5 Changes between 1.1.0b and 1.1.0c [xx XXX xxxx]
7 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
8 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
9 prevent issues where no progress is being made and the peer continually
10 sends unrecognised record types, using up resources processing them.
13 *) Removed automatic addition of RPATH in shared libraries and executables,
14 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
17 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
19 *) Fix Use After Free for large message sizes
21 The patch applied to address CVE-2016-6307 resulted in an issue where if a
22 message larger than approx 16k is received then the underlying buffer to
23 store the incoming message is reallocated and moved. Unfortunately a
24 dangling pointer to the old location is left which results in an attempt to
25 write to the previously freed location. This is likely to result in a
26 crash, however it could potentially lead to execution of arbitrary code.
28 This issue only affects OpenSSL 1.1.0a.
30 This issue was reported to OpenSSL by Robert Święcki.
34 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
36 *) OCSP Status Request extension unbounded memory growth
38 A malicious client can send an excessively large OCSP Status Request
39 extension. If that client continually requests renegotiation, sending a
40 large OCSP Status Request extension each time, then there will be unbounded
41 memory growth on the server. This will eventually lead to a Denial Of
42 Service attack through memory exhaustion. Servers with a default
43 configuration are vulnerable even if they do not support OCSP. Builds using
44 the "no-ocsp" build time option are not affected.
46 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
50 *) SSL_peek() hang on empty record
52 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
53 sends an empty record. This could be exploited by a malicious peer in a
54 Denial Of Service attack.
56 This issue was reported to OpenSSL by Alex Gaynor.
60 *) Excessive allocation of memory in tls_get_message_header() and
61 dtls1_preprocess_fragment()
63 A (D)TLS message includes 3 bytes for its length in the header for the
64 message. This would allow for messages up to 16Mb in length. Messages of
65 this length are excessive and OpenSSL includes a check to ensure that a
66 peer is sending reasonably sized messages in order to avoid too much memory
67 being consumed to service a connection. A flaw in the logic of version
68 1.1.0 means that memory for the message is allocated too early, prior to
69 the excessive message length check. Due to way memory is allocated in
70 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
71 to service a connection. This could lead to a Denial of Service through
72 memory exhaustion. However, the excessive message length check still takes
73 place, and this would cause the connection to immediately fail. Assuming
74 that the application calls SSL_free() on the failed conneciton in a timely
75 manner then the 21Mb of allocated memory will then be immediately freed
76 again. Therefore the excessive memory allocation will be transitory in
77 nature. This then means that there is only a security impact if:
79 1) The application does not call SSL_free() in a timely manner in the event
80 that the connection fails
82 2) The application is working in a constrained environment where there is
83 very little free memory
85 3) The attacker initiates multiple connection attempts such that there are
86 multiple connections in a state where memory has been allocated for the
87 connection; SSL_free() has not yet been called; and there is insufficient
88 memory to service the multiple requests.
90 Except in the instance of (1) above any Denial Of Service is likely to be
91 transitory because as soon as the connection fails the memory is
92 subsequently freed again in the SSL_free() call. However there is an
93 increased risk during this period of application crashes due to the lack of
94 memory - which would then mean a more serious Denial of Service.
96 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
97 (CVE-2016-6307 and CVE-2016-6308)
100 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
101 had to be removed. Primary reason is that vendor assembler can't
102 assemble our modules with -KPIC flag. As result it, assembly
103 support, was not even available as option. But its lack means
104 lack of side-channel resistant code, which is incompatible with
105 security by todays standards. Fortunately gcc is readily available
106 prepackaged option, which we firmly point at...
109 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
111 *) Windows command-line tool supports UTF-8 opt-in option for arguments
112 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
113 (to any value) allows Windows user to access PKCS#12 file generated
114 with Windows CryptoAPI and protected with non-ASCII password, as well
115 as files generated under UTF-8 locale on Linux also protected with
119 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
120 have been disabled by default and removed from DEFAULT, just like RC4.
121 See the RC4 item below to re-enable both.
124 *) The method for finding the storage location for the Windows RAND seed file
125 has changed. First we check %RANDFILE%. If that is not set then we check
126 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
127 all else fails we fall back to C:\.
130 *) The EVP_EncryptUpdate() function has had its return type changed from void
131 to int. A return of 0 indicates and error while a return of 1 indicates
135 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
136 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
137 off the constant time implementation for RSA, DSA and DH have been made
138 no-ops and deprecated.
141 *) Windows RAND implementation was simplified to only get entropy by
142 calling CryptGenRandom(). Various other RAND-related tickets
144 [Joseph Wylie Yandle, Rich Salz]
146 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
147 and OPENSSL_LH_, respectively. The old names are available
148 with API compatibility. They new names are now completely documented.
151 *) Unify TYPE_up_ref(obj) methods signature.
152 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
153 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
154 int (instead of void) like all others TYPE_up_ref() methods.
155 So now these methods also check the return value of CRYPTO_atomic_add(),
156 and the validity of object reference counter.
157 [fdasilvayy@gmail.com]
159 *) With Windows Visual Studio builds, the .pdb files are installed
160 alongside the installed libraries and executables. For a static
161 library installation, ossl_static.pdb is the associate compiler
162 generated .pdb file to be used when linking programs.
165 *) Remove openssl.spec. Packaging files belong with the packagers.
168 *) Automatic Darwin/OSX configuration has had a refresh, it will now
169 recognise x86_64 architectures automatically. You can still decide
170 to build for a different bitness with the environment variable
171 KERNEL_BITS (can be 32 or 64), for example:
173 KERNEL_BITS=32 ./config
177 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
178 256 bit AES and HMAC with SHA256.
181 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
184 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
187 *) To enable users to have their own config files and build file templates,
188 Configure looks in the directory indicated by the environment variable
189 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
190 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
191 name and is used as is.
194 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
195 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
196 X509_CERT_FILE_CTX was removed.
199 *) "shared" builds are now the default. To create only static libraries use
200 the "no-shared" Configure option.
203 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
204 All of these option have not worked for some while and are fundamental
208 *) Make various cleanup routines no-ops and mark them as deprecated. Most
209 global cleanup functions are no longer required because they are handled
210 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
211 Explicitly de-initing can cause problems (e.g. where a library that uses
212 OpenSSL de-inits, but an application is still using it). The affected
213 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
214 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
215 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
219 *) --strict-warnings no longer enables runtime debugging options
220 such as REF_DEBUG. Instead, debug options are automatically
221 enabled with '--debug' builds.
222 [Andy Polyakov, Emilia Käsper]
224 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
225 have been moved out of the public header files. New functions for managing
226 these have been added.
229 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
230 objects have been moved out of the public header files. New
231 functions for managing these have been added.
234 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
235 have been moved out of the public header files. New functions for managing
236 these have been added.
239 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
240 moved out of the public header files. New functions for managing these
244 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
247 *) Removed the mk1mf build scripts.
250 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
251 it is always safe to #include a header now.
254 *) Removed the aged BC-32 config and all its supporting scripts
257 *) Removed support for Ultrix, Netware, and OS/2.
260 *) Add support for HKDF.
263 *) Add support for blake2b and blake2s
266 *) Added support for "pipelining". Ciphers that have the
267 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
268 encryptions/decryptions simultaneously. There are currently no built-in
269 ciphers with this property but the expectation is that engines will be able
270 to offer it to significantly improve throughput. Support has been extended
271 into libssl so that multiple records for a single connection can be
272 processed in one go (for >=TLS 1.1).
275 *) Added the AFALG engine. This is an async capable engine which is able to
276 offload work to the Linux kernel. In this initial version it only supports
277 AES128-CBC. The kernel must be version 4.1.0 or greater.
280 *) OpenSSL now uses a new threading API. It is no longer necessary to
281 set locking callbacks to use OpenSSL in a multi-threaded environment. There
282 are two supported threading models: pthreads and windows threads. It is
283 also possible to configure OpenSSL at compile time for "no-threads". The
284 old threading API should no longer be used. The functions have been
285 replaced with "no-op" compatibility macros.
286 [Alessandro Ghedini, Matt Caswell]
288 *) Modify behavior of ALPN to invoke callback after SNI/servername
289 callback, such that updates to the SSL_CTX affect ALPN.
292 *) Add SSL_CIPHER queries for authentication and key-exchange.
295 *) Changes to the DEFAULT cipherlist:
296 - Prefer (EC)DHE handshakes over plain RSA.
297 - Prefer AEAD ciphers over legacy ciphers.
298 - Prefer ECDSA over RSA when both certificates are available.
299 - Prefer TLSv1.2 ciphers/PRF.
300 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
304 *) Change the ECC default curve list to be this, in order: x25519,
305 secp256r1, secp521r1, secp384r1.
308 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
309 disabled by default. They can be re-enabled using the
310 enable-weak-ssl-ciphers option to Configure.
313 *) If the server has ALPN configured, but supports no protocols that the
314 client advertises, send a fatal "no_application_protocol" alert.
315 This behaviour is SHALL in RFC 7301, though it isn't universally
316 implemented by other servers.
319 *) Add X25519 support.
320 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
321 for public and private key encoding using the format documented in
322 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
323 key generation and key derivation.
325 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
329 *) Deprecate SRP_VBASE_get_by_user.
330 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
331 In order to fix an unavoidable memory leak (CVE-2016-0798),
332 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
333 seed, even if the seed is configured.
335 Users should use SRP_VBASE_get1_by_user instead. Note that in
336 SRP_VBASE_get1_by_user, caller must free the returned value. Note
337 also that even though configuring the SRP seed attempts to hide
338 invalid usernames by continuing the handshake with fake
339 credentials, this behaviour is not constant time and no strong
340 guarantees are made that the handshake is indistinguishable from
341 that of a valid user.
344 *) Configuration change; it's now possible to build dynamic engines
345 without having to build shared libraries and vice versa. This
346 only applies to the engines in engines/, those in crypto/engine/
347 will always be built into libcrypto (i.e. "static").
349 Building dynamic engines is enabled by default; to disable, use
350 the configuration option "disable-dynamic-engine".
352 The only requirements for building dynamic engines are the
353 presence of the DSO module and building with position independent
354 code, so they will also automatically be disabled if configuring
355 with "disable-dso" or "disable-pic".
357 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
358 are also taken away from openssl/opensslconf.h, as they are
362 *) Configuration change; if there is a known flag to compile
363 position independent code, it will always be applied on the
364 libcrypto and libssl object files, and never on the application
365 object files. This means other libraries that use routines from
366 libcrypto / libssl can be made into shared libraries regardless
367 of how OpenSSL was configured.
369 If this isn't desirable, the configuration options "disable-pic"
370 or "no-pic" can be used to disable the use of PIC. This will
371 also disable building shared libraries and dynamic engines.
374 *) Removed JPAKE code. It was experimental and has no wide use.
377 *) The INSTALL_PREFIX Makefile variable has been renamed to
378 DESTDIR. That makes for less confusion on what this variable
379 is for. Also, the configuration option --install_prefix is
383 *) Heartbeat for TLS has been removed and is disabled by default
384 for DTLS; configure with enable-heartbeats. Code that uses the
385 old #define's might need to be updated.
386 [Emilia Käsper, Rich Salz]
388 *) Rename REF_CHECK to REF_DEBUG.
391 *) New "unified" build system
393 The "unified" build system is aimed to be a common system for all
394 platforms we support. With it comes new support for VMS.
396 This system builds supports building in a different directory tree
397 than the source tree. It produces one Makefile (for unix family
398 or lookalikes), or one descrip.mms (for VMS).
400 The source of information to make the Makefile / descrip.mms is
401 small files called 'build.info', holding the necessary
402 information for each directory with source to compile, and a
403 template in Configurations, like unix-Makefile.tmpl or
406 With this change, the library names were also renamed on Windows
407 and on VMS. They now have names that are closer to the standard
408 on Unix, and include the major version number, and in certain
409 cases, the architecture they are built for. See "Notes on shared
410 libraries" in INSTALL.
412 We rely heavily on the perl module Text::Template.
415 *) Added support for auto-initialisation and de-initialisation of the library.
416 OpenSSL no longer requires explicit init or deinit routines to be called,
417 except in certain circumstances. See the OPENSSL_init_crypto() and
418 OPENSSL_init_ssl() man pages for further information.
421 *) The arguments to the DTLSv1_listen function have changed. Specifically the
422 "peer" argument is now expected to be a BIO_ADDR object.
424 *) Rewrite of BIO networking library. The BIO library lacked consistent
425 support of IPv6, and adding it required some more extensive
426 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
427 which hold all types of addresses and chains of address information.
428 It also introduces a new API, with functions like BIO_socket,
429 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
430 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
431 have been adapted accordingly.
434 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
438 *) CRIME protection: disable compression by default, even if OpenSSL is
439 compiled with zlib enabled. Applications can still enable compression
440 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
441 using the SSL_CONF library to configure compression.
444 *) The signature of the session callback configured with
445 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
446 was explicitly marked as 'const unsigned char*' instead of
450 *) Always DPURIFY. Remove the use of uninitialized memory in the
451 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
454 *) Removed many obsolete configuration items, including
455 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
456 MD2_CHAR, MD2_INT, MD2_LONG
458 IDEA_SHORT, IDEA_LONG
459 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
460 [Rich Salz, with advice from Andy Polyakov]
462 *) Many BN internals have been moved to an internal header file.
463 [Rich Salz with help from Andy Polyakov]
465 *) Configuration and writing out the results from it has changed.
466 Files such as Makefile include/openssl/opensslconf.h and are now
467 produced through general templates, such as Makefile.in and
468 crypto/opensslconf.h.in and some help from the perl module
471 Also, the center of configuration information is no longer
472 Makefile. Instead, Configure produces a perl module in
473 configdata.pm which holds most of the config data (in the hash
474 table %config), the target data that comes from the target
475 configuration in one of the Configurations/*.conf files (in
479 *) To clarify their intended purposes, the Configure options
480 --prefix and --openssldir change their semantics, and become more
481 straightforward and less interdependent.
483 --prefix shall be used exclusively to give the location INSTALLTOP
484 where programs, scripts, libraries, include files and manuals are
485 going to be installed. The default is now /usr/local.
487 --openssldir shall be used exclusively to give the default
488 location OPENSSLDIR where certificates, private keys, CRLs are
489 managed. This is also where the default openssl.cnf gets
491 If the directory given with this option is a relative path, the
492 values of both the --prefix value and the --openssldir value will
493 be combined to become OPENSSLDIR.
494 The default for --openssldir is INSTALLTOP/ssl.
496 Anyone who uses --openssldir to specify where OpenSSL is to be
497 installed MUST change to use --prefix instead.
500 *) The GOST engine was out of date and therefore it has been removed. An up
501 to date GOST engine is now being maintained in an external repository.
502 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
503 support for GOST ciphersuites (these are only activated if a GOST engine
507 *) EGD is no longer supported by default; use enable-egd when
509 [Ben Kaduk and Rich Salz]
511 *) The distribution now has Makefile.in files, which are used to
512 create Makefile's when Configure is run. *Configure must be run
513 before trying to build now.*
516 *) The return value for SSL_CIPHER_description() for error conditions
520 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
522 Obtaining and performing DNSSEC validation of TLSA records is
523 the application's responsibility. The application provides
524 the TLSA records of its choice to OpenSSL, and these are then
525 used to authenticate the peer.
527 The TLSA records need not even come from DNS. They can, for
528 example, be used to implement local end-entity certificate or
529 trust-anchor "pinning", where the "pin" data takes the form
530 of TLSA records, which can augment or replace verification
531 based on the usual WebPKI public certification authorities.
534 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
535 continues to support deprecated interfaces in default builds.
536 However, applications are strongly advised to compile their
537 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
538 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
539 or the 1.1.0 releases.
541 In environments in which all applications have been ported to
542 not use any deprecated interfaces OpenSSL's Configure script
543 should be used with the --api=1.1.0 option to entirely remove
544 support for the deprecated features from the library and
545 unconditionally disable them in the installed headers.
546 Essentially the same effect can be achieved with the "no-deprecated"
547 argument to Configure, except that this will always restrict
548 the build to just the latest API, rather than a fixed API
551 As applications are ported to future revisions of the API,
552 they should update their compile-time OPENSSL_API_COMPAT define
553 accordingly, but in most cases should be able to continue to
554 compile with later releases.
556 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
557 0x10000000L and 0x00908000L, respectively. However those
558 versions did not support the OPENSSL_API_COMPAT feature, and
559 so applications are not typically tested for explicit support
560 of just the undeprecated features of either release.
563 *) Add support for setting the minimum and maximum supported protocol.
564 It can bet set via the SSL_set_min_proto_version() and
565 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
566 MaxProtcol. It's recommended to use the new APIs to disable
567 protocols instead of disabling individual protocols using
568 SSL_set_options() or SSL_CONF's Protocol. This change also
569 removes support for disabling TLS 1.2 in the OpenSSL TLS
570 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
573 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
576 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
577 and integrates ECDSA and ECDH functionality into EC. Implementations can
578 now redirect key generation and no longer need to convert to or from
581 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
582 include the ec.h header file instead.
585 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
586 ciphers who are no longer supported and drops support the ephemeral RSA key
587 exchange. The LOW ciphers currently doesn't have any ciphers in it.
590 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
591 opaque. For HMAC_CTX, the following constructors and destructors
594 HMAC_CTX *HMAC_CTX_new(void);
595 void HMAC_CTX_free(HMAC_CTX *ctx);
597 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
598 destroy such methods has been added. See EVP_MD_meth_new(3) and
599 EVP_CIPHER_meth_new(3) for documentation.
602 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
603 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
604 EVP_MD_CTX_reset() should be called instead to reinitialise
605 an already created structure.
606 2) For consistency with the majority of our object creators and
607 destructors, EVP_MD_CTX_(create|destroy) were renamed to
608 EVP_MD_CTX_(new|free). The old names are retained as macros
609 for deprecated builds.
612 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
613 cryptographic operations to be performed asynchronously as long as an
614 asynchronous capable engine is used. See the ASYNC_start_job() man page for
615 further details. Libssl has also had this capability integrated with the
616 introduction of the new mode SSL_MODE_ASYNC and associated error
617 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
618 pages. This work was developed in partnership with Intel Corp.
621 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
622 always enabled now. If you want to disable the support you should
623 exclude it using the list of supported ciphers. This also means that the
624 "-no_ecdhe" option has been removed from s_server.
627 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
628 SSL_{CTX_}set1_curves() which can set a list.
631 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
632 curve you want to support using SSL_{CTX_}set1_curves().
635 *) State machine rewrite. The state machine code has been significantly
636 refactored in order to remove much duplication of code and solve issues
637 with the old code (see ssl/statem/README for further details). This change
638 does have some associated API changes. Notably the SSL_state() function
639 has been removed and replaced by SSL_get_state which now returns an
640 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
641 altogether. The previous handshake states defined in ssl.h and ssl3.h have
645 *) All instances of the string "ssleay" in the public API were replaced
646 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
647 Some error codes related to internal RSA_eay API's were renamed.
650 *) The demo files in crypto/threads were moved to demo/threads.
653 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
655 [Matt Caswell, Rich Salz]
657 *) New ASN.1 embed macro.
659 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
660 structure is not allocated: it is part of the parent. That is instead of
668 This reduces memory fragmentation and make it impossible to accidentally
669 set a mandatory field to NULL.
671 This currently only works for some fields specifically a SEQUENCE, CHOICE,
672 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
673 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
677 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
680 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
681 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
682 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
683 DES and RC4 ciphersuites.
686 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
687 This changes the decoding behaviour for some invalid messages,
688 though the change is mostly in the more lenient direction, and
689 legacy behaviour is preserved as much as possible.
692 *) Fix no-stdio build.
693 [ David Woodhouse <David.Woodhouse@intel.com> and also
694 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
696 *) New testing framework
697 The testing framework has been largely rewritten and is now using
698 perl and the perl modules Test::Harness and an extended variant of
699 Test::More called OpenSSL::Test to do its work. All test scripts in
700 test/ have been rewritten into test recipes, and all direct calls to
701 executables in test/Makefile have become individual recipes using the
702 simplified testing OpenSSL::Test::Simple.
704 For documentation on our testing modules, do:
706 perldoc test/testlib/OpenSSL/Test/Simple.pm
707 perldoc test/testlib/OpenSSL/Test.pm
711 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
712 are used; the latter aborts on memory leaks (usually checked on exit).
713 Some undocumented "set malloc, etc., hooks" functions were removed
714 and others were changed. All are now documented.
717 *) In DSA_generate_parameters_ex, if the provided seed is too short,
719 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
721 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
722 from RFC4279, RFC4785, RFC5487, RFC5489.
724 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
725 original RSA_PSK patch.
728 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
729 era flag was never set throughout the codebase (only read). Also removed
730 SSL3_FLAGS_POP_BUFFER which was only used if
731 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
734 *) Changed the default name options in the "ca", "crl", "req" and "x509"
735 to be "oneline" instead of "compat".
738 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
739 not aware of clients that still exhibit this bug, and the workaround
740 hasn't been working properly for a while.
743 *) The return type of BIO_number_read() and BIO_number_written() as well as
744 the corresponding num_read and num_write members in the BIO structure has
745 changed from unsigned long to uint64_t. On platforms where an unsigned
746 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
750 *) Given the pervasive nature of TLS extensions it is inadvisable to run
751 OpenSSL without support for them. It also means that maintaining
752 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
753 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
756 *) Removed support for the two export grade static DH ciphersuites
757 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
758 were newly added (along with a number of other static DH ciphersuites) to
759 1.0.2. However the two export ones have *never* worked since they were
760 introduced. It seems strange in any case to be adding new export
761 ciphersuites, and given "logjam" it also does not seem correct to fix them.
764 *) Version negotiation has been rewritten. In particular SSLv23_method(),
765 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
766 and turned into macros which simply call the new preferred function names
767 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
768 should use the new names instead. Also as part of this change the ssl23.h
769 header file has been removed.
772 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
773 code and the associated standard is no longer considered fit-for-purpose.
776 *) RT2547 was closed. When generating a private key, try to make the
777 output file readable only by the owner. This behavior change might
778 be noticeable when interacting with other software.
780 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
784 *) Added HTTP GET support to the ocsp command.
787 *) Changed default digest for the dgst and enc commands from MD5 to
791 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
794 *) Added support for TLS extended master secret from
795 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
796 initial patch which was a great help during development.
799 *) All libssl internal structures have been removed from the public header
800 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
801 now redundant). Users should not attempt to access internal structures
802 directly. Instead they should use the provided API functions.
805 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
806 Access to deprecated functions can be re-enabled by running config with
807 "enable-deprecated". In addition applications wishing to use deprecated
808 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
809 will, by default, disable some transitive includes that previously existed
810 in the header files (e.g. ec.h will no longer, by default, include bn.h)
813 *) Added support for OCB mode. OpenSSL has been granted a patent license
814 compatible with the OpenSSL license for use of OCB. Details are available
815 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
816 for OCB can be removed by calling config with no-ocb.
819 *) SSLv2 support has been removed. It still supports receiving a SSLv2
820 compatible client hello.
823 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
824 done while fixing the error code for the key-too-small case.
825 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
827 *) CA.sh has been removmed; use CA.pl instead.
830 *) Removed old DES API.
833 *) Remove various unsupported platforms:
839 Sinix/ReliantUNIX RM400
844 16-bit platforms such as WIN16
847 *) Clean up OPENSSL_NO_xxx #define's
848 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
849 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
850 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
851 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
852 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
853 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
854 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
855 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
856 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
857 Remove MS_STATIC; it's a relic from platforms <32 bits.
860 *) Cleaned up dead code
861 Remove all but one '#ifdef undef' which is to be looked at.
864 *) Clean up calling of xxx_free routines.
865 Just like free(), fix most of the xxx_free routines to accept
866 NULL. Remove the non-null checks from callers. Save much code.
869 *) Add secure heap for storage of private keys (when possible).
870 Add BIO_s_secmem(), CBIGNUM, etc.
871 Contributed by Akamai Technologies under our Corporate CLA.
874 *) Experimental support for a new, fast, unbiased prime candidate generator,
875 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
876 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
878 *) New output format NSS in the sess_id command line tool. This allows
879 exporting the session id and the master key in NSS keylog format.
880 [Martin Kaiser <martin@kaiser.cx>]
882 *) Harmonize version and its documentation. -f flag is used to display
884 [mancha <mancha1@zoho.com>]
886 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
887 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
888 [mancha <mancha1@zoho.com>]
890 *) Fix some double frees. These are not thought to be exploitable.
891 [mancha <mancha1@zoho.com>]
893 *) A missing bounds check in the handling of the TLS heartbeat extension
894 can be used to reveal up to 64k of memory to a connected client or
897 Thanks for Neel Mehta of Google Security for discovering this bug and to
898 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
899 preparing the fix (CVE-2014-0160)
900 [Adam Langley, Bodo Moeller]
902 *) Fix for the attack described in the paper "Recovering OpenSSL
903 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
904 by Yuval Yarom and Naomi Benger. Details can be obtained from:
905 http://eprint.iacr.org/2014/140
907 Thanks to Yuval Yarom and Naomi Benger for discovering this
908 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
909 [Yuval Yarom and Naomi Benger]
911 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
912 this fixes a limitation in previous versions of OpenSSL.
915 *) Experimental encrypt-then-mac support.
917 Experimental support for encrypt then mac from
918 draft-gutmann-tls-encrypt-then-mac-02.txt
920 To enable it set the appropriate extension number (0x42 for the test
921 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
923 For non-compliant peers (i.e. just about everything) this should have no
926 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
930 *) Add EVP support for key wrapping algorithms, to avoid problems with
931 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
932 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
933 algorithms and include tests cases.
936 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
940 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
941 MGF1 digest and OAEP label.
944 *) Make openssl verify return errors.
945 [Chris Palmer <palmer@google.com> and Ben Laurie]
947 *) New function ASN1_TIME_diff to calculate the difference between two
948 ASN1_TIME structures or one structure and the current time.
951 *) Update fips_test_suite to support multiple command line options. New
952 test to induce all self test errors in sequence and check expected
956 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
957 sign or verify all in one operation.
960 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
961 test programs and fips_test_suite. Includes functionality to parse
962 the minimal script output of fipsalgest.pl directly.
965 *) Add authorisation parameter to FIPS_module_mode_set().
968 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
971 *) Use separate DRBG fields for internal and external flags. New function
972 FIPS_drbg_health_check() to perform on demand health checking. Add
973 generation tests to fips_test_suite with reduced health check interval to
974 demonstrate periodic health checking. Add "nodh" option to
975 fips_test_suite to skip very slow DH test.
978 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
982 *) More extensive health check for DRBG checking many more failure modes.
983 New function FIPS_selftest_drbg_all() to handle every possible DRBG
984 combination: call this in fips_test_suite.
987 *) Add support for canonical generation of DSA parameter 'g'. See
990 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
991 POST to handle HMAC cases.
994 *) Add functions FIPS_module_version() and FIPS_module_version_text()
995 to return numerical and string versions of the FIPS module number.
998 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
999 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1000 outside the validated module in the FIPS capable OpenSSL.
1003 *) Minor change to DRBG entropy callback semantics. In some cases
1004 there is no multiple of the block length between min_len and
1005 max_len. Allow the callback to return more than max_len bytes
1006 of entropy but discard any extra: it is the callback's responsibility
1007 to ensure that the extra data discarded does not impact the
1008 requested amount of entropy.
1011 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1012 information in FIPS186-3, SP800-57 and SP800-131A.
1015 *) CCM support via EVP. Interface is very similar to GCM case except we
1016 must supply all data in one chunk (i.e. no update, final) and the
1017 message length must be supplied if AAD is used. Add algorithm test
1021 *) Initial version of POST overhaul. Add POST callback to allow the status
1022 of POST to be monitored and/or failures induced. Modify fips_test_suite
1023 to use callback. Always run all selftests even if one fails.
1026 *) XTS support including algorithm test driver in the fips_gcmtest program.
1027 Note: this does increase the maximum key length from 32 to 64 bytes but
1028 there should be no binary compatibility issues as existing applications
1029 will never use XTS mode.
1032 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1033 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1034 performs algorithm blocking for unapproved PRNG types. Also do not
1035 set PRNG type in FIPS_mode_set(): leave this to the application.
1036 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1037 the standard OpenSSL PRNG: set additional data to a date time vector.
1040 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1041 This shouldn't present any incompatibility problems because applications
1042 shouldn't be using these directly and any that are will need to rethink
1043 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1046 *) Extensive self tests and health checking required by SP800-90 DRBG.
1047 Remove strength parameter from FIPS_drbg_instantiate and always
1048 instantiate at maximum supported strength.
1051 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1054 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1057 *) New function DH_compute_key_padded() to compute a DH key and pad with
1058 leading zeroes if needed: this complies with SP800-56A et al.
1061 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1062 anything, incomplete, subject to change and largely untested at present.
1065 *) Modify fipscanisteronly build option to only build the necessary object
1066 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1069 *) Add experimental option FIPSSYMS to give all symbols in
1070 fipscanister.o and FIPS or fips prefix. This will avoid
1071 conflicts with future versions of OpenSSL. Add perl script
1072 util/fipsas.pl to preprocess assembly language source files
1073 and rename any affected symbols.
1076 *) Add selftest checks and algorithm block of non-fips algorithms in
1077 FIPS mode. Remove DES2 from selftests.
1080 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1081 return internal method without any ENGINE dependencies. Add new
1082 tiny fips sign and verify functions.
1085 *) New build option no-ec2m to disable characteristic 2 code.
1088 *) New build option "fipscanisteronly". This only builds fipscanister.o
1089 and (currently) associated fips utilities. Uses the file Makefile.fips
1090 instead of Makefile.org as the prototype.
1093 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1094 Update fips_gcmtest to use IV generator.
1097 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1098 setting output buffer to NULL. The *Final function must be
1099 called although it will not retrieve any additional data. The tag
1100 can be set or retrieved with a ctrl. The IV length is by default 12
1101 bytes (96 bits) but can be set to an alternative value. If the IV
1102 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1106 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1107 underlying do_cipher function handles all cipher semantics itself
1108 including padding and finalisation. This is useful if (for example)
1109 an ENGINE cipher handles block padding itself. The behaviour of
1110 do_cipher is subtly changed if this flag is set: the return value
1111 is the number of characters written to the output buffer (zero is
1112 no longer an error code) or a negative error code. Also if the
1113 input buffer is NULL and length 0 finalisation should be performed.
1116 *) If a candidate issuer certificate is already part of the constructed
1117 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1120 *) Improve forward-security support: add functions
1122 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1123 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1125 for use by SSL/TLS servers; the callback function will be called whenever a
1126 new session is created, and gets to decide whether the session may be
1127 cached to make it resumable (return 0) or not (return 1). (As by the
1128 SSL/TLS protocol specifications, the session_id sent by the server will be
1129 empty to indicate that the session is not resumable; also, the server will
1130 not generate RFC 4507 (RFC 5077) session tickets.)
1132 A simple reasonable callback implementation is to return is_forward_secure.
1133 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1134 by the SSL/TLS server library, indicating whether it can provide forward
1136 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1138 *) New -verify_name option in command line utilities to set verification
1142 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1143 Add CMAC pkey methods.
1146 *) Experimental renegotiation in s_server -www mode. If the client
1147 browses /reneg connection is renegotiated. If /renegcert it is
1148 renegotiated requesting a certificate.
1151 *) Add an "external" session cache for debugging purposes to s_server. This
1152 should help trace issues which normally are only apparent in deployed
1153 multi-process servers.
1156 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1157 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1158 BIO_set_cipher() and some obscure PEM functions were changed so they
1159 can now return an error. The RAND changes required a change to the
1160 RAND_METHOD structure.
1163 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1164 a gcc attribute to warn if the result of a function is ignored. This
1165 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1166 whose return value is often ignored.
1169 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1170 These allow SCTs (signed certificate timestamps) to be requested and
1171 validated when establishing a connection.
1172 [Rob Percival <robpercival@google.com>]
1174 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1176 *) Prevent padding oracle in AES-NI CBC MAC check
1178 A MITM attacker can use a padding oracle attack to decrypt traffic
1179 when the connection uses an AES CBC cipher and the server support
1182 This issue was introduced as part of the fix for Lucky 13 padding
1183 attack (CVE-2013-0169). The padding check was rewritten to be in
1184 constant time by making sure that always the same bytes are read and
1185 compared against either the MAC or padding bytes. But it no longer
1186 checked that there was enough data to have both the MAC and padding
1189 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1193 *) Fix EVP_EncodeUpdate overflow
1195 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1196 Base64 encoding of binary data. If an attacker is able to supply very large
1197 amounts of input data then a length check can overflow resulting in a heap
1200 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1201 the PEM_write_bio* family of functions. These are mainly used within the
1202 OpenSSL command line applications, so any application which processes data
1203 from an untrusted source and outputs it as a PEM file should be considered
1204 vulnerable to this issue. User applications that call these APIs directly
1205 with large amounts of untrusted data may also be vulnerable.
1207 This issue was reported by Guido Vranken.
1211 *) Fix EVP_EncryptUpdate overflow
1213 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1214 is able to supply very large amounts of input data after a previous call to
1215 EVP_EncryptUpdate() with a partial block then a length check can overflow
1216 resulting in a heap corruption. Following an analysis of all OpenSSL
1217 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1218 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1219 the first called function after an EVP_EncryptInit(), and therefore that
1220 specific call must be safe. The second form is where the length passed to
1221 EVP_EncryptUpdate() can be seen from the code to be some small value and
1222 therefore there is no possibility of an overflow. Since all instances are
1223 one of these two forms, it is believed that there can be no overflows in
1224 internal code due to this problem. It should be noted that
1225 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1226 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1227 of these calls have also been analysed too and it is believed there are no
1228 instances in internal usage where an overflow could occur.
1230 This issue was reported by Guido Vranken.
1234 *) Prevent ASN.1 BIO excessive memory allocation
1236 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1237 a short invalid encoding can cause allocation of large amounts of memory
1238 potentially consuming excessive resources or exhausting memory.
1240 Any application parsing untrusted data through d2i BIO functions is
1241 affected. The memory based functions such as d2i_X509() are *not* affected.
1242 Since the memory based functions are used by the TLS library, TLS
1243 applications are not affected.
1245 This issue was reported by Brian Carpenter.
1251 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1252 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1253 in arbitrary stack data being returned in the buffer.
1255 This issue was reported by Guido Vranken.
1259 *) Modify behavior of ALPN to invoke callback after SNI/servername
1260 callback, such that updates to the SSL_CTX affect ALPN.
1263 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1267 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1268 methods are enabled and ssl2 is disabled the methods return NULL.
1271 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1273 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1274 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1275 provide any "EXPORT" or "LOW" strength ciphers.
1278 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1279 is by default disabled at build-time. Builds that are not configured with
1280 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1281 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1282 will need to explicitly call either of:
1284 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1286 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1288 as appropriate. Even if either of those is used, or the application
1289 explicitly uses the version-specific SSLv2_method() or its client and
1290 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1291 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1292 ciphers, and SSLv2 56-bit DES are no longer available.
1296 *) Fix a double-free in DSA code
1298 A double free bug was discovered when OpenSSL parses malformed DSA private
1299 keys and could lead to a DoS attack or memory corruption for applications
1300 that receive DSA private keys from untrusted sources. This scenario is
1303 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1308 *) Disable SRP fake user seed to address a server memory leak.
1310 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1312 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1313 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1314 was changed to ignore the "fake user" SRP seed, even if the seed
1317 Users should use SRP_VBASE_get1_by_user instead. Note that in
1318 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1319 also that even though configuring the SRP seed attempts to hide
1320 invalid usernames by continuing the handshake with fake
1321 credentials, this behaviour is not constant time and no strong
1322 guarantees are made that the handshake is indistinguishable from
1323 that of a valid user.
1327 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1329 In the BN_hex2bn function the number of hex digits is calculated using an
1330 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1331 large values of |i| this can result in |bn_expand| not allocating any
1332 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1333 field as NULL leading to a subsequent NULL ptr deref. For very large values
1334 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1335 In this case memory is allocated to the internal BIGNUM data field, but it
1336 is insufficiently sized leading to heap corruption. A similar issue exists
1337 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1338 is ever called by user applications with very large untrusted hex/dec data.
1339 This is anticipated to be a rare occurrence.
1341 All OpenSSL internal usage of these functions use data that is not expected
1342 to be untrusted, e.g. config file data or application command line
1343 arguments. If user developed applications generate config file data based
1344 on untrusted data then it is possible that this could also lead to security
1345 consequences. This is also anticipated to be rare.
1347 This issue was reported to OpenSSL by Guido Vranken.
1351 *) Fix memory issues in BIO_*printf functions
1353 The internal |fmtstr| function used in processing a "%s" format string in
1354 the BIO_*printf functions could overflow while calculating the length of a
1355 string and cause an OOB read when printing very long strings.
1357 Additionally the internal |doapr_outch| function can attempt to write to an
1358 OOB memory location (at an offset from the NULL pointer) in the event of a
1359 memory allocation failure. In 1.0.2 and below this could be caused where
1360 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1361 could be in processing a very long "%s" format string. Memory leaks can
1364 The first issue may mask the second issue dependent on compiler behaviour.
1365 These problems could enable attacks where large amounts of untrusted data
1366 is passed to the BIO_*printf functions. If applications use these functions
1367 in this way then they could be vulnerable. OpenSSL itself uses these
1368 functions when printing out human-readable dumps of ASN.1 data. Therefore
1369 applications that print this data could be vulnerable if the data is from
1370 untrusted sources. OpenSSL command line applications could also be
1371 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1372 as command line arguments.
1374 Libssl is not considered directly vulnerable. Additionally certificates etc
1375 received via remote connections via libssl are also unlikely to be able to
1376 trigger these issues because of message size limits enforced within libssl.
1378 This issue was reported to OpenSSL Guido Vranken.
1382 *) Side channel attack on modular exponentiation
1384 A side-channel attack was found which makes use of cache-bank conflicts on
1385 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1386 of RSA keys. The ability to exploit this issue is limited as it relies on
1387 an attacker who has control of code in a thread running on the same
1388 hyper-threaded core as the victim thread which is performing decryptions.
1390 This issue was reported to OpenSSL by Yuval Yarom, The University of
1391 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1392 Nadia Heninger, University of Pennsylvania with more information at
1393 http://cachebleed.info.
1397 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1398 if no keysize is specified with default_bits. This fixes an
1399 omission in an earlier change that changed all RSA/DSA key generation
1400 apps to use 2048 bits by default.
1403 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1404 *) DH small subgroups
1406 Historically OpenSSL only ever generated DH parameters based on "safe"
1407 primes. More recently (in version 1.0.2) support was provided for
1408 generating X9.42 style parameter files such as those required for RFC 5114
1409 support. The primes used in such files may not be "safe". Where an
1410 application is using DH configured with parameters based on primes that are
1411 not "safe" then an attacker could use this fact to find a peer's private
1412 DH exponent. This attack requires that the attacker complete multiple
1413 handshakes in which the peer uses the same private DH exponent. For example
1414 this could be used to discover a TLS server's private DH exponent if it's
1415 reusing the private DH exponent or it's using a static DH ciphersuite.
1417 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1418 TLS. It is not on by default. If the option is not set then the server
1419 reuses the same private DH exponent for the life of the server process and
1420 would be vulnerable to this attack. It is believed that many popular
1421 applications do set this option and would therefore not be at risk.
1423 The fix for this issue adds an additional check where a "q" parameter is
1424 available (as is the case in X9.42 based parameters). This detects the
1425 only known attack, and is the only possible defense for static DH
1426 ciphersuites. This could have some performance impact.
1428 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1429 default and cannot be disabled. This could have some performance impact.
1431 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1435 *) SSLv2 doesn't block disabled ciphers
1437 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1438 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1439 been disabled, provided that the SSLv2 protocol was not also disabled via
1442 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1443 and Sebastian Schinzel.
1447 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1449 *) BN_mod_exp may produce incorrect results on x86_64
1451 There is a carry propagating bug in the x86_64 Montgomery squaring
1452 procedure. No EC algorithms are affected. Analysis suggests that attacks
1453 against RSA and DSA as a result of this defect would be very difficult to
1454 perform and are not believed likely. Attacks against DH are considered just
1455 feasible (although very difficult) because most of the work necessary to
1456 deduce information about a private key may be performed offline. The amount
1457 of resources required for such an attack would be very significant and
1458 likely only accessible to a limited number of attackers. An attacker would
1459 additionally need online access to an unpatched system using the target
1460 private key in a scenario with persistent DH parameters and a private
1461 key that is shared between multiple clients. For example this can occur by
1462 default in OpenSSL DHE based SSL/TLS ciphersuites.
1464 This issue was reported to OpenSSL by Hanno Böck.
1468 *) Certificate verify crash with missing PSS parameter
1470 The signature verification routines will crash with a NULL pointer
1471 dereference if presented with an ASN.1 signature using the RSA PSS
1472 algorithm and absent mask generation function parameter. Since these
1473 routines are used to verify certificate signature algorithms this can be
1474 used to crash any certificate verification operation and exploited in a
1475 DoS attack. Any application which performs certificate verification is
1476 vulnerable including OpenSSL clients and servers which enable client
1479 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1483 *) X509_ATTRIBUTE memory leak
1485 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1486 memory. This structure is used by the PKCS#7 and CMS routines so any
1487 application which reads PKCS#7 or CMS data from untrusted sources is
1488 affected. SSL/TLS is not affected.
1490 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1495 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1496 This changes the decoding behaviour for some invalid messages,
1497 though the change is mostly in the more lenient direction, and
1498 legacy behaviour is preserved as much as possible.
1501 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1503 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1505 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1507 *) Alternate chains certificate forgery
1509 During certificate verification, OpenSSL will attempt to find an
1510 alternative certificate chain if the first attempt to build such a chain
1511 fails. An error in the implementation of this logic can mean that an
1512 attacker could cause certain checks on untrusted certificates to be
1513 bypassed, such as the CA flag, enabling them to use a valid leaf
1514 certificate to act as a CA and "issue" an invalid certificate.
1516 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1520 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1522 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1523 incompatibility in the handling of HMAC. The previous ABI has now been
1527 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1529 *) Malformed ECParameters causes infinite loop
1531 When processing an ECParameters structure OpenSSL enters an infinite loop
1532 if the curve specified is over a specially malformed binary polynomial
1535 This can be used to perform denial of service against any
1536 system which processes public keys, certificate requests or
1537 certificates. This includes TLS clients and TLS servers with
1538 client authentication enabled.
1540 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1544 *) Exploitable out-of-bounds read in X509_cmp_time
1546 X509_cmp_time does not properly check the length of the ASN1_TIME
1547 string and can read a few bytes out of bounds. In addition,
1548 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1551 An attacker can use this to craft malformed certificates and CRLs of
1552 various sizes and potentially cause a segmentation fault, resulting in
1553 a DoS on applications that verify certificates or CRLs. TLS clients
1554 that verify CRLs are affected. TLS clients and servers with client
1555 authentication enabled may be affected if they use custom verification
1558 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1559 independently by Hanno Böck.
1563 *) PKCS7 crash with missing EnvelopedContent
1565 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1566 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1567 with missing content and trigger a NULL pointer dereference on parsing.
1569 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1570 structures from untrusted sources are affected. OpenSSL clients and
1571 servers are not affected.
1573 This issue was reported to OpenSSL by Michal Zalewski (Google).
1577 *) CMS verify infinite loop with unknown hash function
1579 When verifying a signedData message the CMS code can enter an infinite loop
1580 if presented with an unknown hash function OID. This can be used to perform
1581 denial of service against any system which verifies signedData messages using
1583 This issue was reported to OpenSSL by Johannes Bauer.
1587 *) Race condition handling NewSessionTicket
1589 If a NewSessionTicket is received by a multi-threaded client when attempting to
1590 reuse a previous ticket then a race condition can occur potentially leading to
1591 a double free of the ticket data.
1595 *) Only support 256-bit or stronger elliptic curves with the
1596 'ecdh_auto' setting (server) or by default (client). Of supported
1597 curves, prefer P-256 (both).
1600 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1602 *) ClientHello sigalgs DoS fix
1604 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1605 invalid signature algorithms extension a NULL pointer dereference will
1606 occur. This can be exploited in a DoS attack against the server.
1608 This issue was was reported to OpenSSL by David Ramos of Stanford
1611 [Stephen Henson and Matt Caswell]
1613 *) Multiblock corrupted pointer fix
1615 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1616 feature only applies on 64 bit x86 architecture platforms that support AES
1617 NI instructions. A defect in the implementation of "multiblock" can cause
1618 OpenSSL's internal write buffer to become incorrectly set to NULL when
1619 using non-blocking IO. Typically, when the user application is using a
1620 socket BIO for writing, this will only result in a failed connection.
1621 However if some other BIO is used then it is likely that a segmentation
1622 fault will be triggered, thus enabling a potential DoS attack.
1624 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1628 *) Segmentation fault in DTLSv1_listen fix
1630 The DTLSv1_listen function is intended to be stateless and processes the
1631 initial ClientHello from many peers. It is common for user code to loop
1632 over the call to DTLSv1_listen until a valid ClientHello is received with
1633 an associated cookie. A defect in the implementation of DTLSv1_listen means
1634 that state is preserved in the SSL object from one invocation to the next
1635 that can lead to a segmentation fault. Errors processing the initial
1636 ClientHello can trigger this scenario. An example of such an error could be
1637 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1640 This issue was reported to OpenSSL by Per Allansson.
1644 *) Segmentation fault in ASN1_TYPE_cmp fix
1646 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1647 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1648 certificate signature algorithm consistency this can be used to crash any
1649 certificate verification operation and exploited in a DoS attack. Any
1650 application which performs certificate verification is vulnerable including
1651 OpenSSL clients and servers which enable client authentication.
1655 *) Segmentation fault for invalid PSS parameters fix
1657 The signature verification routines will crash with a NULL pointer
1658 dereference if presented with an ASN.1 signature using the RSA PSS
1659 algorithm and invalid parameters. Since these routines are used to verify
1660 certificate signature algorithms this can be used to crash any
1661 certificate verification operation and exploited in a DoS attack. Any
1662 application which performs certificate verification is vulnerable including
1663 OpenSSL clients and servers which enable client authentication.
1665 This issue was was reported to OpenSSL by Brian Carpenter.
1669 *) ASN.1 structure reuse memory corruption fix
1671 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1672 memory corruption via an invalid write. Such reuse is and has been
1673 strongly discouraged and is believed to be rare.
1675 Applications that parse structures containing CHOICE or ANY DEFINED BY
1676 components may be affected. Certificate parsing (d2i_X509 and related
1677 functions) are however not affected. OpenSSL clients and servers are
1682 *) PKCS7 NULL pointer dereferences fix
1684 The PKCS#7 parsing code does not handle missing outer ContentInfo
1685 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1686 missing content and trigger a NULL pointer dereference on parsing.
1688 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1689 otherwise parse PKCS#7 structures from untrusted sources are
1690 affected. OpenSSL clients and servers are not affected.
1692 This issue was reported to OpenSSL by Michal Zalewski (Google).
1696 *) DoS via reachable assert in SSLv2 servers fix
1698 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1699 servers that both support SSLv2 and enable export cipher suites by sending
1700 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1702 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1703 (OpenSSL development team).
1707 *) Empty CKE with client auth and DHE fix
1709 If client auth is used then a server can seg fault in the event of a DHE
1710 ciphersuite being selected and a zero length ClientKeyExchange message
1711 being sent by the client. This could be exploited in a DoS attack.
1715 *) Handshake with unseeded PRNG fix
1717 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1718 with an unseeded PRNG. The conditions are:
1719 - The client is on a platform where the PRNG has not been seeded
1720 automatically, and the user has not seeded manually
1721 - A protocol specific client method version has been used (i.e. not
1722 SSL_client_methodv23)
1723 - A ciphersuite is used that does not require additional random data from
1724 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1726 If the handshake succeeds then the client random that has been used will
1727 have been generated from a PRNG with insufficient entropy and therefore the
1728 output may be predictable.
1730 For example using the following command with an unseeded openssl will
1731 succeed on an unpatched platform:
1733 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1737 *) Use After Free following d2i_ECPrivatekey error fix
1739 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1740 could cause a use after free condition. This, in turn, could cause a double
1741 free in several private key parsing functions (such as d2i_PrivateKey
1742 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1743 for applications that receive EC private keys from untrusted
1744 sources. This scenario is considered rare.
1746 This issue was discovered by the BoringSSL project and fixed in their
1751 *) X509_to_X509_REQ NULL pointer deref fix
1753 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1754 the certificate key is invalid. This function is rarely used in practice.
1756 This issue was discovered by Brian Carpenter.
1760 *) Removed the export ciphers from the DEFAULT ciphers
1763 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1765 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1766 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1767 So far those who have to target multiple platforms would compromise
1768 and argue that binary targeting say ARMv5 would still execute on
1769 ARMv8. "Universal" build resolves this compromise by providing
1770 near-optimal performance even on newer platforms.
1773 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1774 (other platforms pending).
1775 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1777 *) Add support for the SignedCertificateTimestampList certificate and
1778 OCSP response extensions from RFC6962.
1781 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1782 for corner cases. (Certain input points at infinity could lead to
1783 bogus results, with non-infinity inputs mapped to infinity too.)
1786 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1787 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1788 common cases are optimized and there still is room for further
1789 improvements. Vector Permutation AES for Altivec is also added.
1792 *) Add support for little-endian ppc64 Linux target.
1793 [Marcelo Cerri (IBM)]
1795 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1796 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1797 are optimized and there still is room for further improvements.
1798 Both 32- and 64-bit modes are supported.
1799 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1801 *) Improved ARMv7 NEON support.
1804 *) Support for SPARC Architecture 2011 crypto extensions, first
1805 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1806 SHA256/512, MD5, GHASH and modular exponentiation.
1807 [Andy Polyakov, David Miller]
1809 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1811 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1813 *) Support for new and upcoming Intel processors, including AVX2,
1814 BMI and SHA ISA extensions. This includes additional "stitched"
1815 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1818 This work was sponsored by Intel Corp.
1821 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1822 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1823 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1826 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1827 this fixes a limitation in previous versions of OpenSSL.
1830 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1831 MGF1 digest and OAEP label.
1834 *) Add EVP support for key wrapping algorithms, to avoid problems with
1835 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1836 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1837 algorithms and include tests cases.
1840 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1842 [Douglas E. Engert, Steve Henson]
1844 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1845 difference in days and seconds between two tm or ASN1_TIME structures.
1848 *) Add -rev test option to s_server to just reverse order of characters
1849 received by client and send back to server. Also prints an abbreviated
1850 summary of the connection parameters.
1853 *) New option -brief for s_client and s_server to print out a brief summary
1854 of connection parameters.
1857 *) Add callbacks for arbitrary TLS extensions.
1858 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1860 *) New option -crl_download in several openssl utilities to download CRLs
1861 from CRLDP extension in certificates.
1864 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1867 *) New function X509_CRL_diff to generate a delta CRL from the difference
1868 of two full CRLs. Add support to "crl" utility.
1871 *) New functions to set lookup_crls function and to retrieve
1872 X509_STORE from X509_STORE_CTX.
1875 *) Print out deprecated issuer and subject unique ID fields in
1879 *) Extend OCSP I/O functions so they can be used for simple general purpose
1880 HTTP as well as OCSP. New wrapper function which can be used to download
1881 CRLs using the OCSP API.
1884 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1887 *) SSL_CONF* functions. These provide a common framework for application
1888 configuration using configuration files or command lines.
1891 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1892 message callback and prints the results. Needs compile time option
1893 "enable-ssl-trace". New options to s_client and s_server to enable
1897 *) New ctrl and macro to retrieve supported points extensions.
1898 Print out extension in s_server and s_client.
1901 *) New functions to retrieve certificate signature and signature
1905 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1909 *) New Suite B modes for TLS code. These use and enforce the requirements
1910 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1911 only use Suite B curves. The Suite B modes can be set by using the
1912 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1915 *) New chain verification flags for Suite B levels of security. Check
1916 algorithms are acceptable when flags are set in X509_verify_cert.
1919 *) Make tls1_check_chain return a set of flags indicating checks passed
1920 by a certificate chain. Add additional tests to handle client
1921 certificates: checks for matching certificate type and issuer name
1925 *) If an attempt is made to use a signature algorithm not in the peer
1926 preference list abort the handshake. If client has no suitable
1927 signature algorithms in response to a certificate request do not
1928 use the certificate.
1931 *) If server EC tmp key is not in client preference list abort handshake.
1934 *) Add support for certificate stores in CERT structure. This makes it
1935 possible to have different stores per SSL structure or one store in
1936 the parent SSL_CTX. Include distinct stores for certificate chain
1937 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1938 to build and store a certificate chain in CERT structure: returning
1939 an error if the chain cannot be built: this will allow applications
1940 to test if a chain is correctly configured.
1942 Note: if the CERT based stores are not set then the parent SSL_CTX
1943 store is used to retain compatibility with existing behaviour.
1947 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1948 mask based on the current session, check mask when sending client
1949 hello and checking the requested ciphersuite.
1952 *) New ctrls to retrieve and set certificate types in a certificate
1953 request message. Print out received values in s_client. If certificate
1954 types is not set with custom values set sensible values based on
1955 supported signature algorithms.
1958 *) Support for distinct client and server supported signature algorithms.
1961 *) Add certificate callback. If set this is called whenever a certificate
1962 is required by client or server. An application can decide which
1963 certificate chain to present based on arbitrary criteria: for example
1964 supported signature algorithms. Add very simple example to s_server.
1965 This fixes many of the problems and restrictions of the existing client
1966 certificate callback: for example you can now clear an existing
1967 certificate and specify the whole chain.
1970 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1971 the certificate can be used for (if anything). Set valid_flags field
1972 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1973 to have similar checks in it.
1975 Add new "cert_flags" field to CERT structure and include a "strict mode".
1976 This enforces some TLS certificate requirements (such as only permitting
1977 certificate signature algorithms contained in the supported algorithms
1978 extension) which some implementations ignore: this option should be used
1979 with caution as it could cause interoperability issues.
1982 *) Update and tidy signature algorithm extension processing. Work out
1983 shared signature algorithms based on preferences and peer algorithms
1984 and print them out in s_client and s_server. Abort handshake if no
1985 shared signature algorithms.
1988 *) Add new functions to allow customised supported signature algorithms
1989 for SSL and SSL_CTX structures. Add options to s_client and s_server
1993 *) New function SSL_certs_clear() to delete all references to certificates
1994 from an SSL structure. Before this once a certificate had been added
1995 it couldn't be removed.
1998 *) Integrate hostname, email address and IP address checking with certificate
1999 verification. New verify options supporting checking in openssl utility.
2002 *) Fixes and wildcard matching support to hostname and email checking
2003 functions. Add manual page.
2004 [Florian Weimer (Red Hat Product Security Team)]
2006 *) New functions to check a hostname email or IP address against a
2007 certificate. Add options x509 utility to print results of checks against
2011 *) Fix OCSP checking.
2012 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2014 *) Initial experimental support for explicitly trusted non-root CAs.
2015 OpenSSL still tries to build a complete chain to a root but if an
2016 intermediate CA has a trust setting included that is used. The first
2017 setting is used: whether to trust (e.g., -addtrust option to the x509
2021 *) Add -trusted_first option which attempts to find certificates in the
2022 trusted store even if an untrusted chain is also supplied.
2025 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2026 platform support for Linux and Android.
2029 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2032 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2033 When in FIPS mode the approved implementations are used as normal,
2034 when not in FIPS mode the internal unapproved versions are used instead.
2035 This means that the FIPS capable OpenSSL isn't forced to use the
2036 (often lower performance) FIPS implementations outside FIPS mode.
2039 *) Transparently support X9.42 DH parameters when calling
2040 PEM_read_bio_DHparameters. This means existing applications can handle
2041 the new parameter format automatically.
2044 *) Initial experimental support for X9.42 DH parameter format: mainly
2045 to support use of 'q' parameter for RFC5114 parameters.
2048 *) Add DH parameters from RFC5114 including test data to dhtest.
2051 *) Support for automatic EC temporary key parameter selection. If enabled
2052 the most preferred EC parameters are automatically used instead of
2053 hardcoded fixed parameters. Now a server just has to call:
2054 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2055 support ECDH and use the most appropriate parameters.
2058 *) Enhance and tidy EC curve and point format TLS extension code. Use
2059 static structures instead of allocation if default values are used.
2060 New ctrls to set curves we wish to support and to retrieve shared curves.
2061 Print out shared curves in s_server. New options to s_server and s_client
2062 to set list of supported curves.
2065 *) New ctrls to retrieve supported signature algorithms and
2066 supported curve values as an array of NIDs. Extend openssl utility
2067 to print out received values.
2070 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2071 between NIDs and the more common NIST names such as "P-256". Enhance
2072 ecparam utility and ECC method to recognise the NIST names for curves.
2075 *) Enhance SSL/TLS certificate chain handling to support different
2076 chains for each certificate instead of one chain in the parent SSL_CTX.
2079 *) Support for fixed DH ciphersuite client authentication: where both
2080 server and client use DH certificates with common parameters.
2083 *) Support for fixed DH ciphersuites: those requiring DH server
2087 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2089 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2090 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2091 X509_CINF_get_signature were reverted post internal team review.
2093 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2095 *) Build fixes for the Windows and OpenVMS platforms
2096 [Matt Caswell and Richard Levitte]
2098 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2100 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2101 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2102 dereference. This could lead to a Denial Of Service attack. Thanks to
2103 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2107 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2108 dtls1_buffer_record function under certain conditions. In particular this
2109 could occur if an attacker sent repeated DTLS records with the same
2110 sequence number but for the next epoch. The memory leak could be exploited
2111 by an attacker in a Denial of Service attack through memory exhaustion.
2112 Thanks to Chris Mueller for reporting this issue.
2116 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2117 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2118 method would be set to NULL which could later result in a NULL pointer
2119 dereference. Thanks to Frank Schmirler for reporting this issue.
2123 *) Abort handshake if server key exchange message is omitted for ephemeral
2126 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2127 reporting this issue.
2131 *) Remove non-export ephemeral RSA code on client and server. This code
2132 violated the TLS standard by allowing the use of temporary RSA keys in
2133 non-export ciphersuites and could be used by a server to effectively
2134 downgrade the RSA key length used to a value smaller than the server
2135 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2136 INRIA or reporting this issue.
2140 *) Fixed issue where DH client certificates are accepted without verification.
2141 An OpenSSL server will accept a DH certificate for client authentication
2142 without the certificate verify message. This effectively allows a client to
2143 authenticate without the use of a private key. This only affects servers
2144 which trust a client certificate authority which issues certificates
2145 containing DH keys: these are extremely rare and hardly ever encountered.
2146 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2151 *) Ensure that the session ID context of an SSL is updated when its
2152 SSL_CTX is updated via SSL_set_SSL_CTX.
2154 The session ID context is typically set from the parent SSL_CTX,
2155 and can vary with the CTX.
2158 *) Fix various certificate fingerprint issues.
2160 By using non-DER or invalid encodings outside the signed portion of a
2161 certificate the fingerprint can be changed without breaking the signature.
2162 Although no details of the signed portion of the certificate can be changed
2163 this can cause problems with some applications: e.g. those using the
2164 certificate fingerprint for blacklists.
2166 1. Reject signatures with non zero unused bits.
2168 If the BIT STRING containing the signature has non zero unused bits reject
2169 the signature. All current signature algorithms require zero unused bits.
2171 2. Check certificate algorithm consistency.
2173 Check the AlgorithmIdentifier inside TBS matches the one in the
2174 certificate signature. NB: this will result in signature failure
2175 errors for some broken certificates.
2177 Thanks to Konrad Kraszewski from Google for reporting this issue.
2179 3. Check DSA/ECDSA signatures use DER.
2181 Re-encode DSA/ECDSA signatures and compare with the original received
2182 signature. Return an error if there is a mismatch.
2184 This will reject various cases including garbage after signature
2185 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2186 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2187 (negative or with leading zeroes).
2189 Further analysis was conducted and fixes were developed by Stephen Henson
2190 of the OpenSSL core team.
2195 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2196 results on some platforms, including x86_64. This bug occurs at random
2197 with a very low probability, and is not known to be exploitable in any
2198 way, though its exact impact is difficult to determine. Thanks to Pieter
2199 Wuille (Blockstream) who reported this issue and also suggested an initial
2200 fix. Further analysis was conducted by the OpenSSL development team and
2201 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2202 the OpenSSL core team.
2206 *) Do not resume sessions on the server if the negotiated protocol
2207 version does not match the session's version. Resuming with a different
2208 version, while not strictly forbidden by the RFC, is of questionable
2209 sanity and breaks all known clients.
2210 [David Benjamin, Emilia Käsper]
2212 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2213 early CCS messages during renegotiation. (Note that because
2214 renegotiation is encrypted, this early CCS was not exploitable.)
2217 *) Tighten client-side session ticket handling during renegotiation:
2218 ensure that the client only accepts a session ticket if the server sends
2219 the extension anew in the ServerHello. Previously, a TLS client would
2220 reuse the old extension state and thus accept a session ticket if one was
2221 announced in the initial ServerHello.
2223 Similarly, ensure that the client requires a session ticket if one
2224 was advertised in the ServerHello. Previously, a TLS client would
2225 ignore a missing NewSessionTicket message.
2228 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2230 *) SRTP Memory Leak.
2232 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2233 sends a carefully crafted handshake message, to cause OpenSSL to fail
2234 to free up to 64k of memory causing a memory leak. This could be
2235 exploited in a Denial Of Service attack. This issue affects OpenSSL
2236 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2237 whether SRTP is used or configured. Implementations of OpenSSL that
2238 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2240 The fix was developed by the OpenSSL team.
2244 *) Session Ticket Memory Leak.
2246 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2247 integrity of that ticket is first verified. In the event of a session
2248 ticket integrity check failing, OpenSSL will fail to free memory
2249 causing a memory leak. By sending a large number of invalid session
2250 tickets an attacker could exploit this issue in a Denial Of Service
2255 *) Build option no-ssl3 is incomplete.
2257 When OpenSSL is configured with "no-ssl3" as a build option, servers
2258 could accept and complete a SSL 3.0 handshake, and clients could be
2259 configured to send them.
2261 [Akamai and the OpenSSL team]
2263 *) Add support for TLS_FALLBACK_SCSV.
2264 Client applications doing fallback retries should call
2265 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2267 [Adam Langley, Bodo Moeller]
2269 *) Add additional DigestInfo checks.
2271 Re-encode DigestInto in DER and check against the original when
2272 verifying RSA signature: this will reject any improperly encoded
2273 DigestInfo structures.
2275 Note: this is a precautionary measure and no attacks are currently known.
2279 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2281 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2282 SRP code can be overrun an internal buffer. Add sanity check that
2283 g, A, B < N to SRP code.
2285 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2286 Group for discovering this issue.
2290 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2291 TLS 1.0 instead of higher protocol versions when the ClientHello message
2292 is badly fragmented. This allows a man-in-the-middle attacker to force a
2293 downgrade to TLS 1.0 even if both the server and the client support a
2294 higher protocol version, by modifying the client's TLS records.
2296 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2297 researching this issue.
2301 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2302 to a denial of service attack. A malicious server can crash the client
2303 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2304 ciphersuite and sending carefully crafted handshake messages.
2306 Thanks to Felix Gröbert (Google) for discovering and researching this
2311 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2312 to leak memory. This can be exploited through a Denial of Service attack.
2313 Thanks to Adam Langley for discovering and researching this issue.
2317 *) An attacker can force openssl to consume large amounts of memory whilst
2318 processing DTLS handshake messages. This can be exploited through a
2319 Denial of Service attack.
2320 Thanks to Adam Langley for discovering and researching this issue.
2324 *) An attacker can force an error condition which causes openssl to crash
2325 whilst processing DTLS packets due to memory being freed twice. This
2326 can be exploited through a Denial of Service attack.
2327 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2332 *) If a multithreaded client connects to a malicious server using a resumed
2333 session and the server sends an ec point format extension it could write
2334 up to 255 bytes to freed memory.
2336 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2341 *) A malicious server can crash an OpenSSL client with a null pointer
2342 dereference (read) by specifying an SRP ciphersuite even though it was not
2343 properly negotiated with the client. This can be exploited through a
2344 Denial of Service attack.
2346 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2347 discovering and researching this issue.
2351 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2352 X509_name_oneline, X509_name_print_ex et al. to leak some information
2353 from the stack. Applications may be affected if they echo pretty printing
2354 output to the attacker.
2356 Thanks to Ivan Fratric (Google) for discovering this issue.
2358 [Emilia Käsper, and Steve Henson]
2360 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2361 for corner cases. (Certain input points at infinity could lead to
2362 bogus results, with non-infinity inputs mapped to infinity too.)
2365 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2367 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2368 handshake can force the use of weak keying material in OpenSSL
2369 SSL/TLS clients and servers.
2371 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2372 researching this issue. (CVE-2014-0224)
2373 [KIKUCHI Masashi, Steve Henson]
2375 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2376 OpenSSL DTLS client the code can be made to recurse eventually crashing
2379 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2381 [Imre Rad, Steve Henson]
2383 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2384 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2385 client or server. This is potentially exploitable to run arbitrary
2386 code on a vulnerable client or server.
2388 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2389 [Jüri Aedla, Steve Henson]
2391 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2392 are subject to a denial of service attack.
2394 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2395 this issue. (CVE-2014-3470)
2396 [Felix Gröbert, Ivan Fratric, Steve Henson]
2398 *) Harmonize version and its documentation. -f flag is used to display
2400 [mancha <mancha1@zoho.com>]
2402 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2403 in i2d_ECPrivateKey.
2404 [mancha <mancha1@zoho.com>]
2406 *) Fix some double frees. These are not thought to be exploitable.
2407 [mancha <mancha1@zoho.com>]
2409 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2411 *) A missing bounds check in the handling of the TLS heartbeat extension
2412 can be used to reveal up to 64k of memory to a connected client or
2415 Thanks for Neel Mehta of Google Security for discovering this bug and to
2416 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2417 preparing the fix (CVE-2014-0160)
2418 [Adam Langley, Bodo Moeller]
2420 *) Fix for the attack described in the paper "Recovering OpenSSL
2421 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2422 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2423 http://eprint.iacr.org/2014/140
2425 Thanks to Yuval Yarom and Naomi Benger for discovering this
2426 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2427 [Yuval Yarom and Naomi Benger]
2429 *) TLS pad extension: draft-agl-tls-padding-03
2431 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2432 TLS client Hello record length value would otherwise be > 255 and
2433 less that 512 pad with a dummy extension containing zeroes so it
2434 is at least 512 bytes long.
2436 [Adam Langley, Steve Henson]
2438 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2440 *) Fix for TLS record tampering bug. A carefully crafted invalid
2441 handshake could crash OpenSSL with a NULL pointer exception.
2442 Thanks to Anton Johansson for reporting this issues.
2445 *) Keep original DTLS digest and encryption contexts in retransmission
2446 structures so we can use the previous session parameters if they need
2447 to be resent. (CVE-2013-6450)
2450 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2451 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2452 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2453 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2454 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2455 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2456 [Rob Stradling, Adam Langley]
2458 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2460 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2461 supporting platforms or when small records were transferred.
2462 [Andy Polyakov, Steve Henson]
2464 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2466 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2468 This addresses the flaw in CBC record processing discovered by
2469 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2470 at: http://www.isg.rhul.ac.uk/tls/
2472 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2473 Security Group at Royal Holloway, University of London
2474 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2475 Emilia Käsper for the initial patch.
2477 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2479 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2480 ciphersuites which can be exploited in a denial of service attack.
2481 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2482 and detecting this bug and to Wolfgang Ettlinger
2483 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2487 *) Return an error when checking OCSP signatures when key is NULL.
2488 This fixes a DoS attack. (CVE-2013-0166)
2491 *) Make openssl verify return errors.
2492 [Chris Palmer <palmer@google.com> and Ben Laurie]
2494 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2495 the right response is stapled. Also change SSL_get_certificate()
2496 so it returns the certificate actually sent.
2497 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2498 [Rob Stradling <rob.stradling@comodo.com>]
2500 *) Fix possible deadlock when decoding public keys.
2503 *) Don't use TLS 1.0 record version number in initial client hello
2507 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2509 *) Sanity check record length before skipping explicit IV in TLS
2510 1.2, 1.1 and DTLS to fix DoS attack.
2512 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2513 fuzzing as a service testing platform.
2517 *) Initialise tkeylen properly when encrypting CMS messages.
2518 Thanks to Solar Designer of Openwall for reporting this issue.
2521 *) In FIPS mode don't try to use composite ciphers as they are not
2525 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2527 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2528 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2529 mean any application compiled against OpenSSL 1.0.0 headers setting
2530 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2531 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2532 0x10000000L Any application which was previously compiled against
2533 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2534 will need to be recompiled as a result. Letting be results in
2535 inability to disable specifically TLS 1.1 and in client context,
2536 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2539 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2540 disable just protocol X, but all protocols above X *if* there are
2541 protocols *below* X still enabled. In more practical terms it means
2542 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2543 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2544 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2548 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2550 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2551 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2552 in CRYPTO_realloc_clean.
2554 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2555 issue and to Adam Langley <agl@chromium.org> for fixing it.
2557 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2559 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2562 *) Workarounds for some broken servers that "hang" if a client hello
2563 record length exceeds 255 bytes.
2565 1. Do not use record version number > TLS 1.0 in initial client
2566 hello: some (but not all) hanging servers will now work.
2567 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2568 the number of ciphers sent in the client hello. This should be
2569 set to an even number, such as 50, for example by passing:
2570 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2571 Most broken servers should now work.
2572 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2573 TLS 1.2 client support entirely.
2576 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2579 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2581 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2582 STRING form instead of a DigestInfo.
2585 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2586 and the RSA_sign/RSA_verify functions. This was made more apparent when
2587 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2588 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2589 the correct format in RSA_verify so both forms transparently work.
2592 *) Some servers which support TLS 1.0 can choke if we initially indicate
2593 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2594 encrypted premaster secret. As a workaround use the maximum permitted
2595 client version in client hello, this should keep such servers happy
2596 and still work with previous versions of OpenSSL.
2599 *) Add support for TLS/DTLS heartbeats.
2600 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2602 *) Add support for SCTP.
2603 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2605 *) Improved PRNG seeding for VOS.
2606 [Paul Green <Paul.Green@stratus.com>]
2608 *) Extensive assembler packs updates, most notably:
2610 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2611 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2612 - x86_64: bit-sliced AES implementation;
2613 - ARM: NEON support, contemporary platforms optimizations;
2614 - s390x: z196 support;
2615 - *: GHASH and GF(2^m) multiplication implementations;
2619 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2620 (removal of unnecessary code)
2621 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2623 *) Add TLS key material exporter from RFC 5705.
2626 *) Add DTLS-SRTP negotiation from RFC 5764.
2629 *) Add Next Protocol Negotiation,
2630 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2631 disabled with a no-npn flag to config or Configure. Code donated
2633 [Adam Langley <agl@google.com> and Ben Laurie]
2635 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2636 NIST-P256, NIST-P521, with constant-time single point multiplication on
2637 typical inputs. Compiler support for the nonstandard type __uint128_t is
2638 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2639 Code made available under Apache License version 2.0.
2641 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2642 line to include this in your build of OpenSSL, and run "make depend" (or
2643 "make update"). This enables the following EC_METHODs:
2645 EC_GFp_nistp224_method()
2646 EC_GFp_nistp256_method()
2647 EC_GFp_nistp521_method()
2649 EC_GROUP_new_by_curve_name() will automatically use these (while
2650 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2652 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2654 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2655 all platforms. Move ssize_t definition from e_os.h to the public
2656 header file e_os2.h as it now appears in public header file cms.h
2659 *) New -sigopt option to the ca, req and x509 utilities. Additional
2660 signature parameters can be passed using this option and in
2664 *) Add RSA PSS signing function. This will generate and set the
2665 appropriate AlgorithmIdentifiers for PSS based on those in the
2666 corresponding EVP_MD_CTX structure. No application support yet.
2669 *) Support for companion algorithm specific ASN1 signing routines.
2670 New function ASN1_item_sign_ctx() signs a pre-initialised
2671 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2672 the appropriate parameters.
2675 *) Add new algorithm specific ASN1 verification initialisation function
2676 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2677 handling will be the same no matter what EVP_PKEY_METHOD is used.
2678 Add a PSS handler to support verification of PSS signatures: checked
2679 against a number of sample certificates.
2682 *) Add signature printing for PSS. Add PSS OIDs.
2683 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2685 *) Add algorithm specific signature printing. An individual ASN1 method
2686 can now print out signatures instead of the standard hex dump.
2688 More complex signatures (e.g. PSS) can print out more meaningful
2689 information. Include DSA version that prints out the signature
2693 *) Password based recipient info support for CMS library: implementing
2697 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2698 neatly separates the code into cipher and PBE sections and is required
2699 for some algorithms that split PBES2 into separate pieces (such as
2700 password based CMS).
2703 *) Session-handling fixes:
2704 - Fix handling of connections that are resuming with a session ID,
2705 but also support Session Tickets.
2706 - Fix a bug that suppressed issuing of a new ticket if the client
2707 presented a ticket with an expired session.
2708 - Try to set the ticket lifetime hint to something reasonable.
2709 - Make tickets shorter by excluding irrelevant information.
2710 - On the client side, don't ignore renewed tickets.
2711 [Adam Langley, Bodo Moeller (Google)]
2713 *) Fix PSK session representation.
2716 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2718 This work was sponsored by Intel.
2721 *) Add GCM support to TLS library. Some custom code is needed to split
2722 the IV between the fixed (from PRF) and explicit (from TLS record)
2723 portions. This adds all GCM ciphersuites supported by RFC5288 and
2724 RFC5289. Generalise some AES* cipherstrings to include GCM and
2725 add a special AESGCM string for GCM only.
2728 *) Expand range of ctrls for AES GCM. Permit setting invocation
2729 field on decrypt and retrieval of invocation field only on encrypt.
2732 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2733 As required by RFC5289 these ciphersuites cannot be used if for
2734 versions of TLS earlier than 1.2.
2737 *) For FIPS capable OpenSSL interpret a NULL default public key method
2738 as unset and return the appropriate default but do *not* set the default.
2739 This means we can return the appropriate method in applications that
2740 switch between FIPS and non-FIPS modes.
2743 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2744 ENGINE is used then we cannot handle that in the FIPS module so we
2745 keep original code iff non-FIPS operations are allowed.
2748 *) Add -attime option to openssl utilities.
2749 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2751 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2754 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2755 FIPS EC methods unconditionally for now.
2758 *) New build option no-ec2m to disable characteristic 2 code.
2761 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2762 all cases can be covered as some introduce binary incompatibilities.
2765 *) Redirect RSA operations to FIPS module including keygen,
2766 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2769 *) Add similar low level API blocking to ciphers.
2772 *) Low level digest APIs are not approved in FIPS mode: any attempt
2773 to use these will cause a fatal error. Applications that *really* want
2774 to use them can use the private_* version instead.
2777 *) Redirect cipher operations to FIPS module for FIPS builds.
2780 *) Redirect digest operations to FIPS module for FIPS builds.
2783 *) Update build system to add "fips" flag which will link in fipscanister.o
2784 for static and shared library builds embedding a signature if needed.
2787 *) Output TLS supported curves in preference order instead of numerical
2788 order. This is currently hardcoded for the highest order curves first.
2789 This should be configurable so applications can judge speed vs strength.
2792 *) Add TLS v1.2 server support for client authentication.
2795 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2799 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2800 FIPS modules versions.
2803 *) Add TLS v1.2 client side support for client authentication. Keep cache
2804 of handshake records longer as we don't know the hash algorithm to use
2805 until after the certificate request message is received.
2808 *) Initial TLS v1.2 client support. Add a default signature algorithms
2809 extension including all the algorithms we support. Parse new signature
2810 format in client key exchange. Relax some ECC signing restrictions for
2811 TLS v1.2 as indicated in RFC5246.
2814 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2815 to new signature format when needed using client digest preference.
2816 All server ciphersuites should now work correctly in TLS v1.2. No client
2817 support yet and no support for client certificates.
2820 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2821 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2822 ciphersuites. At present only RSA key exchange ciphersuites work with
2823 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2824 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2825 and version checking.
2828 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2829 with this defined it will not be affected by any changes to ssl internal
2830 structures. Add several utility functions to allow openssl application
2831 to work with OPENSSL_NO_SSL_INTERN defined.
2835 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2837 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2840 *) Permit abbreviated handshakes when renegotiating using the function
2841 SSL_renegotiate_abbreviated().
2842 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2844 *) Add call to ENGINE_register_all_complete() to
2845 ENGINE_load_builtin_engines(), so some implementations get used
2846 automatically instead of needing explicit application support.
2849 *) Add support for TLS key exporter as described in RFC5705.
2850 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2852 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2853 a few changes are required:
2855 Add SSL_OP_NO_TLSv1_1 flag.
2856 Add TLSv1_1 methods.
2857 Update version checking logic to handle version 1.1.
2858 Add explicit IV handling (ported from DTLS code).
2859 Add command line options to s_client/s_server.
2862 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2864 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2865 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2866 content decryption and always return the same error. Note: this attack
2867 needs on average 2^20 messages so it only affects automated senders. The
2868 old behaviour can be re-enabled in the CMS code by setting the
2869 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2870 an MMA defence is not necessary.
2871 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2872 this issue. (CVE-2012-0884)
2875 *) Fix CVE-2011-4619: make sure we really are receiving a
2876 client hello before rejecting multiple SGC restarts. Thanks to
2877 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2880 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2882 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2883 Thanks to Antonio Martin, Enterprise Secure Access Research and
2884 Development, Cisco Systems, Inc. for discovering this bug and
2885 preparing a fix. (CVE-2012-0050)
2888 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2890 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2891 of the Vaudenay padding oracle attack on CBC mode encryption
2892 which enables an efficient plaintext recovery attack against
2893 the OpenSSL implementation of DTLS. Their attack exploits timing
2894 differences arising during decryption processing. A research
2895 paper describing this attack can be found at:
2896 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2897 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2898 Security Group at Royal Holloway, University of London
2899 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2900 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2901 for preparing the fix. (CVE-2011-4108)
2902 [Robin Seggelmann, Michael Tuexen]
2904 *) Clear bytes used for block padding of SSL 3.0 records.
2906 [Adam Langley (Google)]
2908 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2909 Kadianakis <desnacked@gmail.com> for discovering this issue and
2910 Adam Langley for preparing the fix. (CVE-2011-4619)
2911 [Adam Langley (Google)]
2913 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2914 [Andrey Kulikov <amdeich@gmail.com>]
2916 *) Prevent malformed RFC3779 data triggering an assertion failure.
2917 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2918 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2919 [Rob Austein <sra@hactrn.net>]
2921 *) Improved PRNG seeding for VOS.
2922 [Paul Green <Paul.Green@stratus.com>]
2924 *) Fix ssl_ciph.c set-up race.
2925 [Adam Langley (Google)]
2927 *) Fix spurious failures in ecdsatest.c.
2928 [Emilia Käsper (Google)]
2930 *) Fix the BIO_f_buffer() implementation (which was mixing different
2931 interpretations of the '..._len' fields).
2932 [Adam Langley (Google)]
2934 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2935 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2936 threads won't reuse the same blinding coefficients.
2938 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2939 lock to call BN_BLINDING_invert_ex, and avoids one use of
2940 BN_BLINDING_update for each BN_BLINDING structure (previously,
2941 the last update always remained unused).
2942 [Emilia Käsper (Google)]
2944 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2945 [Bob Buckholz (Google)]
2947 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2949 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2950 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2951 [Kaspar Brand <ossl@velox.ch>]
2953 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2954 for multi-threaded use of ECDH. (CVE-2011-3210)
2955 [Adam Langley (Google)]
2957 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2960 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2961 signature public key algorithm by using OID xref utilities instead.
2962 Before this you could only use some ECC ciphersuites with SHA1 only.
2965 *) Add protection against ECDSA timing attacks as mentioned in the paper
2966 by Billy Bob Brumley and Nicola Tuveri, see:
2968 http://eprint.iacr.org/2011/232.pdf
2970 [Billy Bob Brumley and Nicola Tuveri]
2972 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2974 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2975 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2977 *) Fix bug in string printing code: if *any* escaping is enabled we must
2978 escape the escape character (backslash) or the resulting string is
2982 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2984 *) Disable code workaround for ancient and obsolete Netscape browsers
2985 and servers: an attacker can use it in a ciphersuite downgrade attack.
2986 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2989 *) Fixed J-PAKE implementation error, originally discovered by
2990 Sebastien Martini, further info and confirmation from Stefan
2991 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2994 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2996 *) Fix extension code to avoid race conditions which can result in a buffer
2997 overrun vulnerability: resumed sessions must not be modified as they can
2998 be shared by multiple threads. CVE-2010-3864
3001 *) Fix WIN32 build system to correctly link an ENGINE directory into
3005 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3007 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3009 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3011 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3013 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3014 context. The operation can be customised via the ctrl mechanism in
3015 case ENGINEs want to include additional functionality.
3018 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3021 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3022 output hashes compatible with older versions of OpenSSL.
3023 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3025 *) Fix compression algorithm handling: if resuming a session use the
3026 compression algorithm of the resumed session instead of determining
3027 it from client hello again. Don't allow server to change algorithm.
3030 *) Add load_crls() function to apps tidying load_certs() too. Add option
3031 to verify utility to allow additional CRLs to be included.
3034 *) Update OCSP request code to permit adding custom headers to the request:
3035 some responders need this.
3038 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3040 [Julia Lawall <julia@diku.dk>]
3042 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3043 needlessly dereferenced structures, used obsolete functions and
3044 didn't handle all updated verify codes correctly.
3047 *) Disable MD2 in the default configuration.
3050 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3051 indicate the initial BIO being pushed or popped. This makes it possible
3052 to determine whether the BIO is the one explicitly called or as a result
3053 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3054 it handles reference counts correctly and doesn't zero out the I/O bio
3055 when it is not being explicitly popped. WARNING: applications which
3056 included workarounds for the old buggy behaviour will need to be modified
3057 or they could free up already freed BIOs.
3060 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3061 renaming to all platforms (within the 0.9.8 branch, this was
3062 done conditionally on Netware platforms to avoid a name clash).
3063 [Guenter <lists@gknw.net>]
3065 *) Add ECDHE and PSK support to DTLS.
3066 [Michael Tuexen <tuexen@fh-muenster.de>]
3068 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3072 *) Add "missing" function EVP_MD_flags() (without this the only way to
3073 retrieve a digest flags is by accessing the structure directly. Update
3074 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3075 or cipher is registered as in the "from" argument. Print out all
3076 registered digests in the dgst usage message instead of manually
3077 attempting to work them out.
3080 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3081 this allows the use of compression and extensions. Change default cipher
3082 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3083 by default unless an application cipher string requests it.
3086 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3087 key ids to find matching certificates and keys but some PKCS#12 files
3088 don't follow the (somewhat unwritten) rules and this strategy fails.
3089 Now just gather all certificates together and the first private key
3090 then look for the first certificate that matches the key.
3093 *) Support use of registered digest and cipher names for dgst and cipher
3094 commands instead of having to add each one as a special case. So now
3101 openssl dgst -sha256 foo
3103 and this works for ENGINE based algorithms too.
3107 *) Update Gost ENGINE to support parameter files.
3108 [Victor B. Wagner <vitus@cryptocom.ru>]
3110 *) Support GeneralizedTime in ca utility.
3111 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3113 *) Enhance the hash format used for certificate directory links. The new
3114 form uses the canonical encoding (meaning equivalent names will work
3115 even if they aren't identical) and uses SHA1 instead of MD5. This form
3116 is incompatible with the older format and as a result c_rehash should
3117 be used to rebuild symbolic links.
3120 *) Make PKCS#8 the default write format for private keys, replacing the
3121 traditional format. This form is standardised, more secure and doesn't
3122 include an implicit MD5 dependency.
3125 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3126 committed to OpenSSL should pass this lot as a minimum.
3129 *) Add session ticket override functionality for use by EAP-FAST.
3130 [Jouni Malinen <j@w1.fi>]
3132 *) Modify HMAC functions to return a value. Since these can be implemented
3133 in an ENGINE errors can occur.
3136 *) Type-checked OBJ_bsearch_ex.
3139 *) Type-checked OBJ_bsearch. Also some constification necessitated
3140 by type-checking. Still to come: TXT_DB, bsearch(?),
3141 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3145 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3146 seconds to a tm structure directly, instead of going through OS
3147 specific date routines. This avoids any issues with OS routines such
3148 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3149 and X509_time_adj_ex() to cover the extended range. The existing
3150 X509_time_adj() is still usable and will no longer have any date issues.
3153 *) Delta CRL support. New use deltas option which will attempt to locate
3154 and search any appropriate delta CRLs available.
3156 This work was sponsored by Google.
3159 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3160 code and add additional score elements. Validate alternate CRL paths
3161 as part of the CRL checking and indicate a new error "CRL path validation
3162 error" in this case. Applications wanting additional details can use
3163 the verify callback and check the new "parent" field. If this is not
3164 NULL CRL path validation is taking place. Existing applications won't
3165 see this because it requires extended CRL support which is off by
3168 This work was sponsored by Google.
3171 *) Support for freshest CRL extension.
3173 This work was sponsored by Google.
3176 *) Initial indirect CRL support. Currently only supported in the CRLs
3177 passed directly and not via lookup. Process certificate issuer
3178 CRL entry extension and lookup CRL entries by bother issuer name
3179 and serial number. Check and process CRL issuer entry in IDP extension.
3181 This work was sponsored by Google.
3184 *) Add support for distinct certificate and CRL paths. The CRL issuer
3185 certificate is validated separately in this case. Only enabled if
3186 an extended CRL support flag is set: this flag will enable additional
3187 CRL functionality in future.
3189 This work was sponsored by Google.
3192 *) Add support for policy mappings extension.
3194 This work was sponsored by Google.
3197 *) Fixes to pathlength constraint, self issued certificate handling,
3198 policy processing to align with RFC3280 and PKITS tests.
3200 This work was sponsored by Google.
3203 *) Support for name constraints certificate extension. DN, email, DNS
3204 and URI types are currently supported.
3206 This work was sponsored by Google.
3209 *) To cater for systems that provide a pointer-based thread ID rather
3210 than numeric, deprecate the current numeric thread ID mechanism and
3211 replace it with a structure and associated callback type. This
3212 mechanism allows a numeric "hash" to be extracted from a thread ID in
3213 either case, and on platforms where pointers are larger than 'long',
3214 mixing is done to help ensure the numeric 'hash' is usable even if it
3215 can't be guaranteed unique. The default mechanism is to use "&errno"
3216 as a pointer-based thread ID to distinguish between threads.
3218 Applications that want to provide their own thread IDs should now use
3219 CRYPTO_THREADID_set_callback() to register a callback that will call
3220 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3222 Note that ERR_remove_state() is now deprecated, because it is tied
3223 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3224 to free the current thread's error state should be replaced by
3225 ERR_remove_thread_state(NULL).
3227 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3228 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3229 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3230 application was previously providing a numeric thread callback that
3231 was inappropriate for distinguishing threads, then uniqueness might
3232 have been obtained with &errno that happened immediately in the
3233 intermediate development versions of OpenSSL; this is no longer the
3234 case, the numeric thread callback will now override the automatic use
3236 [Geoff Thorpe, with help from Bodo Moeller]
3238 *) Initial support for different CRL issuing certificates. This covers a
3239 simple case where the self issued certificates in the chain exist and
3240 the real CRL issuer is higher in the existing chain.
3242 This work was sponsored by Google.
3245 *) Removed effectively defunct crypto/store from the build.
3248 *) Revamp of STACK to provide stronger type-checking. Still to come:
3249 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3250 ASN1_STRING, CONF_VALUE.
3253 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3254 RAM on SSL connections. This option can save about 34k per idle SSL.
3257 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3258 STACK, TXT_DB, bsearch, qsort.
3261 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3262 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3263 support for data, signedData, compressedData, digestedData and
3264 encryptedData, envelopedData types included. Scripts to check against
3265 RFC4134 examples draft and interop and consistency checks of many
3266 content types and variants.
3269 *) Add options to enc utility to support use of zlib compression BIO.
3272 *) Extend mk1mf to support importing of options and assembly language
3273 files from Configure script, currently only included in VC-WIN32.
3274 The assembly language rules can now optionally generate the source
3275 files from the associated perl scripts.
3278 *) Implement remaining functionality needed to support GOST ciphersuites.
3279 Interop testing has been performed using CryptoPro implementations.
3280 [Victor B. Wagner <vitus@cryptocom.ru>]
3282 *) s390x assembler pack.
3285 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3289 *) Implement Opaque PRF Input TLS extension as specified in
3290 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3291 official specification yet and no extension type assignment by
3292 IANA exists, this extension (for now) will have to be explicitly
3293 enabled when building OpenSSL by providing the extension number
3294 to use. For example, specify an option
3296 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3298 to the "config" or "Configure" script to enable the extension,
3299 assuming extension number 0x9527 (which is a completely arbitrary
3300 and unofficial assignment based on the MD5 hash of the Internet
3301 Draft). Note that by doing so, you potentially lose
3302 interoperability with other TLS implementations since these might
3303 be using the same extension number for other purposes.
3305 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3306 opaque PRF input value to use in the handshake. This will create
3307 an interal copy of the length-'len' string at 'src', and will
3308 return non-zero for success.
3310 To get more control and flexibility, provide a callback function
3313 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3314 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3318 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3321 Callback function 'cb' will be called in handshakes, and is
3322 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3323 Argument 'arg' is for application purposes (the value as given to
3324 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3325 be provided to the callback function). The callback function
3326 has to return non-zero to report success: usually 1 to use opaque
3327 PRF input just if possible, or 2 to enforce use of the opaque PRF
3328 input. In the latter case, the library will abort the handshake
3329 if opaque PRF input is not successfully negotiated.
3331 Arguments 'peerinput' and 'len' given to the callback function
3332 will always be NULL and 0 in the case of a client. A server will
3333 see the client's opaque PRF input through these variables if
3334 available (NULL and 0 otherwise). Note that if the server
3335 provides an opaque PRF input, the length must be the same as the
3336 length of the client's opaque PRF input.
3338 Note that the callback function will only be called when creating
3339 a new session (session resumption can resume whatever was
3340 previously negotiated), and will not be called in SSL 2.0
3341 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3342 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3343 for applications that need to enforce opaque PRF input.
3347 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3350 [Victor B. Wagner <vitus@cryptocom.ru>]
3352 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3353 RFC4507bis. The encrypted ticket format is an encrypted encoded
3354 SSL_SESSION structure, that way new session features are automatically
3357 If a client application caches session in an SSL_SESSION structure
3358 support is transparent because tickets are now stored in the encoded
3361 The SSL_CTX structure automatically generates keys for ticket
3362 protection in servers so again support should be possible
3363 with no application modification.
3365 If a client or server wishes to disable RFC4507 support then the option
3366 SSL_OP_NO_TICKET can be set.
3368 Add a TLS extension debugging callback to allow the contents of any client
3369 or server extensions to be examined.
3371 This work was sponsored by Google.
3374 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3375 OpenSSL should now compile cleanly on gcc 4.2
3376 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3378 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3379 support including streaming MAC support: this is required for GOST
3380 ciphersuite support.
3381 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3383 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3384 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3385 to output in BER and PEM format.
3388 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3389 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3390 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3391 ENGINE support for HMAC keys which are unextractable. New -mac and
3392 -macopt options to dgst utility.
3395 *) New option -sigopt to dgst utility. Update dgst to use
3396 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3397 alternative signing parameters such as X9.31 or PSS in the dgst
3401 *) Change ssl_cipher_apply_rule(), the internal function that does
3402 the work each time a ciphersuite string requests enabling
3403 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3404 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3405 the order of disabled ciphersuites such that those ciphersuites
3406 that most recently went from enabled to disabled not only stay
3407 in order with respect to each other, but also have higher priority
3408 than other disabled ciphersuites the next time ciphersuites are
3411 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3412 the same ciphersuites as with "HIGH" alone, but in a specific
3413 order where the PSK ciphersuites come first (since they are the
3414 most recently disabled ciphersuites when "HIGH" is parsed).
3416 Also, change ssl_create_cipher_list() (using this new
3417 funcionality) such that between otherwise identical
3418 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3422 *) Change ssl_create_cipher_list() so that it automatically
3423 arranges the ciphersuites in reasonable order before starting
3424 to process the rule string. Thus, the definition for "DEFAULT"
3425 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3426 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3427 This makes it much easier to arrive at a reasonable default order
3428 in applications for which anonymous ciphers are OK (meaning
3429 that you can't actually use DEFAULT).
3430 [Bodo Moeller; suggested by Victor Duchovni]
3432 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3433 processing) into multiple integers instead of setting
3434 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3435 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3436 (These masks as well as the individual bit definitions are hidden
3437 away into the non-exported interface ssl/ssl_locl.h, so this
3438 change to the definition of the SSL_CIPHER structure shouldn't
3439 affect applications.) This give us more bits for each of these
3440 categories, so there is no longer a need to coagulate AES128 and
3441 AES256 into a single algorithm bit, and to coagulate Camellia128
3442 and Camellia256 into a single algorithm bit, which has led to all
3445 Thus, among other things, the kludge introduced in 0.9.7m and
3446 0.9.8e for masking out AES256 independently of AES128 or masking
3447 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3449 With the change, we also introduce new ciphersuite aliases that
3450 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3454 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3455 Use the leftmost N bytes of the signature input if the input is
3456 larger than the prime q (with N being the size in bytes of q).
3459 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3460 it yet and it is largely untested.
3463 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3466 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3467 some compilers (gcc 4.2 and later) reject their use. Safestack is
3468 reimplemented. Update ASN1 to avoid use of legacy functions.
3471 *) Win32/64 targets are linked with Winsock2.
3474 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3475 to external functions. This can be used to increase CRL handling
3476 efficiency especially when CRLs are very large by (for example) storing
3477 the CRL revoked certificates in a database.
3480 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3481 new CRLs added to a directory can be used. New command line option
3482 -verify_return_error to s_client and s_server. This causes real errors
3483 to be returned by the verify callback instead of carrying on no matter
3484 what. This reflects the way a "real world" verify callback would behave.
3487 *) GOST engine, supporting several GOST algorithms and public key formats.
3488 Kindly donated by Cryptocom.
3491 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3492 partitioned by DP are handled but no indirect CRL or reason partitioning
3493 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3494 selected via a scoring technique which handles IDP and AKID in CRLs.
3497 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3498 will ultimately be used for all verify operations: this will remove the
3499 X509_STORE dependency on certificate verification and allow alternative
3500 lookup methods. X509_STORE based implementations of these two callbacks.
3503 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3504 Modify get_crl() to find a valid (unexpired) CRL if possible.
3507 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3508 this would be called X509_CRL_cmp() but that name is already used by
3509 a function that just compares CRL issuer names. Cache several CRL
3510 extensions in X509_CRL structure and cache CRLDP in X509.
3513 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3514 this maps equivalent X509_NAME structures into a consistent structure.
3515 Name comparison can then be performed rapidly using memcmp().
3518 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3522 *) Allow digests to supply their own micalg string for S/MIME type using
3523 the ctrl EVP_MD_CTRL_MICALG.
3526 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3527 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3528 ctrl. It can then customise the structure before and/or after signing
3532 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3533 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3534 to free up any added signature OIDs.
3537 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3538 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3539 digest and cipher tables. New options added to openssl utility:
3540 list-message-digest-algorithms and list-cipher-algorithms.
3543 *) Change the array representation of binary polynomials: the list
3544 of degrees of non-zero coefficients is now terminated with -1.
3545 Previously it was terminated with 0, which was also part of the
3546 value; thus, the array representation was not applicable to
3547 polynomials where t^0 has coefficient zero. This change makes
3548 the array representation useful in a more general context.
3551 *) Various modifications and fixes to SSL/TLS cipher string
3552 handling. For ECC, the code now distinguishes between fixed ECDH
3553 with RSA certificates on the one hand and with ECDSA certificates
3554 on the other hand, since these are separate ciphersuites. The
3555 unused code for Fortezza ciphersuites has been removed.
3557 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3558 (not "ECDHE"). For consistency with the code for DH
3559 certificates, use of ECDH certificates is now considered ECDH
3560 authentication, not RSA or ECDSA authentication (the latter is
3561 merely the CA's signing algorithm and not actively used in the
3564 The temporary ciphersuite alias "ECCdraft" is no longer
3565 available, and ECC ciphersuites are no longer excluded from "ALL"
3566 and "DEFAULT". The following aliases now exist for RFC 4492
3567 ciphersuites, most of these by analogy with the DH case:
3569 kECDHr - ECDH cert, signed with RSA
3570 kECDHe - ECDH cert, signed with ECDSA
3571 kECDH - ECDH cert (signed with either RSA or ECDSA)
3572 kEECDH - ephemeral ECDH
3573 ECDH - ECDH cert or ephemeral ECDH
3579 AECDH - anonymous ECDH
3580 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3584 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3585 Use correct micalg parameters depending on digest(s) in signed message.
3588 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3589 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3592 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3593 an engine to register a method. Add ENGINE lookups for methods and
3594 functional reference processing.
3597 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3598 EVP_{Sign,Verify}* which allow an application to customise the signature
3602 *) New -resign option to smime utility. This adds one or more signers
3603 to an existing PKCS#7 signedData structure. Also -md option to use an
3604 alternative message digest algorithm for signing.
3607 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3608 create PKCS7 structures containing multiple signers. Update smime
3609 application to support multiple signers.
3612 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3616 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3617 Reorganize PBE internals to lookup from a static table using NIDs,
3618 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3619 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3620 PRF which will be automatically used with PBES2.
3623 *) Replace the algorithm specific calls to generate keys in "req" with the
3627 *) Update PKCS#7 enveloped data routines to use new API. This is now
3628 supported by any public key method supporting the encrypt operation. A
3629 ctrl is added to allow the public key algorithm to examine or modify
3630 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3634 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3635 a default digest type to use. In most cases this will be SHA1 but some
3636 algorithms (such as GOST) need to specify an alternative digest. The
3637 return value indicates how strong the preference is 1 means optional and
3638 2 is mandatory (that is it is the only supported type). Modify
3639 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3640 use the default md. Update openssl utilities to use the default digest
3641 type for signing if it is not explicitly indicated.
3644 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3645 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3646 signing method from the key type. This effectively removes the link
3647 between digests and public key types.
3650 *) Add an OID cross reference table and utility functions. Its purpose is to
3651 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3652 rsaEncryption. This will allow some of the algorithm specific hackery
3653 needed to use the correct OID to be removed.
3656 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3657 structures for PKCS7_sign(). They are now set up by the relevant public
3661 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3664 *) Add support for key derivation (agreement) in the API, DH method and
3668 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3669 public and private key formats. As a side effect these add additional
3670 command line functionality not previously available: DSA signatures can be
3671 generated and verified using pkeyutl and DH key support and generation in
3676 [Oliver Tappe <zooey@hirschkaefer.de>]
3678 *) New make target "install_html_docs" installs HTML renditions of the
3680 [Oliver Tappe <zooey@hirschkaefer.de>]
3682 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3683 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3684 support key and parameter generation and add initial key generation
3685 functionality for RSA.
3688 *) Add functions for main EVP_PKEY_method operations. The undocumented
3689 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3690 EVP_PKEY_{encrypt,decrypt}_old.
3693 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3694 key API, doesn't do much yet.
3697 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3698 public key algorithms. New option to openssl utility:
3699 "list-public-key-algorithms" to print out info.
3702 *) Implement the Supported Elliptic Curves Extension for
3703 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3706 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3707 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3710 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3711 utilities such as rsa, dsa, dsaparam etc except they process any key
3715 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3716 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3717 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3721 *) Initial support for pluggable public key ASN1.
3722 De-spaghettify the public key ASN1 handling. Move public and private
3723 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3724 algorithm specific handling to a single module within the relevant
3725 algorithm directory. Add functions to allow (near) opaque processing
3726 of public and private key structures.
3729 *) Implement the Supported Point Formats Extension for
3730 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3733 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3734 for the psk identity [hint] and the psk callback functions to the
3735 SSL_SESSION, SSL and SSL_CTX structure.
3738 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3742 SSL_CTX_use_psk_identity_hint
3743 SSL_get_psk_identity_hint
3744 SSL_get_psk_identity
3745 SSL_use_psk_identity_hint
3747 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3749 *) Add RFC 3161 compliant time stamp request creation, response generation
3750 and response verification functionality.
3751 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3753 *) Add initial support for TLS extensions, specifically for the server_name
3754 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3755 have new members for a host name. The SSL data structure has an
3756 additional member SSL_CTX *initial_ctx so that new sessions can be
3757 stored in that context to allow for session resumption, even after the
3758 SSL has been switched to a new SSL_CTX in reaction to a client's
3759 server_name extension.
3761 New functions (subject to change):
3763 SSL_get_servername()
3764 SSL_get_servername_type()
3767 New CTRL codes and macros (subject to change):
3769 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3770 - SSL_CTX_set_tlsext_servername_callback()
3771 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3772 - SSL_CTX_set_tlsext_servername_arg()
3773 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3775 openssl s_client has a new '-servername ...' option.
3777 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3778 '-key2 ...', '-servername_fatal' (subject to change). This allows
3779 testing the HostName extension for a specific single host name ('-cert'
3780 and '-key' remain fallbacks for handshakes without HostName
3781 negotiation). If the unrecognized_name alert has to be sent, this by
3782 default is a warning; it becomes fatal with the '-servername_fatal'
3785 [Peter Sylvester, Remy Allais, Christophe Renou]
3787 *) Whirlpool hash implementation is added.
3790 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3791 bn(64,32). Because of instruction set limitations it doesn't have
3792 any negative impact on performance. This was done mostly in order
3793 to make it possible to share assembler modules, such as bn_mul_mont
3794 implementations, between 32- and 64-bit builds without hassle.
3797 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3798 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3802 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3803 dedicated Montgomery multiplication procedure, is introduced.
3804 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3805 "64-bit" performance on certain 32-bit targets.
3808 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3809 in SSL structures. New SSL ctrl to set maximum send fragment size.
3810 Save memory by seeting the I/O buffer sizes dynamically instead of
3811 using the maximum available value.
3814 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3815 in addition to the text details.
3818 *) Very, very preliminary EXPERIMENTAL support for printing of general
3819 ASN1 structures. This currently produces rather ugly output and doesn't
3820 handle several customised structures at all.
3823 *) Integrated support for PVK file format and some related formats such
3824 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3825 these in the 'rsa' and 'dsa' utilities.
3828 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3831 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3832 place for the (very old) "NETSCAPE" format certificates which are now
3833 handled using new ASN1 code equivalents.
3836 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3837 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3838 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3841 *) Modify CRL distribution points extension code to print out previously
3842 unsupported fields. Enhance extension setting code to allow setting of
3846 *) Add print and set support for Issuing Distribution Point CRL extension.
3849 *) Change 'Configure' script to enable Camellia by default.
3852 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3854 *) When rejecting SSL/TLS records due to an incorrect version number, never
3855 update s->server with a new major version number. As of
3856 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3857 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3858 the previous behavior could result in a read attempt at NULL when
3859 receiving specific incorrect SSL/TLS records once record payload
3860 protection is active. (CVE-2010-0740)
3861 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3863 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3864 could be crashed if the relevant tables were not present (e.g. chrooted).
3865 [Tomas Hoger <thoger@redhat.com>]
3867 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3869 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3870 [Martin Olsson, Neel Mehta]
3872 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3873 accommodate for stack sorting, always a write lock!).
3876 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3877 excessive delays in the RAND_poll(): over a minute. As a workaround
3878 include a time check in the inner Heap32Next loop too.
3881 *) The code that handled flushing of data in SSL/TLS originally used the
3882 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3883 the problem outlined in PR#1949. The fix suggested there however can
3884 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3885 of Apache). So instead simplify the code to flush unconditionally.
3886 This should be fine since flushing with no data to flush is a no op.
3889 *) Handle TLS versions 2.0 and later properly and correctly use the
3890 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3891 off ancient servers have a habit of sticking around for a while...
3894 *) Modify compression code so it frees up structures without using the
3895 ex_data callbacks. This works around a problem where some applications
3896 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3897 restarting) then use compression (e.g. SSL with compression) later.
3898 This results in significant per-connection memory leaks and
3899 has caused some security issues including CVE-2008-1678 and
3903 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3904 change when encrypting or decrypting.
3907 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3908 connect and renegotiate with servers which do not support RI.
3909 Until RI is more widely deployed this option is enabled by default.
3912 *) Add "missing" ssl ctrls to clear options and mode.
3915 *) If client attempts to renegotiate and doesn't support RI respond with
3916 a no_renegotiation alert as required by RFC5746. Some renegotiating
3917 TLS clients will continue a connection gracefully when they receive
3918 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3919 waiting for a server hello which it will never receive. Now we treat a
3920 received no_renegotiation alert as a fatal error. This is because
3921 applications requesting a renegotiation might well expect it to succeed
3922 and would have no code in place to handle the server denying it so the
3923 only safe thing to do is to terminate the connection.
3926 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3927 peer supports secure renegotiation and 0 otherwise. Print out peer
3928 renegotiation support in s_client/s_server.
3931 *) Replace the highly broken and deprecated SPKAC certification method with
3932 the updated NID creation version. This should correctly handle UTF8.
3935 *) Implement RFC5746. Re-enable renegotiation but require the extension
3936 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3937 turns out to be a bad idea. It has been replaced by
3938 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3939 SSL_CTX_set_options(). This is really not recommended unless you
3940 know what you are doing.
3941 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3943 *) Fixes to stateless session resumption handling. Use initial_ctx when
3944 issuing and attempting to decrypt tickets in case it has changed during
3945 servername handling. Use a non-zero length session ID when attempting
3946 stateless session resumption: this makes it possible to determine if
3947 a resumption has occurred immediately after receiving server hello
3948 (several places in OpenSSL subtly assume this) instead of later in
3952 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3953 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3954 fixes for a few places where the return code is not checked
3956 [Julia Lawall <julia@diku.dk>]
3958 *) Add --strict-warnings option to Configure script to include devteam
3959 warnings in other configurations.
3962 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3963 makes it possible to install openssl libraries in locations which
3964 have names other than "lib", for example "/usr/lib64" which some
3966 [Steve Henson, based on patch from Jeremy Utley]
3968 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3969 X690 8.9.12 and can produce some misleading textual output of OIDs.
3970 [Steve Henson, reported by Dan Kaminsky]
3972 *) Delete MD2 from algorithm tables. This follows the recommendation in
3973 several standards that it is not used in new applications due to
3974 several cryptographic weaknesses. For binary compatibility reasons
3975 the MD2 API is still compiled in by default.
3978 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3982 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3983 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3985 [Guenter <lists@gknw.net>]
3987 *) Fix the server certificate chain building code to use X509_verify_cert(),
3988 it used to have an ad-hoc builder which was unable to cope with anything
3989 other than a simple chain.
3990 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3992 *) Don't check self signed certificate signatures in X509_verify_cert()
3993 by default (a flag can override this): it just wastes time without
3994 adding any security. As a useful side effect self signed root CAs
3995 with non-FIPS digests are now usable in FIPS mode.
3998 *) In dtls1_process_out_of_seq_message() the check if the current message
3999 is already buffered was missing. For every new message was memory
4000 allocated, allowing an attacker to perform an denial of service attack
4001 with sending out of seq handshake messages until there is no memory
4002 left. Additionally every future messege was buffered, even if the
4003 sequence number made no sense and would be part of another handshake.
4004 So only messages with sequence numbers less than 10 in advance will be
4005 buffered. (CVE-2009-1378)
4006 [Robin Seggelmann, discovered by Daniel Mentz]
4008 *) Records are buffered if they arrive with a future epoch to be
4009 processed after finishing the corresponding handshake. There is
4010 currently no limitation to this buffer allowing an attacker to perform
4011 a DOS attack with sending records with future epochs until there is no
4012 memory left. This patch adds the pqueue_size() function to determine
4013 the size of a buffer and limits the record buffer to 100 entries.
4015 [Robin Seggelmann, discovered by Daniel Mentz]
4017 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4018 parent structure is freed. (CVE-2009-1379)
4021 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4022 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4025 [Ilya O. <vrghost@gmail.com>]
4027 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4029 *) Disable renegotiation completely - this fixes a severe security
4030 problem (CVE-2009-3555) at the cost of breaking all
4031 renegotiation. Renegotiation can be re-enabled by setting
4032 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4033 run-time. This is really not recommended unless you know what
4037 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4039 *) Don't set val to NULL when freeing up structures, it is freed up by
4040 underlying code. If sizeof(void *) > sizeof(long) this can result in
4041 zeroing past the valid field. (CVE-2009-0789)
4042 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4044 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4045 checked correctly. This would allow some invalid signed attributes to
4046 appear to verify correctly. (CVE-2009-0591)
4047 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4049 *) Reject UniversalString and BMPString types with invalid lengths. This
4050 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4051 a legal length. (CVE-2009-0590)
4054 *) Set S/MIME signing as the default purpose rather than setting it
4055 unconditionally. This allows applications to override it at the store
4059 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4060 to handle some structures.
4063 *) Improve efficiency of mem_gets: don't search whole buffer each time
4065 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4067 *) New -hex option for openssl rand.
4070 *) Print out UTF8String and NumericString when parsing ASN1.
4073 *) Support NumericString type for name components.
4076 *) Allow CC in the environment to override the automatically chosen
4077 compiler. Note that nothing is done to ensure flags work with the
4081 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4083 *) Properly check EVP_VerifyFinal() and similar return values
4085 [Ben Laurie, Bodo Moeller, Google Security Team]
4087 *) Enable TLS extensions by default.
4090 *) Allow the CHIL engine to be loaded, whether the application is
4091 multithreaded or not. (This does not release the developer from the
4092 obligation to set up the dynamic locking callbacks.)
4093 [Sander Temme <sander@temme.net>]
4095 *) Use correct exit code if there is an error in dgst command.
4096 [Steve Henson; problem pointed out by Roland Dirlewanger]
4098 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4099 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4102 *) Add experimental JPAKE support, including demo authentication in
4103 s_client and s_server.
4106 *) Set the comparison function in v3_addr_canonize().
4107 [Rob Austein <sra@hactrn.net>]
4109 *) Add support for XMPP STARTTLS in s_client.
4110 [Philip Paeps <philip@freebsd.org>]
4112 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4113 to ensure that even with this option, only ciphersuites in the
4114 server's preference list will be accepted. (Note that the option
4115 applies only when resuming a session, so the earlier behavior was
4116 just about the algorithm choice for symmetric cryptography.)
4119 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4121 *) Fix NULL pointer dereference if a DTLS server received
4122 ChangeCipherSpec as first record (CVE-2009-1386).
4125 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4126 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4129 *) The fix in 0.9.8c that supposedly got rid of unsafe
4130 double-checked locking was incomplete for RSA blinding,
4131 addressing just one layer of what turns out to have been
4132 doubly unsafe triple-checked locking.
4134 So now fix this for real by retiring the MONT_HELPER macro
4135 in crypto/rsa/rsa_eay.c.
4137 [Bodo Moeller; problem pointed out by Marius Schilder]
4139 *) Various precautionary measures:
4141 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4143 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4144 (NB: This would require knowledge of the secret session ticket key
4145 to exploit, in which case you'd be SOL either way.)
4147 - Change bn_nist.c so that it will properly handle input BIGNUMs
4148 outside the expected range.
4150 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4153 [Neel Mehta, Bodo Moeller]
4155 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4156 the load fails. Useful for distros.
4157 [Ben Laurie and the FreeBSD team]
4159 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4162 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4165 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4167 This work was sponsored by Logica.
4170 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4171 keystores. Support for SSL/TLS client authentication too.
4172 Not compiled unless enable-capieng specified to Configure.
4174 This work was sponsored by Logica.
4177 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
4178 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4179 attribute creation routines such as certificate requests and PKCS#12
4183 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4185 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4186 handshake which could lead to a cilent crash as found using the
4187 Codenomicon TLS test suite (CVE-2008-1672)
4188 [Steve Henson, Mark Cox]
4190 *) Fix double free in TLS server name extensions which could lead to
4191 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4194 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4196 Clear the error queue to ensure that error entries left from
4197 older function calls do not interfere with the correct operation.
4198 [Lutz Jaenicke, Erik de Castro Lopo]
4200 *) Remove root CA certificates of commercial CAs:
4202 The OpenSSL project does not recommend any specific CA and does not
4203 have any policy with respect to including or excluding any CA.
4204 Therefore it does not make any sense to ship an arbitrary selection
4205 of root CA certificates with the OpenSSL software.
4208 *) RSA OAEP patches to fix two separate invalid memory reads.
4209 The first one involves inputs when 'lzero' is greater than
4210 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4211 before the beginning of from). The second one involves inputs where
4212 the 'db' section contains nothing but zeroes (there is a one-byte
4213 invalid read after the end of 'db').
4214 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4216 *) Partial backport from 0.9.9-dev:
4218 Introduce bn_mul_mont (dedicated Montgomery multiplication
4219 procedure) as a candidate for BIGNUM assembler implementation.
4220 While 0.9.9-dev uses assembler for various architectures, only
4221 x86_64 is available by default here in the 0.9.8 branch, and
4222 32-bit x86 is available through a compile-time setting.
4224 To try the 32-bit x86 assembler implementation, use Configure
4225 option "enable-montasm" (which exists only for this backport).
4227 As "enable-montasm" for 32-bit x86 disclaims code stability
4228 anyway, in this constellation we activate additional code
4229 backported from 0.9.9-dev for further performance improvements,
4230 namely BN_from_montgomery_word. (To enable this otherwise,
4231 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4233 [Andy Polyakov (backport partially by Bodo Moeller)]
4235 *) Add TLS session ticket callback. This allows an application to set
4236 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4237 values. This is useful for key rollover for example where several key
4238 sets may exist with different names.
4241 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4242 This was broken until now in 0.9.8 releases, such that the only way
4243 a registered ENGINE could be used (assuming it initialises
4244 successfully on the host) was to explicitly set it as the default
4245 for the relevant algorithms. This is in contradiction with 0.9.7
4246 behaviour and the documentation. With this fix, when an ENGINE is
4247 registered into a given algorithm's table of implementations, the
4248 'uptodate' flag is reset so that auto-discovery will be used next
4249 time a new context for that algorithm attempts to select an
4251 [Ian Lister (tweaked by Geoff Thorpe)]
4253 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4254 implementation in the following ways:
4256 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4259 Lack of BER streaming support means one pass streaming processing is
4260 only supported if data is detached: setting the streaming flag is
4261 ignored for embedded content.
4263 CMS support is disabled by default and must be explicitly enabled
4264 with the enable-cms configuration option.
4267 *) Update the GMP engine glue to do direct copies between BIGNUM and
4268 mpz_t when openssl and GMP use the same limb size. Otherwise the
4269 existing "conversion via a text string export" trick is still used.
4270 [Paul Sheer <paulsheer@gmail.com>]
4272 *) Zlib compression BIO. This is a filter BIO which compressed and
4273 uncompresses any data passed through it.
4276 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4277 RFC3394 compatible AES key wrapping.
4280 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4281 sets string data without copying. X509_ALGOR_set0() and
4282 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4283 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4284 from an X509_ATTRIBUTE structure optionally checking it occurs only
4285 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4289 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4290 to get the expected BN_FLG_CONSTTIME behavior.
4291 [Bodo Moeller (Google)]
4295 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4296 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4297 - added some more tests to do_tests.pl
4298 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4299 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4300 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4301 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4302 - various changes to netware.pl to enable gcc-cross builds on Win32
4304 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4305 - various changes to fix missing prototype warnings
4306 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4307 - added AES, WHIRLPOOL and CPUID assembler code to build files
4308 - added missing AES assembler make rules to mk1mf.pl
4309 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4310 [Guenter Knauf <eflash@gmx.net>]
4312 *) Implement certificate status request TLS extension defined in RFC3546.
4313 A client can set the appropriate parameters and receive the encoded
4314 OCSP response via a callback. A server can query the supplied parameters
4315 and set the encoded OCSP response in the callback. Add simplified examples
4316 to s_client and s_server.
4319 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4321 *) Fix various bugs:
4322 + Binary incompatibility of ssl_ctx_st structure
4323 + DTLS interoperation with non-compliant servers
4324 + Don't call get_session_cb() without proposed session
4325 + Fix ia64 assembler code
4326 [Andy Polyakov, Steve Henson]
4328 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4330 *) DTLS Handshake overhaul. There were longstanding issues with
4331 OpenSSL DTLS implementation, which were making it impossible for
4332 RFC 4347 compliant client to communicate with OpenSSL server.
4333 Unfortunately just fixing these incompatibilities would "cut off"
4334 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4335 server keeps tolerating non RFC compliant syntax. The opposite is
4336 not true, 0.9.8f client can not communicate with earlier server.
4337 This update even addresses CVE-2007-4995.
4340 *) Changes to avoid need for function casts in OpenSSL: some compilers
4341 (gcc 4.2 and later) reject their use.
4342 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4345 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4346 RFC4507bis. The encrypted ticket format is an encrypted encoded
4347 SSL_SESSION structure, that way new session features are automatically
4350 If a client application caches session in an SSL_SESSION structure
4351 support is transparent because tickets are now stored in the encoded
4354 The SSL_CTX structure automatically generates keys for ticket
4355 protection in servers so again support should be possible
4356 with no application modification.
4358 If a client or server wishes to disable RFC4507 support then the option
4359 SSL_OP_NO_TICKET can be set.
4361 Add a TLS extension debugging callback to allow the contents of any client
4362 or server extensions to be examined.
4364 This work was sponsored by Google.
4367 *) Add initial support for TLS extensions, specifically for the server_name
4368 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4369 have new members for a host name. The SSL data structure has an
4370 additional member SSL_CTX *initial_ctx so that new sessions can be
4371 stored in that context to allow for session resumption, even after the
4372 SSL has been switched to a new SSL_CTX in reaction to a client's
4373 server_name extension.
4375 New functions (subject to change):
4377 SSL_get_servername()
4378 SSL_get_servername_type()
4381 New CTRL codes and macros (subject to change):
4383 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4384 - SSL_CTX_set_tlsext_servername_callback()
4385 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4386 - SSL_CTX_set_tlsext_servername_arg()
4387 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4389 openssl s_client has a new '-servername ...' option.
4391 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4392 '-key2 ...', '-servername_fatal' (subject to change). This allows
4393 testing the HostName extension for a specific single host name ('-cert'
4394 and '-key' remain fallbacks for handshakes without HostName
4395 negotiation). If the unrecognized_name alert has to be sent, this by
4396 default is a warning; it becomes fatal with the '-servername_fatal'
4399 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4401 *) Add AES and SSE2 assembly language support to VC++ build.
4404 *) Mitigate attack on final subtraction in Montgomery reduction.
4407 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4408 (which previously caused an internal error).
4411 *) Squeeze another 10% out of IGE mode when in != out.
4414 *) AES IGE mode speedup.
4415 [Dean Gaudet (Google)]
4417 *) Add the Korean symmetric 128-bit cipher SEED (see
4418 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4419 add SEED ciphersuites from RFC 4162:
4421 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4422 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4423 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4424 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4426 To minimize changes between patchlevels in the OpenSSL 0.9.8
4427 series, SEED remains excluded from compilation unless OpenSSL
4428 is configured with 'enable-seed'.
4429 [KISA, Bodo Moeller]
4431 *) Mitigate branch prediction attacks, which can be practical if a
4432 single processor is shared, allowing a spy process to extract
4433 information. For detailed background information, see
4434 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4435 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4436 and Necessary Software Countermeasures"). The core of the change
4437 are new versions BN_div_no_branch() and
4438 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4439 respectively, which are slower, but avoid the security-relevant
4440 conditional branches. These are automatically called by BN_div()
4441 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4442 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4443 remove a conditional branch.
4445 BN_FLG_CONSTTIME is the new name for the previous
4446 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4447 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4448 in the exponent causes BN_mod_exp_mont() to use the alternative
4449 implementation in BN_mod_exp_mont_consttime().) The old name
4450 remains as a deprecated alias.
4452 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4453 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4454 constant-time implementations for more than just exponentiation.
4455 Here too the old name is kept as a deprecated alias.
4457 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4458 the BN_BLINDING structure gets an independent copy of the
4459 modulus. This means that the previous "BIGNUM *m" argument to
4460 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4461 essentially becomes "const BIGNUM *m", although we can't actually
4462 change this in the header file before 0.9.9. It allows
4463 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4464 enable BN_FLG_CONSTTIME.
4466 [Matthew D Wood (Intel Corp)]
4468 *) In the SSL/TLS server implementation, be strict about session ID
4469 context matching (which matters if an application uses a single
4470 external cache for different purposes). Previously,
4471 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4472 set. This did ensure strict client verification, but meant that,
4473 with applications using a single external cache for quite
4474 different requirements, clients could circumvent ciphersuite
4475 restrictions for a given session ID context by starting a session
4476 in a different context.
4479 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4480 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4481 authentication-only ciphersuites.
4484 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4485 not complete and could lead to a possible single byte overflow
4486 (CVE-2007-5135) [Ben Laurie]
4488 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4490 *) Since AES128 and AES256 (and similarly Camellia128 and
4491 Camellia256) share a single mask bit in the logic of
4492 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4493 kludge to work properly if AES128 is available and AES256 isn't
4494 (or if Camellia128 is available and Camellia256 isn't).
4497 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4498 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4499 When a point or a seed is encoded in a BIT STRING, we need to
4500 prevent the removal of trailing zero bits to get the proper DER
4501 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4502 of a NamedBitList, for which trailing 0 bits need to be removed.)
4505 *) Have SSL/TLS server implementation tolerate "mismatched" record
4506 protocol version while receiving ClientHello even if the
4507 ClientHello is fragmented. (The server can't insist on the
4508 particular protocol version it has chosen before the ServerHello
4509 message has informed the client about his choice.)
4512 *) Add RFC 3779 support.
4513 [Rob Austein for ARIN, Ben Laurie]
4515 *) Load error codes if they are not already present instead of using a
4516 static variable. This allows them to be cleanly unloaded and reloaded.
4517 Improve header file function name parsing.
4520 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4521 or CAPABILITY handshake as required by RFCs.
4524 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4526 *) Introduce limits to prevent malicious keys being able to
4527 cause a denial of service. (CVE-2006-2940)
4528 [Steve Henson, Bodo Moeller]
4530 *) Fix ASN.1 parsing of certain invalid structures that can result
4531 in a denial of service. (CVE-2006-2937) [Steve Henson]
4533 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4534 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4536 *) Fix SSL client code which could crash if connecting to a
4537 malicious SSLv2 server. (CVE-2006-4343)
4538 [Tavis Ormandy and Will Drewry, Google Security Team]
4540 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
4541 match only those. Before that, "AES256-SHA" would be interpreted
4542 as a pattern and match "AES128-SHA" too (since AES128-SHA got
4543 the same strength classification in 0.9.7h) as we currently only
4544 have a single AES bit in the ciphersuite description bitmap.
4545 That change, however, also applied to ciphersuite strings such as
4546 "RC4-MD5" that intentionally matched multiple ciphersuites --
4547 namely, SSL 2.0 ciphersuites in addition to the more common ones
4548 from SSL 3.0/TLS 1.0.
4550 So we change the selection algorithm again: Naming an explicit
4551 ciphersuite selects this one ciphersuite, and any other similar
4552 ciphersuite (same bitmap) from *other* protocol versions.
4553 Thus, "RC4-MD5" again will properly select both the SSL 2.0
4554 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
4556 Since SSL 2.0 does not have any ciphersuites for which the
4557 128/256 bit distinction would be relevant, this works for now.
4558 The proper fix will be to use different bits for AES128 and
4559 AES256, which would have avoided the problems from the beginning;
4560 however, bits are scarce, so we can only do this in a new release
4561 (not just a patchlevel) when we can change the SSL_CIPHER
4562 definition to split the single 'unsigned long mask' bitmap into
4563 multiple values to extend the available space.
4567 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
4569 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4570 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4572 *) Add AES IGE and biIGE modes.
4575 *) Change the Unix randomness entropy gathering to use poll() when
4576 possible instead of select(), since the latter has some
4577 undesirable limitations.
4578 [Darryl Miles via Richard Levitte and Bodo Moeller]
4580 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
4581 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
4582 cannot be implicitly activated as part of, e.g., the "AES" alias.
4583 However, please upgrade to OpenSSL 0.9.9[-dev] for
4584 non-experimental use of the ECC ciphersuites to get TLS extension
4585 support, which is required for curve and point format negotiation
4586 to avoid potential handshake problems.
4589 *) Disable rogue ciphersuites:
4591 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4592 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4593 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4595 The latter two were purportedly from
4596 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4599 Also deactivate the remaining ciphersuites from
4600 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4601 unofficial, and the ID has long expired.
4604 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
4605 dual-core machines) and other potential thread-safety issues.
4608 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
4609 versions), which is now available for royalty-free use
4610 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
4611 Also, add Camellia TLS ciphersuites from RFC 4132.
4613 To minimize changes between patchlevels in the OpenSSL 0.9.8
4614 series, Camellia remains excluded from compilation unless OpenSSL
4615 is configured with 'enable-camellia'.
4618 *) Disable the padding bug check when compression is in use. The padding
4619 bug check assumes the first packet is of even length, this is not
4620 necessarily true if compresssion is enabled and can result in false
4621 positives causing handshake failure. The actual bug test is ancient
4622 code so it is hoped that implementations will either have fixed it by
4623 now or any which still have the bug do not support compression.
4626 Changes between 0.9.8a and 0.9.8b [04 May 2006]
4628 *) When applying a cipher rule check to see if string match is an explicit
4629 cipher suite and only match that one cipher suite if it is.
4632 *) Link in manifests for VC++ if needed.
4633 [Austin Ziegler <halostatue@gmail.com>]
4635 *) Update support for ECC-based TLS ciphersuites according to
4636 draft-ietf-tls-ecc-12.txt with proposed changes (but without
4637 TLS extensions, which are supported starting with the 0.9.9
4638 branch, not in the OpenSSL 0.9.8 branch).
4641 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
4642 opaque EVP_CIPHER_CTX handling.
4645 *) Fixes and enhancements to zlib compression code. We now only use
4646 "zlib1.dll" and use the default __cdecl calling convention on Win32
4647 to conform with the standards mentioned here:
4648 http://www.zlib.net/DLL_FAQ.txt
4649 Static zlib linking now works on Windows and the new --with-zlib-include
4650 --with-zlib-lib options to Configure can be used to supply the location
4651 of the headers and library. Gracefully handle case where zlib library
4655 *) Several fixes and enhancements to the OID generation code. The old code
4656 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
4657 handle numbers larger than ULONG_MAX, truncated printing and had a
4658 non standard OBJ_obj2txt() behaviour.
4661 *) Add support for building of engines under engine/ as shared libraries
4662 under VC++ build system.
4665 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
4666 Hopefully, we will not see any false combination of paths any more.
4669 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
4671 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4672 (part of SSL_OP_ALL). This option used to disable the
4673 countermeasure against man-in-the-middle protocol-version
4674 rollback in the SSL 2.0 server implementation, which is a bad
4675 idea. (CVE-2005-2969)
4677 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4678 for Information Security, National Institute of Advanced Industrial
4679 Science and Technology [AIST], Japan)]
4681 *) Add two function to clear and return the verify parameter flags.
4684 *) Keep cipherlists sorted in the source instead of sorting them at
4685 runtime, thus removing the need for a lock.
4688 *) Avoid some small subgroup attacks in Diffie-Hellman.
4689 [Nick Mathewson and Ben Laurie]
4691 *) Add functions for well-known primes.
4694 *) Extended Windows CE support.
4695 [Satoshi Nakamura and Andy Polyakov]
4697 *) Initialize SSL_METHOD structures at compile time instead of during
4698 runtime, thus removing the need for a lock.
4701 *) Make PKCS7_decrypt() work even if no certificate is supplied by
4702 attempting to decrypt each encrypted key in turn. Add support to
4706 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
4708 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
4711 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
4714 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
4715 key into the same file any more.
4718 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
4721 *) Add -utf8 command line and config file option to 'ca'.
4722 [Stefan <stf@udoma.org]
4724 *) Removed the macro des_crypt(), as it seems to conflict with some
4725 libraries. Use DES_crypt().
4728 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
4729 involves renaming the source and generated shared-libs for
4730 both. The engines will accept the corrected or legacy ids
4731 ('ncipher' and '4758_cca' respectively) when binding. NB,
4732 this only applies when building 'shared'.
4733 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
4735 *) Add attribute functions to EVP_PKEY structure. Modify
4736 PKCS12_create() to recognize a CSP name attribute and
4737 use it. Make -CSP option work again in pkcs12 utility.
4740 *) Add new functionality to the bn blinding code:
4741 - automatic re-creation of the BN_BLINDING parameters after
4742 a fixed number of uses (currently 32)
4743 - add new function for parameter creation
4744 - introduce flags to control the update behaviour of the
4745 BN_BLINDING parameters
4746 - hide BN_BLINDING structure
4747 Add a second BN_BLINDING slot to the RSA structure to improve
4748 performance when a single RSA object is shared among several
4752 *) Add support for DTLS.
4753 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
4755 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
4756 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
4759 *) Remove buggy and incomplete DH cert support from
4760 ssl/ssl_rsa.c and ssl/s3_both.c
4763 *) Use SHA-1 instead of MD5 as the default digest algorithm for
4764 the apps/openssl applications.
4767 *) Compile clean with "-Wall -Wmissing-prototypes
4768 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
4769 DEBUG_SAFESTACK must also be set.
4772 *) Change ./Configure so that certain algorithms can be disabled by default.
4773 The new counterpiece to "no-xxx" is "enable-xxx".
4775 The patented RC5 and MDC2 algorithms will now be disabled unless
4776 "enable-rc5" and "enable-mdc2", respectively, are specified.
4778 (IDEA remains enabled despite being patented. This is because IDEA
4779 is frequently required for interoperability, and there is no license
4780 fee for non-commercial use. As before, "no-idea" can be used to
4781 avoid this algorithm.)
4785 *) Add processing of proxy certificates (see RFC 3820). This work was
4786 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
4787 EGEE (Enabling Grids for E-science in Europe).
4790 *) RC4 performance overhaul on modern architectures/implementations, such
4791 as Intel P4, IA-64 and AMD64.
4794 *) New utility extract-section.pl. This can be used specify an alternative
4795 section number in a pod file instead of having to treat each file as
4796 a separate case in Makefile. This can be done by adding two lines to the
4799 =for comment openssl_section:XXX
4801 The blank line is mandatory.
4805 *) New arguments -certform, -keyform and -pass for s_client and s_server
4806 to allow alternative format key and certificate files and passphrase
4810 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
4811 update associated structures and add various utility functions.
4813 Add new policy related verify parameters, include policy checking in
4814 standard verify code. Enhance 'smime' application with extra parameters
4815 to support policy checking and print out.
4818 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
4819 Nehemiah processors. These extensions support AES encryption in hardware
4820 as well as RNG (though RNG support is currently disabled).
4821 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
4823 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
4826 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
4827 [Andy Polyakov and a number of other people]
4829 *) Improved PowerPC platform support. Most notably BIGNUM assembler
4830 implementation contributed by IBM.
4831 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
4833 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
4834 exponent rather than 'unsigned long'. There is a corresponding change to
4835 the new 'rsa_keygen' element of the RSA_METHOD structure.
4836 [Jelte Jansen, Geoff Thorpe]
4838 *) Functionality for creating the initial serial number file is now
4839 moved from CA.pl to the 'ca' utility with a new option -create_serial.
4841 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
4842 number file to 1, which is bound to cause problems. To avoid
4843 the problems while respecting compatibility between different 0.9.7
4844 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
4845 CA.pl for serial number initialization. With the new release 0.9.8,
4846 we can fix the problem directly in the 'ca' utility.)
4849 *) Reduced header interdepencies by declaring more opaque objects in
4850 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
4851 give fewer recursive includes, which could break lazy source code - so
4852 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
4853 developers should define this symbol when building and using openssl to
4854 ensure they track the recommended behaviour, interfaces, [etc], but
4855 backwards-compatible behaviour prevails when this isn't defined.
4858 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
4861 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
4862 This will generate a random key of the appropriate length based on the
4863 cipher context. The EVP_CIPHER can provide its own random key generation
4864 routine to support keys of a specific form. This is used in the des and
4865 3des routines to generate a key of the correct parity. Update S/MIME
4866 code to use new functions and hence generate correct parity DES keys.
4867 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
4868 valid (weak or incorrect parity).
4871 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
4872 as looking them up. This is useful when the verified structure may contain
4873 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
4874 present unless the new PKCS7_NO_CRL flag is asserted.
4877 *) Extend ASN1 oid configuration module. It now additionally accepts the
4880 shortName = some long name, 1.2.3.4
4883 *) Reimplemented the BN_CTX implementation. There is now no more static
4884 limitation on the number of variables it can handle nor the depth of the
4885 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
4886 information can now expand as required, and rather than having a single
4887 static array of bignums, BN_CTX now uses a linked-list of such arrays
4888 allowing it to expand on demand whilst maintaining the usefulness of
4889 BN_CTX's "bundling".
4892 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
4893 to allow all RSA operations to function using a single BN_CTX.
4896 *) Preliminary support for certificate policy evaluation and checking. This
4897 is initially intended to pass the tests outlined in "Conformance Testing
4898 of Relying Party Client Certificate Path Processing Logic" v1.07.
4901 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
4902 remained unused and not that useful. A variety of other little bignum
4903 tweaks and fixes have also been made continuing on from the audit (see
4907 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
4908 associated ASN1, EVP and SSL functions and old ASN1 macros.
4911 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
4912 and this should never fail. So the return value from the use of
4913 BN_set_word() (which can fail due to needless expansion) is now deprecated;
4914 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
4917 *) BN_CTX_get() should return zero-valued bignums, providing the same
4918 initialised value as BN_new().
4919 [Geoff Thorpe, suggested by Ulf Möller]
4921 *) Support for inhibitAnyPolicy certificate extension.
4924 *) An audit of the BIGNUM code is underway, for which debugging code is
4925 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
4926 is considered valid when processing BIGNUMs, and causes execution to
4927 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
4928 further steps are taken to deliberately pollute unused data in BIGNUM
4929 structures to try and expose faulty code further on. For now, openssl will
4930 (in its default mode of operation) continue to tolerate the inconsistent
4931 forms that it has tolerated in the past, but authors and packagers should
4932 consider trying openssl and their own applications when compiled with
4933 these debugging symbols defined. It will help highlight potential bugs in
4934 their own code, and will improve the test coverage for OpenSSL itself. At
4935 some point, these tighter rules will become openssl's default to improve
4936 maintainability, though the assert()s and other overheads will remain only
4937 in debugging configurations. See bn.h for more details.
4938 [Geoff Thorpe, Nils Larsch, Ulf Möller]
4940 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
4941 that can only be obtained through BN_CTX_new() (which implicitly
4942 initialises it). The presence of this function only made it possible
4943 to overwrite an existing structure (and cause memory leaks).
4946 *) Because of the callback-based approach for implementing LHASH as a
4947 template type, lh_insert() adds opaque objects to hash-tables and
4948 lh_doall() or lh_doall_arg() are typically used with a destructor callback
4949 to clean up those corresponding objects before destroying the hash table
4950 (and losing the object pointers). So some over-zealous constifications in
4951 LHASH have been relaxed so that lh_insert() does not take (nor store) the
4952 objects as "const" and the lh_doall[_arg] callback wrappers are not
4953 prototyped to have "const" restrictions on the object pointers they are
4954 given (and so aren't required to cast them away any more).
4957 *) The tmdiff.h API was so ugly and minimal that our own timing utility
4958 (speed) prefers to use its own implementation. The two implementations
4959 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
4960 its object type properly exposed (MS_TM) instead of casting to/from "char
4961 *". This may still change yet if someone realises MS_TM and "ms_time_***"
4962 aren't necessarily the greatest nomenclatures - but this is what was used
4963 internally to the implementation so I've used that for now.
4966 *) Ensure that deprecated functions do not get compiled when
4967 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
4968 the self-tests were still using deprecated key-generation functions so
4969 these have been updated also.
4972 *) Reorganise PKCS#7 code to separate the digest location functionality
4973 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
4974 New function PKCS7_set_digest() to set the digest type for PKCS#7
4975 digestedData type. Add additional code to correctly generate the
4976 digestedData type and add support for this type in PKCS7 initialization
4980 *) New function PKCS7_set0_type_other() this initializes a PKCS7
4981 structure of type "other".
4984 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
4985 sure the loop does correctly stop and breaking ("division by zero")
4986 modulus operations are not performed. The (pre-generated) prime
4987 table crypto/bn/bn_prime.h was already correct, but it could not be
4988 re-generated on some platforms because of the "division by zero"
4989 situation in the script.
4990 [Ralf S. Engelschall]
4992 *) Update support for ECC-based TLS ciphersuites according to
4993 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
4994 SHA-1 now is only used for "small" curves (where the
4995 representation of a field element takes up to 24 bytes); for
4996 larger curves, the field element resulting from ECDH is directly
4997 used as premaster secret.
4998 [Douglas Stebila (Sun Microsystems Laboratories)]
5000 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5001 curve secp160r1 to the tests.
5002 [Douglas Stebila (Sun Microsystems Laboratories)]
5004 *) Add the possibility to load symbols globally with DSO.
5005 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5007 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5008 control of the error stack.
5011 *) Add support for STORE in ENGINE.
5014 *) Add the STORE type. The intention is to provide a common interface
5015 to certificate and key stores, be they simple file-based stores, or
5016 HSM-type store, or LDAP stores, or...
5017 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5020 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5021 pass a list of arguments to any function as well as provide a way
5022 for a function to pass data back to the caller.
5025 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5026 works like BUF_strdup() but can be used to duplicate a portion of
5027 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5031 *) Add the function sk_find_ex() which works like sk_find(), but will
5032 return an index to an element even if an exact match couldn't be
5033 found. The index is guaranteed to point at the element where the
5034 searched-for key would be inserted to preserve sorting order.
5037 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5038 takes an extra flags argument for optional functionality. Currently,
5039 the following flags are defined:
5041 OBJ_BSEARCH_VALUE_ON_NOMATCH
5042 This one gets OBJ_bsearch_ex() to return a pointer to the first
5043 element where the comparing function returns a negative or zero
5046 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5047 This one gets OBJ_bsearch_ex() to return a pointer to the first
5048 element where the comparing function returns zero. This is useful
5049 if there are more than one element where the comparing function
5053 *) Make it possible to create self-signed certificates with 'openssl ca'
5054 in such a way that the self-signed certificate becomes part of the
5055 CA database and uses the same mechanisms for serial number generation
5056 as all other certificate signing. The new flag '-selfsign' enables
5057 this functionality. Adapt CA.sh and CA.pl.in.
5060 *) Add functionality to check the public key of a certificate request
5061 against a given private. This is useful to check that a certificate
5062 request can be signed by that key (self-signing).
5065 *) Make it possible to have multiple active certificates with the same
5066 subject in the CA index file. This is done only if the keyword
5067 'unique_subject' is set to 'no' in the main CA section (default
5068 if 'CA_default') of the configuration file. The value is saved
5069 with the database itself in a separate index attribute file,
5070 named like the index file with '.attr' appended to the name.
5073 *) Generate muti valued AVAs using '+' notation in config files for
5077 *) Support for nameConstraints certificate extension.
5080 *) Support for policyConstraints certificate extension.
5083 *) Support for policyMappings certificate extension.
5086 *) Make sure the default DSA_METHOD implementation only uses its
5087 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5088 and change its own handlers to be NULL so as to remove unnecessary
5089 indirection. This lets alternative implementations fallback to the
5090 default implementation more easily.
5093 *) Support for directoryName in GeneralName related extensions
5097 *) Make it possible to link applications using Makefile.shared.
5098 Make that possible even when linking against static libraries!
5101 *) Support for single pass processing for S/MIME signing. This now
5102 means that S/MIME signing can be done from a pipe, in addition
5103 cleartext signing (multipart/signed type) is effectively streaming
5104 and the signed data does not need to be all held in memory.
5106 This is done with a new flag PKCS7_STREAM. When this flag is set
5107 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5108 is done after the data is output (and digests calculated) in
5109 SMIME_write_PKCS7().
5112 *) Add full support for -rpath/-R, both in shared libraries and
5113 applications, at least on the platforms where it's known how
5117 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
5118 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
5119 will now compute a table of multiples of the generator that
5120 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
5121 faster (notably in the case of a single point multiplication,
5122 scalar * generator).
5123 [Nils Larsch, Bodo Moeller]
5125 *) IPv6 support for certificate extensions. The various extensions
5126 which use the IP:a.b.c.d can now take IPv6 addresses using the
5127 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
5131 *) Added an ENGINE that implements RSA by performing private key
5132 exponentiations with the GMP library. The conversions to and from
5133 GMP's mpz_t format aren't optimised nor are any montgomery forms
5134 cached, and on x86 it appears OpenSSL's own performance has caught up.
5135 However there are likely to be other architectures where GMP could
5136 provide a boost. This ENGINE is not built in by default, but it can be
5137 specified at Configure time and should be accompanied by the necessary
5138 linker additions, eg;
5139 ./config -DOPENSSL_USE_GMP -lgmp
5142 *) "openssl engine" will not display ENGINE/DSO load failure errors when
5143 testing availability of engines with "-t" - the old behaviour is
5144 produced by increasing the feature's verbosity with "-tt".
5147 *) ECDSA routines: under certain error conditions uninitialized BN objects
5148 could be freed. Solution: make sure initialization is performed early
5149 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
5153 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
5154 and DH_METHOD (eg. by ENGINE implementations) to override the normal
5155 software implementations. For DSA and DH, parameter generation can
5156 also be overridden by providing the appropriate method callbacks.
5159 *) Change the "progress" mechanism used in key-generation and
5160 primality testing to functions that take a new BN_GENCB pointer in
5161 place of callback/argument pairs. The new API functions have "_ex"
5162 postfixes and the older functions are reimplemented as wrappers for
5163 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
5164 declarations of the old functions to help (graceful) attempts to
5165 migrate to the new functions. Also, the new key-generation API
5166 functions operate on a caller-supplied key-structure and return
5167 success/failure rather than returning a key or NULL - this is to
5168 help make "keygen" another member function of RSA_METHOD etc.
5170 Example for using the new callback interface:
5172 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
5176 BN_GENCB_set(&my_cb, my_callback, my_arg);
5178 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
5179 /* For the meaning of a, b in calls to my_callback(), see the
5180 * documentation of the function that calls the callback.
5181 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
5182 * my_callback should return 1 if it wants BN_is_prime_ex()
5183 * to continue, or 0 to stop.
5188 *) Change the ZLIB compression method to be stateful, and make it
5189 available to TLS with the number defined in
5190 draft-ietf-tls-compression-04.txt.
5193 *) Add the ASN.1 structures and functions for CertificatePair, which
5194 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
5196 CertificatePair ::= SEQUENCE {
5197 forward [0] Certificate OPTIONAL,
5198 reverse [1] Certificate OPTIONAL,
5199 -- at least one of the pair shall be present -- }
5201 Also implement the PEM functions to read and write certificate
5202 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
5204 This needed to be defined, mostly for the sake of the LDAP
5205 attribute crossCertificatePair, but may prove useful elsewhere as
5209 *) Make it possible to inhibit symlinking of shared libraries in
5210 Makefile.shared, for Cygwin's sake.
5213 *) Extend the BIGNUM API by creating a function
5214 void BN_set_negative(BIGNUM *a, int neg);
5215 and a macro that behave like
5216 int BN_is_negative(const BIGNUM *a);
5218 to avoid the need to access 'a->neg' directly in applications.
5221 *) Implement fast modular reduction for pseudo-Mersenne primes
5222 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
5223 EC_GROUP_new_curve_GFp() will now automatically use this
5225 [Nils Larsch <nla@trustcenter.de>]
5227 *) Add new lock type (CRYPTO_LOCK_BN).
5230 *) Change the ENGINE framework to automatically load engines
5231 dynamically from specific directories unless they could be
5232 found to already be built in or loaded. Move all the
5233 current engines except for the cryptodev one to a new
5235 The engines in engines/ are built as shared libraries if
5236 the "shared" options was given to ./Configure or ./config.
5237 Otherwise, they are inserted in libcrypto.a.
5238 /usr/local/ssl/engines is the default directory for dynamic
5239 engines, but that can be overridden at configure time through
5240 the usual use of --prefix and/or --openssldir, and at run
5241 time with the environment variable OPENSSL_ENGINES.
5242 [Geoff Thorpe and Richard Levitte]
5244 *) Add Makefile.shared, a helper makefile to build shared
5245 libraries. Adapt Makefile.org.
5248 *) Add version info to Win32 DLLs.
5249 [Peter 'Luna' Runestig" <peter@runestig.com>]
5251 *) Add new 'medium level' PKCS#12 API. Certificates and keys
5252 can be added using this API to created arbitrary PKCS#12
5253 files while avoiding the low level API.
5255 New options to PKCS12_create(), key or cert can be NULL and
5256 will then be omitted from the output file. The encryption
5257 algorithm NIDs can be set to -1 for no encryption, the mac
5258 iteration count can be set to 0 to omit the mac.
5260 Enhance pkcs12 utility by making the -nokeys and -nocerts
5261 options work when creating a PKCS#12 file. New option -nomac
5262 to omit the mac, NONE can be set for an encryption algorithm.
5263 New code is modified to use the enhanced PKCS12_create()
5264 instead of the low level API.
5267 *) Extend ASN1 encoder to support indefinite length constructed
5268 encoding. This can output sequences tags and octet strings in
5269 this form. Modify pk7_asn1.c to support indefinite length
5270 encoding. This is experimental and needs additional code to
5271 be useful, such as an ASN1 bio and some enhanced streaming
5274 Extend template encode functionality so that tagging is passed
5275 down to the template encoder.
5278 *) Let 'openssl req' fail if an argument to '-newkey' is not
5279 recognized instead of using RSA as a default.
5282 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
5283 As these are not official, they are not included in "ALL";
5284 the "ECCdraft" ciphersuite group alias can be used to select them.
5285 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
5287 *) Add ECDH engine support.
5288 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
5290 *) Add ECDH in new directory crypto/ecdh/.
5291 [Douglas Stebila (Sun Microsystems Laboratories)]
5293 *) Let BN_rand_range() abort with an error after 100 iterations
5294 without success (which indicates a broken PRNG).
5297 *) Change BN_mod_sqrt() so that it verifies that the input value
5298 is really the square of the return value. (Previously,
5299 BN_mod_sqrt would show GIGO behaviour.)
5302 *) Add named elliptic curves over binary fields from X9.62, SECG,
5303 and WAP/WTLS; add OIDs that were still missing.
5305 [Sheueling Chang Shantz and Douglas Stebila
5306 (Sun Microsystems Laboratories)]
5308 *) Extend the EC library for elliptic curves over binary fields
5309 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
5312 EC_GF2m_simple_method
5316 EC_GROUP_new_curve_GF2m
5317 EC_GROUP_set_curve_GF2m
5318 EC_GROUP_get_curve_GF2m
5319 EC_POINT_set_affine_coordinates_GF2m
5320 EC_POINT_get_affine_coordinates_GF2m
5321 EC_POINT_set_compressed_coordinates_GF2m
5323 Point compression for binary fields is disabled by default for
5324 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
5327 As binary polynomials are represented as BIGNUMs, various members
5328 of the EC_GROUP and EC_POINT data structures can be shared
5329 between the implementations for prime fields and binary fields;
5330 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
5331 are essentially identical to their ..._GFp counterparts.
5332 (For simplicity, the '..._GFp' prefix has been dropped from
5333 various internal method names.)
5335 An internal 'field_div' method (similar to 'field_mul' and
5336 'field_sqr') has been added; this is used only for binary fields.
5338 [Sheueling Chang Shantz and Douglas Stebila
5339 (Sun Microsystems Laboratories)]
5341 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
5342 through methods ('mul', 'precompute_mult').
5344 The generic implementations (now internally called 'ec_wNAF_mul'
5345 and 'ec_wNAF_precomputed_mult') remain the default if these
5346 methods are undefined.
5348 [Sheueling Chang Shantz and Douglas Stebila
5349 (Sun Microsystems Laboratories)]
5351 *) New function EC_GROUP_get_degree, which is defined through
5352 EC_METHOD. For curves over prime fields, this returns the bit
5353 length of the modulus.
5355 [Sheueling Chang Shantz and Douglas Stebila
5356 (Sun Microsystems Laboratories)]
5358 *) New functions EC_GROUP_dup, EC_POINT_dup.
5359 (These simply call ..._new and ..._copy).
5361 [Sheueling Chang Shantz and Douglas Stebila
5362 (Sun Microsystems Laboratories)]
5364 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
5365 Polynomials are represented as BIGNUMs (where the sign bit is not
5366 used) in the following functions [macros]:
5369 BN_GF2m_sub [= BN_GF2m_add]
5370 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
5371 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
5372 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
5374 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
5375 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
5376 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
5377 BN_GF2m_cmp [= BN_ucmp]
5379 (Note that only the 'mod' functions are actually for fields GF(2^m).
5380 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
5382 For some functions, an the irreducible polynomial defining a
5383 field can be given as an 'unsigned int[]' with strictly
5384 decreasing elements giving the indices of those bits that are set;
5385 i.e., p[] represents the polynomial
5386 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
5388 p[0] > p[1] > ... > p[k] = 0.
5389 This applies to the following functions:
5394 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
5395 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
5397 BN_GF2m_mod_sqrt_arr
5398 BN_GF2m_mod_solve_quad_arr
5402 Conversion can be performed by the following functions:
5407 bntest.c has additional tests for binary polynomial arithmetic.
5409 Two implementations for BN_GF2m_mod_div() are available.
5410 The default algorithm simply uses BN_GF2m_mod_inv() and
5411 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
5412 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
5413 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
5415 [Sheueling Chang Shantz and Douglas Stebila
5416 (Sun Microsystems Laboratories)]
5418 *) Add new error code 'ERR_R_DISABLED' that can be used when some
5419 functionality is disabled at compile-time.
5420 [Douglas Stebila <douglas.stebila@sun.com>]
5422 *) Change default behaviour of 'openssl asn1parse' so that more
5423 information is visible when viewing, e.g., a certificate:
5425 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
5426 mode the content of non-printable OCTET STRINGs is output in a
5427 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
5428 avoid the appearance of a printable string.
5429 [Nils Larsch <nla@trustcenter.de>]
5431 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
5433 EC_GROUP_set_asn1_flag()
5434 EC_GROUP_get_asn1_flag()
5435 EC_GROUP_set_point_conversion_form()
5436 EC_GROUP_get_point_conversion_form()
5437 These control ASN1 encoding details:
5438 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
5439 has been set to OPENSSL_EC_NAMED_CURVE.
5440 - Points are encoded in uncompressed form by default; options for
5441 asn1_for are as for point2oct, namely
5442 POINT_CONVERSION_COMPRESSED
5443 POINT_CONVERSION_UNCOMPRESSED
5444 POINT_CONVERSION_HYBRID
5446 Also add 'seed' and 'seed_len' members to EC_GROUP with access
5449 EC_GROUP_get0_seed()
5450 EC_GROUP_get_seed_len()
5451 This is used only for ASN1 purposes (so far).
5452 [Nils Larsch <nla@trustcenter.de>]
5454 *) Add 'field_type' member to EC_METHOD, which holds the NID
5455 of the appropriate field type OID. The new function
5456 EC_METHOD_get_field_type() returns this value.
5457 [Nils Larsch <nla@trustcenter.de>]
5462 EC_POINT_point2hex()
5463 EC_POINT_hex2point()
5464 providing useful interfaces to EC_POINT_point2oct() and
5465 EC_POINT_oct2point().
5466 [Nils Larsch <nla@trustcenter.de>]
5468 *) Change internals of the EC library so that the functions
5469 EC_GROUP_set_generator()
5470 EC_GROUP_get_generator()
5471 EC_GROUP_get_order()
5472 EC_GROUP_get_cofactor()
5473 are implemented directly in crypto/ec/ec_lib.c and not dispatched
5474 to methods, which would lead to unnecessary code duplication when
5475 adding different types of curves.
5476 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
5478 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
5479 arithmetic, and such that modified wNAFs are generated
5480 (which avoid length expansion in many cases).
5483 *) Add a function EC_GROUP_check_discriminant() (defined via
5484 EC_METHOD) that verifies that the curve discriminant is non-zero.
5486 Add a function EC_GROUP_check() that makes some sanity tests
5487 on a EC_GROUP, its generator and order. This includes
5488 EC_GROUP_check_discriminant().
5489 [Nils Larsch <nla@trustcenter.de>]
5491 *) Add ECDSA in new directory crypto/ecdsa/.
5493 Add applications 'openssl ecparam' and 'openssl ecdsa'
5494 (these are based on 'openssl dsaparam' and 'openssl dsa').
5496 ECDSA support is also included in various other files across the
5497 library. Most notably,
5498 - 'openssl req' now has a '-newkey ecdsa:file' option;
5499 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
5500 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
5501 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
5502 them suitable for ECDSA where domain parameters must be
5503 extracted before the specific public key;
5504 - ECDSA engine support has been added.
5505 [Nils Larsch <nla@trustcenter.de>]
5507 *) Include some named elliptic curves, and add OIDs from X9.62,
5508 SECG, and WAP/WTLS. Each curve can be obtained from the new
5510 EC_GROUP_new_by_curve_name(),
5511 and the list of available named curves can be obtained with
5512 EC_get_builtin_curves().
5513 Also add a 'curve_name' member to EC_GROUP objects, which can be
5515 EC_GROUP_set_curve_name()
5516 EC_GROUP_get_curve_name()
5517 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
5519 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5520 was actually never needed) and in BN_mul(). The removal in BN_mul()
5521 required a small change in bn_mul_part_recursive() and the addition
5522 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5523 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5524 bn_sub_words() and bn_add_words() except they take arrays with
5528 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
5530 *) Cleanse PEM buffers before freeing them since they may contain
5532 [Benjamin Bennett <ben@psc.edu>]
5534 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5535 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5536 authentication-only ciphersuites.
5539 *) Since AES128 and AES256 share a single mask bit in the logic of
5540 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5541 kludge to work properly if AES128 is available and AES256 isn't.
5544 *) Expand security boundary to match 1.1.1 module.
5547 *) Remove redundant features: hash file source, editing of test vectors
5548 modify fipsld to use external fips_premain.c signature.
5551 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
5552 run algorithm test programs.
5555 *) Make algorithm test programs more tolerant of whitespace.
5558 *) Have SSL/TLS server implementation tolerate "mismatched" record
5559 protocol version while receiving ClientHello even if the
5560 ClientHello is fragmented. (The server can't insist on the
5561 particular protocol version it has chosen before the ServerHello
5562 message has informed the client about his choice.)
5565 *) Load error codes if they are not already present instead of using a
5566 static variable. This allows them to be cleanly unloaded and reloaded.
5569 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
5571 *) Introduce limits to prevent malicious keys being able to
5572 cause a denial of service. (CVE-2006-2940)
5573 [Steve Henson, Bodo Moeller]
5575 *) Fix ASN.1 parsing of certain invalid structures that can result
5576 in a denial of service. (CVE-2006-2937) [Steve Henson]
5578 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5579 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5581 *) Fix SSL client code which could crash if connecting to a
5582 malicious SSLv2 server. (CVE-2006-4343)
5583 [Tavis Ormandy and Will Drewry, Google Security Team]
5585 *) Change ciphersuite string processing so that an explicit
5586 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
5587 will no longer include "AES128-SHA"), and any other similar
5588 ciphersuite (same bitmap) from *other* protocol versions (so that
5589 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
5590 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
5591 changes from 0.9.8b and 0.9.8d.
5594 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
5596 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5597 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5599 *) Change the Unix randomness entropy gathering to use poll() when
5600 possible instead of select(), since the latter has some
5601 undesirable limitations.
5602 [Darryl Miles via Richard Levitte and Bodo Moeller]
5604 *) Disable rogue ciphersuites:
5606 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5607 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5608 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5610 The latter two were purportedly from
5611 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5614 Also deactive the remaining ciphersuites from
5615 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5616 unofficial, and the ID has long expired.
5619 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5620 dual-core machines) and other potential thread-safety issues.
5623 Changes between 0.9.7i and 0.9.7j [04 May 2006]
5625 *) Adapt fipsld and the build system to link against the validated FIPS
5626 module in FIPS mode.
5629 *) Fixes for VC++ 2005 build under Windows.
5632 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
5633 from a Windows bash shell such as MSYS. It is autodetected from the
5634 "config" script when run from a VC++ environment. Modify standard VC++
5635 build to use fipscanister.o from the GNU make build.
5638 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
5640 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
5641 The value now differs depending on if you build for FIPS or not.
5642 BEWARE! A program linked with a shared FIPSed libcrypto can't be
5643 safely run with a non-FIPSed libcrypto, as it may crash because of
5644 the difference induced by this change.
5647 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
5649 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5650 (part of SSL_OP_ALL). This option used to disable the
5651 countermeasure against man-in-the-middle protocol-version
5652 rollback in the SSL 2.0 server implementation, which is a bad
5653 idea. (CVE-2005-2969)
5655 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5656 for Information Security, National Institute of Advanced Industrial
5657 Science and Technology [AIST], Japan)]
5659 *) Minimal support for X9.31 signatures and PSS padding modes. This is
5660 mainly for FIPS compliance and not fully integrated at this stage.
5663 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
5664 the exponentiation using a fixed-length exponent. (Otherwise,
5665 the information leaked through timing could expose the secret key
5666 after many signatures; cf. Bleichenbacher's attack on DSA with
5670 *) Make a new fixed-window mod_exp implementation the default for
5671 RSA, DSA, and DH private-key operations so that the sequence of
5672 squares and multiplies and the memory access pattern are
5673 independent of the particular secret key. This will mitigate
5674 cache-timing and potential related attacks.
5676 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
5677 and this is automatically used by BN_mod_exp_mont() if the new flag
5678 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
5679 will use this BN flag for private exponents unless the flag
5680 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
5681 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
5683 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
5685 *) Change the client implementation for SSLv23_method() and
5686 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
5687 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
5688 (Previously, the SSL 2.0 backwards compatible Client Hello
5689 message format would be used even with SSL_OP_NO_SSLv2.)
5692 *) Add support for smime-type MIME parameter in S/MIME messages which some
5696 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
5697 a threadsafe manner. Modify rsa code to use new function and add calls
5698 to dsa and dh code (which had race conditions before).
5701 *) Include the fixed error library code in the C error file definitions
5702 instead of fixing them up at runtime. This keeps the error code
5703 structures constant.
5706 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
5708 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
5711 *) Fixes for newer kerberos headers. NB: the casts are needed because
5712 the 'length' field is signed on one version and unsigned on another
5713 with no (?) obvious way to tell the difference, without these VC++
5714 complains. Also the "definition" of FAR (blank) is no longer included
5715 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
5716 some needed definitions.
5719 *) Undo Cygwin change.
5722 *) Added support for proxy certificates according to RFC 3820.
5723 Because they may be a security thread to unaware applications,
5724 they must be explicitly allowed in run-time. See
5725 docs/HOWTO/proxy_certificates.txt for further information.
5728 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
5730 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
5731 server and client random values. Previously
5732 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
5733 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
5735 This change has negligible security impact because:
5737 1. Server and client random values still have 24 bytes of pseudo random
5740 2. Server and client random values are sent in the clear in the initial
5743 3. The master secret is derived using the premaster secret (48 bytes in
5744 size for static RSA ciphersuites) as well as client server and random
5747 The OpenSSL team would like to thank the UK NISCC for bringing this issue
5750 [Stephen Henson, reported by UK NISCC]
5752 *) Use Windows randomness collection on Cygwin.
5755 *) Fix hang in EGD/PRNGD query when communication socket is closed
5756 prematurely by EGD/PRNGD.
5757 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
5759 *) Prompt for pass phrases when appropriate for PKCS12 input format.
5762 *) Back-port of selected performance improvements from development
5763 branch, as well as improved support for PowerPC platforms.
5766 *) Add lots of checks for memory allocation failure, error codes to indicate
5767 failure and freeing up memory if a failure occurs.
5768 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
5770 *) Add new -passin argument to dgst.
5773 *) Perform some character comparisons of different types in X509_NAME_cmp:
5774 this is needed for some certificates that re-encode DNs into UTF8Strings
5775 (in violation of RFC3280) and can't or won't issue name rollover
5779 *) Make an explicit check during certificate validation to see that
5780 the CA setting in each certificate on the chain is correct. As a
5781 side effect always do the following basic checks on extensions,
5782 not just when there's an associated purpose to the check:
5784 - if there is an unhandled critical extension (unless the user
5785 has chosen to ignore this fault)
5786 - if the path length has been exceeded (if one is set at all)
5787 - that certain extensions fit the associated purpose (if one has
5791 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
5793 *) Avoid a race condition when CRLs are checked in a multi threaded
5794 environment. This would happen due to the reordering of the revoked
5795 entries during signature checking and serial number lookup. Now the
5796 encoding is cached and the serial number sort performed under a lock.
5797 Add new STACK function sk_is_sorted().
5800 *) Add Delta CRL to the extension code.
5803 *) Various fixes to s3_pkt.c so alerts are sent properly.
5804 [David Holmes <d.holmes@f5.com>]
5806 *) Reduce the chances of duplicate issuer name and serial numbers (in
5807 violation of RFC3280) using the OpenSSL certificate creation utilities.
5808 This is done by creating a random 64 bit value for the initial serial
5809 number when a serial number file is created or when a self signed
5810 certificate is created using 'openssl req -x509'. The initial serial
5811 number file is created using 'openssl x509 -next_serial' in CA.pl
5812 rather than being initialized to 1.
5815 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
5817 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5818 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5819 [Joe Orton, Steve Henson]
5821 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
5823 [Joe Orton, Steve Henson]
5825 *) Make it possible to have multiple active certificates with the same
5826 subject in the CA index file. This is done only if the keyword
5827 'unique_subject' is set to 'no' in the main CA section (default
5828 if 'CA_default') of the configuration file. The value is saved
5829 with the database itself in a separate index attribute file,
5830 named like the index file with '.attr' appended to the name.
5833 *) X509 verify fixes. Disable broken certificate workarounds when
5834 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
5835 keyUsage extension present. Don't accept CRLs with unhandled critical
5836 extensions: since verify currently doesn't process CRL extensions this
5837 rejects a CRL with *any* critical extensions. Add new verify error codes
5841 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
5842 A clarification of RFC2560 will require the use of OCTET STRINGs and
5843 some implementations cannot handle the current raw format. Since OpenSSL
5844 copies and compares OCSP nonces as opaque blobs without any attempt at
5845 parsing them this should not create any compatibility issues.
5848 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
5849 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
5850 this HMAC (and other) operations are several times slower than OpenSSL
5854 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
5855 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5857 *) Use the correct content when signing type "other".
5860 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
5862 *) Fix various bugs revealed by running the NISCC test suite:
5864 Stop out of bounds reads in the ASN1 code when presented with
5865 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5867 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
5869 If verify callback ignores invalid public key errors don't try to check
5870 certificate signature with the NULL public key.
5874 *) New -ignore_err option in ocsp application to stop the server
5875 exiting on the first error in a request.
5878 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5879 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5883 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5884 extra data after the compression methods not only for TLS 1.0
5885 but also for SSL 3.0 (as required by the specification).
5886 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5888 *) Change X509_certificate_type() to mark the key as exported/exportable
5889 when it's 512 *bits* long, not 512 bytes.
5892 *) Change AES_cbc_encrypt() so it outputs exact multiple of
5893 blocks during encryption.
5896 *) Various fixes to base64 BIO and non blocking I/O. On write
5897 flushes were not handled properly if the BIO retried. On read
5898 data was not being buffered properly and had various logic bugs.
5899 This also affects blocking I/O when the data being decoded is a
5903 *) Various S/MIME bugfixes and compatibility changes:
5904 output correct application/pkcs7 MIME type if
5905 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
5906 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
5907 of files as .eml work). Correctly handle very long lines in MIME
5911 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
5913 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5914 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5915 a protocol version number mismatch like a decryption error
5916 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5919 *) Turn on RSA blinding by default in the default implementation
5920 to avoid a timing attack. Applications that don't want it can call
5921 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5922 They would be ill-advised to do so in most cases.
5923 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5925 *) Change RSA blinding code so that it works when the PRNG is not
5926 seeded (in this case, the secret RSA exponent is abused as
5927 an unpredictable seed -- if it is not unpredictable, there
5928 is no point in blinding anyway). Make RSA blinding thread-safe
5929 by remembering the creator's thread ID in rsa->blinding and
5930 having all other threads use local one-time blinding factors
5931 (this requires more computation than sharing rsa->blinding, but
5932 avoids excessive locking; and if an RSA object is not shared
5933 between threads, blinding will still be very fast).
5936 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
5937 ENGINE as defaults for all supported algorithms irrespective of
5938 the 'flags' parameter. 'flags' is now honoured, so applications
5939 should make sure they are passing it correctly.
5942 *) Target "mingw" now allows native Windows code to be generated in
5943 the Cygwin environment as well as with the MinGW compiler.
5946 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
5948 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5949 via timing by performing a MAC computation even if incorrect
5950 block cipher padding has been found. This is a countermeasure
5951 against active attacks where the attacker has to distinguish
5952 between bad padding and a MAC verification error. (CVE-2003-0078)
5954 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5955 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5956 Martin Vuagnoux (EPFL, Ilion)]
5958 *) Make the no-err option work as intended. The intention with no-err
5959 is not to have the whole error stack handling routines removed from
5960 libcrypto, it's only intended to remove all the function name and
5961 reason texts, thereby removing some of the footprint that may not
5962 be interesting if those errors aren't displayed anyway.
5964 NOTE: it's still possible for any application or module to have it's
5965 own set of error texts inserted. The routines are there, just not
5966 used by default when no-err is given.
5969 *) Add support for FreeBSD on IA64.
5970 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
5972 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
5973 Kerberos function mit_des_cbc_cksum(). Before this change,
5974 the value returned by DES_cbc_cksum() was like the one from
5975 mit_des_cbc_cksum(), except the bytes were swapped.
5976 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
5978 *) Allow an application to disable the automatic SSL chain building.
5979 Before this a rather primitive chain build was always performed in
5980 ssl3_output_cert_chain(): an application had no way to send the
5981 correct chain if the automatic operation produced an incorrect result.
5983 Now the chain builder is disabled if either:
5985 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
5987 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
5989 The reasoning behind this is that an application would not want the
5990 auto chain building to take place if extra chain certificates are
5991 present and it might also want a means of sending no additional
5992 certificates (for example the chain has two certificates and the
5996 *) Add the possibility to build without the ENGINE framework.
5997 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
5999 *) Under Win32 gmtime() can return NULL: check return value in
6000 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6003 *) DSA routines: under certain error conditions uninitialized BN objects
6004 could be freed. Solution: make sure initialization is performed early
6005 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6006 Nils Larsch <nla@trustcenter.de> via PR#459)
6009 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6010 checked on reconnect on the client side, therefore session resumption
6011 could still fail with a "ssl session id is different" error. This
6012 behaviour is masked when SSL_OP_ALL is used due to
6013 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6014 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6015 followup to PR #377.
6018 *) IA-32 assembler support enhancements: unified ELF targets, support
6019 for SCO/Caldera platforms, fix for Cygwin shared build.
6022 *) Add support for FreeBSD on sparc64. As a consequence, support for
6023 FreeBSD on non-x86 processors is separate from x86 processors on
6024 the config script, much like the NetBSD support.
6025 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6027 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6029 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6032 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6033 code (06) was taken as the first octet of the session ID and the last
6034 octet was ignored consequently. As a result SSLv2 client side session
6035 caching could not have worked due to the session ID mismatch between
6037 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6041 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6042 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6046 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6047 seems that in spite of existing for more than a year, many application
6048 author have done nothing to provide the necessary callbacks, which
6049 means that this particular engine will not work properly anywhere.
6050 This is a very unfortunate situation which forces us, in the name
6051 of usability, to give the hw_ncipher.c a static lock, which is part
6053 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6054 appear in 0.9.8 or later. We EXPECT application authors to have
6055 dealt properly with this when 0.9.8 is released (unless we actually
6056 make such changes in the libcrypto locking code that changes will
6057 have to be made anyway).
6060 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6061 octets have been read, EOF or an error occurs. Without this change
6062 some truncated ASN1 structures will not produce an error.
6065 *) Disable Heimdal support, since it hasn't been fully implemented.
6066 Still give the possibility to force the use of Heimdal, but with
6067 warnings and a request that patches get sent to openssl-dev.
6070 *) Add the VC-CE target, introduce the WINCE sysname, and add
6071 INSTALL.WCE and appropriate conditionals to make it build.
6072 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6074 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6075 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6076 edit numbers of the version.
6077 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6079 *) Introduce safe string copy and catenation functions
6080 (BUF_strlcpy() and BUF_strlcat()).
6081 [Ben Laurie (CHATS) and Richard Levitte]
6083 *) Avoid using fixed-size buffers for one-line DNs.
6084 [Ben Laurie (CHATS)]
6086 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6087 resizing buffers containing secrets, and use where appropriate.
6088 [Ben Laurie (CHATS)]
6090 *) Avoid using fixed size buffers for configuration file location.
6091 [Ben Laurie (CHATS)]
6093 *) Avoid filename truncation for various CA files.
6094 [Ben Laurie (CHATS)]
6096 *) Use sizeof in preference to magic numbers.
6097 [Ben Laurie (CHATS)]
6099 *) Avoid filename truncation in cert requests.
6100 [Ben Laurie (CHATS)]
6102 *) Add assertions to check for (supposedly impossible) buffer
6104 [Ben Laurie (CHATS)]
6106 *) Don't cache truncated DNS entries in the local cache (this could
6107 potentially lead to a spoofing attack).
6108 [Ben Laurie (CHATS)]
6110 *) Fix various buffers to be large enough for hex/decimal
6111 representations in a platform independent manner.
6112 [Ben Laurie (CHATS)]
6114 *) Add CRYPTO_realloc_clean() to avoid information leakage when
6115 resizing buffers containing secrets, and use where appropriate.
6116 [Ben Laurie (CHATS)]
6118 *) Add BIO_indent() to avoid much slightly worrying code to do
6120 [Ben Laurie (CHATS)]
6122 *) Convert sprintf()/BIO_puts() to BIO_printf().
6123 [Ben Laurie (CHATS)]
6125 *) buffer_gets() could terminate with the buffer only half
6127 [Ben Laurie (CHATS)]
6129 *) Add assertions to prevent user-supplied crypto functions from
6130 overflowing internal buffers by having large block sizes, etc.
6131 [Ben Laurie (CHATS)]
6133 *) New OPENSSL_assert() macro (similar to assert(), but enabled
6135 [Ben Laurie (CHATS)]
6137 *) Eliminate unused copy of key in RC4.
6138 [Ben Laurie (CHATS)]
6140 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
6141 [Ben Laurie (CHATS)]
6143 *) Fix off-by-one error in EGD path.
6144 [Ben Laurie (CHATS)]
6146 *) If RANDFILE path is too long, ignore instead of truncating.
6147 [Ben Laurie (CHATS)]
6149 *) Eliminate unused and incorrectly sized X.509 structure
6151 [Ben Laurie (CHATS)]
6153 *) Eliminate unused and dangerous function knumber().
6154 [Ben Laurie (CHATS)]
6156 *) Eliminate unused and dangerous structure, KSSL_ERR.
6157 [Ben Laurie (CHATS)]
6159 *) Protect against overlong session ID context length in an encoded
6160 session object. Since these are local, this does not appear to be
6162 [Ben Laurie (CHATS)]
6164 *) Change from security patch (see 0.9.6e below) that did not affect
6165 the 0.9.6 release series:
6167 Remote buffer overflow in SSL3 protocol - an attacker could
6168 supply an oversized master key in Kerberos-enabled versions.
6170 [Ben Laurie (CHATS)]
6172 *) Change the SSL kerb5 codes to match RFC 2712.
6175 *) Make -nameopt work fully for req and add -reqopt switch.
6176 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
6178 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
6179 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
6181 *) Make sure tests can be performed even if the corresponding algorithms
6182 have been removed entirely. This was also the last step to make
6183 OpenSSL compilable with DJGPP under all reasonable conditions.
6184 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
6186 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
6187 to allow version independent disabling of normally unselected ciphers,
6188 which may be activated as a side-effect of selecting a single cipher.
6190 (E.g., cipher list string "RSA" enables ciphersuites that are left
6191 out of "ALL" because they do not provide symmetric encryption.
6192 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
6193 [Lutz Jaenicke, Bodo Moeller]
6195 *) Add appropriate support for separate platform-dependent build
6196 directories. The recommended way to make a platform-dependent
6197 build directory is the following (tested on Linux), maybe with
6200 # Place yourself outside of the OpenSSL source tree. In
6201 # this example, the environment variable OPENSSL_SOURCE
6202 # is assumed to contain the absolute OpenSSL source directory.
6203 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
6204 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
6205 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
6206 mkdir -p `dirname $F`
6207 ln -s $OPENSSL_SOURCE/$F $F
6210 To be absolutely sure not to disturb the source tree, a "make clean"
6211 is a good thing. If it isn't successful, don't worry about it,
6212 it probably means the source directory is very clean.
6215 *) Make sure any ENGINE control commands make local copies of string
6216 pointers passed to them whenever necessary. Otherwise it is possible
6217 the caller may have overwritten (or deallocated) the original string
6218 data when a later ENGINE operation tries to use the stored values.
6219 [Götz Babin-Ebell <babinebell@trustcenter.de>]
6221 *) Improve diagnostics in file reading and command-line digests.
6222 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
6224 *) Add AES modes CFB and OFB to the object database. Correct an
6225 error in AES-CFB decryption.
6228 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
6229 allows existing EVP_CIPHER_CTX structures to be reused after
6230 calling EVP_*Final(). This behaviour is used by encryption
6231 BIOs and some applications. This has the side effect that
6232 applications must explicitly clean up cipher contexts with
6233 EVP_CIPHER_CTX_cleanup() or they will leak memory.
6236 *) Check the values of dna and dnb in bn_mul_recursive before calling
6237 bn_mul_comba (a non zero value means the a or b arrays do not contain
6238 n2 elements) and fallback to bn_mul_normal if either is not zero.
6241 *) Fix escaping of non-ASCII characters when using the -subj option
6242 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
6245 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
6246 form for "surname", serialNumber has no short form.
6247 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
6248 therefore remove "mail" short name for "internet 7".
6249 The OID for unique identifiers in X509 certificates is
6250 x500UniqueIdentifier, not uniqueIdentifier.
6251 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
6254 *) Add an "init" command to the ENGINE config module and auto initialize
6255 ENGINEs. Without any "init" command the ENGINE will be initialized
6256 after all ctrl commands have been executed on it. If init=1 the
6257 ENGINE is initailized at that point (ctrls before that point are run
6258 on the uninitialized ENGINE and after on the initialized one). If
6259 init=0 then the ENGINE will not be iniatialized at all.
6262 *) Fix the 'app_verify_callback' interface so that the user-defined
6263 argument is actually passed to the callback: In the
6264 SSL_CTX_set_cert_verify_callback() prototype, the callback
6265 declaration has been changed from
6268 int (*cb)(X509_STORE_CTX *,void *);
6269 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
6270 i=s->ctx->app_verify_callback(&ctx)
6271 has been changed into
6272 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
6274 To update applications using SSL_CTX_set_cert_verify_callback(),
6275 a dummy argument can be added to their callback functions.
6276 [D. K. Smetters <smetters@parc.xerox.com>]
6278 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
6279 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
6281 *) Add and OPENSSL_LOAD_CONF define which will cause
6282 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
6283 This allows older applications to transparently support certain
6284 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
6285 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
6286 load the config file and OPENSSL_add_all_algorithms_conf() which will
6287 always load it have also been added.
6290 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
6291 Adjust NIDs and EVP layer.
6292 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6294 *) Config modules support in openssl utility.
6296 Most commands now load modules from the config file,
6297 though in a few (such as version) this isn't done
6298 because it couldn't be used for anything.
6300 In the case of ca and req the config file used is
6301 the same as the utility itself: that is the -config
6302 command line option can be used to specify an
6306 *) Move default behaviour from OPENSSL_config(). If appname is NULL
6307 use "openssl_conf" if filename is NULL use default openssl config file.
6310 *) Add an argument to OPENSSL_config() to allow the use of an alternative
6311 config section name. Add a new flag to tolerate a missing config file
6312 and move code to CONF_modules_load_file().
6315 *) Support for crypto accelerator cards from Accelerated Encryption
6316 Processing, www.aep.ie. (Use engine 'aep')
6317 The support was copied from 0.9.6c [engine] and adapted/corrected
6318 to work with the new engine framework.
6319 [AEP Inc. and Richard Levitte]
6321 *) Support for SureWare crypto accelerator cards from Baltimore
6322 Technologies. (Use engine 'sureware')
6323 The support was copied from 0.9.6c [engine] and adapted
6324 to work with the new engine framework.
6327 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
6328 make the newer ENGINE framework commands for the CHIL engine work.
6329 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
6331 *) Make it possible to produce shared libraries on ReliantUNIX.
6332 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
6334 *) Add the configuration target debug-linux-ppro.
6335 Make 'openssl rsa' use the general key loading routines
6336 implemented in apps.c, and make those routines able to
6337 handle the key format FORMAT_NETSCAPE and the variant
6339 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6341 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6342 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6344 *) Add -keyform to rsautl, and document -engine.
6345 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
6347 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
6348 BIO_R_NO_SUCH_FILE error code rather than the generic
6349 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
6352 *) Add new functions
6354 ERR_peek_last_error_line
6355 ERR_peek_last_error_line_data.
6356 These are similar to
6359 ERR_peek_error_line_data,
6360 but report on the latest error recorded rather than the first one
6361 still in the error queue.
6362 [Ben Laurie, Bodo Moeller]
6364 *) default_algorithms option in ENGINE config module. This allows things
6366 default_algorithms = ALL
6367 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
6370 *) Preliminary ENGINE config module.
6373 *) New experimental application configuration code.
6376 *) Change the AES code to follow the same name structure as all other
6377 symmetric ciphers, and behave the same way. Move everything to
6378 the directory crypto/aes, thereby obsoleting crypto/rijndael.
6379 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
6381 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
6382 [Ben Laurie and Theo de Raadt]
6384 *) Add option to output public keys in req command.
6385 [Massimiliano Pala madwolf@openca.org]
6387 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
6388 (up to about 10% better than before for P-192 and P-224).
6391 *) New functions/macros
6393 SSL_CTX_set_msg_callback(ctx, cb)
6394 SSL_CTX_set_msg_callback_arg(ctx, arg)
6395 SSL_set_msg_callback(ssl, cb)
6396 SSL_set_msg_callback_arg(ssl, arg)
6398 to request calling a callback function
6400 void cb(int write_p, int version, int content_type,
6401 const void *buf, size_t len, SSL *ssl, void *arg)
6403 whenever a protocol message has been completely received
6404 (write_p == 0) or sent (write_p == 1). Here 'version' is the
6405 protocol version according to which the SSL library interprets
6406 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
6407 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
6408 the content type as defined in the SSL 3.0/TLS 1.0 protocol
6409 specification (change_cipher_spec(20), alert(21), handshake(22)).
6410 'buf' and 'len' point to the actual message, 'ssl' to the
6411 SSL object, and 'arg' is the application-defined value set by
6412 SSL[_CTX]_set_msg_callback_arg().
6414 'openssl s_client' and 'openssl s_server' have new '-msg' options
6415 to enable a callback that displays all protocol messages.
6418 *) Change the shared library support so shared libraries are built as
6419 soon as the corresponding static library is finished, and thereby get
6420 openssl and the test programs linked against the shared library.
6421 This still only happens when the keyword "shard" has been given to
6422 the configuration scripts.
6424 NOTE: shared library support is still an experimental thing, and
6425 backward binary compatibility is still not guaranteed.
6426 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
6428 *) Add support for Subject Information Access extension.
6429 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6431 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
6432 additional bytes when new memory had to be allocated, not just
6433 when reusing an existing buffer.
6436 *) New command line and configuration option 'utf8' for the req command.
6437 This allows field values to be specified as UTF8 strings.
6440 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
6441 runs for the former and machine-readable output for the latter.
6444 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
6445 of the e-mail address in the DN (i.e., it will go into a certificate
6446 extension only). The new configuration file option 'email_in_dn = no'
6447 has the same effect.
6448 [Massimiliano Pala madwolf@openca.org]
6450 *) Change all functions with names starting with des_ to be starting
6451 with DES_ instead. Add wrappers that are compatible with libdes,
6452 but are named _ossl_old_des_*. Finally, add macros that map the
6453 des_* symbols to the corresponding _ossl_old_des_* if libdes
6454 compatibility is desired. If OpenSSL 0.9.6c compatibility is
6455 desired, the des_* symbols will be mapped to DES_*, with one
6458 Since we provide two compatibility mappings, the user needs to
6459 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
6460 compatibility is desired. The default (i.e., when that macro
6461 isn't defined) is OpenSSL 0.9.6c compatibility.
6463 There are also macros that enable and disable the support of old
6464 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
6465 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
6466 are defined, the default will apply: to support the old des routines.
6468 In either case, one must include openssl/des.h to get the correct
6469 definitions. Do not try to just include openssl/des_old.h, that
6472 NOTE: This is a major break of an old API into a new one. Software
6473 authors are encouraged to switch to the DES_ style functions. Some
6474 time in the future, des_old.h and the libdes compatibility functions
6475 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
6476 default), and then completely removed.
6479 *) Test for certificates which contain unsupported critical extensions.
6480 If such a certificate is found during a verify operation it is
6481 rejected by default: this behaviour can be overridden by either
6482 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
6483 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
6484 X509_supported_extension() has also been added which returns 1 if a
6485 particular extension is supported.
6488 *) Modify the behaviour of EVP cipher functions in similar way to digests
6489 to retain compatibility with existing code.
6492 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
6493 compatibility with existing code. In particular the 'ctx' parameter does
6494 not have to be to be initialized before the call to EVP_DigestInit() and
6495 it is tidied up after a call to EVP_DigestFinal(). New function
6496 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
6497 EVP_MD_CTX_copy() changed to not require the destination to be
6498 initialized valid and new function EVP_MD_CTX_copy_ex() added which
6499 requires the destination to be valid.
6501 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
6502 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
6505 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
6506 so that complete 'Handshake' protocol structures are kept in memory
6507 instead of overwriting 'msg_type' and 'length' with 'body' data.
6510 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
6511 [Massimo Santin via Richard Levitte]
6513 *) Major restructuring to the underlying ENGINE code. This includes
6514 reduction of linker bloat, separation of pure "ENGINE" manipulation
6515 (initialisation, etc) from functionality dealing with implementations
6516 of specific crypto iterfaces. This change also introduces integrated
6517 support for symmetric ciphers and digest implementations - so ENGINEs
6518 can now accelerate these by providing EVP_CIPHER and EVP_MD
6519 implementations of their own. This is detailed in crypto/engine/README
6520 as it couldn't be adequately described here. However, there are a few
6521 API changes worth noting - some RSA, DSA, DH, and RAND functions that
6522 were changed in the original introduction of ENGINE code have now
6523 reverted back - the hooking from this code to ENGINE is now a good
6524 deal more passive and at run-time, operations deal directly with
6525 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
6526 dereferencing through an ENGINE pointer any more. Also, the ENGINE
6527 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
6528 they were not being used by the framework as there is no concept of a
6529 BIGNUM_METHOD and they could not be generalised to the new
6530 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
6531 ENGINE_cpy() has been removed as it cannot be consistently defined in
6535 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
6538 *) Change mkdef.pl to sort symbols that get the same entry number,
6539 and make sure the automatically generated functions ERR_load_*
6540 become part of libeay.num as well.
6543 *) New function SSL_renegotiate_pending(). This returns true once
6544 renegotiation has been requested (either SSL_renegotiate() call
6545 or HelloRequest/ClientHello received from the peer) and becomes
6546 false once a handshake has been completed.
6547 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
6548 sends a HelloRequest, but does not ensure that a handshake takes
6549 place. SSL_renegotiate_pending() is useful for checking if the
6550 client has followed the request.)
6553 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
6554 By default, clients may request session resumption even during
6555 renegotiation (if session ID contexts permit); with this option,
6556 session resumption is possible only in the first handshake.
6558 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
6559 more bits available for options that should not be part of
6560 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
6563 *) Add some demos for certificate and certificate request creation.
6566 *) Make maximum certificate chain size accepted from the peer application
6567 settable (SSL*_get/set_max_cert_list()), as proposed by
6568 "Douglas E. Engert" <deengert@anl.gov>.
6571 *) Add support for shared libraries for Unixware-7
6572 (Boyd Lynn Gerber <gerberb@zenez.com>).
6575 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
6576 be done prior to destruction. Use this to unload error strings from
6577 ENGINEs that load their own error strings. NB: This adds two new API
6578 functions to "get" and "set" this destroy handler in an ENGINE.
6581 *) Alter all existing ENGINE implementations (except "openssl" and
6582 "openbsd") to dynamically instantiate their own error strings. This
6583 makes them more flexible to be built both as statically-linked ENGINEs
6584 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
6585 Also, add stub code to each that makes building them as self-contained
6586 shared-libraries easier (see README.ENGINE).
6589 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
6590 implementations into applications that are completely implemented in
6591 self-contained shared-libraries. The "dynamic" ENGINE exposes control
6592 commands that can be used to configure what shared-library to load and
6593 to control aspects of the way it is handled. Also, made an update to
6594 the README.ENGINE file that brings its information up-to-date and
6595 provides some information and instructions on the "dynamic" ENGINE
6596 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
6599 *) Make it possible to unload ranges of ERR strings with a new
6600 "ERR_unload_strings" function.
6603 *) Add a copy() function to EVP_MD.
6606 *) Make EVP_MD routines take a context pointer instead of just the
6607 md_data void pointer.
6610 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
6611 that the digest can only process a single chunk of data
6612 (typically because it is provided by a piece of
6613 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
6614 is only going to provide a single chunk of data, and hence the
6615 framework needn't accumulate the data for oneshot drivers.
6618 *) As with "ERR", make it possible to replace the underlying "ex_data"
6619 functions. This change also alters the storage and management of global
6620 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
6621 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
6622 index counters. The API functions that use this state have been changed
6623 to take a "class_index" rather than pointers to the class's local STACK
6624 and counter, and there is now an API function to dynamically create new
6625 classes. This centralisation allows us to (a) plug a lot of the
6626 thread-safety problems that existed, and (b) makes it possible to clean
6627 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
6628 such data would previously have always leaked in application code and
6629 workarounds were in place to make the memory debugging turn a blind eye
6630 to it. Application code that doesn't use this new function will still
6631 leak as before, but their memory debugging output will announce it now
6632 rather than letting it slide.
6634 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
6635 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
6636 has a return value to indicate success or failure.
6639 *) Make it possible to replace the underlying "ERR" functions such that the
6640 global state (2 LHASH tables and 2 locks) is only used by the "default"
6641 implementation. This change also adds two functions to "get" and "set"
6642 the implementation prior to it being automatically set the first time
6643 any other ERR function takes place. Ie. an application can call "get",
6644 pass the return value to a module it has just loaded, and that module
6645 can call its own "set" function using that value. This means the
6646 module's "ERR" operations will use (and modify) the error state in the
6647 application and not in its own statically linked copy of OpenSSL code.
6650 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
6651 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
6652 the operation, and provides a more encapsulated way for external code
6653 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
6654 to use these functions rather than manually incrementing the counts.
6656 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
6659 *) Add EVP test program.
6662 *) Add symmetric cipher support to ENGINE. Expect the API to change!
6665 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
6666 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
6667 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
6668 These allow a CRL to be built without having to access X509_CRL fields
6669 directly. Modify 'ca' application to use new functions.
6672 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
6673 bug workarounds. Rollback attack detection is a security feature.
6674 The problem will only arise on OpenSSL servers when TLSv1 is not
6675 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
6676 Software authors not wanting to support TLSv1 will have special reasons
6677 for their choice and can explicitly enable this option.
6678 [Bodo Moeller, Lutz Jaenicke]
6680 *) Rationalise EVP so it can be extended: don't include a union of
6681 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
6682 (similar to those existing for EVP_CIPHER_CTX).
6687 EVP_MD_CTX_init(&md); /* new function call */
6688 EVP_DigestInit(&md, EVP_sha1());
6689 EVP_DigestUpdate(&md, in, len);
6690 EVP_DigestFinal(&md, out, NULL);
6691 EVP_MD_CTX_cleanup(&md); /* new function call */
6695 *) Make DES key schedule conform to the usual scheme, as well as
6696 correcting its structure. This means that calls to DES functions
6697 now have to pass a pointer to a des_key_schedule instead of a
6698 plain des_key_schedule (which was actually always a pointer
6701 des_key_schedule ks;
6703 des_set_key_checked(..., &ks);
6704 des_ncbc_encrypt(..., &ks, ...);
6706 (Note that a later change renames 'des_...' into 'DES_...'.)
6709 *) Initial reduction of linker bloat: the use of some functions, such as
6710 PEM causes large amounts of unused functions to be linked in due to
6711 poor organisation. For example pem_all.c contains every PEM function
6712 which has a knock on effect of linking in large amounts of (unused)
6713 ASN1 code. Grouping together similar functions and splitting unrelated
6714 functions prevents this.
6717 *) Cleanup of EVP macros.
6720 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
6721 correct _ecb suffix.
6724 *) Add initial OCSP responder support to ocsp application. The
6725 revocation information is handled using the text based index
6726 use by the ca application. The responder can either handle
6727 requests generated internally, supplied in files (for example
6728 via a CGI script) or using an internal minimal server.
6731 *) Add configuration choices to get zlib compression for TLS.
6734 *) Changes to Kerberos SSL for RFC 2712 compliance:
6735 1. Implemented real KerberosWrapper, instead of just using
6736 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
6737 2. Implemented optional authenticator field of KerberosWrapper.
6739 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
6740 and authenticator structs; see crypto/krb5/.
6742 Generalized Kerberos calls to support multiple Kerberos libraries.
6743 [Vern Staats <staatsvr@asc.hpc.mil>,
6744 Jeffrey Altman <jaltman@columbia.edu>
6745 via Richard Levitte]
6747 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
6748 already does with RSA. testdsa.h now has 'priv_key/pub_key'
6749 values for each of the key sizes rather than having just
6750 parameters (and 'speed' generating keys each time).
6753 *) Speed up EVP routines.
6756 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
6757 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
6758 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
6759 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
6761 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
6762 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
6763 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
6766 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
6768 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
6771 *) Added the OS2-EMX target.
6772 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
6774 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
6775 to support NCONF routines in extension code. New function CONF_set_nconf()
6776 to allow functions which take an NCONF to also handle the old LHASH
6777 structure: this means that the old CONF compatible routines can be
6778 retained (in particular wrt extensions) without having to duplicate the
6779 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
6782 *) Enhance the general user interface with mechanisms for inner control
6783 and with possibilities to have yes/no kind of prompts.
6786 *) Change all calls to low level digest routines in the library and
6787 applications to use EVP. Add missing calls to HMAC_cleanup() and
6788 don't assume HMAC_CTX can be copied using memcpy().
6789 [Verdon Walker <VWalker@novell.com>, Steve Henson]
6791 *) Add the possibility to control engines through control names but with
6792 arbitrary arguments instead of just a string.
6793 Change the key loaders to take a UI_METHOD instead of a callback
6794 function pointer. NOTE: this breaks binary compatibility with earlier
6795 versions of OpenSSL [engine].
6796 Adapt the nCipher code for these new conditions and add a card insertion
6800 *) Enhance the general user interface with mechanisms to better support
6801 dialog box interfaces, application-defined prompts, the possibility
6802 to use defaults (for example default passwords from somewhere else)
6803 and interrupts/cancellations.
6806 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
6807 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
6810 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
6811 tidy up some unnecessarily weird code in 'sk_new()').
6812 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
6814 *) Change the key loading routines for ENGINEs to use the same kind
6815 callback (pem_password_cb) as all other routines that need this
6819 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
6820 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
6821 than this minimum value is recommended.
6824 *) New random seeder for OpenVMS, using the system process statistics
6825 that are easily reachable.
6828 *) Windows apparently can't transparently handle global
6829 variables defined in DLLs. Initialisations such as:
6831 const ASN1_ITEM *it = &ASN1_INTEGER_it;
6833 won't compile. This is used by the any applications that need to
6834 declare their own ASN1 modules. This was fixed by adding the option
6835 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
6836 needed for static libraries under Win32.
6839 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
6840 setting of purpose and trust fields. New X509_STORE trust and
6841 purpose functions and tidy up setting in other SSL functions.
6844 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
6845 structure. These are inherited by X509_STORE_CTX when it is
6846 initialised. This allows various defaults to be set in the
6847 X509_STORE structure (such as flags for CRL checking and custom
6848 purpose or trust settings) for functions which only use X509_STORE_CTX
6849 internally such as S/MIME.
6851 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
6852 trust settings if they are not set in X509_STORE. This allows X509_STORE
6853 purposes and trust (in S/MIME for example) to override any set by default.
6855 Add command line options for CRL checking to smime, s_client and s_server
6859 *) Initial CRL based revocation checking. If the CRL checking flag(s)
6860 are set then the CRL is looked up in the X509_STORE structure and
6861 its validity and signature checked, then if the certificate is found
6862 in the CRL the verify fails with a revoked error.
6864 Various new CRL related callbacks added to X509_STORE_CTX structure.
6866 Command line options added to 'verify' application to support this.
6868 This needs some additional work, such as being able to handle multiple
6869 CRLs with different times, extension based lookup (rather than just
6870 by subject name) and ultimately more complete V2 CRL extension
6874 *) Add a general user interface API (crypto/ui/). This is designed
6875 to replace things like des_read_password and friends (backward
6876 compatibility functions using this new API are provided).
6877 The purpose is to remove prompting functions from the DES code
6878 section as well as provide for prompting through dialog boxes in
6879 a window system and the like.
6882 *) Add "ex_data" support to ENGINE so implementations can add state at a
6883 per-structure level rather than having to store it globally.
6886 *) Make it possible for ENGINE structures to be copied when retrieved by
6887 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
6888 This causes the "original" ENGINE structure to act like a template,
6889 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
6890 operational state can be localised to each ENGINE structure, despite the
6891 fact they all share the same "methods". New ENGINE structures returned in
6892 this case have no functional references and the return value is the single
6893 structural reference. This matches the single structural reference returned
6894 by ENGINE_by_id() normally, when it is incremented on the pre-existing
6898 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
6899 needs to match any other type at all we need to manually clear the
6903 *) Changes to the "openssl engine" utility to include;
6904 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
6905 about an ENGINE's available control commands.
6906 - executing control commands from command line arguments using the
6907 '-pre' and '-post' switches. '-post' is only used if '-t' is
6908 specified and the ENGINE is successfully initialised. The syntax for
6909 the individual commands are colon-separated, for example;
6910 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
6913 *) New dynamic control command support for ENGINEs. ENGINEs can now
6914 declare their own commands (numbers), names (strings), descriptions,
6915 and input types for run-time discovery by calling applications. A
6916 subset of these commands are implicitly classed as "executable"
6917 depending on their input type, and only these can be invoked through
6918 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
6919 can be based on user input, config files, etc). The distinction is
6920 that "executable" commands cannot return anything other than a boolean
6921 result and can only support numeric or string input, whereas some
6922 discoverable commands may only be for direct use through
6923 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
6924 pointers, or other custom uses. The "executable" commands are to
6925 support parameterisations of ENGINE behaviour that can be
6926 unambiguously defined by ENGINEs and used consistently across any
6927 OpenSSL-based application. Commands have been added to all the
6928 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
6929 control over shared-library paths without source code alterations.
6932 *) Changed all ENGINE implementations to dynamically allocate their
6933 ENGINEs rather than declaring them statically. Apart from this being
6934 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
6935 this also allows the implementations to compile without using the
6936 internal engine_int.h header.
6939 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
6940 'const' value. Any code that should be able to modify a RAND_METHOD
6941 should already have non-const pointers to it (ie. they should only
6942 modify their own ones).
6945 *) Made a variety of little tweaks to the ENGINE code.
6946 - "atalla" and "ubsec" string definitions were moved from header files
6947 to C code. "nuron" string definitions were placed in variables
6948 rather than hard-coded - allowing parameterisation of these values
6949 later on via ctrl() commands.
6950 - Removed unused "#if 0"'d code.
6951 - Fixed engine list iteration code so it uses ENGINE_free() to release
6952 structural references.
6953 - Constified the RAND_METHOD element of ENGINE structures.
6954 - Constified various get/set functions as appropriate and added
6955 missing functions (including a catch-all ENGINE_cpy that duplicates
6956 all ENGINE values onto a new ENGINE except reference counts/state).
6957 - Removed NULL parameter checks in get/set functions. Setting a method
6958 or function to NULL is a way of cancelling out a previously set
6959 value. Passing a NULL ENGINE parameter is just plain stupid anyway
6960 and doesn't justify the extra error symbols and code.
6961 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
6962 flags from engine_int.h to engine.h.
6963 - Changed prototypes for ENGINE handler functions (init(), finish(),
6964 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
6967 *) Implement binary inversion algorithm for BN_mod_inverse in addition
6968 to the algorithm using long division. The binary algorithm can be
6969 used only if the modulus is odd. On 32-bit systems, it is faster
6970 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
6971 roughly 5-15% for 256-bit moduli), so we use it only for moduli
6972 up to 450 bits. In 64-bit environments, the binary algorithm
6973 appears to be advantageous for much longer moduli; here we use it
6974 for moduli up to 2048 bits.
6977 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
6978 could not support the combine flag in choice fields.
6981 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
6982 extensions from a certificate request to the certificate.
6985 *) Allow multiple 'certopt' and 'nameopt' options to be separated
6986 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
6987 file: this allows the display of the certificate about to be
6988 signed to be customised, to allow certain fields to be included
6989 or excluded and extension details. The old system didn't display
6990 multicharacter strings properly, omitted fields not in the policy
6991 and couldn't display additional details such as extensions.
6994 *) Function EC_POINTs_mul for multiple scalar multiplication
6995 of an arbitrary number of elliptic curve points
6996 \sum scalars[i]*points[i],
6997 optionally including the generator defined for the EC_GROUP:
6998 scalar*generator + \sum scalars[i]*points[i].
7000 EC_POINT_mul is a simple wrapper function for the typical case
7001 that the point list has just one item (besides the optional
7005 *) First EC_METHODs for curves over GF(p):
7007 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7008 operations and provides various method functions that can also
7009 operate with faster implementations of modular arithmetic.
7011 EC_GFp_mont_method() reuses most functions that are part of
7012 EC_GFp_simple_method, but uses Montgomery arithmetic.
7014 [Bodo Moeller; point addition and point doubling
7015 implementation directly derived from source code provided by
7016 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7018 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7019 crypto/ec/ec_lib.c):
7021 Curves are EC_GROUP objects (with an optional group generator)
7022 based on EC_METHODs that are built into the library.
7024 Points are EC_POINT objects based on EC_GROUP objects.
7026 Most of the framework would be able to handle curves over arbitrary
7027 finite fields, but as there are no obvious types for fields other
7028 than GF(p), some functions are limited to that for now.
7031 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7032 that the file contains a complete HTTP response.
7035 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7036 change the def and num file printf format specifier from "%-40sXXX"
7037 to "%-39s XXX". The latter will always guarantee a space after the
7038 field while the former will cause them to run together if the field
7039 is 40 of more characters long.
7042 *) Constify the cipher and digest 'method' functions and structures
7043 and modify related functions to take constant EVP_MD and EVP_CIPHER
7047 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7048 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7051 *) Modify EVP_Digest*() routines so they now return values. Although the
7052 internal software routines can never fail additional hardware versions
7056 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7058 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7059 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7062 ERR_R_NESTED_ASN1_ERROR
7064 ERR_R_MISSING_ASN1_EOS
7065 were 4 .. 9, conflicting with
7066 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7068 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7069 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7071 Add new error code 'ERR_R_INTERNAL_ERROR'.
7074 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7078 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7079 sets the subject name for a new request or supersedes the
7080 subject name in a given request. Formats that can be parsed are
7081 'CN=Some Name, OU=myOU, C=IT'
7083 'CN=Some Name/OU=myOU/C=IT'.
7085 Add options '-batch' and '-verbose' to 'openssl req'.
7086 [Massimiliano Pala <madwolf@hackmasters.net>]
7088 *) Introduce the possibility to access global variables through
7089 functions on platform were that's the best way to handle exporting
7090 global variables in shared libraries. To enable this functionality,
7091 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7092 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7093 is normally done by Configure or something similar).
7095 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7096 in the source file (foo.c) like this:
7098 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7099 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7101 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7102 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7104 OPENSSL_DECLARE_GLOBAL(int,foo);
7105 #define foo OPENSSL_GLOBAL_REF(foo)
7106 OPENSSL_DECLARE_GLOBAL(double,bar);
7107 #define bar OPENSSL_GLOBAL_REF(bar)
7109 The #defines are very important, and therefore so is including the
7110 header file everywhere where the defined globals are used.
7112 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
7113 of ASN.1 items, but that structure is a bit different.
7115 The largest change is in util/mkdef.pl which has been enhanced with
7116 better and easier to understand logic to choose which symbols should
7117 go into the Windows .def files as well as a number of fixes and code
7118 cleanup (among others, algorithm keywords are now sorted
7119 lexicographically to avoid constant rewrites).
7122 *) In BN_div() keep a copy of the sign of 'num' before writing the
7123 result to 'rm' because if rm==num the value will be overwritten
7124 and produce the wrong result if 'num' is negative: this caused
7125 problems with BN_mod() and BN_nnmod().
7128 *) Function OCSP_request_verify(). This checks the signature on an
7129 OCSP request and verifies the signer certificate. The signer
7130 certificate is just checked for a generic purpose and OCSP request
7134 *) Add OCSP_check_validity() function to check the validity of OCSP
7135 responses. OCSP responses are prepared in real time and may only
7136 be a few seconds old. Simply checking that the current time lies
7137 between thisUpdate and nextUpdate max reject otherwise valid responses
7138 caused by either OCSP responder or client clock inaccuracy. Instead
7139 we allow thisUpdate and nextUpdate to fall within a certain period of
7140 the current time. The age of the response can also optionally be
7141 checked. Two new options -validity_period and -status_age added to
7145 *) If signature or public key algorithm is unrecognized print out its
7146 OID rather that just UNKNOWN.
7149 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
7150 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
7151 ID to be generated from the issuer certificate alone which can then be
7152 passed to OCSP_id_issuer_cmp().
7155 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
7156 ASN1 modules to export functions returning ASN1_ITEM pointers
7157 instead of the ASN1_ITEM structures themselves. This adds several
7158 new macros which allow the underlying ASN1 function/structure to
7159 be accessed transparently. As a result code should not use ASN1_ITEM
7160 references directly (such as &X509_it) but instead use the relevant
7161 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
7162 use of the new ASN1 code on platforms where exporting structures
7163 is problematical (for example in shared libraries) but exporting
7164 functions returning pointers to structures is not.
7167 *) Add support for overriding the generation of SSL/TLS session IDs.
7168 These callbacks can be registered either in an SSL_CTX or per SSL.
7169 The purpose of this is to allow applications to control, if they wish,
7170 the arbitrary values chosen for use as session IDs, particularly as it
7171 can be useful for session caching in multiple-server environments. A
7172 command-line switch for testing this (and any client code that wishes
7173 to use such a feature) has been added to "s_server".
7174 [Geoff Thorpe, Lutz Jaenicke]
7176 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
7177 of the form '#if defined(...) || defined(...) || ...' and
7178 '#if !defined(...) && !defined(...) && ...'. This also avoids
7179 the growing number of special cases it was previously handling.
7182 *) Make all configuration macros available for application by making
7183 sure they are available in opensslconf.h, by giving them names starting
7184 with "OPENSSL_" to avoid conflicts with other packages and by making
7185 sure e_os2.h will cover all platform-specific cases together with
7187 Additionally, it is now possible to define configuration/platform-
7188 specific names (called "system identities"). In the C code, these
7189 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
7190 macro with the name beginning with "OPENSSL_SYS_", which is determined
7191 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
7195 *) New option -set_serial to 'req' and 'x509' this allows the serial
7196 number to use to be specified on the command line. Previously self
7197 signed certificates were hard coded with serial number 0 and the
7198 CA options of 'x509' had to use a serial number in a file which was
7202 *) New options to 'ca' utility to support V2 CRL entry extensions.
7203 Currently CRL reason, invalidity date and hold instruction are
7204 supported. Add new CRL extensions to V3 code and some new objects.
7207 *) New function EVP_CIPHER_CTX_set_padding() this is used to
7208 disable standard block padding (aka PKCS#5 padding) in the EVP
7209 API, which was previously mandatory. This means that the data is
7210 not padded in any way and so the total length much be a multiple
7211 of the block size, otherwise an error occurs.
7214 *) Initial (incomplete) OCSP SSL support.
7217 *) New function OCSP_parse_url(). This splits up a URL into its host,
7218 port and path components: primarily to parse OCSP URLs. New -url
7219 option to ocsp utility.
7222 *) New nonce behavior. The return value of OCSP_check_nonce() now
7223 reflects the various checks performed. Applications can decide
7224 whether to tolerate certain situations such as an absent nonce
7225 in a response when one was present in a request: the ocsp application
7226 just prints out a warning. New function OCSP_add1_basic_nonce()
7227 this is to allow responders to include a nonce in a response even if
7228 the request is nonce-less.
7231 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
7232 skipped when using openssl x509 multiple times on a single input file,
7233 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
7236 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
7237 set string type: to handle setting ASN1_TIME structures. Fix ca
7238 utility to correctly initialize revocation date of CRLs.
7241 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
7242 the clients preferred ciphersuites and rather use its own preferences.
7243 Should help to work around M$ SGC (Server Gated Cryptography) bug in
7244 Internet Explorer by ensuring unchanged hash method during stepup.
7245 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
7248 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
7249 to aes and add a new 'exist' option to print out symbols that don't
7253 *) Additional options to ocsp utility to allow flags to be set and
7254 additional certificates supplied.
7257 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
7258 OCSP client a number of certificate to only verify the response
7262 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
7263 handle the new API. Currently only ECB, CBC modes supported. Add new
7266 Add TLS AES ciphersuites as described in RFC3268, "Advanced
7267 Encryption Standard (AES) Ciphersuites for Transport Layer
7268 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
7269 not enabled by default and were not part of the "ALL" ciphersuite
7270 alias because they were not yet official; they could be
7271 explicitly requested by specifying the "AESdraft" ciphersuite
7272 group alias. In the final release of OpenSSL 0.9.7, the group
7273 alias is called "AES" and is part of "ALL".)
7274 [Ben Laurie, Steve Henson, Bodo Moeller]
7276 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
7277 request to response.
7280 *) Functions for OCSP responders. OCSP_request_onereq_count(),
7281 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
7282 extract information from a certificate request. OCSP_response_create()
7283 creates a response and optionally adds a basic response structure.
7284 OCSP_basic_add1_status() adds a complete single response to a basic
7285 response and returns the OCSP_SINGLERESP structure just added (to allow
7286 extensions to be included for example). OCSP_basic_add1_cert() adds a
7287 certificate to a basic response and OCSP_basic_sign() signs a basic
7288 response with various flags. New helper functions ASN1_TIME_check()
7289 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
7290 (converts ASN1_TIME to GeneralizedTime).
7293 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
7294 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
7295 structure from a certificate. X509_pubkey_digest() digests the public_key
7296 contents: this is used in various key identifiers.
7299 *) Make sk_sort() tolerate a NULL argument.
7300 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
7302 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
7303 passed by the function are trusted implicitly. If any of them signed the
7304 response then it is assumed to be valid and is not verified.
7307 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
7308 to data. This was previously part of the PKCS7 ASN1 code. This
7309 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
7310 [Steve Henson, reported by Kenneth R. Robinette
7311 <support@securenetterm.com>]
7313 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
7314 routines: without these tracing memory leaks is very painful.
7315 Fix leaks in PKCS12 and PKCS7 routines.
7318 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
7319 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
7320 effectively meant GeneralizedTime would never be used. Now it
7321 is initialised to -1 but X509_time_adj() now has to check the value
7322 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
7323 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
7324 [Steve Henson, reported by Kenneth R. Robinette
7325 <support@securenetterm.com>]
7327 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
7328 result in a zero length in the ASN1_INTEGER structure which was
7329 not consistent with the structure when d2i_ASN1_INTEGER() was used
7330 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
7331 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
7332 where it did not print out a minus for negative ASN1_INTEGER.
7335 *) Add summary printout to ocsp utility. The various functions which
7336 convert status values to strings have been renamed to:
7337 OCSP_response_status_str(), OCSP_cert_status_str() and
7338 OCSP_crl_reason_str() and are no longer static. New options
7339 to verify nonce values and to disable verification. OCSP response
7340 printout format cleaned up.
7343 *) Add additional OCSP certificate checks. These are those specified
7344 in RFC2560. This consists of two separate checks: the CA of the
7345 certificate being checked must either be the OCSP signer certificate
7346 or the issuer of the OCSP signer certificate. In the latter case the
7347 OCSP signer certificate must contain the OCSP signing extended key
7348 usage. This check is performed by attempting to match the OCSP
7349 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
7350 in the OCSP_CERTID structures of the response.
7353 *) Initial OCSP certificate verification added to OCSP_basic_verify()
7354 and related routines. This uses the standard OpenSSL certificate
7355 verify routines to perform initial checks (just CA validity) and
7356 to obtain the certificate chain. Then additional checks will be
7357 performed on the chain. Currently the root CA is checked to see
7358 if it is explicitly trusted for OCSP signing. This is used to set
7359 a root CA as a global signing root: that is any certificate that
7360 chains to that CA is an acceptable OCSP signing certificate.
7363 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
7364 extensions from a separate configuration file.
7365 As when reading extensions from the main configuration file,
7366 the '-extensions ...' option may be used for specifying the
7368 [Massimiliano Pala <madwolf@comune.modena.it>]
7370 *) New OCSP utility. Allows OCSP requests to be generated or
7371 read. The request can be sent to a responder and the output
7372 parsed, outputed or printed in text form. Not complete yet:
7373 still needs to check the OCSP response validity.
7376 *) New subcommands for 'openssl ca':
7377 'openssl ca -status <serial>' prints the status of the cert with
7378 the given serial number (according to the index file).
7379 'openssl ca -updatedb' updates the expiry status of certificates
7381 [Massimiliano Pala <madwolf@comune.modena.it>]
7383 *) New '-newreq-nodes' command option to CA.pl. This is like
7384 '-newreq', but calls 'openssl req' with the '-nodes' option
7385 so that the resulting key is not encrypted.
7386 [Damien Miller <djm@mindrot.org>]
7388 *) New configuration for the GNU Hurd.
7389 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
7391 *) Initial code to implement OCSP basic response verify. This
7392 is currently incomplete. Currently just finds the signer's
7393 certificate and verifies the signature on the response.
7396 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
7397 value of OPENSSLDIR. This is available via the new '-d' option
7398 to 'openssl version', and is also included in 'openssl version -a'.
7401 *) Allowing defining memory allocation callbacks that will be given
7402 file name and line number information in additional arguments
7403 (a const char* and an int). The basic functionality remains, as
7404 well as the original possibility to just replace malloc(),
7405 realloc() and free() by functions that do not know about these
7406 additional arguments. To register and find out the current
7407 settings for extended allocation functions, the following
7408 functions are provided:
7410 CRYPTO_set_mem_ex_functions
7411 CRYPTO_set_locked_mem_ex_functions
7412 CRYPTO_get_mem_ex_functions
7413 CRYPTO_get_locked_mem_ex_functions
7415 These work the same way as CRYPTO_set_mem_functions and friends.
7416 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
7417 extended allocation function is enabled.
7418 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
7419 a conventional allocation function is enabled.
7420 [Richard Levitte, Bodo Moeller]
7422 *) Finish off removing the remaining LHASH function pointer casts.
7423 There should no longer be any prototype-casting required when using
7424 the LHASH abstraction, and any casts that remain are "bugs". See
7425 the callback types and macros at the head of lhash.h for details
7426 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
7429 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
7430 If /dev/[u]random devices are not available or do not return enough
7431 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
7433 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
7434 /etc/entropy will be queried once each in this sequence, querying stops
7435 when enough entropy was collected without querying more sockets.
7438 *) Change the Unix RAND_poll() variant to be able to poll several
7439 random devices, as specified by DEVRANDOM, until a sufficient amount
7440 of data has been collected. We spend at most 10 ms on each file
7441 (select timeout) and read in non-blocking mode. DEVRANDOM now
7442 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
7443 (previously it was just the string "/dev/urandom"), so on typical
7444 platforms the 10 ms delay will never occur.
7445 Also separate out the Unix variant to its own file, rand_unix.c.
7446 For VMS, there's a currently-empty rand_vms.c.
7449 *) Move OCSP client related routines to ocsp_cl.c. These
7450 provide utility functions which an application needing
7451 to issue a request to an OCSP responder and analyse the
7452 response will typically need: as opposed to those which an
7453 OCSP responder itself would need which will be added later.
7455 OCSP_request_sign() signs an OCSP request with an API similar
7456 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
7457 response. OCSP_response_get1_basic() extracts basic response
7458 from response. OCSP_resp_find_status(): finds and extracts status
7459 information from an OCSP_CERTID structure (which will be created
7460 when the request structure is built). These are built from lower
7461 level functions which work on OCSP_SINGLERESP structures but
7462 won't normally be used unless the application wishes to examine
7463 extensions in the OCSP response for example.
7465 Replace nonce routines with a pair of functions.
7466 OCSP_request_add1_nonce() adds a nonce value and optionally
7467 generates a random value. OCSP_check_nonce() checks the
7468 validity of the nonce in an OCSP response.
7471 *) Change function OCSP_request_add() to OCSP_request_add0_id().
7472 This doesn't copy the supplied OCSP_CERTID and avoids the
7473 need to free up the newly created id. Change return type
7474 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
7475 This can then be used to add extensions to the request.
7476 Deleted OCSP_request_new(), since most of its functionality
7477 is now in OCSP_REQUEST_new() (and the case insensitive name
7478 clash) apart from the ability to set the request name which
7479 will be added elsewhere.
7482 *) Update OCSP API. Remove obsolete extensions argument from
7483 various functions. Extensions are now handled using the new
7484 OCSP extension code. New simple OCSP HTTP function which
7485 can be used to send requests and parse the response.
7488 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
7489 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
7490 uses the special reorder version of SET OF to sort the attributes
7491 and reorder them to match the encoded order. This resolves a long
7492 standing problem: a verify on a PKCS7 structure just after signing
7493 it used to fail because the attribute order did not match the
7494 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
7495 it uses the received order. This is necessary to tolerate some broken
7496 software that does not order SET OF. This is handled by encoding
7497 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
7498 to produce the required SET OF.
7501 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
7502 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
7503 files to get correct declarations of the ASN.1 item variables.
7506 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
7507 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
7508 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
7509 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
7510 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
7511 ASN1_ITEM and no wrapper functions.
7514 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
7515 replace the old function pointer based I/O routines. Change most of
7516 the *_d2i_bio() and *_d2i_fp() functions to use these.
7519 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
7520 lines, recognice more "algorithms" that can be deselected, and make
7521 it complain about algorithm deselection that isn't recognised.
7524 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
7525 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
7526 to use new functions. Add NO_ASN1_OLD which can be set to remove
7527 some old style ASN1 functions: this can be used to determine if old
7528 code will still work when these eventually go away.
7531 *) New extension functions for OCSP structures, these follow the
7532 same conventions as certificates and CRLs.
7535 *) New function X509V3_add1_i2d(). This automatically encodes and
7536 adds an extension. Its behaviour can be customised with various
7537 flags to append, replace or delete. Various wrappers added for
7538 certificates and CRLs.
7541 *) Fix to avoid calling the underlying ASN1 print routine when
7542 an extension cannot be parsed. Correct a typo in the
7543 OCSP_SERVICELOC extension. Tidy up print OCSP format.
7546 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
7547 entries for variables.
7550 *) Add functionality to apps/openssl.c for detecting locking
7551 problems: As the program is single-threaded, all we have
7552 to do is register a locking callback using an array for
7553 storing which locks are currently held by the program.
7556 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
7557 SSL_get_ex_data_X509_STORE_idx(), which is used in
7558 ssl_verify_cert_chain() and thus can be called at any time
7559 during TLS/SSL handshakes so that thread-safety is essential.
7560 Unfortunately, the ex_data design is not at all suited
7561 for multi-threaded use, so it probably should be abolished.
7564 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
7565 [Broadcom, tweaked and integrated by Geoff Thorpe]
7567 *) Move common extension printing code to new function
7568 X509V3_print_extensions(). Reorganise OCSP print routines and
7569 implement some needed OCSP ASN1 functions. Add OCSP extensions.
7572 *) New function X509_signature_print() to remove duplication in some
7576 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
7577 set (this was treated exactly the same as SET OF previously). This
7578 is used to reorder the STACK representing the structure to match the
7579 encoding. This will be used to get round a problem where a PKCS7
7580 structure which was signed could not be verified because the STACK
7581 order did not reflect the encoded order.
7584 *) Reimplement the OCSP ASN1 module using the new code.
7587 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
7588 for its ASN1 operations. The old style function pointers still exist
7589 for now but they will eventually go away.
7592 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
7593 completely replaces the old ASN1 functionality with a table driven
7594 encoder and decoder which interprets an ASN1_ITEM structure describing
7595 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
7596 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
7597 has also been converted to the new form.
7600 *) Change BN_mod_exp_recp so that negative moduli are tolerated
7601 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
7602 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
7603 for negative moduli.
7606 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
7607 of not touching the result's sign bit.
7610 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
7614 *) Changed the LHASH code to use prototypes for callbacks, and created
7615 macros to declare and implement thin (optionally static) functions
7616 that provide type-safety and avoid function pointer casting for the
7617 type-specific callbacks.
7620 *) Added Kerberos Cipher Suites to be used with TLS, as written in
7622 [Veers Staats <staatsvr@asc.hpc.mil>,
7623 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
7625 *) Reformat the FAQ so the different questions and answers can be divided
7626 in sections depending on the subject.
7629 *) Have the zlib compression code load ZLIB.DLL dynamically under
7633 *) New function BN_mod_sqrt for computing square roots modulo a prime
7634 (using the probabilistic Tonelli-Shanks algorithm unless
7635 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
7636 be handled deterministically).
7637 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7639 *) Make BN_mod_inverse faster by explicitly handling small quotients
7640 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
7641 512 bits], about 30% for larger ones [1024 or 2048 bits].)
7644 *) New function BN_kronecker.
7647 *) Fix BN_gcd so that it works on negative inputs; the result is
7648 positive unless both parameters are zero.
7649 Previously something reasonably close to an infinite loop was
7650 possible because numbers could be growing instead of shrinking
7651 in the implementation of Euclid's algorithm.
7654 *) Fix BN_is_word() and BN_is_one() macros to take into account the
7655 sign of the number in question.
7657 Fix BN_is_word(a,w) to work correctly for w == 0.
7659 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
7660 because its test if the absolute value of 'a' equals 'w'.
7661 Note that BN_abs_is_word does *not* handle w == 0 reliably;
7662 it exists mostly for use in the implementations of BN_is_zero(),
7663 BN_is_one(), and BN_is_word().
7666 *) New function BN_swap.
7669 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
7670 the exponentiation functions are more likely to produce reasonable
7671 results on negative inputs.
7674 *) Change BN_mod_mul so that the result is always non-negative.
7675 Previously, it could be negative if one of the factors was negative;
7676 I don't think anyone really wanted that behaviour.
7679 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
7680 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
7681 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
7682 and add new functions:
7691 BN_mod_lshift1_quick
7695 These functions always generate non-negative results.
7697 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
7698 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
7700 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
7701 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
7702 be reduced modulo m.
7703 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
7706 The following entry accidentally appeared in the CHANGES file
7707 distributed with OpenSSL 0.9.7. The modifications described in
7708 it do *not* apply to OpenSSL 0.9.7.
7710 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
7711 was actually never needed) and in BN_mul(). The removal in BN_mul()
7712 required a small change in bn_mul_part_recursive() and the addition
7713 of the functions bn_cmp_part_words(), bn_sub_part_words() and
7714 bn_add_part_words(), which do the same thing as bn_cmp_words(),
7715 bn_sub_words() and bn_add_words() except they take arrays with
7720 *) In 'openssl passwd', verify passwords read from the terminal
7721 unless the '-salt' option is used (which usually means that
7722 verification would just waste user's time since the resulting
7723 hash is going to be compared with some given password hash)
7724 or the new '-noverify' option is used.
7726 This is an incompatible change, but it does not affect
7727 non-interactive use of 'openssl passwd' (passwords on the command
7728 line, '-stdin' option, '-in ...' option) and thus should not
7732 *) Remove all references to RSAref, since there's no more need for it.
7735 *) Make DSO load along a path given through an environment variable
7736 (SHLIB_PATH) with shl_load().
7739 *) Constify the ENGINE code as a result of BIGNUM constification.
7740 Also constify the RSA code and most things related to it. In a
7741 few places, most notable in the depth of the ASN.1 code, ugly
7742 casts back to non-const were required (to be solved at a later
7746 *) Make it so the openssl application has all engines loaded by default.
7749 *) Constify the BIGNUM routines a little more.
7752 *) Add the following functions:
7754 ENGINE_load_cswift()
7756 ENGINE_load_atalla()
7758 ENGINE_load_builtin_engines()
7760 That way, an application can itself choose if external engines that
7761 are built-in in OpenSSL shall ever be used or not. The benefit is
7762 that applications won't have to be linked with libdl or other dso
7763 libraries unless it's really needed.
7765 Changed 'openssl engine' to load all engines on demand.
7766 Changed the engine header files to avoid the duplication of some
7767 declarations (they differed!).
7770 *) 'openssl engine' can now list capabilities.
7773 *) Better error reporting in 'openssl engine'.
7776 *) Never call load_dh_param(NULL) in s_server.
7779 *) Add engine application. It can currently list engines by name and
7780 identity, and test if they are actually available.
7783 *) Improve RPM specification file by forcing symbolic linking and making
7784 sure the installed documentation is also owned by root.root.
7785 [Damien Miller <djm@mindrot.org>]
7787 *) Give the OpenSSL applications more possibilities to make use of
7788 keys (public as well as private) handled by engines.
7791 *) Add OCSP code that comes from CertCo.
7794 *) Add VMS support for the Rijndael code.
7797 *) Added untested support for Nuron crypto accelerator.
7800 *) Add support for external cryptographic devices. This code was
7801 previously distributed separately as the "engine" branch.
7802 [Geoff Thorpe, Richard Levitte]
7804 *) Rework the filename-translation in the DSO code. It is now possible to
7805 have far greater control over how a "name" is turned into a filename
7806 depending on the operating environment and any oddities about the
7807 different shared library filenames on each system.
7810 *) Support threads on FreeBSD-elf in Configure.
7813 *) Fix for SHA1 assembly problem with MASM: it produces
7814 warnings about corrupt line number information when assembling
7815 with debugging information. This is caused by the overlapping
7817 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
7820 NCONF_get_number() has no error checking at all. As a replacement,
7821 NCONF_get_number_e() is defined (_e for "error checking") and is
7822 promoted strongly. The old NCONF_get_number is kept around for
7823 binary backward compatibility.
7824 Make it possible for methods to load from something other than a BIO,
7825 by providing a function pointer that is given a name instead of a BIO.
7826 For example, this could be used to load configuration data from an
7830 *) Fix for non blocking accept BIOs. Added new I/O special reason
7831 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
7832 with non blocking I/O was not possible because no retry code was
7833 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
7837 *) Added the beginnings of Rijndael support.
7840 *) Fix for bug in DirectoryString mask setting. Add support for
7841 X509_NAME_print_ex() in 'req' and X509_print_ex() function
7842 to allow certificate printing to more controllable, additional
7843 'certopt' option to 'x509' to allow new printing options to be
7847 *) Clean old EAY MD5 hack from e_os.h.
7850 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
7852 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7853 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7854 [Joe Orton, Steve Henson]
7856 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
7858 *) Fix additional bug revealed by the NISCC test suite:
7860 Stop bug triggering large recursion when presented with
7861 certain ASN.1 tags (CVE-2003-0851)
7864 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
7866 *) Fix various bugs revealed by running the NISCC test suite:
7868 Stop out of bounds reads in the ASN1 code when presented with
7869 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7871 If verify callback ignores invalid public key errors don't try to check
7872 certificate signature with the NULL public key.
7876 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7877 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7881 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7882 extra data after the compression methods not only for TLS 1.0
7883 but also for SSL 3.0 (as required by the specification).
7884 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7886 *) Change X509_certificate_type() to mark the key as exported/exportable
7887 when it's 512 *bits* long, not 512 bytes.
7890 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
7892 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7893 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7894 a protocol version number mismatch like a decryption error
7895 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7898 *) Turn on RSA blinding by default in the default implementation
7899 to avoid a timing attack. Applications that don't want it can call
7900 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7901 They would be ill-advised to do so in most cases.
7902 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7904 *) Change RSA blinding code so that it works when the PRNG is not
7905 seeded (in this case, the secret RSA exponent is abused as
7906 an unpredictable seed -- if it is not unpredictable, there
7907 is no point in blinding anyway). Make RSA blinding thread-safe
7908 by remembering the creator's thread ID in rsa->blinding and
7909 having all other threads use local one-time blinding factors
7910 (this requires more computation than sharing rsa->blinding, but
7911 avoids excessive locking; and if an RSA object is not shared
7912 between threads, blinding will still be very fast).
7915 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
7917 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7918 via timing by performing a MAC computation even if incorrrect
7919 block cipher padding has been found. This is a countermeasure
7920 against active attacks where the attacker has to distinguish
7921 between bad padding and a MAC verification error. (CVE-2003-0078)
7923 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7924 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7925 Martin Vuagnoux (EPFL, Ilion)]
7927 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
7929 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
7930 memory from it's contents. This is done with a counter that will
7931 place alternating values in each byte. This can be used to solve
7932 two issues: 1) the removal of calls to memset() by highly optimizing
7933 compilers, and 2) cleansing with other values than 0, since those can
7934 be read through on certain media, for example a swap space on disk.
7937 *) Bugfix: client side session caching did not work with external caching,
7938 because the session->cipher setting was not restored when reloading
7939 from the external cache. This problem was masked, when
7940 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
7941 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
7944 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
7945 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
7946 [Zeev Lieber <zeev-l@yahoo.com>]
7948 *) Undo an undocumented change introduced in 0.9.6e which caused
7949 repeated calls to OpenSSL_add_all_ciphers() and
7950 OpenSSL_add_all_digests() to be ignored, even after calling
7954 *) Change the default configuration reader to deal with last line not
7955 being properly terminated.
7958 *) Change X509_NAME_cmp() so it applies the special rules on handling
7959 DN values that are of type PrintableString, as well as RDNs of type
7960 emailAddress where the value has the type ia5String.
7961 [stefank@valicert.com via Richard Levitte]
7963 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
7964 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
7965 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
7966 the bitwise-OR of the two for use by the majority of applications
7967 wanting this behaviour, and update the docs. The documented
7968 behaviour and actual behaviour were inconsistent and had been
7969 changing anyway, so this is more a bug-fix than a behavioural
7971 [Geoff Thorpe, diagnosed by Nadav Har'El]
7973 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
7974 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
7977 *) Fix initialization code race conditions in
7978 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
7979 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
7980 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
7981 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
7982 ssl2_get_cipher_by_char(),
7983 ssl3_get_cipher_by_char().
7984 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
7986 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
7987 the cached sessions are flushed, as the remove_cb() might use ex_data
7988 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
7989 (see [openssl.org #212]).
7990 [Geoff Thorpe, Lutz Jaenicke]
7992 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
7993 length, instead of the encoding length to d2i_ASN1_OBJECT.
7996 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
7998 *) [In 0.9.6g-engine release:]
7999 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8000 [Lynn Gazis <lgazis@rainbow.com>]
8002 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8004 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8005 and get fix the header length calculation.
8006 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8007 Alon Kantor <alonk@checkpoint.com> (and others),
8010 *) Use proper error handling instead of 'assertions' in buffer
8011 overflow checks added in 0.9.6e. This prevents DoS (the
8012 assertions could call abort()).
8013 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8015 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8017 *) Add various sanity checks to asn1_get_length() to reject
8018 the ASN1 length bytes if they exceed sizeof(long), will appear
8019 negative or the content length exceeds the length of the
8021 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8023 *) Fix cipher selection routines: ciphers without encryption had no flags
8024 for the cipher strength set and where therefore not handled correctly
8025 by the selection routines (PR #130).
8028 *) Fix EVP_dsa_sha macro.
8032 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8033 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8034 that was added in OpenSSL 0.9.6d.
8036 As the countermeasure turned out to be incompatible with some
8037 broken SSL implementations, the new option is part of SSL_OP_ALL.
8038 SSL_OP_ALL is usually employed when compatibility with weird SSL
8039 implementations is desired (e.g. '-bugs' option to 's_client' and
8040 's_server'), so the new option is automatically set in many
8044 *) Changes in security patch:
8046 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8047 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8048 Air Force Materiel Command, USAF, under agreement number
8051 *) Add various sanity checks to asn1_get_length() to reject
8052 the ASN1 length bytes if they exceed sizeof(long), will appear
8053 negative or the content length exceeds the length of the
8054 supplied buffer. (CVE-2002-0659)
8055 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8057 *) Assertions for various potential buffer overflows, not known to
8059 [Ben Laurie (CHATS)]
8061 *) Various temporary buffers to hold ASCII versions of integers were
8062 too small for 64 bit platforms. (CVE-2002-0655)
8063 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8065 *) Remote buffer overflow in SSL3 protocol - an attacker could
8066 supply an oversized session ID to a client. (CVE-2002-0656)
8067 [Ben Laurie (CHATS)]
8069 *) Remote buffer overflow in SSL2 protocol - an attacker could
8070 supply an oversized client master key. (CVE-2002-0656)
8071 [Ben Laurie (CHATS)]
8073 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8075 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8076 encoded as NULL) with id-dsa-with-sha1.
8077 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8079 *) Check various X509_...() return values in apps/req.c.
8080 [Nils Larsch <nla@trustcenter.de>]
8082 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8083 an end-of-file condition would erroneously be flagged, when the CRLF
8084 was just at the end of a processed block. The bug was discovered when
8085 processing data through a buffering memory BIO handing the data to a
8086 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8087 <ptsekov@syntrex.com> and Nedelcho Stanev.
8090 *) Implement a countermeasure against a vulnerability recently found
8091 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8092 before application data chunks to avoid the use of known IVs
8093 with data potentially chosen by the attacker.
8096 *) Fix length checks in ssl3_get_client_hello().
8099 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8100 to prevent ssl3_read_internal() from incorrectly assuming that
8101 ssl3_read_bytes() found application data while handshake
8102 processing was enabled when in fact s->s3->in_read_app_data was
8103 merely automatically cleared during the initial handshake.
8104 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8106 *) Fix object definitions for Private and Enterprise: they were not
8107 recognized in their shortname (=lowercase) representation. Extend
8108 obj_dat.pl to issue an error when using undefined keywords instead
8109 of silently ignoring the problem (Svenning Sorensen
8110 <sss@sss.dnsalias.net>).
8113 *) Fix DH_generate_parameters() so that it works for 'non-standard'
8114 generators, i.e. generators other than 2 and 5. (Previously, the
8115 code did not properly initialise the 'add' and 'rem' values to
8116 BN_generate_prime().)
8118 In the new general case, we do not insist that 'generator' is
8119 actually a primitive root: This requirement is rather pointless;
8120 a generator of the order-q subgroup is just as good, if not
8124 *) Map new X509 verification errors to alerts. Discovered and submitted by
8125 Tom Wu <tom@arcot.com>.
8128 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
8129 returning non-zero before the data has been completely received
8130 when using non-blocking I/O.
8131 [Bodo Moeller; problem pointed out by John Hughes]
8133 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
8134 [Ben Laurie, Lutz Jaenicke]
8136 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
8137 Yoram Zahavi <YoramZ@gilian.com>).
8140 *) Add information about CygWin 1.3 and on, and preserve proper
8141 configuration for the versions before that.
8142 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
8144 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
8145 check whether we deal with a copy of a session and do not delete from
8146 the cache in this case. Problem reported by "Izhar Shoshani Levi"
8147 <izhar@checkpoint.com>.
8150 *) Do not store session data into the internal session cache, if it
8151 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
8152 flag is set). Proposed by Aslam <aslam@funk.com>.
8155 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
8159 *) [In 0.9.6d-engine release:]
8160 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
8161 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
8163 *) Add the configuration target linux-s390x.
8164 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
8166 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
8167 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
8168 variable as an indication that a ClientHello message has been
8169 received. As the flag value will be lost between multiple
8170 invocations of ssl3_accept when using non-blocking I/O, the
8171 function may not be aware that a handshake has actually taken
8172 place, thus preventing a new session from being added to the
8175 To avoid this problem, we now set s->new_session to 2 instead of
8176 using a local variable.
8177 [Lutz Jaenicke, Bodo Moeller]
8179 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
8180 if the SSL_R_LENGTH_MISMATCH error is detected.
8181 [Geoff Thorpe, Bodo Moeller]
8183 *) New 'shared_ldflag' column in Configure platform table.
8186 *) Fix EVP_CIPHER_mode macro.
8187 ["Dan S. Camper" <dan@bti.net>]
8189 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
8190 type, we must throw them away by setting rr->length to 0.
8191 [D P Chang <dpc@qualys.com>]
8193 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
8195 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
8196 <Dominikus.Scherkl@biodata.com>. (The previous implementation
8197 worked incorrectly for those cases where range = 10..._2 and
8198 3*range is two bits longer than range.)
8201 *) Only add signing time to PKCS7 structures if it is not already
8205 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
8206 OBJ_ld_ce should be OBJ_id_ce.
8207 Also some ip-pda OIDs in crypto/objects/objects.txt were
8208 incorrect (cf. RFC 3039).
8209 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
8211 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
8212 returns early because it has nothing to do.
8213 [Andy Schneider <andy.schneider@bjss.co.uk>]
8215 *) [In 0.9.6c-engine release:]
8216 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
8217 [Andy Schneider <andy.schneider@bjss.co.uk>]
8219 *) [In 0.9.6c-engine release:]
8220 Add support for Cryptographic Appliance's keyserver technology.
8221 (Use engine 'keyclient')
8222 [Cryptographic Appliances and Geoff Thorpe]
8224 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
8225 is called via tools/c89.sh because arguments have to be
8226 rearranged (all '-L' options must appear before the first object
8228 [Richard Shapiro <rshapiro@abinitio.com>]
8230 *) [In 0.9.6c-engine release:]
8231 Add support for Broadcom crypto accelerator cards, backported
8233 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
8235 *) [In 0.9.6c-engine release:]
8236 Add support for SureWare crypto accelerator cards from
8237 Baltimore Technologies. (Use engine 'sureware')
8238 [Baltimore Technologies and Mark Cox]
8240 *) [In 0.9.6c-engine release:]
8241 Add support for crypto accelerator cards from Accelerated
8242 Encryption Processing, www.aep.ie. (Use engine 'aep')
8243 [AEP Inc. and Mark Cox]
8245 *) Add a configuration entry for gcc on UnixWare.
8246 [Gary Benson <gbenson@redhat.com>]
8248 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
8249 messages are stored in a single piece (fixed-length part and
8250 variable-length part combined) and fix various bugs found on the way.
8253 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
8254 instead. BIO_gethostbyname() does not know what timeouts are
8255 appropriate, so entries would stay in cache even when they have
8257 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
8259 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
8260 faced with a pathologically small ClientHello fragment that does
8261 not contain client_version: Instead of aborting with an error,
8262 simply choose the highest available protocol version (i.e.,
8263 TLS 1.0 unless it is disabled). In practice, ClientHello
8264 messages are never sent like this, but this change gives us
8265 strictly correct behaviour at least for TLS.
8268 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
8269 never resets s->method to s->ctx->method when called from within
8270 one of the SSL handshake functions.
8271 [Bodo Moeller; problem pointed out by Niko Baric]
8273 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
8274 (sent using the client's version number) if client_version is
8275 smaller than the protocol version in use. Also change
8276 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
8277 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
8278 the client will at least see that alert.
8281 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
8285 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
8286 client receives HelloRequest while in a handshake.
8287 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
8289 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
8290 should end in 'break', not 'goto end' which circumvents various
8291 cleanups done in state SSL_ST_OK. But session related stuff
8292 must be disabled for SSL_ST_OK in the case that we just sent a
8295 Also avoid some overhead by not calling ssl_init_wbio_buffer()
8296 before just sending a HelloRequest.
8297 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
8299 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
8300 reveal whether illegal block cipher padding was found or a MAC
8301 verification error occurred. (Neither SSLerr() codes nor alerts
8302 are directly visible to potential attackers, but the information
8303 may leak via logfiles.)
8305 Similar changes are not required for the SSL 2.0 implementation
8306 because the number of padding bytes is sent in clear for SSL 2.0,
8307 and the extra bytes are just ignored. However ssl/s2_pkt.c
8308 failed to verify that the purported number of padding bytes is in
8312 *) Add OpenUNIX-8 support including shared libraries
8313 (Boyd Lynn Gerber <gerberb@zenez.com>).
8316 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
8317 'wristwatch attack' using huge encoding parameters (cf.
8318 James H. Manger's CRYPTO 2001 paper). Note that the
8319 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
8320 encoding parameters and hence was not vulnerable.
8323 *) BN_sqr() bug fix.
8324 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
8326 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
8327 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
8328 followed by modular reduction.
8329 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
8331 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
8332 equivalent based on BN_pseudo_rand() instead of BN_rand().
8335 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
8336 This function was broken, as the check for a new client hello message
8337 to handle SGC did not allow these large messages.
8338 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
8341 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
8344 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
8345 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
8348 *) Rework the configuration and shared library support for Tru64 Unix.
8349 The configuration part makes use of modern compiler features and
8350 still retains old compiler behavior for those that run older versions
8351 of the OS. The shared library support part includes a variant that
8352 uses the RPATH feature, and is available through the special
8353 configuration target "alpha-cc-rpath", which will never be selected
8355 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
8357 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
8358 with the same message size as in ssl3_get_certificate_request().
8359 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
8360 messages might inadvertently be reject as too long.
8361 [Petr Lampa <lampa@fee.vutbr.cz>]
8363 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
8366 *) Modified SSL library such that the verify_callback that has been set
8367 specificly for an SSL object with SSL_set_verify() is actually being
8368 used. Before the change, a verify_callback set with this function was
8369 ignored and the verify_callback() set in the SSL_CTX at the time of
8370 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
8371 to allow the necessary settings.
8374 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
8375 explicitly to NULL, as at least on Solaris 8 this seems not always to be
8376 done automatically (in contradiction to the requirements of the C
8377 standard). This made problems when used from OpenSSH.
8380 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
8381 dh->length and always used
8383 BN_rand_range(priv_key, dh->p).
8385 BN_rand_range() is not necessary for Diffie-Hellman, and this
8386 specific range makes Diffie-Hellman unnecessarily inefficient if
8387 dh->length (recommended exponent length) is much smaller than the
8388 length of dh->p. We could use BN_rand_range() if the order of
8389 the subgroup was stored in the DH structure, but we only have
8394 BN_rand(priv_key, l, ...)
8396 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
8402 RSA_eay_public_encrypt
8403 RSA_eay_private_decrypt
8404 RSA_eay_private_encrypt (signing)
8405 RSA_eay_public_decrypt (signature verification)
8407 (default implementations for RSA_public_encrypt,
8408 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
8409 always reject numbers >= n.
8412 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
8413 to synchronize access to 'locking_thread'. This is necessary on
8414 systems where access to 'locking_thread' (an 'unsigned long'
8415 variable) is not atomic.
8418 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
8419 *before* setting the 'crypto_lock_rand' flag. The previous code had
8420 a race condition if 0 is a valid thread ID.
8421 [Travis Vitek <vitek@roguewave.com>]
8423 *) Add support for shared libraries under Irix.
8424 [Albert Chin-A-Young <china@thewrittenword.com>]
8426 *) Add configuration option to build on Linux on both big-endian and
8428 [Ralf Baechle <ralf@uni-koblenz.de>]
8430 *) Add the possibility to create shared libraries on HP-UX.
8433 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
8435 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
8436 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
8437 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
8438 PRNG state recovery was possible based on the output of
8439 one PRNG request appropriately sized to gain knowledge on
8440 'md' followed by enough consecutive 1-byte PRNG requests
8441 to traverse all of 'state'.
8443 1. When updating 'md_local' (the current thread's copy of 'md')
8444 during PRNG output generation, hash all of the previous
8445 'md_local' value, not just the half used for PRNG output.
8447 2. Make the number of bytes from 'state' included into the hash
8448 independent from the number of PRNG bytes requested.
8450 The first measure alone would be sufficient to avoid
8451 Markku-Juhani's attack. (Actually it had never occurred
8452 to me that the half of 'md_local' used for chaining was the
8453 half from which PRNG output bytes were taken -- I had always
8454 assumed that the secret half would be used.) The second
8455 measure makes sure that additional data from 'state' is never
8456 mixed into 'md_local' in small portions; this heuristically
8457 further strengthens the PRNG.
8460 *) Fix crypto/bn/asm/mips3.s.
8463 *) When only the key is given to "enc", the IV is undefined. Print out
8464 an error message in this case.
8467 *) Handle special case when X509_NAME is empty in X509 printing routines.
8470 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
8471 positive and less than q.
8474 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
8475 used: it isn't thread safe and the add_lock_callback should handle
8477 [Paul Rose <Paul.Rose@bridge.com>]
8479 *) Verify that incoming data obeys the block size in
8480 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
8484 [Ulf Möller, Bodo Möller]
8486 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
8487 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
8488 when fixing the server behaviour for backwards-compatible 'client
8489 hello' messages. (Note that the attack is impractical against
8490 SSL 3.0 and TLS 1.0 anyway because length and version checking
8491 means that the probability of guessing a valid ciphertext is
8492 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
8495 Before 0.9.5, the countermeasure (hide the error by generating a
8496 random 'decryption result') did not work properly because
8497 ERR_clear_error() was missing, meaning that SSL_get_error() would
8498 detect the supposedly ignored error.
8500 Both problems are now fixed.
8503 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
8504 (previously it was 1024).
8507 *) Fix for compatibility mode trust settings: ignore trust settings
8508 unless some valid trust or reject settings are present.
8511 *) Fix for blowfish EVP: its a variable length cipher.
8514 *) Fix various bugs related to DSA S/MIME verification. Handle missing
8515 parameters in DSA public key structures and return an error in the
8516 DSA routines if parameters are absent.
8519 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
8520 in the current directory if neither $RANDFILE nor $HOME was set.
8521 RAND_file_name() in 0.9.6a returned NULL in this case. This has
8522 caused some confusion to Windows users who haven't defined $HOME.
8523 Thus RAND_file_name() is changed again: e_os.h can define a
8524 DEFAULT_HOME, which will be used if $HOME is not set.
8525 For Windows, we use "C:"; on other platforms, we still require
8526 environment variables.
8528 *) Move 'if (!initialized) RAND_poll()' into regions protected by
8529 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
8530 having multiple threads call RAND_poll() concurrently.
8533 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
8534 combination of a flag and a thread ID variable.
8535 Otherwise while one thread is in ssleay_rand_bytes (which sets the
8536 flag), *other* threads can enter ssleay_add_bytes without obeying
8537 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
8538 that they do not hold after the first thread unsets add_do_not_lock).
8541 *) Change bctest again: '-x' expressions are not available in all
8545 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
8547 *) Fix a couple of memory leaks in PKCS7_dataDecode()
8548 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
8550 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
8551 the default extension for executables, if any. Also, make the perl
8552 scripts that use symlink() to test if it really exists and use "cp"
8553 if it doesn't. All this made OpenSSL compilable and installable in
8557 *) Fix for asn1_GetSequence() for indefinite length constructed data.
8558 If SEQUENCE is length is indefinite just set c->slen to the total
8559 amount of data available.
8560 [Steve Henson, reported by shige@FreeBSD.org]
8561 [This change does not apply to 0.9.7.]
8563 *) Change bctest to avoid here-documents inside command substitution
8564 (workaround for FreeBSD /bin/sh bug).
8565 For compatibility with Ultrix, avoid shell functions (introduced
8566 in the bctest version that searches along $PATH).
8569 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
8570 with des_encrypt() defined on some operating systems, like Solaris
8574 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
8575 On the Importance of Eliminating Errors in Cryptographic
8576 Computations, J. Cryptology 14 (2001) 2, 101-119,
8577 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
8580 *) MIPS assembler BIGNUM division bug fix.
8583 *) Disabled incorrect Alpha assembler code.
8586 *) Fix PKCS#7 decode routines so they correctly update the length
8587 after reading an EOC for the EXPLICIT tag.
8589 [This change does not apply to 0.9.7.]
8591 *) Fix bug in PKCS#12 key generation routines. This was triggered
8592 if a 3DES key was generated with a 0 initial byte. Include
8593 PKCS12_BROKEN_KEYGEN compilation option to retain the old
8594 (but broken) behaviour.
8597 *) Enhance bctest to search for a working bc along $PATH and print
8599 [Tim Rice <tim@multitalents.net> via Richard Levitte]
8601 *) Fix memory leaks in err.c: free err_data string if necessary;
8602 don't write to the wrong index in ERR_set_error_data.
8605 *) Implement ssl23_peek (analogous to ssl23_read), which previously
8609 *) Replace rdtsc with _emit statements for VC++ version 5.
8610 [Jeremy Cooper <jeremy@baymoo.org>]
8612 *) Make it possible to reuse SSLv2 sessions.
8615 *) In copy_email() check for >= 0 as a return value for
8616 X509_NAME_get_index_by_NID() since 0 is a valid index.
8617 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
8619 *) Avoid coredump with unsupported or invalid public keys by checking if
8620 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
8621 PKCS7_verify() fails with non detached data.
8624 *) Don't use getenv in library functions when run as setuid/setgid.
8625 New function OPENSSL_issetugid().
8628 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
8629 due to incorrect handling of multi-threading:
8631 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
8633 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
8635 3. Count how many times MemCheck_off() has been called so that
8636 nested use can be treated correctly. This also avoids
8637 inband-signalling in the previous code (which relied on the
8638 assumption that thread ID 0 is impossible).
8641 *) Add "-rand" option also to s_client and s_server.
8644 *) Fix CPU detection on Irix 6.x.
8645 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
8646 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8648 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
8651 [This change does not apply to 0.9.7.]
8653 *) Use the cached encoding of an X509_NAME structure rather than
8654 copying it. This is apparently the reason for the libsafe "errors"
8655 but the code is actually correct.
8658 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
8659 Bleichenbacher's DSA attack.
8660 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
8661 to be set and top=0 forces the highest bit to be set; top=-1 is new
8662 and leaves the highest bit random.
8663 [Ulf Moeller, Bodo Moeller]
8665 *) In the NCONF_...-based implementations for CONF_... queries
8666 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
8667 a temporary CONF structure with the data component set to NULL
8668 (which gives segmentation faults in lh_retrieve).
8669 Instead, use NULL for the CONF pointer in CONF_get_string and
8670 CONF_get_number (which may use environment variables) and directly
8671 return NULL from CONF_get_section.
8674 *) Fix potential buffer overrun for EBCDIC.
8677 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
8678 keyUsage if basicConstraints absent for a CA.
8681 *) Make SMIME_write_PKCS7() write mail header values with a format that
8682 is more generally accepted (no spaces before the semicolon), since
8683 some programs can't parse those values properly otherwise. Also make
8684 sure BIO's that break lines after each write do not create invalid
8688 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
8689 macros previously used would not encode an empty SEQUENCE OF
8690 and break the signature.
8692 [This change does not apply to 0.9.7.]
8694 *) Zero the premaster secret after deriving the master secret in
8698 *) Add some EVP_add_digest_alias registrations (as found in
8699 OpenSSL_add_all_digests()) to SSL_library_init()
8700 aka OpenSSL_add_ssl_algorithms(). This provides improved
8701 compatibility with peers using X.509 certificates
8702 with unconventional AlgorithmIdentifier OIDs.
8705 *) Fix for Irix with NO_ASM.
8706 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
8708 *) ./config script fixes.
8709 [Ulf Moeller, Richard Levitte]
8711 *) Fix 'openssl passwd -1'.
8714 *) Change PKCS12_key_gen_asc() so it can cope with non null
8715 terminated strings whose length is passed in the passlen
8716 parameter, for example from PEM callbacks. This was done
8717 by adding an extra length parameter to asc2uni().
8718 [Steve Henson, reported by <oddissey@samsung.co.kr>]
8720 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
8721 call failed, free the DSA structure.
8724 *) Fix to uni2asc() to cope with zero length Unicode strings.
8725 These are present in some PKCS#12 files.
8728 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
8729 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
8730 when writing a 32767 byte record.
8731 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
8733 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
8734 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
8736 (RSA objects have a reference count access to which is protected
8737 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
8738 so they are meant to be shared between threads.)
8739 [Bodo Moeller, Geoff Thorpe; original patch submitted by
8740 "Reddie, Steven" <Steven.Reddie@ca.com>]
8742 *) Fix a deadlock in CRYPTO_mem_leaks().
8745 *) Use better test patterns in bntest.
8748 *) rand_win.c fix for Borland C.
8751 *) BN_rshift bugfix for n == 0.
8754 *) Add a 'bctest' script that checks for some known 'bc' bugs
8755 so that 'make test' does not abort just because 'bc' is broken.
8758 *) Store verify_result within SSL_SESSION also for client side to
8759 avoid potential security hole. (Re-used sessions on the client side
8760 always resulted in verify_result==X509_V_OK, not using the original
8761 result of the server certificate verification.)
8764 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
8765 SSL3_RT_APPLICATION_DATA, return 0.
8766 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
8770 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
8771 releases, have been re-implemented by renaming the previous
8772 implementations of ssl2_read and ssl3_read to ssl2_read_internal
8773 and ssl3_read_internal, respectively, and adding 'peek' parameters
8774 to them. The new ssl[23]_{read,peek} functions are calls to
8775 ssl[23]_read_internal with the 'peek' flag set appropriately.
8776 A 'peek' parameter has also been added to ssl3_read_bytes, which
8777 does the actual work for ssl3_read_internal.
8780 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
8781 the method-specific "init()" handler. Also clean up ex_data after
8782 calling the method-specific "finish()" handler. Previously, this was
8783 happening the other way round.
8786 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
8787 The previous value, 12, was not always sufficient for BN_mod_exp().
8790 *) Make sure that shared libraries get the internal name engine with
8791 the full version number and not just 0. This should mark the
8792 shared libraries as not backward compatible. Of course, this should
8793 be changed again when we can guarantee backward binary compatibility.
8796 *) Fix typo in get_cert_by_subject() in by_dir.c
8797 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
8799 *) Rework the system to generate shared libraries:
8801 - Make note of the expected extension for the shared libraries and
8802 if there is a need for symbolic links from for example libcrypto.so.0
8803 to libcrypto.so.0.9.7. There is extended info in Configure for
8806 - Make as few rebuilds of the shared libraries as possible.
8808 - Still avoid linking the OpenSSL programs with the shared libraries.
8810 - When installing, install the shared libraries separately from the
8814 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
8816 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
8817 and not in SSL_clear because the latter is also used by the
8818 accept/connect functions; previously, the settings made by
8819 SSL_set_read_ahead would be lost during the handshake.
8820 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
8822 *) Correct util/mkdef.pl to be selective about disabled algorithms.
8823 Previously, it would create entries for disabled algorithms no
8827 *) Added several new manual pages for SSL_* function.
8830 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
8832 *) In ssl23_get_client_hello, generate an error message when faced
8833 with an initial SSL 3.0/TLS record that is too small to contain the
8834 first two bytes of the ClientHello message, i.e. client_version.
8835 (Note that this is a pathologic case that probably has never happened
8836 in real life.) The previous approach was to use the version number
8837 from the record header as a substitute; but our protocol choice
8838 should not depend on that one because it is not authenticated
8839 by the Finished messages.
8842 *) More robust randomness gathering functions for Windows.
8843 [Jeffrey Altman <jaltman@columbia.edu>]
8845 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
8846 not set then we don't setup the error code for issuer check errors
8847 to avoid possibly overwriting other errors which the callback does
8848 handle. If an application does set the flag then we assume it knows
8849 what it is doing and can handle the new informational codes
8853 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
8854 a general "ANY" type, as such it should be able to decode anything
8855 including tagged types. However it didn't check the class so it would
8856 wrongly interpret tagged types in the same way as their universal
8857 counterpart and unknown types were just rejected. Changed so that the
8858 tagged and unknown types are handled in the same way as a SEQUENCE:
8859 that is the encoding is stored intact. There is also a new type
8860 "V_ASN1_OTHER" which is used when the class is not universal, in this
8861 case we have no idea what the actual type is so we just lump them all
8865 *) On VMS, stdout may very well lead to a file that is written to
8866 in a record-oriented fashion. That means that every write() will
8867 write a separate record, which will be read separately by the
8868 programs trying to read from it. This can be very confusing.
8870 The solution is to put a BIO filter in the way that will buffer
8871 text until a linefeed is reached, and then write everything a
8872 line at a time, so every record written will be an actual line,
8873 not chunks of lines and not (usually doesn't happen, but I've
8874 seen it once) several lines in one record. BIO_f_linebuffer() is
8877 Currently, it's a VMS-only method, because that's where it has
8878 been tested well enough.
8881 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
8882 it can return incorrect results.
8883 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
8884 but it was in 0.9.6-beta[12].)
8887 *) Disable the check for content being present when verifying detached
8888 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
8889 include zero length content when signing messages.
8892 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
8893 BIO_ctrl (for BIO pairs).
8896 *) Add DSO method for VMS.
8899 *) Bug fix: Montgomery multiplication could produce results with the
8903 *) Add RPM specification openssl.spec and modify it to build three
8904 packages. The default package contains applications, application
8905 documentation and run-time libraries. The devel package contains
8906 include files, static libraries and function documentation. The
8907 doc package contains the contents of the doc directory. The original
8908 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
8911 *) Add a large number of documentation files for many SSL routines.
8912 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8914 *) Add a configuration entry for Sony News 4.
8915 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
8917 *) Don't set the two most significant bits to one when generating a
8918 random number < q in the DSA library.
8921 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
8922 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
8923 the underlying transport is blocking) if a handshake took place.
8924 (The default behaviour is needed by applications such as s_client
8925 and s_server that use select() to determine when to use SSL_read;
8926 but for applications that know in advance when to expect data, it
8927 just makes things more complicated.)
8930 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
8934 *) Add a few more EBCDIC conditionals that make `req' and `x509'
8935 work better on such systems.
8936 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8938 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
8939 Update PKCS12_parse() so it copies the friendlyName and the
8940 keyid to the certificates aux info.
8943 *) Fix bug in PKCS7_verify() which caused an infinite loop
8944 if there was more than one signature.
8945 [Sven Uszpelkat <su@celocom.de>]
8947 *) Major change in util/mkdef.pl to include extra information
8948 about each symbol, as well as presenting variables as well
8949 as functions. This change means that there's n more need
8950 to rebuild the .num files when some algorithms are excluded.
8953 *) Allow the verify time to be set by an application,
8954 rather than always using the current time.
8957 *) Phase 2 verify code reorganisation. The certificate
8958 verify code now looks up an issuer certificate by a
8959 number of criteria: subject name, authority key id
8960 and key usage. It also verifies self signed certificates
8961 by the same criteria. The main comparison function is
8962 X509_check_issued() which performs these checks.
8964 Lot of changes were necessary in order to support this
8965 without completely rewriting the lookup code.
8967 Authority and subject key identifier are now cached.
8969 The LHASH 'certs' is X509_STORE has now been replaced
8970 by a STACK_OF(X509_OBJECT). This is mainly because an
8971 LHASH can't store or retrieve multiple objects with
8972 the same hash value.
8974 As a result various functions (which were all internal
8975 use only) have changed to handle the new X509_STORE
8976 structure. This will break anything that messed round
8977 with X509_STORE internally.
8979 The functions X509_STORE_add_cert() now checks for an
8980 exact match, rather than just subject name.
8982 The X509_STORE API doesn't directly support the retrieval
8983 of multiple certificates matching a given criteria, however
8984 this can be worked round by performing a lookup first
8985 (which will fill the cache with candidate certificates)
8986 and then examining the cache for matches. This is probably
8987 the best we can do without throwing out X509_LOOKUP
8988 entirely (maybe later...).
8990 The X509_VERIFY_CTX structure has been enhanced considerably.
8992 All certificate lookup operations now go via a get_issuer()
8993 callback. Although this currently uses an X509_STORE it
8994 can be replaced by custom lookups. This is a simple way
8995 to bypass the X509_STORE hackery necessary to make this
8996 work and makes it possible to use more efficient techniques
8997 in future. A very simple version which uses a simple
8998 STACK for its trusted certificate store is also provided
8999 using X509_STORE_CTX_trusted_stack().
9001 The verify_cb() and verify() callbacks now have equivalents
9002 in the X509_STORE_CTX structure.
9004 X509_STORE_CTX also has a 'flags' field which can be used
9005 to customise the verify behaviour.
9008 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9009 excludes S/MIME capabilities.
9012 *) When a certificate request is read in keep a copy of the
9013 original encoding of the signed data and use it when outputting
9014 again. Signatures then use the original encoding rather than
9015 a decoded, encoded version which may cause problems if the
9016 request is improperly encoded.
9019 *) For consistency with other BIO_puts implementations, call
9020 buffer_write(b, ...) directly in buffer_puts instead of calling
9023 In BIO_puts, increment b->num_write as in BIO_write.
9024 [Peter.Sylvester@EdelWeb.fr]
9026 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9027 BN_zero, we may not return a BIGNUM with an array consisting of
9031 *) Avoid calling abort() from within the library when problems are
9032 detected, except if preprocessor symbols have been defined
9033 (such as REF_CHECK, BN_DEBUG etc.).
9036 *) New openssl application 'rsautl'. This utility can be
9037 used for low level RSA operations. DER public key
9038 BIO/fp routines also added.
9041 *) New Configure entry and patches for compiling on QNX 4.
9042 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9044 *) A demo state-machine implementation was sponsored by
9045 Nuron (http://www.nuron.com/) and is now available in
9046 demos/state_machine.
9049 *) New options added to the 'dgst' utility for signature
9050 generation and verification.
9053 *) Unrecognized PKCS#7 content types are now handled via a
9054 catch all ASN1_TYPE structure. This allows unsupported
9055 types to be stored as a "blob" and an application can
9056 encode and decode it manually.
9059 *) Fix various signed/unsigned issues to make a_strex.c
9061 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9063 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9064 length if passed a buffer. ASN1_INTEGER_to_BN failed
9065 if passed a NULL BN and its argument was negative.
9066 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9068 *) Modification to PKCS#7 encoding routines to output definite
9069 length encoding. Since currently the whole structures are in
9070 memory there's not real point in using indefinite length
9071 constructed encoding. However if OpenSSL is compiled with
9072 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9075 *) Added BIO_vprintf() and BIO_vsnprintf().
9078 *) Added more prefixes to parse for in the the strings written
9079 through a logging bio, to cover all the levels that are available
9080 through syslog. The prefixes are now:
9082 PANIC, EMERG, EMR => LOG_EMERG
9083 ALERT, ALR => LOG_ALERT
9084 CRIT, CRI => LOG_CRIT
9085 ERROR, ERR => LOG_ERR
9086 WARNING, WARN, WAR => LOG_WARNING
9087 NOTICE, NOTE, NOT => LOG_NOTICE
9088 INFO, INF => LOG_INFO
9089 DEBUG, DBG => LOG_DEBUG
9091 and as before, if none of those prefixes are present at the
9092 beginning of the string, LOG_ERR is chosen.
9094 On Win32, the LOG_* levels are mapped according to this:
9096 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9097 LOG_WARNING => EVENTLOG_WARNING_TYPE
9098 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9102 *) Made it possible to reconfigure with just the configuration
9103 argument "reconf" or "reconfigure". The command line arguments
9104 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9105 and are retrieved from there when reconfiguring.
9109 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9111 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9114 *) The obj_dat.pl script was messing up the sorting of object
9115 names. The reason was that it compared the quoted version
9116 of strings as a result "OCSP" > "OCSP Signing" because
9117 " > SPACE. Changed script to store unquoted versions of
9118 names and add quotes on output. It was also omitting some
9119 names from the lookup table if they were given a default
9120 value (that is if SN is missing it is given the same
9121 value as LN and vice versa), these are now added on the
9122 grounds that if an object has a name we should be able to
9123 look it up. Finally added warning output when duplicate
9124 short or long names are found.
9127 *) Changes needed for Tandem NSK.
9128 [Scott Uroff <scott@xypro.com>]
9130 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
9131 RSA_padding_check_SSLv23(), special padding was never detected
9132 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
9133 version rollback attacks was not effective.
9135 In s23_clnt.c, don't use special rollback-attack detection padding
9136 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
9137 client; similarly, in s23_srvr.c, don't do the rollback check if
9138 SSL 2.0 is the only protocol enabled in the server.
9141 *) Make it possible to get hexdumps of unprintable data with 'openssl
9142 asn1parse'. By implication, the functions ASN1_parse_dump() and
9143 BIO_dump_indent() are added.
9146 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
9147 these print out strings and name structures based on various
9148 flags including RFC2253 support and proper handling of
9149 multibyte characters. Added options to the 'x509' utility
9150 to allow the various flags to be set.
9153 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
9154 Also change the functions X509_cmp_current_time() and
9155 X509_gmtime_adj() work with an ASN1_TIME structure,
9156 this will enable certificates using GeneralizedTime in validity
9157 dates to be checked.
9160 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
9161 negative public key encodings) on by default,
9162 NO_NEG_PUBKEY_BUG can be set to disable it.
9165 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
9166 content octets. An i2c_ASN1_OBJECT is unnecessary because
9167 the encoding can be trivially obtained from the structure.
9170 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
9171 not read locks (CRYPTO_r_[un]lock).
9174 *) A first attempt at creating official support for shared
9175 libraries through configuration. I've kept it so the
9176 default is static libraries only, and the OpenSSL programs
9177 are always statically linked for now, but there are
9178 preparations for dynamic linking in place.
9179 This has been tested on Linux and Tru64.
9182 *) Randomness polling function for Win9x, as described in:
9183 Peter Gutmann, Software Generation of Practically Strong
9187 *) Fix so PRNG is seeded in req if using an already existing
9191 *) New options to smime application. -inform and -outform
9192 allow alternative formats for the S/MIME message including
9193 PEM and DER. The -content option allows the content to be
9194 specified separately. This should allow things like Netscape
9195 form signing output easier to verify.
9198 *) Fix the ASN1 encoding of tags using the 'long form'.
9201 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
9202 STRING types. These convert content octets to and from the
9203 underlying type. The actual tag and length octets are
9204 already assumed to have been read in and checked. These
9205 are needed because all other string types have virtually
9206 identical handling apart from the tag. By having versions
9207 of the ASN1 functions that just operate on content octets
9208 IMPLICIT tagging can be handled properly. It also allows
9209 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
9210 and ASN1_INTEGER are identical apart from the tag.
9213 *) Change the handling of OID objects as follows:
9215 - New object identifiers are inserted in objects.txt, following
9216 the syntax given in objects.README.
9217 - objects.pl is used to process obj_mac.num and create a new
9219 - obj_dat.pl is used to create a new obj_dat.h, using the data in
9222 This is currently kind of a hack, and the perl code in objects.pl
9223 isn't very elegant, but it works as I intended. The simplest way
9224 to check that it worked correctly is to look in obj_dat.h and
9225 check the array nid_objs and make sure the objects haven't moved
9226 around (this is important!). Additions are OK, as well as
9227 consistent name changes.
9230 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
9233 *) Addition of the command line parameter '-rand file' to 'openssl req'.
9234 The given file adds to whatever has already been seeded into the
9235 random pool through the RANDFILE configuration file option or
9236 environment variable, or the default random state file.
9239 *) mkstack.pl now sorts each macro group into lexical order.
9240 Previously the output order depended on the order the files
9241 appeared in the directory, resulting in needless rewriting
9245 *) Patches to make OpenSSL compile under Win32 again. Mostly
9246 work arounds for the VC++ problem that it treats func() as
9247 func(void). Also stripped out the parts of mkdef.pl that
9248 added extra typesafe functions: these no longer exist.
9251 *) Reorganisation of the stack code. The macros are now all
9252 collected in safestack.h . Each macro is defined in terms of
9253 a "stack macro" of the form SKM_<name>(type, a, b). The
9254 DEBUG_SAFESTACK is now handled in terms of function casts,
9255 this has the advantage of retaining type safety without the
9256 use of additional functions. If DEBUG_SAFESTACK is not defined
9257 then the non typesafe macros are used instead. Also modified the
9258 mkstack.pl script to handle the new form. Needs testing to see
9259 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
9260 the default if no major problems. Similar behaviour for ASN1_SET_OF
9261 and PKCS12_STACK_OF.
9264 *) When some versions of IIS use the 'NET' form of private key the
9265 key derivation algorithm is different. Normally MD5(password) is
9266 used as a 128 bit RC4 key. In the modified case
9267 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
9268 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
9269 as the old Netscape_RSA functions except they have an additional
9270 'sgckey' parameter which uses the modified algorithm. Also added
9271 an -sgckey command line option to the rsa utility. Thanks to
9272 Adrian Peck <bertie@ncipher.com> for posting details of the modified
9273 algorithm to openssl-dev.
9276 *) The evp_local.h macros were using 'c.##kname' which resulted in
9277 invalid expansion on some systems (SCO 5.0.5 for example).
9278 Corrected to 'c.kname'.
9279 [Phillip Porch <root@theporch.com>]
9281 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
9282 a STACK of email addresses from a certificate or request, these look
9283 in the subject name and the subject alternative name extensions and
9284 omit any duplicate addresses.
9287 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
9288 This makes DSA verification about 2 % faster.
9291 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
9292 (meaning that now 2^5 values will be precomputed, which is only 4 KB
9293 plus overhead for 1024 bit moduli).
9294 This makes exponentiations about 0.5 % faster for 1024 bit
9295 exponents (as measured by "openssl speed rsa2048").
9298 *) Rename memory handling macros to avoid conflicts with other
9300 Malloc => OPENSSL_malloc
9301 Malloc_locked => OPENSSL_malloc_locked
9302 Realloc => OPENSSL_realloc
9303 Free => OPENSSL_free
9306 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
9307 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
9310 *) CygWin32 support.
9311 [John Jarvie <jjarvie@newsguy.com>]
9313 *) The type-safe stack code has been rejigged. It is now only compiled
9314 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
9315 by default all type-specific stack functions are "#define"d back to
9316 standard stack functions. This results in more streamlined output
9317 but retains the type-safety checking possibilities of the original
9321 *) The STACK code has been cleaned up, and certain type declarations
9322 that didn't make a lot of sense have been brought in line. This has
9323 also involved a cleanup of sorts in safestack.h to more correctly
9324 map type-safe stack functions onto their plain stack counterparts.
9325 This work has also resulted in a variety of "const"ifications of
9326 lots of the code, especially "_cmp" operations which should normally
9327 be prototyped with "const" parameters anyway.
9330 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
9331 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
9332 (The PRNG state consists of two parts, the large pool 'state' and 'md',
9333 where all of 'md' is used each time the PRNG is used, but 'state'
9334 is used only indexed by a cyclic counter. As entropy may not be
9335 well distributed from the beginning, 'md' is important as a
9336 chaining variable. However, the output function chains only half
9337 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
9338 all of 'md', and seeding with STATE_SIZE dummy bytes will result
9339 in all of 'state' being rewritten, with the new values depending
9340 on virtually all of 'md'. This overcomes the 80 bit limitation.)
9343 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
9344 the handshake is continued after ssl_verify_cert_chain();
9345 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
9346 can lead to 'unexplainable' connection aborts later.
9347 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
9349 *) Major EVP API cipher revision.
9350 Add hooks for extra EVP features. This allows various cipher
9351 parameters to be set in the EVP interface. Support added for variable
9352 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
9353 setting of RC2 and RC5 parameters.
9355 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
9358 Remove lots of duplicated code from the EVP library. For example *every*
9359 cipher init() function handles the 'iv' in the same way according to the
9360 cipher mode. They also all do nothing if the 'key' parameter is NULL and
9361 for CFB and OFB modes they zero ctx->num.
9363 New functionality allows removal of S/MIME code RC2 hack.
9365 Most of the routines have the same form and so can be declared in terms
9368 By shifting this to the top level EVP_CipherInit() it can be removed from
9369 all individual ciphers. If the cipher wants to handle IVs or keys
9370 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
9373 Change lots of functions like EVP_EncryptUpdate() to now return a
9374 value: although software versions of the algorithms cannot fail
9375 any installed hardware versions can.
9378 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
9379 this option is set, tolerate broken clients that send the negotiated
9380 protocol version number instead of the requested protocol version
9384 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
9385 i.e. non-zero for export ciphersuites, zero otherwise.
9386 Previous versions had this flag inverted, inconsistent with
9387 rsa_tmp_cb (..._TMP_RSA_CB).
9388 [Bodo Moeller; problem reported by Amit Chopra]
9390 *) Add missing DSA library text string. Work around for some IIS
9391 key files with invalid SEQUENCE encoding.
9394 *) Add a document (doc/standards.txt) that list all kinds of standards
9395 and so on that are implemented in OpenSSL.
9398 *) Enhance c_rehash script. Old version would mishandle certificates
9399 with the same subject name hash and wouldn't handle CRLs at all.
9400 Added -fingerprint option to crl utility, to support new c_rehash
9404 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
9407 *) Fix for SSL server purpose checking. Server checking was
9408 rejecting certificates which had extended key usage present
9409 but no ssl client purpose.
9410 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
9412 *) Make PKCS#12 code work with no password. The PKCS#12 spec
9413 is a little unclear about how a blank password is handled.
9414 Since the password in encoded as a BMPString with terminating
9415 double NULL a zero length password would end up as just the
9416 double NULL. However no password at all is different and is
9417 handled differently in the PKCS#12 key generation code. NS
9418 treats a blank password as zero length. MSIE treats it as no
9419 password on export: but it will try both on import. We now do
9420 the same: PKCS12_parse() tries zero length and no password if
9421 the password is set to "" or NULL (NULL is now a valid password:
9422 it wasn't before) as does the pkcs12 application.
9425 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
9426 perror when PEM_read_bio_X509_REQ fails, the error message must
9427 be obtained from the error queue.
9430 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
9431 it in ERR_remove_state if appropriate, and change ERR_get_state
9432 accordingly to avoid race conditions (this is necessary because
9433 thread_hash is no longer constant once set).
9436 *) Bugfix for linux-elf makefile.one.
9439 *) RSA_get_default_method() will now cause a default
9440 RSA_METHOD to be chosen if one doesn't exist already.
9441 Previously this was only set during a call to RSA_new()
9442 or RSA_new_method(NULL) meaning it was possible for
9443 RSA_get_default_method() to return NULL.
9446 *) Added native name translation to the existing DSO code
9447 that will convert (if the flag to do so is set) filenames
9448 that are sufficiently small and have no path information
9449 into a canonical native form. Eg. "blah" converted to
9450 "libblah.so" or "blah.dll" etc.
9453 *) New function ERR_error_string_n(e, buf, len) which is like
9454 ERR_error_string(e, buf), but writes at most 'len' bytes
9455 including the 0 terminator. For ERR_error_string_n, 'buf'
9457 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
9459 *) CONF library reworked to become more general. A new CONF
9460 configuration file reader "class" is implemented as well as a
9461 new functions (NCONF_*, for "New CONF") to handle it. The now
9462 old CONF_* functions are still there, but are reimplemented to
9463 work in terms of the new functions. Also, a set of functions
9464 to handle the internal storage of the configuration data is
9465 provided to make it easier to write new configuration file
9466 reader "classes" (I can definitely see something reading a
9467 configuration file in XML format, for example), called _CONF_*,
9468 or "the configuration storage API"...
9470 The new configuration file reading functions are:
9472 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
9473 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
9475 NCONF_default, NCONF_WIN32
9477 NCONF_dump_fp, NCONF_dump_bio
9479 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
9480 NCONF_new creates a new CONF object. This works in the same way
9481 as other interfaces in OpenSSL, like the BIO interface.
9482 NCONF_dump_* dump the internal storage of the configuration file,
9483 which is useful for debugging. All other functions take the same
9484 arguments as the old CONF_* functions wth the exception of the
9485 first that must be a `CONF *' instead of a `LHASH *'.
9487 To make it easer to use the new classes with the old CONF_* functions,
9488 the function CONF_set_default_method is provided.
9491 *) Add '-tls1' option to 'openssl ciphers', which was already
9492 mentioned in the documentation but had not been implemented.
9493 (This option is not yet really useful because even the additional
9494 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
9497 *) Initial DSO code added into libcrypto for letting OpenSSL (and
9498 OpenSSL-based applications) load shared libraries and bind to
9499 them in a portable way.
9500 [Geoff Thorpe, with contributions from Richard Levitte]
9502 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
9504 *) Make sure _lrotl and _lrotr are only used with MSVC.
9506 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
9507 (the default implementation of RAND_status).
9509 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
9510 to '-clrext' (= clear extensions), as intended and documented.
9511 [Bodo Moeller; inconsistency pointed out by Michael Attili
9514 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
9515 was larger than the MD block size.
9516 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
9518 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
9519 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
9520 using the passed key: if the passed key was a private key the result
9521 of X509_print(), for example, would be to print out all the private key
9525 *) des_quad_cksum() byte order bug fix.
9526 [Ulf Möller, using the problem description in krb4-0.9.7, where
9527 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
9529 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
9531 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
9533 *) For easily testing in shell scripts whether some command
9534 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
9535 returns with exit code 0 iff no command of the given name is available.
9536 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
9537 the output goes to stdout and nothing is printed to stderr.
9538 Additional arguments are always ignored.
9540 Since for each cipher there is a command of the same name,
9541 the 'no-cipher' compilation switches can be tested this way.
9543 ('openssl no-XXX' is not able to detect pseudo-commands such
9544 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
9547 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
9550 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
9551 is set; it will be thrown away anyway because each handshake creates
9553 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
9554 to parameters -- in previous versions (since OpenSSL 0.9.3) the
9555 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
9556 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
9559 *) New s_client option -ign_eof: EOF at stdin is ignored, and
9560 'Q' and 'R' lose their special meanings (quit/renegotiate).
9561 This is part of what -quiet does; unlike -quiet, -ign_eof
9562 does not suppress any output.
9565 *) Add compatibility options to the purpose and trust code. The
9566 purpose X509_PURPOSE_ANY is "any purpose" which automatically
9567 accepts a certificate or CA, this was the previous behaviour,
9568 with all the associated security issues.
9570 X509_TRUST_COMPAT is the old trust behaviour: only and
9571 automatically trust self signed roots in certificate store. A
9572 new trust setting X509_TRUST_DEFAULT is used to specify that
9573 a purpose has no associated trust setting and it should instead
9574 use the value in the default purpose.
9577 *) Fix the PKCS#8 DSA private key code so it decodes keys again
9578 and fix a memory leak.
9581 *) In util/mkerr.pl (which implements 'make errors'), preserve
9582 reason strings from the previous version of the .c file, as
9583 the default to have only downcase letters (and digits) in
9584 automatically generated reasons codes is not always appropriate.
9587 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
9588 using strerror. Previously, ERR_reason_error_string() returned
9589 library names as reason strings for SYSerr; but SYSerr is a special
9590 case where small numbers are errno values, not library numbers.
9593 *) Add '-dsaparam' option to 'openssl dhparam' application. This
9594 converts DSA parameters into DH parameters. (When creating parameters,
9595 DSA_generate_parameters is used.)
9598 *) Include 'length' (recommended exponent length) in C code generated
9599 by 'openssl dhparam -C'.
9602 *) The second argument to set_label in perlasm was already being used
9603 so couldn't be used as a "file scope" flag. Moved to third argument
9607 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
9608 instead of RAND_bytes for encryption IVs and salts.
9611 *) Include RAND_status() into RAND_METHOD instead of implementing
9612 it only for md_rand.c Otherwise replacing the PRNG by calling
9613 RAND_set_rand_method would be impossible.
9616 *) Don't let DSA_generate_key() enter an infinite loop if the random
9617 number generation fails.
9620 *) New 'rand' application for creating pseudo-random output.
9623 *) Added configuration support for Linux/IA64
9624 [Rolf Haberrecker <rolf@suse.de>]
9626 *) Assembler module support for Mingw32.
9629 *) Shared library support for HPUX (in shlib/).
9630 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
9632 *) Shared library support for Solaris gcc.
9633 [Lutz Behnke <behnke@trustcenter.de>]
9635 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
9637 *) PKCS7_encrypt() was adding text MIME headers twice because they
9638 were added manually and by SMIME_crlf_copy().
9641 *) In bntest.c don't call BN_rand with zero bits argument.
9642 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
9644 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
9645 case was implemented. This caused BN_div_recp() to fail occasionally.
9648 *) Add an optional second argument to the set_label() in the perl
9649 assembly language builder. If this argument exists and is set
9650 to 1 it signals that the assembler should use a symbol whose
9651 scope is the entire file, not just the current function. This
9652 is needed with MASM which uses the format label:: for this scope.
9653 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
9655 *) Change the ASN1 types so they are typedefs by default. Before
9656 almost all types were #define'd to ASN1_STRING which was causing
9657 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
9661 *) Change names of new functions to the new get1/get0 naming
9662 convention: After 'get1', the caller owns a reference count
9663 and has to call ..._free; 'get0' returns a pointer to some
9664 data structure without incrementing reference counters.
9665 (Some of the existing 'get' functions increment a reference
9666 counter, some don't.)
9667 Similarly, 'set1' and 'add1' functions increase reference
9668 counters or duplicate objects.
9671 *) Allow for the possibility of temp RSA key generation failure:
9672 the code used to assume it always worked and crashed on failure.
9675 *) Fix potential buffer overrun problem in BIO_printf().
9676 [Ulf Möller, using public domain code by Patrick Powell; problem
9677 pointed out by David Sacerdote <das33@cornell.edu>]
9679 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
9680 RAND_egd() and RAND_status(). In the command line application,
9681 the EGD socket can be specified like a seed file using RANDFILE
9685 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
9686 Some CAs (e.g. Verisign) distribute certificates in this form.
9689 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
9690 list to exclude them. This means that no special compilation option
9691 is needed to use anonymous DH: it just needs to be included in the
9695 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
9696 EVP_MD_type. The old functionality is available in a new macro called
9697 EVP_MD_md(). Change code that uses it and update docs.
9700 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
9701 where the 'void *' argument is replaced by a function pointer argument.
9702 Previously 'void *' was abused to point to functions, which works on
9703 many platforms, but is not correct. As these functions are usually
9704 called by macros defined in OpenSSL header files, most source code
9705 should work without changes.
9708 *) <openssl/opensslconf.h> (which is created by Configure) now contains
9709 sections with information on -D... compiler switches used for
9710 compiling the library so that applications can see them. To enable
9711 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
9712 must be defined. E.g.,
9713 #define OPENSSL_ALGORITHM_DEFINES
9714 #include <openssl/opensslconf.h>
9715 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
9716 [Richard Levitte, Ulf and Bodo Möller]
9718 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
9722 *) Change the 'other' type in certificate aux info to a STACK_OF
9723 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
9724 the required ASN1 format: arbitrary types determined by an OID.
9727 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
9728 argument to 'req'. This is not because the function is newer or
9729 better than others it just uses the work 'NEW' in the certificate
9730 request header lines. Some software needs this.
9733 *) Reorganise password command line arguments: now passwords can be
9734 obtained from various sources. Delete the PEM_cb function and make
9735 it the default behaviour: i.e. if the callback is NULL and the
9736 usrdata argument is not NULL interpret it as a null terminated pass
9737 phrase. If usrdata and the callback are NULL then the pass phrase
9738 is prompted for as usual.
9741 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
9742 the support is automatically enabled. The resulting binaries will
9743 autodetect the card and use it if present.
9744 [Ben Laurie and Compaq Inc.]
9746 *) Work around for Netscape hang bug. This sends certificate request
9747 and server done in one record. Since this is perfectly legal in the
9748 SSL/TLS protocol it isn't a "bug" option and is on by default. See
9749 the bugs/SSLv3 entry for more info.
9752 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
9755 *) Add -rand argument to smime and pkcs12 applications and read/write
9759 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
9762 *) Add command line password options to the remaining applications.
9765 *) Bug fix for BN_div_recp() for numerators with an even number of
9769 *) More tests in bntest.c, and changed test_bn output.
9772 *) ./config recognizes MacOS X now.
9775 *) Bug fix for BN_div() when the first words of num and divsor are
9776 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
9779 *) Add support for various broken PKCS#8 formats, and command line
9780 options to produce them.
9783 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
9784 get temporary BIGNUMs from a BN_CTX.
9787 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
9791 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
9792 include a #define from the old name to the new. The original intent
9793 was that statically linked binaries could for example just call
9794 SSLeay_add_all_ciphers() to just add ciphers to the table and not
9795 link with digests. This never worked because SSLeay_add_all_digests()
9796 and SSLeay_add_all_ciphers() were in the same source file so calling
9797 one would link with the other. They are now in separate source files.
9800 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
9803 *) Use a less unusual form of the Miller-Rabin primality test (it used
9804 a binary algorithm for exponentiation integrated into the Miller-Rabin
9805 loop, our standard modexp algorithms are faster).
9808 *) Support for the EBCDIC character set completed.
9809 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
9811 *) Source code cleanups: use const where appropriate, eliminate casts,
9812 use void * instead of char * in lhash.
9815 *) Bugfix: ssl3_send_server_key_exchange was not restartable
9816 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
9817 this the server could overwrite ephemeral keys that the client
9821 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
9822 using 50 iterations of the Rabin-Miller test.
9824 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
9825 iterations of the Rabin-Miller test as required by the appendix
9826 to FIPS PUB 186[-1]) instead of DSA_is_prime.
9827 As BN_is_prime_fasttest includes trial division, DSA parameter
9828 generation becomes much faster.
9830 This implies a change for the callback functions in DSA_is_prime
9831 and DSA_generate_parameters: The callback function is called once
9832 for each positive witness in the Rabin-Miller test, not just
9833 occasionally in the inner loop; and the parameters to the
9834 callback function now provide an iteration count for the outer
9835 loop rather than for the current invocation of the inner loop.
9836 DSA_generate_parameters additionally can call the callback
9837 function with an 'iteration count' of -1, meaning that a
9838 candidate has passed the trial division test (when q is generated
9839 from an application-provided seed, trial division is skipped).
9842 *) New function BN_is_prime_fasttest that optionally does trial
9843 division before starting the Rabin-Miller test and has
9844 an additional BN_CTX * argument (whereas BN_is_prime always
9845 has to allocate at least one BN_CTX).
9846 'callback(1, -1, cb_arg)' is called when a number has passed the
9847 trial division stage.
9850 *) Fix for bug in CRL encoding. The validity dates weren't being handled
9854 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
9857 *) New function BN_pseudo_rand().
9860 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
9861 bignum version of BN_from_montgomery() with the working code from
9862 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
9866 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
9867 made it impossible to use the same SSL_SESSION data structure in
9868 SSL2 clients in multiple threads.
9871 *) The return value of RAND_load_file() no longer counts bytes obtained
9872 by stat(). RAND_load_file(..., -1) is new and uses the complete file
9873 to seed the PRNG (previously an explicit byte count was required).
9874 [Ulf Möller, Bodo Möller]
9876 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
9877 used (char *) instead of (void *) and had casts all over the place.
9880 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
9883 *) Retain source code compatibility for BN_prime_checks macro:
9884 BN_is_prime(..., BN_prime_checks, ...) now uses
9885 BN_prime_checks_for_size to determine the appropriate number of
9886 Rabin-Miller iterations.
9889 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
9890 DH_CHECK_P_NOT_SAFE_PRIME.
9891 (Check if this is true? OpenPGP calls them "strong".)
9894 *) Merge the functionality of "dh" and "gendh" programs into a new program
9895 "dhparam". The old programs are retained for now but will handle DH keys
9896 (instead of parameters) in future.
9899 *) Make the ciphers, s_server and s_client programs check the return values
9900 when a new cipher list is set.
9903 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
9904 ciphers. Before when the 56bit ciphers were enabled the sorting was
9907 The syntax for the cipher sorting has been extended to support sorting by
9908 cipher-strength (using the strength_bits hard coded in the tables).
9909 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
9911 Fix a bug in the cipher-command parser: when supplying a cipher command
9912 string with an "undefined" symbol (neither command nor alphanumeric
9913 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
9914 an error is flagged.
9916 Due to the strength-sorting extension, the code of the
9917 ssl_create_cipher_list() function was completely rearranged. I hope that
9918 the readability was also increased :-)
9919 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9921 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
9922 for the first serial number and places 2 in the serial number file. This
9923 avoids problems when the root CA is created with serial number zero and
9924 the first user certificate has the same issuer name and serial number
9928 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
9929 the new code. Add documentation for this stuff.
9932 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
9933 X509_*() to X509at_*() on the grounds that they don't handle X509
9934 structures and behave in an analogous way to the X509v3 functions:
9935 they shouldn't be called directly but wrapper functions should be used
9938 So we also now have some wrapper functions that call the X509at functions
9939 when passed certificate requests. (TO DO: similar things can be done with
9940 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
9941 things. Some of these need some d2i or i2d and print functionality
9942 because they handle more complex structures.)
9945 *) Add missing #ifndefs that caused missing symbols when building libssl
9946 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
9947 NO_RSA in ssl/s2*.c.
9948 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
9950 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
9951 has a return value which indicates the quality of the random data
9952 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
9953 error queue. New function RAND_pseudo_bytes() generates output that is
9954 guaranteed to be unique but not unpredictable. RAND_add is like
9955 RAND_seed, but takes an extra argument for an entropy estimate
9956 (RAND_seed always assumes full entropy).
9959 *) Do more iterations of Rabin-Miller probable prime test (specifically,
9960 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
9961 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
9962 in crypto/bn/bn_prime.c for the complete table). This guarantees a
9963 false-positive rate of at most 2^-80 for random input.
9966 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
9969 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
9970 in the 0.9.5 release), this returns the chain
9971 from an X509_CTX structure with a dup of the stack and all
9972 the X509 reference counts upped: so the stack will exist
9973 after X509_CTX_cleanup() has been called. Modify pkcs12.c
9976 Also make SSL_SESSION_print() print out the verify return
9980 *) Add manpage for the pkcs12 command. Also change the default
9981 behaviour so MAC iteration counts are used unless the new
9982 -nomaciter option is used. This improves file security and
9983 only older versions of MSIE (4.0 for example) need it.
9986 *) Honor the no-xxx Configure options when creating .DEF files.
9989 *) Add PKCS#10 attributes to field table: challengePassword,
9990 unstructuredName and unstructuredAddress. These are taken from
9991 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
9992 international characters are used.
9994 More changes to X509_ATTRIBUTE code: allow the setting of types
9995 based on strings. Remove the 'loc' parameter when adding
9996 attributes because these will be a SET OF encoding which is sorted
10000 *) Initial changes to the 'req' utility to allow request generation
10001 automation. This will allow an application to just generate a template
10002 file containing all the field values and have req construct the
10005 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10006 used all over the place including certificate requests and PKCS#7
10007 structures. They are currently handled manually where necessary with
10008 some primitive wrappers for PKCS#7. The new functions behave in a
10009 manner analogous to the X509 extension functions: they allow
10010 attributes to be looked up by NID and added.
10012 Later something similar to the X509V3 code would be desirable to
10013 automatically handle the encoding, decoding and printing of the
10014 more complex types. The string types like challengePassword can
10015 be handled by the string table functions.
10017 Also modified the multi byte string table handling. Now there is
10018 a 'global mask' which masks out certain types. The table itself
10019 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10020 is useful when for example there is only one permissible type
10021 (as in countryName) and using the mask might result in no valid
10025 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10026 SSL_get_peer_finished to allow applications to obtain the latest
10027 Finished messages sent to the peer or expected from the peer,
10028 respectively. (SSL_get_peer_finished is usually the Finished message
10029 actually received from the peer, otherwise the protocol will be aborted.)
10031 As the Finished message are message digests of the complete handshake
10032 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10033 be used for external authentication procedures when the authentication
10034 provided by SSL/TLS is not desired or is not enough.
10037 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10038 the host supports BWX extension and if Compaq C is present on the
10039 $PATH. Just exploiting of the BWX extension results in 20-30%
10040 performance kick for some algorithms, e.g. DES and RC4 to mention
10041 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10045 *) Add support for MS "fast SGC". This is arguably a violation of the
10046 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10047 weak crypto and after checking the certificate is SGC a second one
10048 with strong crypto. MS SGC stops the first handshake after receiving
10049 the server certificate message and sends a second client hello. Since
10050 a server will typically do all the time consuming operations before
10051 expecting any further messages from the client (server key exchange
10052 is the most expensive) there is little difference between the two.
10054 To get OpenSSL to support MS SGC we have to permit a second client
10055 hello message after we have sent server done. In addition we have to
10056 reset the MAC if we do get this second client hello.
10059 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10060 if a DER encoded private key is RSA or DSA traditional format. Changed
10061 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10062 format DER encoded private key. Newer code should use PKCS#8 format which
10063 has the key type encoded in the ASN1 structure. Added DER private key
10064 support to pkcs8 application.
10067 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10068 ciphersuites has been selected (as required by the SSL 3/TLS 1
10069 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10070 is set, we interpret this as a request to violate the specification
10071 (the worst that can happen is a handshake failure, and 'correct'
10072 behaviour would result in a handshake failure anyway).
10075 *) In SSL_CTX_add_session, take into account that there might be multiple
10076 SSL_SESSION structures with the same session ID (e.g. when two threads
10077 concurrently obtain them from an external cache).
10078 The internal cache can handle only one SSL_SESSION with a given ID,
10079 so if there's a conflict, we now throw out the old one to achieve
10083 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10084 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10085 some routines that use cipher OIDs: some ciphers do not have OIDs
10086 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10090 *) Simplify the trust setting structure and code. Now we just have
10091 two sequences of OIDs for trusted and rejected settings. These will
10092 typically have values the same as the extended key usage extension
10093 and any application specific purposes.
10095 The trust checking code now has a default behaviour: it will just
10096 check for an object with the same NID as the passed id. Functions can
10097 be provided to override either the default behaviour or the behaviour
10098 for a given id. SSL client, server and email already have functions
10099 in place for compatibility: they check the NID and also return "trusted"
10100 if the certificate is self signed.
10103 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10104 traditional format into an EVP_PKEY structure.
10107 *) Add a password callback function PEM_cb() which either prompts for
10108 a password if usr_data is NULL or otherwise assumes it is a null
10109 terminated password. Allow passwords to be passed on command line
10110 environment or config files in a few more utilities.
10113 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
10114 keys. Add some short names for PKCS#8 PBE algorithms and allow them
10115 to be specified on the command line for the pkcs8 and pkcs12 utilities.
10116 Update documentation.
10119 *) Support for ASN1 "NULL" type. This could be handled before by using
10120 ASN1_TYPE but there wasn't any function that would try to read a NULL
10121 and produce an error if it couldn't. For compatibility we also have
10122 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
10123 don't allocate anything because they don't need to.
10126 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
10128 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
10130 *) Rebuild of the memory allocation routines used by OpenSSL code and
10131 possibly others as well. The purpose is to make an interface that
10132 provide hooks so anyone can build a separate set of allocation and
10133 deallocation routines to be used by OpenSSL, for example memory
10134 pool implementations, or something else, which was previously hard
10135 since Malloc(), Realloc() and Free() were defined as macros having
10136 the values malloc, realloc and free, respectively (except for Win32
10137 compilations). The same is provided for memory debugging code.
10138 OpenSSL already comes with functionality to find memory leaks, but
10139 this gives people a chance to debug other memory problems.
10141 With these changes, a new set of functions and macros have appeared:
10143 CRYPTO_set_mem_debug_functions() [F]
10144 CRYPTO_get_mem_debug_functions() [F]
10145 CRYPTO_dbg_set_options() [F]
10146 CRYPTO_dbg_get_options() [F]
10147 CRYPTO_malloc_debug_init() [M]
10149 The memory debug functions are NULL by default, unless the library
10150 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
10151 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
10152 gives the standard debugging functions that come with OpenSSL) or
10153 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
10154 provided by the library user) must be used. When the standard
10155 debugging functions are used, CRYPTO_dbg_set_options can be used to
10156 request additional information:
10157 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
10158 the CRYPTO_MDEBUG_xxx macro when compiling the library.
10160 Also, things like CRYPTO_set_mem_functions will always give the
10161 expected result (the new set of functions is used for allocation
10162 and deallocation) at all times, regardless of platform and compiler
10165 To finish it up, some functions that were never use in any other
10166 way than through macros have a new API and new semantic:
10168 CRYPTO_dbg_malloc()
10169 CRYPTO_dbg_realloc()
10172 All macros of value have retained their old syntax.
10173 [Richard Levitte and Bodo Moeller]
10175 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
10176 ordering of SMIMECapabilities wasn't in "strength order" and there
10177 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
10181 *) Some ASN1 types with illegal zero length encoding (INTEGER,
10182 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
10183 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
10185 *) Merge in my S/MIME library for OpenSSL. This provides a simple
10186 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
10187 functionality to handle multipart/signed properly) and a utility
10188 called 'smime' to call all this stuff. This is based on code I
10189 originally wrote for Celo who have kindly allowed it to be
10190 included in OpenSSL.
10193 *) Add variants des_set_key_checked and des_set_key_unchecked of
10194 des_set_key (aka des_key_sched). Global variable des_check_key
10195 decides which of these is called by des_set_key; this way
10196 des_check_key behaves as it always did, but applications and
10197 the library itself, which was buggy for des_check_key == 1,
10198 have a cleaner way to pick the version they need.
10201 *) New function PKCS12_newpass() which changes the password of a
10205 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
10206 dynamic mix. In both cases the ids can be used as an index into the
10207 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
10208 functions so they accept a list of the field values and the
10209 application doesn't need to directly manipulate the X509_TRUST
10213 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
10217 *) Modify the way the V3 extension code looks up extensions. This now
10218 works in a similar way to the object code: we have some "standard"
10219 extensions in a static table which is searched with OBJ_bsearch()
10220 and the application can add dynamic ones if needed. The file
10221 crypto/x509v3/ext_dat.h now has the info: this file needs to be
10222 updated whenever a new extension is added to the core code and kept
10223 in ext_nid order. There is a simple program 'tabtest.c' which checks
10224 this. New extensions are not added too often so this file can readily
10225 be maintained manually.
10227 There are two big advantages in doing things this way. The extensions
10228 can be looked up immediately and no longer need to be "added" using
10229 X509V3_add_standard_extensions(): this function now does nothing.
10230 [Side note: I get *lots* of email saying the extension code doesn't
10231 work because people forget to call this function]
10232 Also no dynamic allocation is done unless new extensions are added:
10233 so if we don't add custom extensions there is no need to call
10234 X509V3_EXT_cleanup().
10237 *) Modify enc utility's salting as follows: make salting the default. Add a
10238 magic header, so unsalted files fail gracefully instead of just decrypting
10239 to garbage. This is because not salting is a big security hole, so people
10240 should be discouraged from doing it.
10243 *) Fixes and enhancements to the 'x509' utility. It allowed a message
10244 digest to be passed on the command line but it only used this
10245 parameter when signing a certificate. Modified so all relevant
10246 operations are affected by the digest parameter including the
10247 -fingerprint and -x509toreq options. Also -x509toreq choked if a
10248 DSA key was used because it didn't fix the digest.
10251 *) Initial certificate chain verify code. Currently tests the untrusted
10252 certificates for consistency with the verify purpose (which is set
10253 when the X509_STORE_CTX structure is set up) and checks the pathlength.
10255 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
10256 this is because it will reject chains with invalid extensions whereas
10257 every previous version of OpenSSL and SSLeay made no checks at all.
10259 Trust code: checks the root CA for the relevant trust settings. Trust
10260 settings have an initial value consistent with the verify purpose: e.g.
10261 if the verify purpose is for SSL client use it expects the CA to be
10262 trusted for SSL client use. However the default value can be changed to
10263 permit custom trust settings: one example of this would be to only trust
10264 certificates from a specific "secure" set of CAs.
10266 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
10267 which should be used for version portability: especially since the
10268 verify structure is likely to change more often now.
10270 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
10271 to set them. If not set then assume SSL clients will verify SSL servers
10274 Two new options to the verify program: -untrusted allows a set of
10275 untrusted certificates to be passed in and -purpose which sets the
10276 intended purpose of the certificate. If a purpose is set then the
10277 new chain verify code is used to check extension consistency.
10280 *) Support for the authority information access extension.
10283 *) Modify RSA and DSA PEM read routines to transparently handle
10284 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
10285 public keys in a format compatible with certificate
10286 SubjectPublicKeyInfo structures. Unfortunately there were already
10287 functions called *_PublicKey_* which used various odd formats so
10288 these are retained for compatibility: however the DSA variants were
10289 never in a public release so they have been deleted. Changed dsa/rsa
10290 utilities to handle the new format: note no releases ever handled public
10291 keys so we should be OK.
10293 The primary motivation for this change is to avoid the same fiasco
10294 that dogs private keys: there are several incompatible private key
10295 formats some of which are standard and some OpenSSL specific and
10296 require various evil hacks to allow partial transparent handling and
10297 even then it doesn't work with DER formats. Given the option anything
10298 other than PKCS#8 should be dumped: but the other formats have to
10299 stay in the name of compatibility.
10301 With public keys and the benefit of hindsight one standard format
10302 is used which works with EVP_PKEY, RSA or DSA structures: though
10303 it clearly returns an error if you try to read the wrong kind of key.
10305 Added a -pubkey option to the 'x509' utility to output the public key.
10306 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
10307 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
10308 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
10309 that do the same as the EVP_PKEY_assign_*() except they up the
10310 reference count of the added key (they don't "swallow" the
10314 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
10315 CRLs would fail if the file contained no certificates or no CRLs:
10316 added a new function to read in both types and return the number
10317 read: this means that if none are read it will be an error. The
10318 DER versions of the certificate and CRL reader would always fail
10319 because it isn't possible to mix certificates and CRLs in DER format
10320 without choking one or the other routine. Changed this to just read
10321 a certificate: this is the best we can do. Also modified the code
10322 in apps/verify.c to take notice of return codes: it was previously
10323 attempting to read in certificates from NULL pointers and ignoring
10324 any errors: this is one reason why the cert and CRL reader seemed
10325 to work. It doesn't check return codes from the default certificate
10326 routines: these may well fail if the certificates aren't installed.
10329 *) Code to support otherName option in GeneralName.
10332 *) First update to verify code. Change the verify utility
10333 so it warns if it is passed a self signed certificate:
10334 for consistency with the normal behaviour. X509_verify
10335 has been modified to it will now verify a self signed
10336 certificate if *exactly* the same certificate appears
10337 in the store: it was previously impossible to trust a
10338 single self signed certificate. This means that:
10339 openssl verify ss.pem
10340 now gives a warning about a self signed certificate but
10341 openssl verify -CAfile ss.pem ss.pem
10345 *) For servers, store verify_result in SSL_SESSION data structure
10346 (and add it to external session representation).
10347 This is needed when client certificate verifications fails,
10348 but an application-provided verification callback (set by
10349 SSL_CTX_set_cert_verify_callback) allows accepting the session
10350 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
10351 but returns 1): When the session is reused, we have to set
10352 ssl->verify_result to the appropriate error code to avoid
10354 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
10356 *) Fix a bug in the new PKCS#7 code: it didn't consider the
10357 case in PKCS7_dataInit() where the signed PKCS7 structure
10358 didn't contain any existing data because it was being created.
10359 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
10361 *) Add a salt to the key derivation routines in enc.c. This
10362 forms the first 8 bytes of the encrypted file. Also add a
10363 -S option to allow a salt to be input on the command line.
10366 *) New function X509_cmp(). Oddly enough there wasn't a function
10367 to compare two certificates. We do this by working out the SHA1
10368 hash and comparing that. X509_cmp() will be needed by the trust
10372 *) SSL_get1_session() is like SSL_get_session(), but increments
10373 the reference count in the SSL_SESSION returned.
10374 [Geoff Thorpe <geoff@eu.c2.net>]
10376 *) Fix for 'req': it was adding a null to request attributes.
10377 Also change the X509_LOOKUP and X509_INFO code to handle
10378 certificate auxiliary information.
10381 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
10385 *) Add the possibility to add extra information to the memory leak
10386 detecting output, to form tracebacks, showing from where each
10387 allocation was originated: CRYPTO_push_info("constant string") adds
10388 the string plus current file name and line number to a per-thread
10389 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
10390 is like calling CYRPTO_pop_info() until the stack is empty.
10391 Also updated memory leak detection code to be multi-thread-safe.
10394 *) Add options -text and -noout to pkcs7 utility and delete the
10395 encryption options which never did anything. Update docs.
10398 *) Add options to some of the utilities to allow the pass phrase
10399 to be included on either the command line (not recommended on
10400 OSes like Unix) or read from the environment. Update the
10401 manpages and fix a few bugs.
10404 *) Add a few manpages for some of the openssl commands.
10407 *) Fix the -revoke option in ca. It was freeing up memory twice,
10408 leaking and not finding already revoked certificates.
10411 *) Extensive changes to support certificate auxiliary information.
10412 This involves the use of X509_CERT_AUX structure and X509_AUX
10413 functions. An X509_AUX function such as PEM_read_X509_AUX()
10414 can still read in a certificate file in the usual way but it
10415 will also read in any additional "auxiliary information". By
10416 doing things this way a fair degree of compatibility can be
10417 retained: existing certificates can have this information added
10418 using the new 'x509' options.
10420 Current auxiliary information includes an "alias" and some trust
10421 settings. The trust settings will ultimately be used in enhanced
10422 certificate chain verification routines: currently a certificate
10423 can only be trusted if it is self signed and then it is trusted
10427 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
10428 The problem was that one of the replacement routines had not been working
10429 since SSLeay releases. For now the offending routine has been replaced
10430 with non-optimised assembler. Even so, this now gives around 95%
10431 performance improvement for 1024 bit RSA signs.
10434 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
10435 handling. Most clients have the effective key size in bits equal to
10436 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
10437 A few however don't do this and instead use the size of the decrypted key
10438 to determine the RC2 key length and the AlgorithmIdentifier to determine
10439 the effective key length. In this case the effective key length can still
10440 be 40 bits but the key length can be 168 bits for example. This is fixed
10441 by manually forcing an RC2 key into the EVP_PKEY structure because the
10442 EVP code can't currently handle unusual RC2 key sizes: it always assumes
10443 the key length and effective key length are equal.
10446 *) Add a bunch of functions that should simplify the creation of
10447 X509_NAME structures. Now you should be able to do:
10448 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
10449 and have it automatically work out the correct field type and fill in
10450 the structures. The more adventurous can try:
10451 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
10452 and it will (hopefully) work out the correct multibyte encoding.
10455 *) Change the 'req' utility to use the new field handling and multibyte
10456 copy routines. Before the DN field creation was handled in an ad hoc
10457 way in req, ca, and x509 which was rather broken and didn't support
10458 BMPStrings or UTF8Strings. Since some software doesn't implement
10459 BMPStrings or UTF8Strings yet, they can be enabled using the config file
10460 using the dirstring_type option. See the new comment in the default
10461 openssl.cnf for more info.
10464 *) Make crypto/rand/md_rand.c more robust:
10465 - Assure unique random numbers after fork().
10466 - Make sure that concurrent threads access the global counter and
10467 md serializably so that we never lose entropy in them
10468 or use exactly the same state in multiple threads.
10469 Access to the large state is not always serializable because
10470 the additional locking could be a performance killer, and
10471 md should be large enough anyway.
10474 *) New file apps/app_rand.c with commonly needed functionality
10475 for handling the random seed file.
10477 Use the random seed file in some applications that previously did not:
10479 dsaparam -genkey (which also ignored its '-rand' option),
10482 x509 (when signing).
10483 Except on systems with /dev/urandom, it is crucial to have a random
10484 seed file at least for key creation, DSA signing, and for DH exchanges;
10485 for RSA signatures we could do without one.
10487 gendh and gendsa (unlike genrsa) used to read only the first byte
10488 of each file listed in the '-rand' option. The function as previously
10489 found in genrsa is now in app_rand.c and is used by all programs
10490 that support '-rand'.
10493 *) In RAND_write_file, use mode 0600 for creating files;
10494 don't just chmod when it may be too late.
10497 *) Report an error from X509_STORE_load_locations
10498 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
10501 *) New function ASN1_mbstring_copy() this copies a string in either
10502 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
10503 into an ASN1_STRING type. A mask of permissible types is passed
10504 and it chooses the "minimal" type to use or an error if not type
10508 *) Add function equivalents to the various macros in asn1.h. The old
10509 macros are retained with an M_ prefix. Code inside the library can
10510 use the M_ macros. External code (including the openssl utility)
10511 should *NOT* in order to be "shared library friendly".
10514 *) Add various functions that can check a certificate's extensions
10515 to see if it usable for various purposes such as SSL client,
10516 server or S/MIME and CAs of these types. This is currently
10517 VERY EXPERIMENTAL but will ultimately be used for certificate chain
10518 verification. Also added a -purpose flag to x509 utility to
10519 print out all the purposes.
10522 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
10526 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
10527 for, obtain and decode and extension and obtain its critical flag.
10528 This allows all the necessary extension code to be handled in a
10529 single function call.
10532 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
10533 platforms. See crypto/rc4/rc4_enc.c for further details.
10536 *) New -noout option to asn1parse. This causes no output to be produced
10537 its main use is when combined with -strparse and -out to extract data
10538 from a file (which may not be in ASN.1 format).
10541 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
10542 when producing the local key id.
10543 [Richard Levitte <levitte@stacken.kth.se>]
10545 *) New option -dhparam in s_server. This allows a DH parameter file to be
10546 stated explicitly. If it is not stated then it tries the first server
10547 certificate file. The previous behaviour hard coded the filename
10551 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
10552 a public key to be input or output. For example:
10553 openssl rsa -in key.pem -pubout -out pubkey.pem
10554 Also added necessary DSA public key functions to handle this.
10557 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
10558 in the message. This was handled by allowing
10559 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
10560 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
10562 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
10563 to the end of the strings whereas this didn't. This would cause problems
10564 if strings read with d2i_ASN1_bytes() were later modified.
10565 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
10567 *) Fix for base64 decode bug. When a base64 bio reads only one line of
10568 data and it contains EOF it will end up returning an error. This is
10569 caused by input 46 bytes long. The cause is due to the way base64
10570 BIOs find the start of base64 encoded data. They do this by trying a
10571 trial decode on each line until they find one that works. When they
10572 do a flag is set and it starts again knowing it can pass all the
10573 data directly through the decoder. Unfortunately it doesn't reset
10574 the context it uses. This means that if EOF is reached an attempt
10575 is made to pass two EOFs through the context and this causes the
10576 resulting error. This can also cause other problems as well. As is
10577 usual with these problems it takes *ages* to find and the fix is
10578 trivial: move one line.
10579 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
10581 *) Ugly workaround to get s_client and s_server working under Windows. The
10582 old code wouldn't work because it needed to select() on sockets and the
10583 tty (for keypresses and to see if data could be written). Win32 only
10584 supports select() on sockets so we select() with a 1s timeout on the
10585 sockets and then see if any characters are waiting to be read, if none
10586 are present then we retry, we also assume we can always write data to
10587 the tty. This isn't nice because the code then blocks until we've
10588 received a complete line of data and it is effectively polling the
10589 keyboard at 1s intervals: however it's quite a bit better than not
10590 working at all :-) A dedicated Windows application might handle this
10591 with an event loop for example.
10594 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
10595 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
10596 will be called when RSA_sign() and RSA_verify() are used. This is useful
10597 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
10598 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
10599 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
10600 This necessitated the support of an extra signature type NID_md5_sha1
10601 for SSL signatures and modifications to the SSL library to use it instead
10602 of calling RSA_public_decrypt() and RSA_private_encrypt().
10605 *) Add new -verify -CAfile and -CApath options to the crl program, these
10606 will lookup a CRL issuers certificate and verify the signature in a
10607 similar way to the verify program. Tidy up the crl program so it
10608 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
10609 less strict. It will now permit CRL extensions even if it is not
10610 a V2 CRL: this will allow it to tolerate some broken CRLs.
10613 *) Initialize all non-automatic variables each time one of the openssl
10614 sub-programs is started (this is necessary as they may be started
10615 multiple times from the "OpenSSL>" prompt).
10616 [Lennart Bang, Bodo Moeller]
10618 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
10619 removing all other RSA functionality (this is what NO_RSA does). This
10620 is so (for example) those in the US can disable those operations covered
10621 by the RSA patent while allowing storage and parsing of RSA keys and RSA
10625 *) Non-copying interface to BIO pairs.
10626 (still largely untested)
10629 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
10630 ASCII string. This was handled independently in various places before.
10633 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
10634 UTF8 strings a character at a time.
10637 *) Use client_version from client hello to select the protocol
10638 (s23_srvr.c) and for RSA client key exchange verification
10639 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
10642 *) Add various utility functions to handle SPKACs, these were previously
10643 handled by poking round in the structure internals. Added new function
10644 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
10645 print, verify and generate SPKACs. Based on an original idea from
10646 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
10649 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
10652 *) Allow the config file extension section to be overwritten on the
10653 command line. Based on an original idea from Massimiliano Pala
10654 <madwolf@comune.modena.it>. The new option is called -extensions
10655 and can be applied to ca, req and x509. Also -reqexts to override
10656 the request extensions in req and -crlexts to override the crl extensions
10660 *) Add new feature to the SPKAC handling in ca. Now you can include
10661 the same field multiple times by preceding it by "XXXX." for example:
10664 this is the same syntax as used in the req config file.
10667 *) Allow certificate extensions to be added to certificate requests. These
10668 are specified in a 'req_extensions' option of the req section of the
10669 config file. They can be printed out with the -text option to req but
10670 are otherwise ignored at present.
10673 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
10674 data read consists of only the final block it would not decrypted because
10675 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
10676 A misplaced 'break' also meant the decrypted final block might not be
10677 copied until the next read.
10680 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
10681 a few extra parameters to the DH structure: these will be useful if
10682 for example we want the value of 'q' or implement X9.42 DH.
10685 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
10686 provides hooks that allow the default DSA functions or functions on a
10687 "per key" basis to be replaced. This allows hardware acceleration and
10688 hardware key storage to be handled without major modification to the
10689 library. Also added low level modexp hooks and CRYPTO_EX structure and
10690 associated functions.
10693 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
10694 as "read only": it can't be written to and the buffer it points to will
10695 not be freed. Reading from a read only BIO is much more efficient than
10696 a normal memory BIO. This was added because there are several times when
10697 an area of memory needs to be read from a BIO. The previous method was
10698 to create a memory BIO and write the data to it, this results in two
10699 copies of the data and an O(n^2) reading algorithm. There is a new
10700 function BIO_new_mem_buf() which creates a read only memory BIO from
10701 an area of memory. Also modified the PKCS#7 routines to use read only
10705 *) Bugfix: ssl23_get_client_hello did not work properly when called in
10706 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
10707 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
10708 but a retry condition occurred while trying to read the rest.
10711 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
10712 NID_pkcs7_encrypted by default: this was wrong since this should almost
10713 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
10714 the encrypted data type: this is a more sensible place to put it and it
10715 allows the PKCS#12 code to be tidied up that duplicated this
10719 *) Changed obj_dat.pl script so it takes its input and output files on
10720 the command line. This should avoid shell escape redirection problems
10724 *) Initial support for certificate extension requests, these are included
10725 in things like Xenroll certificate requests. Included functions to allow
10726 extensions to be obtained and added.
10729 *) -crlf option to s_client and s_server for sending newlines as
10730 CRLF (as required by many protocols).
10733 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
10735 *) Install libRSAglue.a when OpenSSL is built with RSAref.
10736 [Ralf S. Engelschall]
10738 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
10739 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
10741 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
10745 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
10746 DH parameters/keys (q is lost during that conversion, but the resulting
10747 DH parameters contain its length).
10749 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
10750 much faster than DH_generate_parameters (which creates parameters
10751 where p = 2*q + 1), and also the smaller q makes DH computations
10752 much more efficient (160-bit exponentiation instead of 1024-bit
10753 exponentiation); so this provides a convenient way to support DHE
10754 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
10755 utter importance to use
10756 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10758 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
10759 when such DH parameters are used, because otherwise small subgroup
10760 attacks may become possible!
10763 *) Avoid memory leak in i2d_DHparams.
10766 *) Allow the -k option to be used more than once in the enc program:
10767 this allows the same encrypted message to be read by multiple recipients.
10770 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
10771 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
10772 it will always use the numerical form of the OID, even if it has a short
10776 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
10777 method only got called if p,q,dmp1,dmq1,iqmp components were present,
10778 otherwise bn_mod_exp was called. In the case of hardware keys for example
10779 no private key components need be present and it might store extra data
10780 in the RSA structure, which cannot be accessed from bn_mod_exp.
10781 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
10782 private key operations.
10785 *) Added support for SPARC Linux.
10788 *) pem_password_cb function type incompatibly changed from
10789 typedef int pem_password_cb(char *buf, int size, int rwflag);
10791 ....(char *buf, int size, int rwflag, void *userdata);
10792 so that applications can pass data to their callbacks:
10793 The PEM[_ASN1]_{read,write}... functions and macros now take an
10794 additional void * argument, which is just handed through whenever
10795 the password callback is called.
10796 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
10798 New function SSL_CTX_set_default_passwd_cb_userdata.
10800 Compatibility note: As many C implementations push function arguments
10801 onto the stack in reverse order, the new library version is likely to
10802 interoperate with programs that have been compiled with the old
10803 pem_password_cb definition (PEM_whatever takes some data that
10804 happens to be on the stack as its last argument, and the callback
10805 just ignores this garbage); but there is no guarantee whatsoever that
10808 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
10809 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
10810 problems not only on Windows, but also on some Unix platforms.
10811 To avoid problematic command lines, these definitions are now in an
10812 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
10813 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
10816 *) MIPS III/IV assembler module is reimplemented.
10819 *) More DES library cleanups: remove references to srand/rand and
10820 delete an unused file.
10823 *) Add support for the the free Netwide assembler (NASM) under Win32,
10824 since not many people have MASM (ml) and it can be hard to obtain.
10825 This is currently experimental but it seems to work OK and pass all
10826 the tests. Check out INSTALL.W32 for info.
10829 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
10830 without temporary keys kept an extra copy of the server key,
10831 and connections with temporary keys did not free everything in case
10835 *) New function RSA_check_key and new openssl rsa option -check
10836 for verifying the consistency of RSA keys.
10837 [Ulf Moeller, Bodo Moeller]
10839 *) Various changes to make Win32 compile work:
10840 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
10841 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
10842 comparison" warnings.
10843 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
10846 *) Add a debugging option to PKCS#5 v2 key generation function: when
10847 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
10848 derived keys are printed to stderr.
10851 *) Copy the flags in ASN1_STRING_dup().
10852 [Roman E. Pavlov <pre@mo.msk.ru>]
10854 *) The x509 application mishandled signing requests containing DSA
10855 keys when the signing key was also DSA and the parameters didn't match.
10857 It was supposed to omit the parameters when they matched the signing key:
10858 the verifying software was then supposed to automatically use the CA's
10859 parameters if they were absent from the end user certificate.
10861 Omitting parameters is no longer recommended. The test was also
10862 the wrong way round! This was probably due to unusual behaviour in
10863 EVP_cmp_parameters() which returns 1 if the parameters match.
10864 This meant that parameters were omitted when they *didn't* match and
10865 the certificate was useless. Certificates signed with 'ca' didn't have
10867 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
10869 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
10870 The interface is as follows:
10871 Applications can use
10872 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
10873 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
10874 "off" is now the default.
10875 The library internally uses
10876 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
10877 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
10878 to disable memory-checking temporarily.
10880 Some inconsistent states that previously were possible (and were
10881 even the default) are now avoided.
10883 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
10884 with each memory chunk allocated; this is occasionally more helpful
10885 than just having a counter.
10887 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
10889 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
10893 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
10894 which largely parallels "options", but is for changing API behaviour,
10895 whereas "options" are about protocol behaviour.
10896 Initial "mode" flags are:
10898 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
10899 a single record has been written.
10900 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
10901 retries use the same buffer location.
10902 (But all of the contents must be
10906 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
10909 *) Fix problems with no-hmac etc.
10910 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
10912 *) New functions RSA_get_default_method(), RSA_set_method() and
10913 RSA_get_method(). These allows replacement of RSA_METHODs without having
10914 to mess around with the internals of an RSA structure.
10917 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
10918 Also really enable memory leak checks in openssl.c and in some
10920 [Chad C. Mulligan, Bodo Moeller]
10922 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
10923 up the length of negative integers. This has now been simplified to just
10924 store the length when it is first determined and use it later, rather
10925 than trying to keep track of where data is copied and updating it to
10927 [Steve Henson, reported by Brien Wheeler
10928 <bwheeler@authentica-security.com>]
10930 *) Add a new function PKCS7_signatureVerify. This allows the verification
10931 of a PKCS#7 signature but with the signing certificate passed to the
10932 function itself. This contrasts with PKCS7_dataVerify which assumes the
10933 certificate is present in the PKCS#7 structure. This isn't always the
10934 case: certificates can be omitted from a PKCS#7 structure and be
10935 distributed by "out of band" means (such as a certificate database).
10938 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
10939 function prototypes in pem.h, also change util/mkdef.pl to add the
10940 necessary function names.
10943 *) mk1mf.pl (used by Windows builds) did not properly read the
10944 options set by Configure in the top level Makefile, and Configure
10945 was not even able to write more than one option correctly.
10946 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
10949 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
10950 file to be loaded from a BIO or FILE pointer. The BIO version will
10951 for example allow memory BIOs to contain config info.
10954 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
10955 Whoever hopes to achieve shared-library compatibility across versions
10956 must use this, not the compile-time macro.
10957 (Exercise 0.9.4: Which is the minimum library version required by
10959 Note: All this applies only to multi-threaded programs, others don't
10963 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
10964 through a BIO pair triggered the default case, i.e.
10965 SSLerr(...,SSL_R_UNKNOWN_STATE).
10968 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
10969 can use the SSL library even if none of the specific BIOs is
10973 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
10974 for the encoded length.
10975 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
10977 *) Add initial documentation of the X509V3 functions.
10980 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
10981 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
10982 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
10983 secure PKCS#8 private key format with a high iteration count.
10986 *) Fix determination of Perl interpreter: A perl or perl5
10987 _directory_ in $PATH was also accepted as the interpreter.
10988 [Ralf S. Engelschall]
10990 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
10991 wrong with it but it was very old and did things like calling
10992 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
10993 unusual formatting.
10996 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
10997 to use the new extension code.
11000 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11001 with macros. This should make it easier to change their form, add extra
11002 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11006 *) Add to configuration table a new entry that can specify an alternative
11007 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11008 according to Mark Crispin <MRC@Panda.COM>.
11012 *) DES CBC did not update the IV. Weird.
11015 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11016 Changing the behaviour of the former might break existing programs --
11017 where IV updating is needed, des_ncbc_encrypt can be used.
11020 *) When bntest is run from "make test" it drives bc to check its
11021 calculations, as well as internally checking them. If an internal check
11022 fails, it needs to cause bc to give a non-zero result or make test carries
11023 on without noticing the failure. Fixed.
11026 *) DES library cleanups.
11029 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11030 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11031 ciphers. NOTE: although the key derivation function has been verified
11032 against some published test vectors it has not been extensively tested
11033 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11037 *) Instead of "mkdir -p", which is not fully portable, use new
11038 Perl script "util/mkdir-p.pl".
11041 *) Rewrite the way password based encryption (PBE) is handled. It used to
11042 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11043 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11044 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11045 the 'parameter' field of the AlgorithmIdentifier is passed to the
11046 underlying key generation function so it must do its own ASN1 parsing.
11047 This has also changed the EVP_PBE_CipherInit() function which now has a
11048 'parameter' argument instead of literal salt and iteration count values
11049 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11052 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11053 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11054 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11055 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11056 value was just used as a "magic string" and not used directly its
11057 value doesn't matter.
11060 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11064 *) "linux-sparc64" configuration (ultrapenguin).
11065 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11066 "linux-sparc" configuration.
11067 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11069 *) config now generates no-xxx options for missing ciphers.
11072 *) Support the EBCDIC character set (work in progress).
11073 File ebcdic.c not yet included because it has a different license.
11074 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11076 *) Support BS2000/OSD-POSIX.
11077 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11079 *) Make callbacks for key generation use void * instead of char *.
11082 *) Make S/MIME samples compile (not yet tested).
11085 *) Additional typesafe stacks.
11088 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11092 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11094 *) New configuration variant "sco5-gcc".
11096 *) Updated some demos.
11097 [Sean O Riordain, Wade Scholine]
11099 *) Add missing BIO_free at exit of pkcs12 application.
11102 *) Fix memory leak in conf.c.
11105 *) Updates for Win32 to assembler version of MD5.
11108 *) Set #! path to perl in apps/der_chop to where we found it
11109 instead of using a fixed path.
11112 *) SHA library changes for irix64-mips4-cc.
11115 *) Improvements for VMS support.
11119 Changes between 0.9.2b and 0.9.3 [24 May 1999]
11121 *) Bignum library bug fix. IRIX 6 passes "make test" now!
11122 This also avoids the problems with SC4.2 and unpatched SC5.
11123 [Andy Polyakov <appro@fy.chalmers.se>]
11125 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
11126 These are required because of the typesafe stack would otherwise break
11127 existing code. If old code used a structure member which used to be STACK
11128 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
11129 sk_num or sk_value it would produce an error because the num, data members
11130 are not present in STACK_OF. Now it just produces a warning. sk_set
11131 replaces the old method of assigning a value to sk_value
11132 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
11133 that does this will no longer work (and should use sk_set instead) but
11134 this could be regarded as a "questionable" behaviour anyway.
11137 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
11138 correctly handle encrypted S/MIME data.
11141 *) Change type of various DES function arguments from des_cblock
11142 (which means, in function argument declarations, pointer to char)
11143 to des_cblock * (meaning pointer to array with 8 char elements),
11144 which allows the compiler to do more typechecking; it was like
11145 that back in SSLeay, but with lots of ugly casts.
11147 Introduce new type const_des_cblock.
11150 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
11151 problems: find RecipientInfo structure that matches recipient certificate
11152 and initialise the ASN1 structures properly based on passed cipher.
11155 *) Belatedly make the BN tests actually check the results.
11158 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
11159 to and from BNs: it was completely broken. New compilation option
11160 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
11161 key elements as negative integers.
11164 *) Reorganize and speed up MD5.
11165 [Andy Polyakov <appro@fy.chalmers.se>]
11168 [Richard Levitte <richard@levitte.org>]
11170 *) New option -out to asn1parse to allow the parsed structure to be
11171 output to a file. This is most useful when combined with the -strparse
11172 option to examine the output of things like OCTET STRINGS.
11175 *) Make SSL library a little more fool-proof by not requiring any longer
11176 that SSL_set_{accept,connect}_state be called before
11177 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
11178 in many applications because usually everything *appeared* to work as
11179 intended anyway -- now it really works as intended).
11182 *) Move openssl.cnf out of lib/.
11185 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
11186 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
11187 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
11188 [Ralf S. Engelschall]
11190 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
11191 handle PKCS#7 enveloped data properly.
11192 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
11194 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
11195 copying pointers. The cert_st handling is changed by this in
11196 various ways (and thus what used to be known as ctx->default_cert
11197 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
11198 any longer when s->cert does not give us what we need).
11199 ssl_cert_instantiate becomes obsolete by this change.
11200 As soon as we've got the new code right (possibly it already is?),
11201 we have solved a couple of bugs of the earlier code where s->cert
11202 was used as if it could not have been shared with other SSL structures.
11204 Note that using the SSL API in certain dirty ways now will result
11205 in different behaviour than observed with earlier library versions:
11206 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
11207 does not influence s as it used to.
11209 In order to clean up things more thoroughly, inside SSL_SESSION
11210 we don't use CERT any longer, but a new structure SESS_CERT
11211 that holds per-session data (if available); currently, this is
11212 the peer's certificate chain and, for clients, the server's certificate
11213 and temporary key. CERT holds only those values that can have
11214 meaningful defaults in an SSL_CTX.
11217 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
11218 from the internal representation. Various PKCS#7 fixes: remove some
11219 evil casts and set the enc_dig_alg field properly based on the signing
11223 *) Allow PKCS#12 password to be set from the command line or the
11224 environment. Let 'ca' get its config file name from the environment
11225 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
11229 *) Allow certificate policies extension to use an IA5STRING for the
11230 organization field. This is contrary to the PKIX definition but
11231 VeriSign uses it and IE5 only recognises this form. Document 'x509'
11235 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
11236 without disallowing inline assembler and the like for non-pedantic builds.
11239 *) Support Borland C++ builder.
11240 [Janez Jere <jj@void.si>, modified by Ulf Möller]
11242 *) Support Mingw32.
11245 *) SHA-1 cleanups and performance enhancements.
11246 [Andy Polyakov <appro@fy.chalmers.se>]
11248 *) Sparc v8plus assembler for the bignum library.
11249 [Andy Polyakov <appro@fy.chalmers.se>]
11251 *) Accept any -xxx and +xxx compiler options in Configure.
11254 *) Update HPUX configuration.
11257 *) Add missing sk_<type>_unshift() function to safestack.h
11258 [Ralf S. Engelschall]
11260 *) New function SSL_CTX_use_certificate_chain_file that sets the
11261 "extra_cert"s in addition to the certificate. (This makes sense
11262 only for "PEM" format files, as chains as a whole are not
11266 *) Support verify_depth from the SSL API.
11267 x509_vfy.c had what can be considered an off-by-one-error:
11268 Its depth (which was not part of the external interface)
11269 was actually counting the number of certificates in a chain;
11270 now it really counts the depth.
11273 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
11274 instead of X509err, which often resulted in confusing error
11275 messages since the error codes are not globally unique
11276 (e.g. an alleged error in ssl3_accept when a certificate
11277 didn't match the private key).
11279 *) New function SSL_CTX_set_session_id_context that allows to set a default
11280 value (so that you don't need SSL_set_session_id_context for each
11281 connection using the SSL_CTX).
11284 *) OAEP decoding bug fix.
11287 *) Support INSTALL_PREFIX for package builders, as proposed by
11291 *) New Configure options "threads" and "no-threads". For systems
11292 where the proper compiler options are known (currently Solaris
11293 and Linux), "threads" is the default.
11296 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
11299 *) Install various scripts to $(OPENSSLDIR)/misc, not to
11300 $(INSTALLTOP)/bin -- they shouldn't clutter directories
11301 such as /usr/local/bin.
11304 *) "make linux-shared" to build shared libraries.
11305 [Niels Poppe <niels@netbox.org>]
11307 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
11310 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
11311 extension adding in x509 utility.
11314 *) Remove NOPROTO sections and error code comments.
11317 *) Partial rewrite of the DEF file generator to now parse the ANSI
11321 *) New Configure options --prefix=DIR and --openssldir=DIR.
11324 *) Complete rewrite of the error code script(s). It is all now handled
11325 by one script at the top level which handles error code gathering,
11326 header rewriting and C source file generation. It should be much better
11327 than the old method: it now uses a modified version of Ulf's parser to
11328 read the ANSI prototypes in all header files (thus the old K&R definitions
11329 aren't needed for error creation any more) and do a better job of
11330 translating function codes into names. The old 'ASN1 error code imbedded
11331 in a comment' is no longer necessary and it doesn't use .err files which
11332 have now been deleted. Also the error code call doesn't have to appear all
11333 on one line (which resulted in some large lines...).
11336 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
11339 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
11340 0 (which usually indicates a closed connection), but continue reading.
11343 *) Fix some race conditions.
11346 *) Add support for CRL distribution points extension. Add Certificate
11347 Policies and CRL distribution points documentation.
11350 *) Move the autogenerated header file parts to crypto/opensslconf.h.
11353 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
11354 8 of keying material. Merlin has also confirmed interop with this fix
11355 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
11356 [Merlin Hughes <merlin@baltimore.ie>]
11358 *) Fix lots of warnings.
11359 [Richard Levitte <levitte@stacken.kth.se>]
11361 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
11362 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
11363 [Richard Levitte <levitte@stacken.kth.se>]
11365 *) Fix problems with sizeof(long) == 8.
11366 [Andy Polyakov <appro@fy.chalmers.se>]
11368 *) Change functions to ANSI C.
11371 *) Fix typos in error codes.
11372 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
11374 *) Remove defunct assembler files from Configure.
11377 *) SPARC v8 assembler BIGNUM implementation.
11378 [Andy Polyakov <appro@fy.chalmers.se>]
11380 *) Support for Certificate Policies extension: both print and set.
11381 Various additions to support the r2i method this uses.
11384 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
11385 return a const string when you are expecting an allocated buffer.
11388 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
11389 types DirectoryString and DisplayText.
11392 *) Add code to allow r2i extensions to access the configuration database,
11393 add an LHASH database driver and add several ctx helper functions.
11396 *) Fix an evil bug in bn_expand2() which caused various BN functions to
11397 fail when they extended the size of a BIGNUM.
11400 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
11401 support typesafe stack.
11404 *) Fix typo in SSL_[gs]et_options().
11405 [Nils Frostberg <nils@medcom.se>]
11407 *) Delete various functions and files that belonged to the (now obsolete)
11408 old X509V3 handling code.
11411 *) New Configure option "rsaref".
11414 *) Don't auto-generate pem.h.
11417 *) Introduce type-safe ASN.1 SETs.
11420 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
11421 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
11423 *) Introduce type-safe STACKs. This will almost certainly break lots of code
11424 that links with OpenSSL (well at least cause lots of warnings), but fear
11425 not: the conversion is trivial, and it eliminates loads of evil casts. A
11426 few STACKed things have been converted already. Feel free to convert more.
11427 In the fullness of time, I'll do away with the STACK type altogether.
11430 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
11431 specified in <certfile> by updating the entry in the index.txt file.
11432 This way one no longer has to edit the index.txt file manually for
11433 revoking a certificate. The -revoke option does the gory details now.
11434 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
11436 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
11437 `-text' option at all and this way the `-noout -text' combination was
11438 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
11439 [Ralf S. Engelschall]
11441 *) Make sure a corresponding plain text error message exists for the
11442 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
11443 verify callback function determined that a certificate was revoked.
11444 [Ralf S. Engelschall]
11446 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
11447 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
11448 all available cipers including rc5, which was forgotten until now.
11449 In order to let the testing shell script know which algorithms
11450 are available, a new (up to now undocumented) command
11451 "openssl list-cipher-commands" is used.
11454 *) Bugfix: s_client occasionally would sleep in select() when
11455 it should have checked SSL_pending() first.
11458 *) New functions DSA_do_sign and DSA_do_verify to provide access to
11459 the raw DSA values prior to ASN.1 encoding.
11462 *) Tweaks to Configure
11463 [Niels Poppe <niels@netbox.org>]
11465 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
11469 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
11472 *) New config option to avoid instructions that are illegal on the 80386.
11473 The default code is faster, but requires at least a 486.
11476 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
11477 SSL2_SERVER_VERSION (not used at all) macros, which are now the
11478 same as SSL2_VERSION anyway.
11481 *) New "-showcerts" option for s_client.
11484 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
11485 application. Various cleanups and fixes.
11488 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
11489 modify error routines to work internally. Add error codes and PBE init
11490 to library startup routines.
11493 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
11494 packing functions to asn1 and evp. Changed function names and error
11495 codes along the way.
11498 *) PKCS12 integration: and so it begins... First of several patches to
11499 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
11500 objects to objects.h
11503 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
11504 and display support for Thawte strong extranet extension.
11507 *) Add LinuxPPC support.
11508 [Jeff Dubrule <igor@pobox.org>]
11510 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
11511 bn_div_words in alpha.s.
11512 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
11514 *) Make sure the RSA OAEP test is skipped under -DRSAref because
11515 OAEP isn't supported when OpenSSL is built with RSAref.
11516 [Ulf Moeller <ulf@fitug.de>]
11518 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
11519 so they no longer are missing under -DNOPROTO.
11520 [Soren S. Jorvang <soren@t.dk>]
11523 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
11525 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
11526 doesn't work when the session is reused. Coming soon!
11529 *) Fix a security hole, that allows sessions to be reused in the wrong
11530 context thus bypassing client cert protection! All software that uses
11531 client certs and session caches in multiple contexts NEEDS PATCHING to
11532 allow session reuse! A fuller solution is in the works.
11533 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
11535 *) Some more source tree cleanups (removed obsolete files
11536 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
11537 permission on "config" script to be executable) and a fix for the INSTALL
11539 [Ulf Moeller <ulf@fitug.de>]
11541 *) Remove some legacy and erroneous uses of malloc, free instead of
11543 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
11545 *) Make rsa_oaep_test return non-zero on error.
11546 [Ulf Moeller <ulf@fitug.de>]
11548 *) Add support for native Solaris shared libraries. Configure
11549 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
11550 if someone would make that last step automatic.
11551 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
11553 *) ctx_size was not built with the right compiler during "make links". Fixed.
11556 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
11557 except NULL ciphers". This means the default cipher list will no longer
11558 enable NULL ciphers. They need to be specifically enabled e.g. with
11559 the string "DEFAULT:eNULL".
11562 *) Fix to RSA private encryption routines: if p < q then it would
11563 occasionally produce an invalid result. This will only happen with
11564 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
11567 *) Be less restrictive and allow also `perl util/perlpath.pl
11568 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
11569 because this way one can also use an interpreter named `perl5' (which is
11570 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
11571 installed as `perl').
11572 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11574 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
11575 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11577 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
11578 advapi32.lib to Win32 build and change the pem test comparison
11579 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
11580 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
11581 and crypto/des/ede_cbcm_enc.c.
11584 *) DES quad checksum was broken on big-endian architectures. Fixed.
11587 *) Comment out two functions in bio.h that aren't implemented. Fix up the
11588 Win32 test batch file so it (might) work again. The Win32 test batch file
11589 is horrible: I feel ill....
11592 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
11593 in e_os.h. Audit of header files to check ANSI and non ANSI
11594 sections: 10 functions were absent from non ANSI section and not exported
11595 from Windows DLLs. Fixed up libeay.num for new functions.
11598 *) Make `openssl version' output lines consistent.
11599 [Ralf S. Engelschall]
11601 *) Fix Win32 symbol export lists for BIO functions: Added
11602 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
11603 to ms/libeay{16,32}.def.
11604 [Ralf S. Engelschall]
11606 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
11607 fine under Unix and passes some trivial tests I've now added. But the
11608 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
11609 added to make sure no one expects that this stuff really works in the
11610 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
11611 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
11613 [Ralf S. Engelschall]
11615 *) Fix the generation of two part addresses in perl.
11616 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
11618 *) Add config entry for Linux on MIPS.
11619 [John Tobey <jtobey@channel1.com>]
11621 *) Make links whenever Configure is run, unless we are on Windoze.
11624 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
11625 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
11629 *) Add a useful kludge to allow package maintainers to specify compiler and
11630 other platforms details on the command line without having to patch the
11631 Configure script everytime: One now can use ``perl Configure
11632 <id>:<details>'', i.e. platform ids are allowed to have details appended
11633 to them (separated by colons). This is treated as there would be a static
11634 pre-configured entry in Configure's %table under key <id> with value
11635 <details> and ``perl Configure <id>'' is called. So, when you want to
11636 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
11637 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
11638 now, which overrides the FreeBSD-elf entry on-the-fly.
11639 [Ralf S. Engelschall]
11641 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
11644 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
11645 on the `perl Configure ...' command line. This way one can compile
11646 OpenSSL libraries with Position Independent Code (PIC) which is needed
11647 for linking it into DSOs.
11648 [Ralf S. Engelschall]
11650 *) Remarkably, export ciphers were totally broken and no-one had noticed!
11654 *) Cleaned up the LICENSE document: The official contact for any license
11655 questions now is the OpenSSL core team under openssl-core@openssl.org.
11656 And add a paragraph about the dual-license situation to make sure people
11657 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
11658 to the OpenSSL toolkit.
11659 [Ralf S. Engelschall]
11661 *) General source tree makefile cleanups: Made `making xxx in yyy...'
11662 display consistent in the source tree and replaced `/bin/rm' by `rm'.
11663 Additionally cleaned up the `make links' target: Remove unnecessary
11664 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
11665 to speed processing and no longer clutter the display with confusing
11666 stuff. Instead only the actually done links are displayed.
11667 [Ralf S. Engelschall]
11669 *) Permit null encryption ciphersuites, used for authentication only. It used
11670 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
11671 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
11675 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
11676 signed attributes when verifying signatures (this would break them),
11677 the detached data encoding was wrong and public keys obtained using
11678 X509_get_pubkey() weren't freed.
11681 *) Add text documentation for the BUFFER functions. Also added a work around
11682 to a Win95 console bug. This was triggered by the password read stuff: the
11683 last character typed gets carried over to the next fread(). If you were
11684 generating a new cert request using 'req' for example then the last
11685 character of the passphrase would be CR which would then enter the first
11689 *) Added the new `Includes OpenSSL Cryptography Software' button as
11690 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
11691 button and can be used by applications based on OpenSSL to show the
11692 relationship to the OpenSSL project.
11693 [Ralf S. Engelschall]
11695 *) Remove confusing variables in function signatures in files
11696 ssl/ssl_lib.c and ssl/ssl.h.
11697 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11699 *) Don't install bss_file.c under PREFIX/include/
11700 [Lennart Bong <lob@kulthea.stacken.kth.se>]
11702 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
11703 functions that return function pointers and has support for NT specific
11704 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
11705 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
11706 unsigned to signed types: this was killing the Win32 compile.
11709 *) Add new certificate file to stack functions,
11710 SSL_add_dir_cert_subjects_to_stack() and
11711 SSL_add_file_cert_subjects_to_stack(). These largely supplant
11712 SSL_load_client_CA_file(), and can be used to add multiple certs easily
11713 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
11714 This means that Apache-SSL and similar packages don't have to mess around
11715 to add as many CAs as they want to the preferred list.
11718 *) Experiment with doxygen documentation. Currently only partially applied to
11720 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
11721 openssl.doxy as the configuration file.
11724 *) Get rid of remaining C++-style comments which strict C compilers hate.
11725 [Ralf S. Engelschall, pointed out by Carlos Amengual]
11727 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
11728 compiled in by default: it has problems with large keys.
11731 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
11732 DH private keys and/or callback functions which directly correspond to
11733 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
11734 is needed for applications which have to configure certificates on a
11735 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
11737 For the RSA certificate situation is makes no difference, but
11738 for the DSA certificate situation this fixes the "no shared cipher"
11739 problem where the OpenSSL cipher selection procedure failed because the
11740 temporary keys were not overtaken from the context and the API provided
11741 no way to reconfigure them.
11742 The new functions now let applications reconfigure the stuff and they
11743 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
11744 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
11745 non-public-API function ssl_cert_instantiate() is used as a helper
11746 function and also to reduce code redundancy inside ssl_rsa.c.
11747 [Ralf S. Engelschall]
11749 *) Move s_server -dcert and -dkey options out of the undocumented feature
11750 area because they are useful for the DSA situation and should be
11751 recognized by the users.
11752 [Ralf S. Engelschall]
11754 *) Fix the cipher decision scheme for export ciphers: the export bits are
11755 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
11756 SSL_EXP_MASK. So, the original variable has to be used instead of the
11757 already masked variable.
11758 [Richard Levitte <levitte@stacken.kth.se>]
11760 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
11761 [Richard Levitte <levitte@stacken.kth.se>]
11763 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
11764 from `int' to `unsigned int' because it's a length and initialized by
11765 EVP_DigestFinal() which expects an `unsigned int *'.
11766 [Richard Levitte <levitte@stacken.kth.se>]
11768 *) Don't hard-code path to Perl interpreter on shebang line of Configure
11769 script. Instead use the usual Shell->Perl transition trick.
11770 [Ralf S. Engelschall]
11772 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
11773 (in addition to RSA certificates) to match the behaviour of `openssl dsa
11774 -noout -modulus' as it's already the case for `openssl rsa -noout
11775 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
11776 currently the public key is printed (a decision which was already done by
11777 `openssl dsa -modulus' in the past) which serves a similar purpose.
11778 Additionally the NO_RSA no longer completely removes the whole -modulus
11779 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
11781 [Ralf S. Engelschall]
11783 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
11784 BIO. See the source (crypto/evp/bio_ok.c) for more info.
11785 [Arne Ansper <arne@ats.cyber.ee>]
11787 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
11788 to be added. Now both 'req' and 'ca' can use new objects defined in the
11792 *) Add cool BIO that does syslog (or event log on NT).
11793 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
11795 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
11796 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
11797 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
11798 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
11801 *) Add preliminary config info for new extension code.
11804 *) Make RSA_NO_PADDING really use no padding.
11805 [Ulf Moeller <ulf@fitug.de>]
11807 *) Generate errors when private/public key check is done.
11810 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
11811 for some CRL extensions and new objects added.
11814 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
11815 key usage extension and fuller support for authority key id.
11818 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
11819 padding method for RSA, which is recommended for new applications in PKCS
11820 #1 v2.0 (RFC 2437, October 1998).
11821 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
11822 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
11823 against Bleichbacher's attack on RSA.
11824 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
11827 *) Updates to the new SSL compression code
11828 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11830 *) Fix so that the version number in the master secret, when passed
11831 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
11832 (because the server will not accept higher), that the version number
11833 is 0x03,0x01, not 0x03,0x00
11834 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11836 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
11837 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
11838 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
11841 *) Support for RAW extensions where an arbitrary extension can be
11842 created by including its DER encoding. See apps/openssl.cnf for
11846 *) Make sure latest Perl versions don't interpret some generated C array
11847 code as Perl array code in the crypto/err/err_genc.pl script.
11848 [Lars Weber <3weber@informatik.uni-hamburg.de>]
11850 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
11851 not many people have the assembler. Various Win32 compilation fixes and
11852 update to the INSTALL.W32 file with (hopefully) more accurate Win32
11853 build instructions.
11856 *) Modify configure script 'Configure' to automatically create crypto/date.h
11857 file under Win32 and also build pem.h from pem.org. New script
11858 util/mkfiles.pl to create the MINFO file on environments that can't do a
11859 'make files': perl util/mkfiles.pl >MINFO should work.
11862 *) Major rework of DES function declarations, in the pursuit of correctness
11863 and purity. As a result, many evil casts evaporated, and some weirdness,
11864 too. You may find this causes warnings in your code. Zapping your evil
11865 casts will probably fix them. Mostly.
11868 *) Fix for a typo in asn1.h. Bug fix to object creation script
11869 obj_dat.pl. It considered a zero in an object definition to mean
11870 "end of object": none of the objects in objects.h have any zeros
11871 so it wasn't spotted.
11872 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
11874 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
11875 Masking (CBCM). In the absence of test vectors, the best I have been able
11876 to do is check that the decrypt undoes the encrypt, so far. Send me test
11877 vectors if you have them.
11880 *) Correct calculation of key length for export ciphers (too much space was
11881 allocated for null ciphers). This has not been tested!
11884 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
11885 message is now correct (it understands "crypto" and "ssl" on its
11886 command line). There is also now an "update" option. This will update
11887 the util/ssleay.num and util/libeay.num files with any new functions.
11889 perl util/mkdef.pl crypto ssl update
11890 it will update them.
11893 *) Overhauled the Perl interface (perl/*):
11894 - ported BN stuff to OpenSSL's different BN library
11895 - made the perl/ source tree CVS-aware
11896 - renamed the package from SSLeay to OpenSSL (the files still contain
11897 their history because I've copied them in the repository)
11898 - removed obsolete files (the test scripts will be replaced
11899 by better Test::Harness variants in the future)
11900 [Ralf S. Engelschall]
11902 *) First cut for a very conservative source tree cleanup:
11903 1. merge various obsolete readme texts into doc/ssleay.txt
11904 where we collect the old documents and readme texts.
11905 2. remove the first part of files where I'm already sure that we no
11906 longer need them because of three reasons: either they are just temporary
11907 files which were left by Eric or they are preserved original files where
11908 I've verified that the diff is also available in the CVS via "cvs diff
11909 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
11910 the crypto/md/ stuff).
11911 [Ralf S. Engelschall]
11913 *) More extension code. Incomplete support for subject and issuer alt
11914 name, issuer and authority key id. Change the i2v function parameters
11915 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
11916 what that's for :-) Fix to ASN1 macro which messed up
11917 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
11920 *) Preliminary support for ENUMERATED type. This is largely copied from the
11924 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
11925 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11927 *) Make sure `make rehash' target really finds the `openssl' program.
11928 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
11930 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
11931 like to hear about it if this slows down other processors.
11934 *) Add CygWin32 platform information to Configure script.
11935 [Alan Batie <batie@aahz.jf.intel.com>]
11937 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
11938 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
11940 *) New program nseq to manipulate netscape certificate sequences
11943 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
11947 *) Fixes to BN code. Previously the default was to define BN_RECURSION
11948 but the BN code had some problems that would cause failures when
11949 doing certificate verification and some other functions.
11950 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
11952 *) Add ASN1 and PEM code to support netscape certificate sequences.
11955 *) Add ASN1 and PEM code to support netscape certificate sequences.
11958 *) Add several PKIX and private extended key usage OIDs.
11961 *) Modify the 'ca' program to handle the new extension code. Modify
11962 openssl.cnf for new extension format, add comments.
11965 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
11966 and add a sample to openssl.cnf so req -x509 now adds appropriate
11970 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
11971 error code, add initial support to X509_print() and x509 application.
11974 *) Takes a deep breath and start adding X509 V3 extension support code. Add
11975 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
11976 stuff is currently isolated and isn't even compiled yet.
11979 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
11980 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
11981 Removed the versions check from X509 routines when loading extensions:
11982 this allows certain broken certificates that don't set the version
11983 properly to be processed.
11986 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
11987 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
11988 can still be regenerated with "make depend".
11991 *) Spelling mistake in C version of CAST-128.
11992 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
11994 *) Changes to the error generation code. The perl script err-code.pl
11995 now reads in the old error codes and retains the old numbers, only
11996 adding new ones if necessary. It also only changes the .err files if new
11997 codes are added. The makefiles have been modified to only insert errors
11998 when needed (to avoid needlessly modifying header files). This is done
11999 by only inserting errors if the .err file is newer than the auto generated
12000 C file. To rebuild all the error codes from scratch (the old behaviour)
12001 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12002 or delete all the .err files.
12005 *) CAST-128 was incorrectly implemented for short keys. The C version has
12006 been fixed, but is untested. The assembler versions are also fixed, but
12007 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12008 to regenerate it if needed.
12009 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12010 Hagino <itojun@kame.net>]
12012 *) File was opened incorrectly in randfile.c.
12013 [Ulf Möller <ulf@fitug.de>]
12015 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12016 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12017 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12018 al: it's just almost always a UTCTime. Note this patch adds new error
12019 codes so do a "make errors" if there are problems.
12022 *) Correct Linux 1 recognition in config.
12023 [Ulf Möller <ulf@fitug.de>]
12025 *) Remove pointless MD5 hash when using DSA keys in ca.
12026 [Anonymous <nobody@replay.com>]
12028 *) Generate an error if given an empty string as a cert directory. Also
12029 generate an error if handed NULL (previously returned 0 to indicate an
12030 error, but didn't set one).
12031 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12033 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12036 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12037 parameters. This was causing a warning which killed off the Win32 compile.
12040 *) Remove C++ style comments from crypto/bn/bn_local.h.
12041 [Neil Costigan <neil.costigan@celocom.com>]
12043 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12044 based on a text string, looking up short and long names and finally
12045 "dot" format. The "dot" format stuff didn't work. Added new function
12046 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12047 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12048 OID is not part of the table.
12051 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12052 X509_LOOKUP_by_alias().
12055 *) Sort openssl functions by name.
12058 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12059 encryption from sample DSA keys (in case anyone is interested the password
12063 *) Make _all_ *_free functions accept a NULL pointer.
12064 [Frans Heymans <fheymans@isaserver.be>]
12066 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12068 [Anonymous <nobody@replay.com>]
12070 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12071 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12073 *) Don't blow it for numeric -newkey arguments to apps/req.
12074 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12076 *) Temp key "for export" tests were wrong in s3_srvr.c.
12077 [Anonymous <nobody@replay.com>]
12079 *) Add prototype for temp key callback functions
12080 SSL_CTX_set_tmp_{rsa,dh}_callback().
12083 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12084 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12087 *) X509_name_add_entry() freed the wrong thing after an error.
12088 [Arne Ansper <arne@ats.cyber.ee>]
12090 *) rsa_eay.c would attempt to free a NULL context.
12091 [Arne Ansper <arne@ats.cyber.ee>]
12093 *) BIO_s_socket() had a broken should_retry() on Windoze.
12094 [Arne Ansper <arne@ats.cyber.ee>]
12096 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12097 [Arne Ansper <arne@ats.cyber.ee>]
12099 *) Make sure the already existing X509_STORE->depth variable is initialized
12100 in X509_STORE_new(), but document the fact that this variable is still
12101 unused in the certificate verification process.
12102 [Ralf S. Engelschall]
12104 *) Fix the various library and apps files to free up pkeys obtained from
12105 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12108 *) Fix reference counting in X509_PUBKEY_get(). This makes
12109 demos/maurice/example2.c work, amongst others, probably.
12110 [Steve Henson and Ben Laurie]
12112 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
12113 `openssl' and second, the shortcut symlinks for the `openssl <command>'
12114 are no longer created. This way we have a single and consistent command
12115 line interface `openssl <command>', similar to `cvs <command>'.
12116 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
12118 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
12119 BIT STRING wrapper always have zero unused bits.
12122 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
12125 *) Make the top-level INSTALL documentation easier to understand.
12128 *) Makefiles updated to exit if an error occurs in a sub-directory
12129 make (including if user presses ^C) [Paul Sutton]
12131 *) Make Montgomery context stuff explicit in RSA data structure.
12134 *) Fix build order of pem and err to allow for generated pem.h.
12137 *) Fix renumbering bug in X509_NAME_delete_entry().
12140 *) Enhanced the err-ins.pl script so it makes the error library number
12141 global and can add a library name. This is needed for external ASN1 and
12142 other error libraries.
12145 *) Fixed sk_insert which never worked properly.
12148 *) Fix ASN1 macros so they can handle indefinite length constructed
12149 EXPLICIT tags. Some non standard certificates use these: they can now
12153 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
12154 into a single doc/ssleay.txt bundle. This way the information is still
12155 preserved but no longer messes up this directory. Now it's new room for
12156 the new set of documentation files.
12157 [Ralf S. Engelschall]
12159 *) SETs were incorrectly DER encoded. This was a major pain, because they
12160 shared code with SEQUENCEs, which aren't coded the same. This means that
12161 almost everything to do with SETs or SEQUENCEs has either changed name or
12162 number of arguments.
12163 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
12165 *) Fix test data to work with the above.
12168 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
12169 was already fixed by Eric for 0.9.1 it seems.
12170 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
12172 *) Autodetect FreeBSD3.
12175 *) Fix various bugs in Configure. This affects the following platforms:
12179 unixware-2.0-pentium
12183 *) Eliminate generated files from CVS. Reorder tests to regenerate files
12184 before they are needed.
12187 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
12191 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
12193 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
12194 changed SSLeay to OpenSSL in version strings.
12195 [Ralf S. Engelschall]
12197 *) Some fixups to the top-level documents.
12200 *) Fixed the nasty bug where rsaref.h was not found under compile-time
12201 because the symlink to include/ was missing.
12202 [Ralf S. Engelschall]
12204 *) Incorporated the popular no-RSA/DSA-only patches
12205 which allow to compile a RSA-free SSLeay.
12206 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
12208 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
12209 when "ssleay" is still not found.
12210 [Ralf S. Engelschall]
12212 *) Added more platforms to Configure: Cray T3E, HPUX 11,
12213 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
12215 *) Updated the README file.
12216 [Ralf S. Engelschall]
12218 *) Added various .cvsignore files in the CVS repository subdirs
12219 to make a "cvs update" really silent.
12220 [Ralf S. Engelschall]
12222 *) Recompiled the error-definition header files and added
12223 missing symbols to the Win32 linker tables.
12224 [Ralf S. Engelschall]
12226 *) Cleaned up the top-level documents;
12227 o new files: CHANGES and LICENSE
12228 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
12229 o merged COPYRIGHT into LICENSE
12230 o removed obsolete TODO file
12231 o renamed MICROSOFT to INSTALL.W32
12232 [Ralf S. Engelschall]
12234 *) Removed dummy files from the 0.9.1b source tree:
12235 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
12236 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
12237 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
12238 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
12239 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
12240 [Ralf S. Engelschall]
12242 *) Added various platform portability fixes.
12245 *) The Genesis of the OpenSSL rpject:
12246 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
12247 Young and Tim J. Hudson created while they were working for C2Net until
12249 [The OpenSSL Project]
12252 Changes between 0.9.0b and 0.9.1b [not released]
12254 *) Updated a few CA certificates under certs/
12257 *) Changed some BIGNUM api stuff.
12260 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
12261 DGUX x86, Linux Alpha, etc.
12264 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
12265 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
12269 *) Add -strparse option to asn1pars program which parses nested
12271 [Dr Stephen Henson <shenson@bigfoot.com>]
12273 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
12276 *) DSA fix for "ca" program.
12279 *) Added "-genkey" option to "dsaparam" program.
12282 *) Added RIPE MD160 (rmd160) message digest.
12285 *) Added -a (all) option to "ssleay version" command.
12288 *) Added PLATFORM define which is the id given to Configure.
12291 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
12294 *) Extended the ASN.1 parser routines.
12297 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
12300 *) Added a BN_CTX to the BN library.
12303 *) Fixed the weak key values in DES library
12306 *) Changed API in EVP library for cipher aliases.
12309 *) Added support for RC2/64bit cipher.
12312 *) Converted the lhash library to the crypto/mem.c functions.
12315 *) Added more recognized ASN.1 object ids.
12318 *) Added more RSA padding checks for SSL/TLS.
12321 *) Added BIO proxy/filter functionality.
12324 *) Added extra_certs to SSL_CTX which can be used
12325 send extra CA certificates to the client in the CA cert chain sending
12326 process. It can be configured with SSL_CTX_add_extra_chain_cert().
12329 *) Now Fortezza is denied in the authentication phase because
12330 this is key exchange mechanism is not supported by SSLeay at all.
12333 *) Additional PKCS1 checks.
12336 *) Support the string "TLSv1" for all TLS v1 ciphers.
12339 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
12340 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
12343 *) Fixed a few memory leaks.
12346 *) Fixed various code and comment typos.
12349 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
12350 bytes sent in the client random.
12351 [Edward Bishop <ebishop@spyglass.com>]