576 of calling RSA_public_decrypt() and RSA_private_encrypt().
10579 *) Add new -verify -CAfile and -CApath options to the crl program, these
10580 will lookup a CRL issuers certificate and verify the signature in a
10581 similar way to the verify program. Tidy up the crl program so it
10582 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10583 less strict. It will now permit CRL extensions even if it is not
10584 a V2 CRL: this will allow it to tolerate some broken CRLs.
10587 *) Initialize all non-automatic variables each time one of the openssl
10588 sub-programs is started (this is necessary as they may be started
10589 multiple times from the "OpenSSL>" prompt).
10590 [Lennart Bang, Bodo Moeller]
10592 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10593 removing all other RSA functionality (this is what NO_RSA does). This
10594 is so (for example) those in the US can disable those operations covered
10595 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10599 *) Non-copying interface to BIO pairs.
10600 (still largely untested)
10603 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10604 ASCII string. This was handled independently in various places before.
10607 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10608 UTF8 strings a character at a time.
10611 *) Use client_version from client hello to select the protocol
10612 (s23_srvr.c) and for RSA client key exchange verification
10613 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10616 *) Add various utility functions to handle SPKACs, these were previously
10617 handled by poking round in the structure internals. Added new function
10618 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10619 print, verify and generate SPKACs. Based on an original idea from
10620 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10623 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10626 *) Allow the config file extension section to be overwritten on the
10627 command line. Based on an original idea from Massimiliano Pala
10628 <madwolf@comune.modena.it>. The new option is called -extensions
10629 and can be applied to ca, req and x509. Also -reqexts to override
10630 the request extensions in req and -crlexts to override the crl extensions
10634 *) Add new feature to the SPKAC handling in ca. Now you can include
10635 the same field multiple times by preceding it by "XXXX." for example:
10638 this is the same syntax as used in the req config file.
10641 *) Allow certificate extensions to be added to certificate requests. These
10642 are specified in a 'req_extensions' option of the req section of the
10643 config file. They can be printed out with the -text option to req but
10644 are otherwise ignored at present.
10647 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10648 data read consists of only the final block it would not decrypted because
10649 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10650 A misplaced 'break' also meant the decrypted final block might not be
10651 copied until the next read.
10654 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10655 a few extra parameters to the DH structure: these will be useful if
10656 for example we want the value of 'q' or implement X9.42 DH.
10659 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10660 provides hooks that allow the default DSA functions or functions on a
10661 "per key" basis to be replaced. This allows hardware acceleration and
10662 hardware key storage to be handled without major modification to the
10663 library. Also added low level modexp hooks and CRYPTO_EX structure and
10664 associated functions.
10667 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10668 as "read only": it can't be written to and the buffer it points to will
10669 not be freed. Reading from a read only BIO is much more efficient than
10670 a normal memory BIO. This was added because there are several times when
10671 an area of memory needs to be read from a BIO. The previous method was
10672 to create a memory BIO and write the data to it, this results in two
10673 copies of the data and an O(n^2) reading algorithm. There is a new
10674 function BIO_new_mem_buf() which creates a read only memory BIO from
10675 an area of memory. Also modified the PKCS#7 routines to use read only
10679 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10680 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10681 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10682 but a retry condition occurred while trying to read the rest.
10685 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10686 NID_pkcs7_encrypted by default: this was wrong since this should almost
10687 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10688 the encrypted data type: this is a more sensible place to put it and it
10689 allows the PKCS#12 code to be tidied up that duplicated this
10693 *) Changed obj_dat.pl script so it takes its input and output files on
10694 the command line. This should avoid shell escape redirection problems
10698 *) Initial support for certificate extension requests, these are included
10699 in things like Xenroll certificate requests. Included functions to allow
10700 extensions to be obtained and added.
10703 *) -crlf option to s_client and s_server for sending newlines as
10704 CRLF (as required by many protocols).
10707 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10709 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10710 [Ralf S. Engelschall]
10712 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10713 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10715 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10719 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10720 DH parameters/keys (q is lost during that conversion, but the resulting
10721 DH parameters contain its length).
10723 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10724 much faster than DH_generate_parameters (which creates parameters
10725 where p = 2*q + 1), and also the smaller q makes DH computations
10726 much more efficient (160-bit exponentiation instead of 1024-bit
10727 exponentiation); so this provides a convenient way to support DHE
10728 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10729 utter importance to use
10730 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10732 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10733 when such DH parameters are used, because otherwise small subgroup
10734 attacks may become possible!
10737 *) Avoid memory leak in i2d_DHparams.
10740 *) Allow the -k option to be used more than once in the enc program:
10741 this allows the same encrypted message to be read by multiple recipients.
10744 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10745 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10746 it will always use the numerical form of the OID, even if it has a short
10750 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10751 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10752 otherwise bn_mod_exp was called. In the case of hardware keys for example
10753 no private key components need be present and it might store extra data
10754 in the RSA structure, which cannot be accessed from bn_mod_exp.
10755 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10756 private key operations.
10759 *) Added support for SPARC Linux.
10762 *) pem_password_cb function type incompatibly changed from
10763 typedef int pem_password_cb(char *buf, int size, int rwflag);
10765 ....(char *buf, int size, int rwflag, void *userdata);
10766 so that applications can pass data to their callbacks:
10767 The PEM[_ASN1]_{read,write}... functions and macros now take an
10768 additional void * argument, which is just handed through whenever
10769 the password callback is called.
10770 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10772 New function SSL_CTX_set_default_passwd_cb_userdata.
10774 Compatibility note: As many C implementations push function arguments
10775 onto the stack in reverse order, the new library version is likely to
10776 interoperate with programs that have been compiled with the old
10777 pem_password_cb definition (PEM_whatever takes some data that
10778 happens to be on the stack as its last argument, and the callback
10779 just ignores this garbage); but there is no guarantee whatsoever that
10782 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10783 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10784 problems not only on Windows, but also on some Unix platforms.
10785 To avoid problematic command lines, these definitions are now in an
10786 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10787 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10790 *) MIPS III/IV assembler module is reimplemented.
10793 *) More DES library cleanups: remove references to srand/rand and
10794 delete an unused file.
10797 *) Add support for the the free Netwide assembler (NASM) under Win32,
10798 since not many people have MASM (ml) and it can be hard to obtain.
10799 This is currently experimental but it seems to work OK and pass all
10800 the tests. Check out INSTALL.W32 for info.
10803 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10804 without temporary keys kept an extra copy of the server key,
10805 and connections with temporary keys did not free everything in case
10809 *) New function RSA_check_key and new openssl rsa option -check
10810 for verifying the consistency of RSA keys.
10811 [Ulf Moeller, Bodo Moeller]
10813 *) Various changes to make Win32 compile work:
10814 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10815 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10816 comparison" warnings.
10817 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10820 *) Add a debugging option to PKCS#5 v2 key generation function: when
10821 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10822 derived keys are printed to stderr.
10825 *) Copy the flags in ASN1_STRING_dup().
10826 [Roman E. Pavlov <pre@mo.msk.ru>]
10828 *) The x509 application mishandled signing requests containing DSA
10829 keys when the signing key was also DSA and the parameters didn't match.
10831 It was supposed to omit the parameters when they matched the signing key:
10832 the verifying software was then supposed to automatically use the CA's
10833 parameters if they were absent from the end user certificate.
10835 Omitting parameters is no longer recommended. The test was also
10836 the wrong way round! This was probably due to unusual behaviour in
10837 EVP_cmp_parameters() which returns 1 if the parameters match.
10838 This meant that parameters were omitted when they *didn't* match and
10839 the certificate was useless. Certificates signed with 'ca' didn't have
10841 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10843 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10844 The interface is as follows:
10845 Applications can use
10846 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10847 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10848 "off" is now the default.
10849 The library internally uses
10850 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10851 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10852 to disable memory-checking temporarily.
10854 Some inconsistent states that previously were possible (and were
10855 even the default) are now avoided.
10857 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10858 with each memory chunk allocated; this is occasionally more helpful
10859 than just having a counter.
10861 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10863 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10867 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10868 which largely parallels "options", but is for changing API behaviour,
10869 whereas "options" are about protocol behaviour.
10870 Initial "mode" flags are:
10872 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10873 a single record has been written.
10874 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10875 retries use the same buffer location.
10876 (But all of the contents must be
10880 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10883 *) Fix problems with no-hmac etc.
10884 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10886 *) New functions RSA_get_default_method(), RSA_set_method() and
10887 RSA_get_method(). These allows replacement of RSA_METHODs without having
10888 to mess around with the internals of an RSA structure.
10891 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10892 Also really enable memory leak checks in openssl.c and in some
10894 [Chad C. Mulligan, Bodo Moeller]
10896 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10897 up the length of negative integers. This has now been simplified to just
10898 store the length when it is first determined and use it later, rather
10899 than trying to keep track of where data is copied and updating it to
10901 [Steve Henson, reported by Brien Wheeler
10902 <bwheeler@authentica-security.com>]
10904 *) Add a new function PKCS7_signatureVerify. This allows the verification
10905 of a PKCS#7 signature but with the signing certificate passed to the
10906 function itself. This contrasts with PKCS7_dataVerify which assumes the
10907 certificate is present in the PKCS#7 structure. This isn't always the
10908 case: certificates can be omitted from a PKCS#7 structure and be
10909 distributed by "out of band" means (such as a certificate database).
10912 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10913 function prototypes in pem.h, also change util/mkdef.pl to add the
10914 necessary function names.
10917 *) mk1mf.pl (used by Windows builds) did not properly read the
10918 options set by Configure in the top level Makefile, and Configure
10919 was not even able to write more than one option correctly.
10920 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10923 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10924 file to be loaded from a BIO or FILE pointer. The BIO version will
10925 for example allow memory BIOs to contain config info.
10928 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10929 Whoever hopes to achieve shared-library compatibility across versions
10930 must use this, not the compile-time macro.
10931 (Exercise 0.9.4: Which is the minimum library version required by
10933 Note: All this applies only to multi-threaded programs, others don't
10937 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10938 through a BIO pair triggered the default case, i.e.
10939 SSLerr(...,SSL_R_UNKNOWN_STATE).
10942 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10943 can use the SSL library even if none of the specific BIOs is
10947 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10948 for the encoded length.
10949 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10951 *) Add initial documentation of the X509V3 functions.
10954 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10955 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10956 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10957 secure PKCS#8 private key format with a high iteration count.
10960 *) Fix determination of Perl interpreter: A perl or perl5
10961 _directory_ in $PATH was also accepted as the interpreter.
10962 [Ralf S. Engelschall]
10964 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10965 wrong with it but it was very old and did things like calling
10966 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10967 unusual formatting.
10970 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10971 to use the new extension code.
10974 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
10975 with macros. This should make it easier to change their form, add extra
10976 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
10980 *) Add to configuration table a new entry that can specify an alternative
10981 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
10982 according to Mark Crispin <MRC@Panda.COM>.
10986 *) DES CBC did not update the IV. Weird.
10989 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
10990 Changing the behaviour of the former might break existing programs --
10991 where IV updating is needed, des_ncbc_encrypt can be used.
10994 *) When bntest is run from "make test" it drives bc to check its
10995 calculations, as well as internally checking them. If an internal check
10996 fails, it needs to cause bc to give a non-zero result or make test carries
10997 on without noticing the failure. Fixed.
11000 *) DES library cleanups.
11003 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11004 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11005 ciphers. NOTE: although the key derivation function has been verified
11006 against some published test vectors it has not been extensively tested
11007 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11011 *) Instead of "mkdir -p", which is not fully portable, use new
11012 Perl script "util/mkdir-p.pl".
11015 *) Rewrite the way password based encryption (PBE) is handled. It used to
11016 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11017 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11018 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11019 the 'parameter' field of the AlgorithmIdentifier is passed to the
11020 underlying key generation function so it must do its own ASN1 parsing.
11021 This has also changed the EVP_PBE_CipherInit() function which now has a
11022 'parameter' argument instead of literal salt and iteration count values
11023 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11026 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11027 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11028 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11029 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11030 value was just used as a "magic string" and not used directly its
11031 value doesn't matter.
11034 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11038 *) "linux-sparc64" configuration (ultrapenguin).
11039 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11040 "linux-sparc" configuration.
11041 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11043 *) config now generates no-xxx options for missing ciphers.
11046 *) Support the EBCDIC character set (work in progress).
11047 File ebcdic.c not yet included because it has a different license.
11048 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11050 *) Support BS2000/OSD-POSIX.
11051 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11053 *) Make callbacks for key generation use void * instead of char *.
11056 *) Make S/MIME samples compile (not yet tested).
11059 *) Additional typesafe stacks.
11062 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11066 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11068 *) New configuration variant "sco5-gcc".
11070 *) Updated some demos.
11071 [Sean O Riordain, Wade Scholine]
11073 *) Add missing BIO_free at exit of pkcs12 application.
11076 *) Fix memory leak in conf.c.
11079 *) Updates for Win32 to assembler version of MD5.
11082 *) Set #! path to perl in apps/der_chop to where we found it
11083 instead of using a fixed path.
11086 *) SHA library changes for irix64-mips4-cc.
11089 *) Improvements for VMS support.
11093 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11095 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11096 This also avoids the problems with SC4.2 and unpatched SC5.
11097 [Andy Polyakov <appro@fy.chalmers.se>]
11099 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11100 These are required because of the typesafe stack would otherwise break
11101 existing code. If old code used a structure member which used to be STACK
11102 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11103 sk_num or sk_value it would produce an error because the num, data members
11104 are not present in STACK_OF. Now it just produces a warning. sk_set
11105 replaces the old method of assigning a value to sk_value
11106 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11107 that does this will no longer work (and should use sk_set instead) but
11108 this could be regarded as a "questionable" behaviour anyway.
11111 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11112 correctly handle encrypted S/MIME data.
11115 *) Change type of various DES function arguments from des_cblock
11116 (which means, in function argument declarations, pointer to char)
11117 to des_cblock * (meaning pointer to array with 8 char elements),
11118 which allows the compiler to do more typechecking; it was like
11119 that back in SSLeay, but with lots of ugly casts.
11121 Introduce new type const_des_cblock.
11124 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11125 problems: find RecipientInfo structure that matches recipient certificate
11126 and initialise the ASN1 structures properly based on passed cipher.
11129 *) Belatedly make the BN tests actually check the results.
11132 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11133 to and from BNs: it was completely broken. New compilation option
11134 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11135 key elements as negative integers.
11138 *) Reorganize and speed up MD5.
11139 [Andy Polyakov <appro@fy.chalmers.se>]
11142 [Richard Levitte <richard@levitte.org>]
11144 *) New option -out to asn1parse to allow the parsed structure to be
11145 output to a file. This is most useful when combined with the -strparse
11146 option to examine the output of things like OCTET STRINGS.
11149 *) Make SSL library a little more fool-proof by not requiring any longer
11150 that SSL_set_{accept,connect}_state be called before
11151 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11152 in many applications because usually everything *appeared* to work as
11153 intended anyway -- now it really works as intended).
11156 *) Move openssl.cnf out of lib/.
11159 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11160 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11161 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11162 [Ralf S. Engelschall]
11164 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11165 handle PKCS#7 enveloped data properly.
11166 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11168 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11169 copying pointers. The cert_st handling is changed by this in
11170 various ways (and thus what used to be known as ctx->default_cert
11171 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11172 any longer when s->cert does not give us what we need).
11173 ssl_cert_instantiate becomes obsolete by this change.
11174 As soon as we've got the new code right (possibly it already is?),
11175 we have solved a couple of bugs of the earlier code where s->cert
11176 was used as if it could not have been shared with other SSL structures.
11178 Note that using the SSL API in certain dirty ways now will result
11179 in different behaviour than observed with earlier library versions:
11180 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11181 does not influence s as it used to.
11183 In order to clean up things more thoroughly, inside SSL_SESSION
11184 we don't use CERT any longer, but a new structure SESS_CERT
11185 that holds per-session data (if available); currently, this is
11186 the peer's certificate chain and, for clients, the server's certificate
11187 and temporary key. CERT holds only those values that can have
11188 meaningful defaults in an SSL_CTX.
11191 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11192 from the internal representation. Various PKCS#7 fixes: remove some
11193 evil casts and set the enc_dig_alg field properly based on the signing
11197 *) Allow PKCS#12 password to be set from the command line or the
11198 environment. Let 'ca' get its config file name from the environment
11199 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11203 *) Allow certificate policies extension to use an IA5STRING for the
11204 organization field. This is contrary to the PKIX definition but
11205 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11209 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11210 without disallowing inline assembler and the like for non-pedantic builds.
11213 *) Support Borland C++ builder.
11214 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11216 *) Support Mingw32.
11219 *) SHA-1 cleanups and performance enhancements.
11220 [Andy Polyakov <appro@fy.chalmers.se>]
11222 *) Sparc v8plus assembler for the bignum library.
11223 [Andy Polyakov <appro@fy.chalmers.se>]
11225 *) Accept any -xxx and +xxx compiler options in Configure.
11228 *) Update HPUX configuration.
11231 *) Add missing sk_<type>_unshift() function to safestack.h
11232 [Ralf S. Engelschall]
11234 *) New function SSL_CTX_use_certificate_chain_file that sets the
11235 "extra_cert"s in addition to the certificate. (This makes sense
11236 only for "PEM" format files, as chains as a whole are not
11240 *) Support verify_depth from the SSL API.
11241 x509_vfy.c had what can be considered an off-by-one-error:
11242 Its depth (which was not part of the external interface)
11243 was actually counting the number of certificates in a chain;
11244 now it really counts the depth.
11247 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11248 instead of X509err, which often resulted in confusing error
11249 messages since the error codes are not globally unique
11250 (e.g. an alleged error in ssl3_accept when a certificate
11251 didn't match the private key).
11253 *) New function SSL_CTX_set_session_id_context that allows to set a default
11254 value (so that you don't need SSL_set_session_id_context for each
11255 connection using the SSL_CTX).
11258 *) OAEP decoding bug fix.
11261 *) Support INSTALL_PREFIX for package builders, as proposed by
11265 *) New Configure options "threads" and "no-threads". For systems
11266 where the proper compiler options are known (currently Solaris
11267 and Linux), "threads" is the default.
11270 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11273 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11274 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11275 such as /usr/local/bin.
11278 *) "make linux-shared" to build shared libraries.
11279 [Niels Poppe <niels@netbox.org>]
11281 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11284 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11285 extension adding in x509 utility.
11288 *) Remove NOPROTO sections and error code comments.
11291 *) Partial rewrite of the DEF file generator to now parse the ANSI
11295 *) New Configure options --prefix=DIR and --openssldir=DIR.
11298 *) Complete rewrite of the error code script(s). It is all now handled
11299 by one script at the top level which handles error code gathering,
11300 header rewriting and C source file generation. It should be much better
11301 than the old method: it now uses a modified version of Ulf's parser to
11302 read the ANSI prototypes in all header files (thus the old K&R definitions
11303 aren't needed for error creation any more) and do a better job of
11304 translating function codes into names. The old 'ASN1 error code imbedded
11305 in a comment' is no longer necessary and it doesn't use .err files which
11306 have now been deleted. Also the error code call doesn't have to appear all
11307 on one line (which resulted in some large lines...).
11310 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11313 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11314 0 (which usually indicates a closed connection), but continue reading.
11317 *) Fix some race conditions.
11320 *) Add support for CRL distribution points extension. Add Certificate
11321 Policies and CRL distribution points documentation.
11324 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11327 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11328 8 of keying material. Merlin has also confirmed interop with this fix
11329 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11330 [Merlin Hughes <merlin@baltimore.ie>]
11332 *) Fix lots of warnings.
11333 [Richard Levitte <levitte@stacken.kth.se>]
11335 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11336 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11337 [Richard Levitte <levitte@stacken.kth.se>]
11339 *) Fix problems with sizeof(long) == 8.
11340 [Andy Polyakov <appro@fy.chalmers.se>]
11342 *) Change functions to ANSI C.
11345 *) Fix typos in error codes.
11346 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11348 *) Remove defunct assembler files from Configure.
11351 *) SPARC v8 assembler BIGNUM implementation.
11352 [Andy Polyakov <appro@fy.chalmers.se>]
11354 *) Support for Certificate Policies extension: both print and set.
11355 Various additions to support the r2i method this uses.
11358 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11359 return a const string when you are expecting an allocated buffer.
11362 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11363 types DirectoryString and DisplayText.
11366 *) Add code to allow r2i extensions to access the configuration database,
11367 add an LHASH database driver and add several ctx helper functions.
11370 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11371 fail when they extended the size of a BIGNUM.
11374 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11375 support typesafe stack.
11378 *) Fix typo in SSL_[gs]et_options().
11379 [Nils Frostberg <nils@medcom.se>]
11381 *) Delete various functions and files that belonged to the (now obsolete)
11382 old X509V3 handling code.
11385 *) New Configure option "rsaref".
11388 *) Don't auto-generate pem.h.
11391 *) Introduce type-safe ASN.1 SETs.
11394 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11395 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11397 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11398 that links with OpenSSL (well at least cause lots of warnings), but fear
11399 not: the conversion is trivial, and it eliminates loads of evil casts. A
11400 few STACKed things have been converted already. Feel free to convert more.
11401 In the fullness of time, I'll do away with the STACK type altogether.
11404 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11405 specified in <certfile> by updating the entry in the index.txt file.
11406 This way one no longer has to edit the index.txt file manually for
11407 revoking a certificate. The -revoke option does the gory details now.
11408 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11410 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11411 `-text' option at all and this way the `-noout -text' combination was
11412 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11413 [Ralf S. Engelschall]
11415 *) Make sure a corresponding plain text error message exists for the
11416 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11417 verify callback function determined that a certificate was revoked.
11418 [Ralf S. Engelschall]
11420 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11421 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11422 all available cipers including rc5, which was forgotten until now.
11423 In order to let the testing shell script know which algorithms
11424 are available, a new (up to now undocumented) command
11425 "openssl list-cipher-commands" is used.
11428 *) Bugfix: s_client occasionally would sleep in select() when
11429 it should have checked SSL_pending() first.
11432 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11433 the raw DSA values prior to ASN.1 encoding.
11436 *) Tweaks to Configure
11437 [Niels Poppe <niels@netbox.org>]
11439 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11443 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11446 *) New config option to avoid instructions that are illegal on the 80386.
11447 The default code is faster, but requires at least a 486.
11450 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11451 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11452 same as SSL2_VERSION anyway.
11455 *) New "-showcerts" option for s_client.
11458 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11459 application. Various cleanups and fixes.
11462 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11463 modify error routines to work internally. Add error codes and PBE init
11464 to library startup routines.
11467 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11468 packing functions to asn1 and evp. Changed function names and error
11469 codes along the way.
11472 *) PKCS12 integration: and so it begins... First of several patches to
11473 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11474 objects to objects.h
11477 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11478 and display support for Thawte strong extranet extension.
11481 *) Add LinuxPPC support.
11482 [Jeff Dubrule <igor@pobox.org>]
11484 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11485 bn_div_words in alpha.s.
11486 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11488 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11489 OAEP isn't supported when OpenSSL is built with RSAref.
11490 [Ulf Moeller <ulf@fitug.de>]
11492 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11493 so they no longer are missing under -DNOPROTO.
11494 [Soren S. Jorvang <soren@t.dk>]
11497 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11499 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11500 doesn't work when the session is reused. Coming soon!
11503 *) Fix a security hole, that allows sessions to be reused in the wrong
11504 context thus bypassing client cert protection! All software that uses
11505 client certs and session caches in multiple contexts NEEDS PATCHING to
11506 allow session reuse! A fuller solution is in the works.
11507 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11509 *) Some more source tree cleanups (removed obsolete files
11510 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11511 permission on "config" script to be executable) and a fix for the INSTALL
11513 [Ulf Moeller <ulf@fitug.de>]
11515 *) Remove some legacy and erroneous uses of malloc, free instead of
11517 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11519 *) Make rsa_oaep_test return non-zero on error.
11520 [Ulf Moeller <ulf@fitug.de>]
11522 *) Add support for native Solaris shared libraries. Configure
11523 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11524 if someone would make that last step automatic.
11525 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11527 *) ctx_size was not built with the right compiler during "make links". Fixed.
11530 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11531 except NULL ciphers". This means the default cipher list will no longer
11532 enable NULL ciphers. They need to be specifically enabled e.g. with
11533 the string "DEFAULT:eNULL".
11536 *) Fix to RSA private encryption routines: if p < q then it would
11537 occasionally produce an invalid result. This will only happen with
11538 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11541 *) Be less restrictive and allow also `perl util/perlpath.pl
11542 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11543 because this way one can also use an interpreter named `perl5' (which is
11544 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11545 installed as `perl').
11546 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11548 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11549 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11551 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11552 advapi32.lib to Win32 build and change the pem test comparison
11553 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11554 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11555 and crypto/des/ede_cbcm_enc.c.
11558 *) DES quad checksum was broken on big-endian architectures. Fixed.
11561 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11562 Win32 test batch file so it (might) work again. The Win32 test batch file
11563 is horrible: I feel ill....
11566 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11567 in e_os.h. Audit of header files to check ANSI and non ANSI
11568 sections: 10 functions were absent from non ANSI section and not exported
11569 from Windows DLLs. Fixed up libeay.num for new functions.
11572 *) Make `openssl version' output lines consistent.
11573 [Ralf S. Engelschall]
11575 *) Fix Win32 symbol export lists for BIO functions: Added
11576 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11577 to ms/libeay{16,32}.def.
11578 [Ralf S. Engelschall]
11580 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11581 fine under Unix and passes some trivial tests I've now added. But the
11582 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11583 added to make sure no one expects that this stuff really works in the
11584 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11585 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11587 [Ralf S. Engelschall]
11589 *) Fix the generation of two part addresses in perl.
11590 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11592 *) Add config entry for Linux on MIPS.
11593 [John Tobey <jtobey@channel1.com>]
11595 *) Make links whenever Configure is run, unless we are on Windoze.
11598 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11599 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11603 *) Add a useful kludge to allow package maintainers to specify compiler and
11604 other platforms details on the command line without having to patch the
11605 Configure script everytime: One now can use ``perl Configure
11606 <id>:<details>'', i.e. platform ids are allowed to have details appended
11607 to them (separated by colons). This is treated as there would be a static
11608 pre-configured entry in Configure's %table under key <id> with value
11609 <details> and ``perl Configure <id>'' is called. So, when you want to
11610 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11611 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11612 now, which overrides the FreeBSD-elf entry on-the-fly.
11613 [Ralf S. Engelschall]
11615 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11618 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11619 on the `perl Configure ...' command line. This way one can compile
11620 OpenSSL libraries with Position Independent Code (PIC) which is needed
11621 for linking it into DSOs.
11622 [Ralf S. Engelschall]
11624 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11628 *) Cleaned up the LICENSE document: The official contact for any license
11629 questions now is the OpenSSL core team under openssl-core@openssl.org.
11630 And add a paragraph about the dual-license situation to make sure people
11631 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11632 to the OpenSSL toolkit.
11633 [Ralf S. Engelschall]
11635 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11636 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11637 Additionally cleaned up the `make links' target: Remove unnecessary
11638 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11639 to speed processing and no longer clutter the display with confusing
11640 stuff. Instead only the actually done links are displayed.
11641 [Ralf S. Engelschall]
11643 *) Permit null encryption ciphersuites, used for authentication only. It used
11644 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11645 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11649 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11650 signed attributes when verifying signatures (this would break them),
11651 the detached data encoding was wrong and public keys obtained using
11652 X509_get_pubkey() weren't freed.
11655 *) Add text documentation for the BUFFER functions. Also added a work around
11656 to a Win95 console bug. This was triggered by the password read stuff: the
11657 last character typed gets carried over to the next fread(). If you were
11658 generating a new cert request using 'req' for example then the last
11659 character of the passphrase would be CR which would then enter the first
11663 *) Added the new `Includes OpenSSL Cryptography Software' button as
11664 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11665 button and can be used by applications based on OpenSSL to show the
11666 relationship to the OpenSSL project.
11667 [Ralf S. Engelschall]
11669 *) Remove confusing variables in function signatures in files
11670 ssl/ssl_lib.c and ssl/ssl.h.
11671 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11673 *) Don't install bss_file.c under PREFIX/include/
11674 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11676 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11677 functions that return function pointers and has support for NT specific
11678 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11679 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11680 unsigned to signed types: this was killing the Win32 compile.
11683 *) Add new certificate file to stack functions,
11684 SSL_add_dir_cert_subjects_to_stack() and
11685 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11686 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11687 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11688 This means that Apache-SSL and similar packages don't have to mess around
11689 to add as many CAs as they want to the preferred list.
11692 *) Experiment with doxygen documentation. Currently only partially applied to
11694 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11695 openssl.doxy as the configuration file.
11698 *) Get rid of remaining C++-style comments which strict C compilers hate.
11699 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11701 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11702 compiled in by default: it has problems with large keys.
11705 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11706 DH private keys and/or callback functions which directly correspond to
11707 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11708 is needed for applications which have to configure certificates on a
11709 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11711 For the RSA certificate situation is makes no difference, but
11712 for the DSA certificate situation this fixes the "no shared cipher"
11713 problem where the OpenSSL cipher selection procedure failed because the
11714 temporary keys were not overtaken from the context and the API provided
11715 no way to reconfigure them.
11716 The new functions now let applications reconfigure the stuff and they
11717 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11718 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11719 non-public-API function ssl_cert_instantiate() is used as a helper
11720 function and also to reduce code redundancy inside ssl_rsa.c.
11721 [Ralf S. Engelschall]
11723 *) Move s_server -dcert and -dkey options out of the undocumented feature
11724 area because they are useful for the DSA situation and should be
11725 recognized by the users.
11726 [Ralf S. Engelschall]
11728 *) Fix the cipher decision scheme for export ciphers: the export bits are
11729 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11730 SSL_EXP_MASK. So, the original variable has to be used instead of the
11731 already masked variable.
11732 [Richard Levitte <levitte@stacken.kth.se>]
11734 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11735 [Richard Levitte <levitte@stacken.kth.se>]
11737 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11738 from `int' to `unsigned int' because it's a length and initialized by
11739 EVP_DigestFinal() which expects an `unsigned int *'.
11740 [Richard Levitte <levitte@stacken.kth.se>]
11742 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11743 script. Instead use the usual Shell->Perl transition trick.
11744 [Ralf S. Engelschall]
11746 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11747 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11748 -noout -modulus' as it's already the case for `openssl rsa -noout
11749 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11750 currently the public key is printed (a decision which was already done by
11751 `openssl dsa -modulus' in the past) which serves a similar purpose.
11752 Additionally the NO_RSA no longer completely removes the whole -modulus
11753 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11755 [Ralf S. Engelschall]
11757 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11758 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11759 [Arne Ansper <arne@ats.cyber.ee>]
11761 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11762 to be added. Now both 'req' and 'ca' can use new objects defined in the
11766 *) Add cool BIO that does syslog (or event log on NT).
11767 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11769 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11770 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11771 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11772 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11775 *) Add preliminary config info for new extension code.
11778 *) Make RSA_NO_PADDING really use no padding.
11779 [Ulf Moeller <ulf@fitug.de>]
11781 *) Generate errors when private/public key check is done.
11784 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11785 for some CRL extensions and new objects added.
11788 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11789 key usage extension and fuller support for authority key id.
11792 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11793 padding method for RSA, which is recommended for new applications in PKCS
11794 #1 v2.0 (RFC 2437, October 1998).
11795 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11796 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11797 against Bleichbacher's attack on RSA.
11798 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11801 *) Updates to the new SSL compression code
11802 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11804 *) Fix so that the version number in the master secret, when passed
11805 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11806 (because the server will not accept higher), that the version number
11807 is 0x03,0x01, not 0x03,0x00
11808 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11810 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11811 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11812 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11815 *) Support for RAW extensions where an arbitrary extension can be
11816 created by including its DER encoding. See apps/openssl.cnf for
11820 *) Make sure latest Perl versions don't interpret some generated C array
11821 code as Perl array code in the crypto/err/err_genc.pl script.
11822 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11824 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11825 not many people have the assembler. Various Win32 compilation fixes and
11826 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11827 build instructions.
11830 *) Modify configure script 'Configure' to automatically create crypto/date.h
11831 file under Win32 and also build pem.h from pem.org. New script
11832 util/mkfiles.pl to create the MINFO file on environments that can't do a
11833 'make files': perl util/mkfiles.pl >MINFO should work.
11836 *) Major rework of DES function declarations, in the pursuit of correctness
11837 and purity. As a result, many evil casts evaporated, and some weirdness,
11838 too. You may find this causes warnings in your code. Zapping your evil
11839 casts will probably fix them. Mostly.
11842 *) Fix for a typo in asn1.h. Bug fix to object creation script
11843 obj_dat.pl. It considered a zero in an object definition to mean
11844 "end of object": none of the objects in objects.h have any zeros
11845 so it wasn't spotted.
11846 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11848 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11849 Masking (CBCM). In the absence of test vectors, the best I have been able
11850 to do is check that the decrypt undoes the encrypt, so far. Send me test
11851 vectors if you have them.
11854 *) Correct calculation of key length for export ciphers (too much space was
11855 allocated for null ciphers). This has not been tested!
11858 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11859 message is now correct (it understands "crypto" and "ssl" on its
11860 command line). There is also now an "update" option. This will update
11861 the util/ssleay.num and util/libeay.num files with any new functions.
11863 perl util/mkdef.pl crypto ssl update
11864 it will update them.
11867 *) Overhauled the Perl interface (perl/*):
11868 - ported BN stuff to OpenSSL's different BN library
11869 - made the perl/ source tree CVS-aware
11870 - renamed the package from SSLeay to OpenSSL (the files still contain
11871 their history because I've copied them in the repository)
11872 - removed obsolete files (the test scripts will be replaced
11873 by better Test::Harness variants in the future)
11874 [Ralf S. Engelschall]
11876 *) First cut for a very conservative source tree cleanup:
11877 1. merge various obsolete readme texts into doc/ssleay.txt
11878 where we collect the old documents and readme texts.
11879 2. remove the first part of files where I'm already sure that we no
11880 longer need them because of three reasons: either they are just temporary
11881 files which were left by Eric or they are preserved original files where
11882 I've verified that the diff is also available in the CVS via "cvs diff
11883 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11884 the crypto/md/ stuff).
11885 [Ralf S. Engelschall]
11887 *) More extension code. Incomplete support for subject and issuer alt
11888 name, issuer and authority key id. Change the i2v function parameters
11889 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11890 what that's for :-) Fix to ASN1 macro which messed up
11891 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11894 *) Preliminary support for ENUMERATED type. This is largely copied from the
11898 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11899 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11901 *) Make sure `make rehash' target really finds the `openssl' program.
11902 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11904 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11905 like to hear about it if this slows down other processors.
11908 *) Add CygWin32 platform information to Configure script.
11909 [Alan Batie <batie@aahz.jf.intel.com>]
11911 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11912 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11914 *) New program nseq to manipulate netscape certificate sequences
11917 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11921 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11922 but the BN code had some problems that would cause failures when
11923 doing certificate verification and some other functions.
11924 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11926 *) Add ASN1 and PEM code to support netscape certificate sequences.
11929 *) Add ASN1 and PEM code to support netscape certificate sequences.
11932 *) Add several PKIX and private extended key usage OIDs.
11935 *) Modify the 'ca' program to handle the new extension code. Modify
11936 openssl.cnf for new extension format, add comments.
11939 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11940 and add a sample to openssl.cnf so req -x509 now adds appropriate
11944 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11945 error code, add initial support to X509_print() and x509 application.
11948 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11949 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11950 stuff is currently isolated and isn't even compiled yet.
11953 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11954 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11955 Removed the versions check from X509 routines when loading extensions:
11956 this allows certain broken certificates that don't set the version
11957 properly to be processed.
11960 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11961 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11962 can still be regenerated with "make depend".
11965 *) Spelling mistake in C version of CAST-128.
11966 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11968 *) Changes to the error generation code. The perl script err-code.pl
11969 now reads in the old error codes and retains the old numbers, only
11970 adding new ones if necessary. It also only changes the .err files if new
11971 codes are added. The makefiles have been modified to only insert errors
11972 when needed (to avoid needlessly modifying header files). This is done
11973 by only inserting errors if the .err file is newer than the auto generated
11974 C file. To rebuild all the error codes from scratch (the old behaviour)
11975 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
11976 or delete all the .err files.
11979 *) CAST-128 was incorrectly implemented for short keys. The C version has
11980 been fixed, but is untested. The assembler versions are also fixed, but
11981 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
11982 to regenerate it if needed.
11983 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
11984 Hagino <itojun@kame.net>]
11986 *) File was opened incorrectly in randfile.c.
11987 [Ulf Möller <ulf@fitug.de>]
11989 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
11990 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
11991 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
11992 al: it's just almost always a UTCTime. Note this patch adds new error
11993 codes so do a "make errors" if there are problems.
11996 *) Correct Linux 1 recognition in config.
11997 [Ulf Möller <ulf@fitug.de>]
11999 *) Remove pointless MD5 hash when using DSA keys in ca.
12000 [Anonymous <nobody@replay.com>]
12002 *) Generate an error if given an empty string as a cert directory. Also
12003 generate an error if handed NULL (previously returned 0 to indicate an
12004 error, but didn't set one).
12005 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12007 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12010 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12011 parameters. This was causing a warning which killed off the Win32 compile.
12014 *) Remove C++ style comments from crypto/bn/bn_local.h.
12015 [Neil Costigan <neil.costigan@celocom.com>]
12017 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12018 based on a text string, looking up short and long names and finally
12019 "dot" format. The "dot" format stuff didn't work. Added new function
12020 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12021 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12022 OID is not part of the table.
12025 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12026 X509_LOOKUP_by_alias().
12029 *) Sort openssl functions by name.
12032 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12033 encryption from sample DSA keys (in case anyone is interested the password
12037 *) Make _all_ *_free functions accept a NULL pointer.
12038 [Frans Heymans <fheymans@isaserver.be>]
12040 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12042 [Anonymous <nobody@replay.com>]
12044 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12045 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12047 *) Don't blow it for numeric -newkey arguments to apps/req.
12048 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12050 *) Temp key "for export" tests were wrong in s3_srvr.c.
12051 [Anonymous <nobody@replay.com>]
12053 *) Add prototype for temp key callback functions
12054 SSL_CTX_set_tmp_{rsa,dh}_callback().
12057 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12058 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12061 *) X509_name_add_entry() freed the wrong thing after an error.
12062 [Arne Ansper <arne@ats.cyber.ee>]
12064 *) rsa_eay.c would attempt to free a NULL context.
12065 [Arne Ansper <arne@ats.cyber.ee>]
12067 *) BIO_s_socket() had a broken should_retry() on Windoze.
12068 [Arne Ansper <arne@ats.cyber.ee>]
12070 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12071 [Arne Ansper <arne@ats.cyber.ee>]
12073 *) Make sure the already existing X509_STORE->depth variable is initialized
12074 in X509_STORE_new(), but document the fact that this variable is still
12075 unused in the certificate verification process.
12076 [Ralf S. Engelschall]
12078 *) Fix the various library and apps files to free up pkeys obtained from
12079 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12082 *) Fix reference counting in X509_PUBKEY_get(). This makes
12083 demos/maurice/example2.c work, amongst others, probably.
12084 [Steve Henson and Ben Laurie]
12086 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12087 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12088 are no longer created. This way we have a single and consistent command
12089 line interface `openssl <command>', similar to `cvs <command>'.
12090 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12092 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12093 BIT STRING wrapper always have zero unused bits.
12096 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12099 *) Make the top-level INSTALL documentation easier to understand.
12102 *) Makefiles updated to exit if an error occurs in a sub-directory
12103 make (including if user presses ^C) [Paul Sutton]
12105 *) Make Montgomery context stuff explicit in RSA data structure.
12108 *) Fix build order of pem and err to allow for generated pem.h.
12111 *) Fix renumbering bug in X509_NAME_delete_entry().
12114 *) Enhanced the err-ins.pl script so it makes the error library number
12115 global and can add a library name. This is needed for external ASN1 and
12116 other error libraries.
12119 *) Fixed sk_insert which never worked properly.
12122 *) Fix ASN1 macros so they can handle indefinite length constructed
12123 EXPLICIT tags. Some non standard certificates use these: they can now
12127 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12128 into a single doc/ssleay.txt bundle. This way the information is still
12129 preserved but no longer messes up this directory. Now it's new room for
12130 the new set of documentation files.
12131 [Ralf S. Engelschall]
12133 *) SETs were incorrectly DER encoded. This was a major pain, because they
12134 shared code with SEQUENCEs, which aren't coded the same. This means that
12135 almost everything to do with SETs or SEQUENCEs has either changed name or
12136 number of arguments.
12137 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12139 *) Fix test data to work with the above.
12142 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12143 was already fixed by Eric for 0.9.1 it seems.
12144 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12146 *) Autodetect FreeBSD3.
12149 *) Fix various bugs in Configure. This affects the following platforms:
12153 unixware-2.0-pentium
12157 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12158 before they are needed.
12161 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12165 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12167 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12168 changed SSLeay to OpenSSL in version strings.
12169 [Ralf S. Engelschall]
12171 *) Some fixups to the top-level documents.
12174 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12175 because the symlink to include/ was missing.
12176 [Ralf S. Engelschall]
12178 *) Incorporated the popular no-RSA/DSA-only patches
12179 which allow to compile a RSA-free SSLeay.
12180 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12182 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12183 when "ssleay" is still not found.
12184 [Ralf S. Engelschall]
12186 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12187 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12189 *) Updated the README file.
12190 [Ralf S. Engelschall]
12192 *) Added various .cvsignore files in the CVS repository subdirs
12193 to make a "cvs update" really silent.
12194 [Ralf S. Engelschall]
12196 *) Recompiled the error-definition header files and added
12197 missing symbols to the Win32 linker tables.
12198 [Ralf S. Engelschall]
12200 *) Cleaned up the top-level documents;
12201 o new files: CHANGES and LICENSE
12202 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12203 o merged COPYRIGHT into LICENSE
12204 o removed obsolete TODO file
12205 o renamed MICROSOFT to INSTALL.W32
12206 [Ralf S. Engelschall]
12208 *) Removed dummy files from the 0.9.1b source tree:
12209 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12210 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12211 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12212 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12213 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12214 [Ralf S. Engelschall]
12216 *) Added various platform portability fixes.
12219 *) The Genesis of the OpenSSL rpject:
12220 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12221 Young and Tim J. Hudson created while they were working for C2Net until
12223 [The OpenSSL Project]
12226 Changes between 0.9.0b and 0.9.1b [not released]
12228 *) Updated a few CA certificates under certs/
12231 *) Changed some BIGNUM api stuff.
12234 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12235 DGUX x86, Linux Alpha, etc.
12238 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12239 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12243 *) Add -strparse option to asn1pars program which parses nested
12245 [Dr Stephen Henson <shenson@bigfoot.com>]
12247 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12250 *) DSA fix for "ca" program.
12253 *) Added "-genkey" option to "dsaparam" program.
12256 *) Added RIPE MD160 (rmd160) message digest.
12259 *) Added -a (all) option to "ssleay version" command.
12262 *) Added PLATFORM define which is the id given to Configure.
12265 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12268 *) Extended the ASN.1 parser routines.
12271 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12274 *) Added a BN_CTX to the BN library.
12277 *) Fixed the weak key values in DES library
12280 *) Changed API in EVP library for cipher aliases.
12283 *) Added support for RC2/64bit cipher.
12286 *) Converted the lhash library to the crypto/mem.c functions.
12289 *) Added more recognized ASN.1 object ids.
12292 *) Added more RSA padding checks for SSL/TLS.
12295 *) Added BIO proxy/filter functionality.
12298 *) Added extra_certs to SSL_CTX which can be used
12299 send extra CA certificates to the client in the CA cert chain sending
12300 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12303 *) Now Fortezza is denied in the authentication phase because
12304 this is key exchange mechanism is not supported by SSLeay at all.
12307 *) Additional PKCS1 checks.
12310 *) Support the string "TLSv1" for all TLS v1 ciphers.
12313 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12314 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12317 *) Fixed a few memory leaks.
12320 *) Fixed various code and comment typos.
12323 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12324 bytes sent in the client random.
12325 [Edward Bishop <ebishop@spyglass.com>]