5 Changes between 1.0.2h and 1.1.0 [xx XXX 2016]
7 *) The method for finding the storage location for the Windows RAND seed file
8 has changed. First we check %RANDFILE%. If that is not set then we check
9 the directories %TMP%, %TEMP%, %USERPROFILE%, %SYSTEMROOT% and %HOME% in
10 that order. If all else fails we fall back to "C:".
13 *) The EVP_EncryptUpdate() function has had its return type changed from void
14 to int. A return of 0 indicates and error while a return of 1 indicates
18 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
19 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
20 off the constant time implementation for RSA, DSA and DH have been made
21 no-ops and deprecated.
24 *) Windows RAND implementation was simplified to only get entropy by
25 calling CryptGenRandom(). Various other RAND-related tickets
27 [Joseph Wylie Yandle, Rich Salz]
29 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
30 and OPENSSL_LH_, respectively. The old names are available
31 with API compatibility. They new names are now completely documented.
34 *) Unify TYPE_up_ref(obj) methods signature.
35 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
36 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
37 int (instead of void) like all others TYPE_up_ref() methods.
38 So now these methods also check the return value of CRYPTO_atomic_add(),
39 and the validity of object reference counter.
40 [fdasilvayy@gmail.com]
42 *) With Windows Visual Studio builds, the .pdb files are installed
43 alongside the installed libraries and executables. For a static
44 library installation, ossl_static.pdb is the associate compiler
45 generated .pdb file to be used when linking programs.
48 *) Remove openssl.spec. Packaging files belong with the packagers.
51 *) Automatic Darwin/OSX configuration has had a refresh, it will now
52 recognise x86_64 architectures automatically. You can still decide
53 to build for a different bitness with the environment variable
54 KERNEL_BITS (can be 32 or 64), for example:
56 KERNEL_BITS=32 ./config
60 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
61 256 bit AES and HMAC with SHA256.
64 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
67 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
70 *) To enable users to have their own config files and build file templates,
71 Configure looks in the directory indicated by the environment variable
72 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
73 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
74 name and is used as is.
77 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
78 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
79 X509_CERT_FILE_CTX was removed.
82 *) "shared" builds are now the default. To create only static libraries use
83 the "no-shared" Configure option.
86 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
87 All of these option have not worked for some while and are fundamental
91 *) Make various cleanup routines no-ops and mark them as deprecated. Most
92 global cleanup functions are no longer required because they are handled
93 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
94 Explicitly de-initing can cause problems (e.g. where a library that uses
95 OpenSSL de-inits, but an application is still using it). The affected
96 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
97 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
98 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
102 *) --strict-warnings no longer enables runtime debugging options
103 such as REF_DEBUG. Instead, debug options are automatically
104 enabled with '--debug' builds.
105 [Andy Polyakov, Emilia Käsper]
107 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
108 have been moved out of the public header files. New functions for managing
109 these have been added.
112 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
113 objects have been moved out of the public header files. New
114 functions for managing these have been added.
117 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
118 have been moved out of the public header files. New functions for managing
119 these have been added.
122 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
123 moved out of the public header files. New functions for managing these
127 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
130 *) Removed the mk1mf build scripts.
133 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
134 it is always safe to #include a header now.
137 *) Removed the aged BC-32 config and all its supporting scripts
140 *) Removed support for Ultrix, Netware, and OS/2.
143 *) Add support for HKDF.
146 *) Add support for blake2b and blake2s
149 *) Added support for "pipelining". Ciphers that have the
150 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
151 encryptions/decryptions simultaneously. There are currently no built-in
152 ciphers with this property but the expectation is that engines will be able
153 to offer it to significantly improve throughput. Support has been extended
154 into libssl so that multiple records for a single connection can be
155 processed in one go (for >=TLS 1.1).
158 *) Added the AFALG engine. This is an async capable engine which is able to
159 offload work to the Linux kernel. In this initial version it only supports
160 AES128-CBC. The kernel must be version 4.1.0 or greater.
163 *) OpenSSL now uses a new threading API. It is no longer necessary to
164 set locking callbacks to use OpenSSL in a multi-threaded environment. There
165 are two supported threading models: pthreads and windows threads. It is
166 also possible to configure OpenSSL at compile time for "no-threads". The
167 old threading API should no longer be used. The functions have been
168 replaced with "no-op" compatibility macros.
169 [Alessandro Ghedini, Matt Caswell]
171 *) Modify behavior of ALPN to invoke callback after SNI/servername
172 callback, such that updates to the SSL_CTX affect ALPN.
175 *) Add SSL_CIPHER queries for authentication and key-exchange.
178 *) Changes to the DEFAULT cipherlist:
179 - Prefer (EC)DHE handshakes over plain RSA.
180 - Prefer AEAD ciphers over legacy ciphers.
181 - Prefer ECDSA over RSA when both certificates are available.
182 - Prefer TLSv1.2 ciphers/PRF.
183 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
187 *) Change the ECC default curve list to be this, in order: x25519,
188 secp256r1, secp521r1, secp384r1.
191 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
192 disabled by default. They can be re-enabled using the
193 enable-weak-ssl-ciphers option to Configure.
196 *) If the server has ALPN configured, but supports no protocols that the
197 client advertises, send a fatal "no_application_protocol" alert.
198 This behaviour is SHALL in RFC 7301, though it isn't universally
199 implemented by other servers.
202 *) Add X25519 support.
203 Integrate support for X25519 into EC library. This includes support
204 for public and private key encoding using the format documented in
205 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
206 OID from that draft, encodes public keys using little endian
207 format in the ECPoint structure and private keys using
208 little endian form in the privateKey field of the ECPrivateKey
209 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
212 Note: the current version supports key generation, public and
213 private key encoding and ECDH key agreement using the EC API.
214 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
218 *) Deprecate SRP_VBASE_get_by_user.
219 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
220 In order to fix an unavoidable memory leak (CVE-2016-0798),
221 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
222 seed, even if the seed is configured.
224 Users should use SRP_VBASE_get1_by_user instead. Note that in
225 SRP_VBASE_get1_by_user, caller must free the returned value. Note
226 also that even though configuring the SRP seed attempts to hide
227 invalid usernames by continuing the handshake with fake
228 credentials, this behaviour is not constant time and no strong
229 guarantees are made that the handshake is indistinguishable from
230 that of a valid user.
233 *) Configuration change; it's now possible to build dynamic engines
234 without having to build shared libraries and vice versa. This
235 only applies to the engines in engines/, those in crypto/engine/
236 will always be built into libcrypto (i.e. "static").
238 Building dynamic engines is enabled by default; to disable, use
239 the configuration option "disable-dynamic-engine".
241 The only requirements for building dynamic engines are the
242 presence of the DSO module and building with position independent
243 code, so they will also automatically be disabled if configuring
244 with "disable-dso" or "disable-pic".
246 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
247 are also taken away from openssl/opensslconf.h, as they are
251 *) Configuration change; if there is a known flag to compile
252 position independent code, it will always be applied on the
253 libcrypto and libssl object files, and never on the application
254 object files. This means other libraries that use routines from
255 libcrypto / libssl can be made into shared libraries regardless
256 of how OpenSSL was configured.
258 If this isn't desirable, the configuration options "disable-pic"
259 or "no-pic" can be used to disable the use of PIC. This will
260 also disable building shared libraries and dynamic engines.
263 *) Removed JPAKE code. It was experimental and has no wide use.
266 *) The INSTALL_PREFIX Makefile variable has been renamed to
267 DESTDIR. That makes for less confusion on what this variable
268 is for. Also, the configuration option --install_prefix is
272 *) Heartbeat for TLS has been removed and is disabled by default
273 for DTLS; configure with enable-heartbeats. Code that uses the
274 old #define's might need to be updated.
275 [Emilia Käsper, Rich Salz]
277 *) Rename REF_CHECK to REF_DEBUG.
280 *) New "unified" build system
282 The "unified" build system is aimed to be a common system for all
283 platforms we support. With it comes new support for VMS.
285 This system builds supports building in a different directory tree
286 than the source tree. It produces one Makefile (for unix family
287 or lookalikes), or one descrip.mms (for VMS).
289 The source of information to make the Makefile / descrip.mms is
290 small files called 'build.info', holding the necessary
291 information for each directory with source to compile, and a
292 template in Configurations, like unix-Makefile.tmpl or
295 We rely heavily on the perl module Text::Template.
298 *) Added support for auto-initialisation and de-initialisation of the library.
299 OpenSSL no longer requires explicit init or deinit routines to be called,
300 except in certain circumstances. See the OPENSSL_init_crypto() and
301 OPENSSL_init_ssl() man pages for further information.
304 *) The arguments to the DTLSv1_listen function have changed. Specifically the
305 "peer" argument is now expected to be a BIO_ADDR object.
307 *) Rewrite of BIO networking library. The BIO library lacked consistent
308 support of IPv6, and adding it required some more extensive
309 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
310 which hold all types of addresses and chains of address information.
311 It also introduces a new API, with functions like BIO_socket,
312 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
313 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
314 have been adapted accordingly.
317 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
321 *) CRIME protection: disable compression by default, even if OpenSSL is
322 compiled with zlib enabled. Applications can still enable compression
323 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
324 using the SSL_CONF library to configure compression.
327 *) The signature of the session callback configured with
328 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
329 was explicitly marked as 'const unsigned char*' instead of
333 *) Always DPURIFY. Remove the use of uninitialized memory in the
334 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
337 *) Removed many obsolete configuration items, including
338 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
339 MD2_CHAR, MD2_INT, MD2_LONG
341 IDEA_SHORT, IDEA_LONG
342 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
343 [Rich Salz, with advice from Andy Polyakov]
345 *) Many BN internals have been moved to an internal header file.
346 [Rich Salz with help from Andy Polyakov]
348 *) Configuration and writing out the results from it has changed.
349 Files such as Makefile include/openssl/opensslconf.h and are now
350 produced through general templates, such as Makefile.in and
351 crypto/opensslconf.h.in and some help from the perl module
354 Also, the center of configuration information is no longer
355 Makefile. Instead, Configure produces a perl module in
356 configdata.pm which holds most of the config data (in the hash
357 table %config), the target data that comes from the target
358 configuration in one of the Configurations/*.conf files (in
362 *) To clarify their intended purposes, the Configure options
363 --prefix and --openssldir change their semantics, and become more
364 straightforward and less interdependent.
366 --prefix shall be used exclusively to give the location INSTALLTOP
367 where programs, scripts, libraries, include files and manuals are
368 going to be installed. The default is now /usr/local.
370 --openssldir shall be used exclusively to give the default
371 location OPENSSLDIR where certificates, private keys, CRLs are
372 managed. This is also where the default openssl.cnf gets
374 If the directory given with this option is a relative path, the
375 values of both the --prefix value and the --openssldir value will
376 be combined to become OPENSSLDIR.
377 The default for --openssldir is INSTALLTOP/ssl.
379 Anyone who uses --openssldir to specify where OpenSSL is to be
380 installed MUST change to use --prefix instead.
383 *) The GOST engine was out of date and therefore it has been removed. An up
384 to date GOST engine is now being maintained in an external repository.
385 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
386 support for GOST ciphersuites (these are only activated if a GOST engine
390 *) EGD is no longer supported by default; use enable-egd when
392 [Ben Kaduk and Rich Salz]
394 *) The distribution now has Makefile.in files, which are used to
395 create Makefile's when Configure is run. *Configure must be run
396 before trying to build now.*
399 *) The return value for SSL_CIPHER_description() for error conditions
403 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
405 Obtaining and performing DNSSEC validation of TLSA records is
406 the application's responsibility. The application provides
407 the TLSA records of its choice to OpenSSL, and these are then
408 used to authenticate the peer.
410 The TLSA records need not even come from DNS. They can, for
411 example, be used to implement local end-entity certificate or
412 trust-anchor "pinning", where the "pin" data takes the form
413 of TLSA records, which can augment or replace verification
414 based on the usual WebPKI public certification authorities.
417 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
418 continues to support deprecated interfaces in default builds.
419 However, applications are strongly advised to compile their
420 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
421 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
422 or the 1.1.0 releases.
424 In environments in which all applications have been ported to
425 not use any deprecated interfaces OpenSSL's Configure script
426 should be used with the --api=1.1.0 option to entirely remove
427 support for the deprecated features from the library and
428 unconditionally disable them in the installed headers.
429 Essentially the same effect can be achieved with the "no-deprecated"
430 argument to Configure, except that this will always restrict
431 the build to just the latest API, rather than a fixed API
434 As applications are ported to future revisions of the API,
435 they should update their compile-time OPENSSL_API_COMPAT define
436 accordingly, but in most cases should be able to continue to
437 compile with later releases.
439 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
440 0x10000000L and 0x00908000L, respectively. However those
441 versions did not support the OPENSSL_API_COMPAT feature, and
442 so applications are not typically tested for explicit support
443 of just the undeprecated features of either release.
446 *) Add support for setting the minimum and maximum supported protocol.
447 It can bet set via the SSL_set_min_proto_version() and
448 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
449 MaxProtcol. It's recommended to use the new APIs to disable
450 protocols instead of disabling individual protocols using
451 SSL_set_options() or SSL_CONF's Protocol. This change also
452 removes support for disabling TLS 1.2 in the OpenSSL TLS
453 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
456 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
459 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
460 and integrates ECDSA and ECDH functionality into EC. Implementations can
461 now redirect key generation and no longer need to convert to or from
464 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
465 include the ec.h header file instead.
468 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
469 ciphers who are no longer supported and drops support the ephemeral RSA key
470 exchange. The LOW ciphers currently doesn't have any ciphers in it.
473 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
474 opaque. For HMAC_CTX, the following constructors and destructors
477 HMAC_CTX *HMAC_CTX_new(void);
478 void HMAC_CTX_free(HMAC_CTX *ctx);
480 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
481 destroy such methods has been added. See EVP_MD_meth_new(3) and
482 EVP_CIPHER_meth_new(3) for documentation.
485 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
486 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
487 EVP_MD_CTX_reset() should be called instead to reinitialise
488 an already created structure.
489 2) For consistency with the majority of our object creators and
490 destructors, EVP_MD_CTX_(create|destroy) were renamed to
491 EVP_MD_CTX_(new|free). The old names are retained as macros
492 for deprecated builds.
495 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
496 cryptographic operations to be performed asynchronously as long as an
497 asynchronous capable engine is used. See the ASYNC_start_job() man page for
498 further details. Libssl has also had this capability integrated with the
499 introduction of the new mode SSL_MODE_ASYNC and associated error
500 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
501 pages. This work was developed in partnership with Intel Corp.
504 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
505 always enabled now. If you want to disable the support you should
506 exclude it using the list of supported ciphers. This also means that the
507 "-no_ecdhe" option has been removed from s_server.
510 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
511 SSL_{CTX_}set1_curves() which can set a list.
514 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
515 curve you want to support using SSL_{CTX_}set1_curves().
518 *) State machine rewrite. The state machine code has been significantly
519 refactored in order to remove much duplication of code and solve issues
520 with the old code (see ssl/statem/README for further details). This change
521 does have some associated API changes. Notably the SSL_state() function
522 has been removed and replaced by SSL_get_state which now returns an
523 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
524 altogether. The previous handshake states defined in ssl.h and ssl3.h have
528 *) All instances of the string "ssleay" in the public API were replaced
529 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
530 Some error codes related to internal RSA_eay API's were renamed.
533 *) The demo files in crypto/threads were moved to demo/threads.
536 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
538 [Matt Caswell, Rich Salz]
540 *) New ASN.1 embed macro.
542 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
543 structure is not allocated: it is part of the parent. That is instead of
551 This reduces memory fragmentation and make it impossible to accidentally
552 set a mandatory field to NULL.
554 This currently only works for some fields specifically a SEQUENCE, CHOICE,
555 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
556 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
560 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
563 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
564 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
565 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
566 DES and RC4 ciphersuites.
569 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
570 This changes the decoding behaviour for some invalid messages,
571 though the change is mostly in the more lenient direction, and
572 legacy behaviour is preserved as much as possible.
575 *) Fix no-stdio build.
576 [ David Woodhouse <David.Woodhouse@intel.com> and also
577 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
579 *) New testing framework
580 The testing framework has been largely rewritten and is now using
581 perl and the perl modules Test::Harness and an extended variant of
582 Test::More called OpenSSL::Test to do its work. All test scripts in
583 test/ have been rewritten into test recipes, and all direct calls to
584 executables in test/Makefile have become individual recipes using the
585 simplified testing OpenSSL::Test::Simple.
587 For documentation on our testing modules, do:
589 perldoc test/testlib/OpenSSL/Test/Simple.pm
590 perldoc test/testlib/OpenSSL/Test.pm
594 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
595 are used; the latter aborts on memory leaks (usually checked on exit).
596 Some undocumented "set malloc, etc., hooks" functions were removed
597 and others were changed. All are now documented.
600 *) In DSA_generate_parameters_ex, if the provided seed is too short,
602 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
604 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
605 from RFC4279, RFC4785, RFC5487, RFC5489.
607 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
608 original RSA_PSK patch.
611 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
612 era flag was never set throughout the codebase (only read). Also removed
613 SSL3_FLAGS_POP_BUFFER which was only used if
614 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
617 *) Changed the default name options in the "ca", "crl", "req" and "x509"
618 to be "oneline" instead of "compat".
621 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
622 not aware of clients that still exhibit this bug, and the workaround
623 hasn't been working properly for a while.
626 *) The return type of BIO_number_read() and BIO_number_written() as well as
627 the corresponding num_read and num_write members in the BIO structure has
628 changed from unsigned long to uint64_t. On platforms where an unsigned
629 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
633 *) Given the pervasive nature of TLS extensions it is inadvisable to run
634 OpenSSL without support for them. It also means that maintaining
635 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
636 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
639 *) Removed support for the two export grade static DH ciphersuites
640 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
641 were newly added (along with a number of other static DH ciphersuites) to
642 1.0.2. However the two export ones have *never* worked since they were
643 introduced. It seems strange in any case to be adding new export
644 ciphersuites, and given "logjam" it also does not seem correct to fix them.
647 *) Version negotiation has been rewritten. In particular SSLv23_method(),
648 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
649 and turned into macros which simply call the new preferred function names
650 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
651 should use the new names instead. Also as part of this change the ssl23.h
652 header file has been removed.
655 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
656 code and the associated standard is no longer considered fit-for-purpose.
659 *) RT2547 was closed. When generating a private key, try to make the
660 output file readable only by the owner. This behavior change might
661 be noticeable when interacting with other software.
663 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
667 *) Added HTTP GET support to the ocsp command.
670 *) Changed default digest for the dgst and enc commands from MD5 to
674 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
677 *) Added support for TLS extended master secret from
678 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
679 initial patch which was a great help during development.
682 *) All libssl internal structures have been removed from the public header
683 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
684 now redundant). Users should not attempt to access internal structures
685 directly. Instead they should use the provided API functions.
688 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
689 Access to deprecated functions can be re-enabled by running config with
690 "enable-deprecated". In addition applications wishing to use deprecated
691 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
692 will, by default, disable some transitive includes that previously existed
693 in the header files (e.g. ec.h will no longer, by default, include bn.h)
696 *) Added support for OCB mode. OpenSSL has been granted a patent license
697 compatible with the OpenSSL license for use of OCB. Details are available
698 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
699 for OCB can be removed by calling config with no-ocb.
702 *) SSLv2 support has been removed. It still supports receiving a SSLv2
703 compatible client hello.
706 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
707 done while fixing the error code for the key-too-small case.
708 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
710 *) CA.sh has been removmed; use CA.pl instead.
713 *) Removed old DES API.
716 *) Remove various unsupported platforms:
722 Sinix/ReliantUNIX RM400
727 16-bit platforms such as WIN16
730 *) Clean up OPENSSL_NO_xxx #define's
731 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
732 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
733 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
734 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
735 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
736 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
737 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
738 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
739 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
740 Remove MS_STATIC; it's a relic from platforms <32 bits.
743 *) Cleaned up dead code
744 Remove all but one '#ifdef undef' which is to be looked at.
747 *) Clean up calling of xxx_free routines.
748 Just like free(), fix most of the xxx_free routines to accept
749 NULL. Remove the non-null checks from callers. Save much code.
752 *) Add secure heap for storage of private keys (when possible).
753 Add BIO_s_secmem(), CBIGNUM, etc.
754 Contributed by Akamai Technologies under our Corporate CLA.
757 *) Experimental support for a new, fast, unbiased prime candidate generator,
758 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
759 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
761 *) New output format NSS in the sess_id command line tool. This allows
762 exporting the session id and the master key in NSS keylog format.
763 [Martin Kaiser <martin@kaiser.cx>]
765 *) Harmonize version and its documentation. -f flag is used to display
767 [mancha <mancha1@zoho.com>]
769 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
770 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
771 [mancha <mancha1@zoho.com>]
773 *) Fix some double frees. These are not thought to be exploitable.
774 [mancha <mancha1@zoho.com>]
776 *) A missing bounds check in the handling of the TLS heartbeat extension
777 can be used to reveal up to 64k of memory to a connected client or
780 Thanks for Neel Mehta of Google Security for discovering this bug and to
781 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
782 preparing the fix (CVE-2014-0160)
783 [Adam Langley, Bodo Moeller]
785 *) Fix for the attack described in the paper "Recovering OpenSSL
786 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
787 by Yuval Yarom and Naomi Benger. Details can be obtained from:
788 http://eprint.iacr.org/2014/140
790 Thanks to Yuval Yarom and Naomi Benger for discovering this
791 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
792 [Yuval Yarom and Naomi Benger]
794 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
795 this fixes a limitation in previous versions of OpenSSL.
798 *) Experimental encrypt-then-mac support.
800 Experimental support for encrypt then mac from
801 draft-gutmann-tls-encrypt-then-mac-02.txt
803 To enable it set the appropriate extension number (0x42 for the test
804 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
806 For non-compliant peers (i.e. just about everything) this should have no
809 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
813 *) Add EVP support for key wrapping algorithms, to avoid problems with
814 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
815 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
816 algorithms and include tests cases.
819 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
823 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
824 MGF1 digest and OAEP label.
827 *) Make openssl verify return errors.
828 [Chris Palmer <palmer@google.com> and Ben Laurie]
830 *) New function ASN1_TIME_diff to calculate the difference between two
831 ASN1_TIME structures or one structure and the current time.
834 *) Update fips_test_suite to support multiple command line options. New
835 test to induce all self test errors in sequence and check expected
839 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
840 sign or verify all in one operation.
843 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
844 test programs and fips_test_suite. Includes functionality to parse
845 the minimal script output of fipsalgest.pl directly.
848 *) Add authorisation parameter to FIPS_module_mode_set().
851 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
854 *) Use separate DRBG fields for internal and external flags. New function
855 FIPS_drbg_health_check() to perform on demand health checking. Add
856 generation tests to fips_test_suite with reduced health check interval to
857 demonstrate periodic health checking. Add "nodh" option to
858 fips_test_suite to skip very slow DH test.
861 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
865 *) More extensive health check for DRBG checking many more failure modes.
866 New function FIPS_selftest_drbg_all() to handle every possible DRBG
867 combination: call this in fips_test_suite.
870 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
871 and POST to handle Dual EC cases.
874 *) Add support for canonical generation of DSA parameter 'g'. See
877 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
878 POST to handle HMAC cases.
881 *) Add functions FIPS_module_version() and FIPS_module_version_text()
882 to return numerical and string versions of the FIPS module number.
885 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
886 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
887 outside the validated module in the FIPS capable OpenSSL.
890 *) Minor change to DRBG entropy callback semantics. In some cases
891 there is no multiple of the block length between min_len and
892 max_len. Allow the callback to return more than max_len bytes
893 of entropy but discard any extra: it is the callback's responsibility
894 to ensure that the extra data discarded does not impact the
895 requested amount of entropy.
898 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
899 information in FIPS186-3, SP800-57 and SP800-131A.
902 *) CCM support via EVP. Interface is very similar to GCM case except we
903 must supply all data in one chunk (i.e. no update, final) and the
904 message length must be supplied if AAD is used. Add algorithm test
908 *) Initial version of POST overhaul. Add POST callback to allow the status
909 of POST to be monitored and/or failures induced. Modify fips_test_suite
910 to use callback. Always run all selftests even if one fails.
913 *) XTS support including algorithm test driver in the fips_gcmtest program.
914 Note: this does increase the maximum key length from 32 to 64 bytes but
915 there should be no binary compatibility issues as existing applications
916 will never use XTS mode.
919 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
920 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
921 performs algorithm blocking for unapproved PRNG types. Also do not
922 set PRNG type in FIPS_mode_set(): leave this to the application.
923 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
924 the standard OpenSSL PRNG: set additional data to a date time vector.
927 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
928 This shouldn't present any incompatibility problems because applications
929 shouldn't be using these directly and any that are will need to rethink
930 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
933 *) Extensive self tests and health checking required by SP800-90 DRBG.
934 Remove strength parameter from FIPS_drbg_instantiate and always
935 instantiate at maximum supported strength.
938 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
941 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
944 *) New function DH_compute_key_padded() to compute a DH key and pad with
945 leading zeroes if needed: this complies with SP800-56A et al.
948 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
949 anything, incomplete, subject to change and largely untested at present.
952 *) Modify fipscanisteronly build option to only build the necessary object
953 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
956 *) Add experimental option FIPSSYMS to give all symbols in
957 fipscanister.o and FIPS or fips prefix. This will avoid
958 conflicts with future versions of OpenSSL. Add perl script
959 util/fipsas.pl to preprocess assembly language source files
960 and rename any affected symbols.
963 *) Add selftest checks and algorithm block of non-fips algorithms in
964 FIPS mode. Remove DES2 from selftests.
967 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
968 return internal method without any ENGINE dependencies. Add new
969 tiny fips sign and verify functions.
972 *) New build option no-ec2m to disable characteristic 2 code.
975 *) New build option "fipscanisteronly". This only builds fipscanister.o
976 and (currently) associated fips utilities. Uses the file Makefile.fips
977 instead of Makefile.org as the prototype.
980 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
981 Update fips_gcmtest to use IV generator.
984 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
985 setting output buffer to NULL. The *Final function must be
986 called although it will not retrieve any additional data. The tag
987 can be set or retrieved with a ctrl. The IV length is by default 12
988 bytes (96 bits) but can be set to an alternative value. If the IV
989 length exceeds the maximum IV length (currently 16 bytes) it cannot be
993 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
994 underlying do_cipher function handles all cipher semantics itself
995 including padding and finalisation. This is useful if (for example)
996 an ENGINE cipher handles block padding itself. The behaviour of
997 do_cipher is subtly changed if this flag is set: the return value
998 is the number of characters written to the output buffer (zero is
999 no longer an error code) or a negative error code. Also if the
1000 input buffer is NULL and length 0 finalisation should be performed.
1003 *) If a candidate issuer certificate is already part of the constructed
1004 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1007 *) Improve forward-security support: add functions
1009 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1010 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1012 for use by SSL/TLS servers; the callback function will be called whenever a
1013 new session is created, and gets to decide whether the session may be
1014 cached to make it resumable (return 0) or not (return 1). (As by the
1015 SSL/TLS protocol specifications, the session_id sent by the server will be
1016 empty to indicate that the session is not resumable; also, the server will
1017 not generate RFC 4507 (RFC 5077) session tickets.)
1019 A simple reasonable callback implementation is to return is_forward_secure.
1020 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1021 by the SSL/TLS server library, indicating whether it can provide forward
1023 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1025 *) New -verify_name option in command line utilities to set verification
1029 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1030 Add CMAC pkey methods.
1033 *) Experimental renegotiation in s_server -www mode. If the client
1034 browses /reneg connection is renegotiated. If /renegcert it is
1035 renegotiated requesting a certificate.
1038 *) Add an "external" session cache for debugging purposes to s_server. This
1039 should help trace issues which normally are only apparent in deployed
1040 multi-process servers.
1043 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1044 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1045 BIO_set_cipher() and some obscure PEM functions were changed so they
1046 can now return an error. The RAND changes required a change to the
1047 RAND_METHOD structure.
1050 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1051 a gcc attribute to warn if the result of a function is ignored. This
1052 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1053 whose return value is often ignored.
1056 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1057 These allow SCTs (signed certificate timestamps) to be requested and
1058 validated when establishing a connection.
1059 [Rob Percival <robpercival@google.com>]
1061 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1063 *) Prevent padding oracle in AES-NI CBC MAC check
1065 A MITM attacker can use a padding oracle attack to decrypt traffic
1066 when the connection uses an AES CBC cipher and the server support
1069 This issue was introduced as part of the fix for Lucky 13 padding
1070 attack (CVE-2013-0169). The padding check was rewritten to be in
1071 constant time by making sure that always the same bytes are read and
1072 compared against either the MAC or padding bytes. But it no longer
1073 checked that there was enough data to have both the MAC and padding
1076 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1080 *) Fix EVP_EncodeUpdate overflow
1082 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1083 Base64 encoding of binary data. If an attacker is able to supply very large
1084 amounts of input data then a length check can overflow resulting in a heap
1087 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1088 the PEM_write_bio* family of functions. These are mainly used within the
1089 OpenSSL command line applications, so any application which processes data
1090 from an untrusted source and outputs it as a PEM file should be considered
1091 vulnerable to this issue. User applications that call these APIs directly
1092 with large amounts of untrusted data may also be vulnerable.
1094 This issue was reported by Guido Vranken.
1098 *) Fix EVP_EncryptUpdate overflow
1100 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1101 is able to supply very large amounts of input data after a previous call to
1102 EVP_EncryptUpdate() with a partial block then a length check can overflow
1103 resulting in a heap corruption. Following an analysis of all OpenSSL
1104 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1105 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1106 the first called function after an EVP_EncryptInit(), and therefore that
1107 specific call must be safe. The second form is where the length passed to
1108 EVP_EncryptUpdate() can be seen from the code to be some small value and
1109 therefore there is no possibility of an overflow. Since all instances are
1110 one of these two forms, it is believed that there can be no overflows in
1111 internal code due to this problem. It should be noted that
1112 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1113 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1114 of these calls have also been analysed too and it is believed there are no
1115 instances in internal usage where an overflow could occur.
1117 This issue was reported by Guido Vranken.
1121 *) Prevent ASN.1 BIO excessive memory allocation
1123 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1124 a short invalid encoding can cause allocation of large amounts of memory
1125 potentially consuming excessive resources or exhausting memory.
1127 Any application parsing untrusted data through d2i BIO functions is
1128 affected. The memory based functions such as d2i_X509() are *not* affected.
1129 Since the memory based functions are used by the TLS library, TLS
1130 applications are not affected.
1132 This issue was reported by Brian Carpenter.
1138 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1139 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1140 in arbitrary stack data being returned in the buffer.
1142 This issue was reported by Guido Vranken.
1146 *) Modify behavior of ALPN to invoke callback after SNI/servername
1147 callback, such that updates to the SSL_CTX affect ALPN.
1150 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1154 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1155 methods are enabled and ssl2 is disabled the methods return NULL.
1158 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1160 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1161 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1162 provide any "EXPORT" or "LOW" strength ciphers.
1165 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1166 is by default disabled at build-time. Builds that are not configured with
1167 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1168 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1169 will need to explicitly call either of:
1171 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1173 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1175 as appropriate. Even if either of those is used, or the application
1176 explicitly uses the version-specific SSLv2_method() or its client and
1177 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1178 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1179 ciphers, and SSLv2 56-bit DES are no longer available.
1183 *) Fix a double-free in DSA code
1185 A double free bug was discovered when OpenSSL parses malformed DSA private
1186 keys and could lead to a DoS attack or memory corruption for applications
1187 that receive DSA private keys from untrusted sources. This scenario is
1190 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1195 *) Disable SRP fake user seed to address a server memory leak.
1197 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1199 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1200 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1201 was changed to ignore the "fake user" SRP seed, even if the seed
1204 Users should use SRP_VBASE_get1_by_user instead. Note that in
1205 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1206 also that even though configuring the SRP seed attempts to hide
1207 invalid usernames by continuing the handshake with fake
1208 credentials, this behaviour is not constant time and no strong
1209 guarantees are made that the handshake is indistinguishable from
1210 that of a valid user.
1214 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1216 In the BN_hex2bn function the number of hex digits is calculated using an
1217 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1218 large values of |i| this can result in |bn_expand| not allocating any
1219 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1220 field as NULL leading to a subsequent NULL ptr deref. For very large values
1221 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1222 In this case memory is allocated to the internal BIGNUM data field, but it
1223 is insufficiently sized leading to heap corruption. A similar issue exists
1224 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1225 is ever called by user applications with very large untrusted hex/dec data.
1226 This is anticipated to be a rare occurrence.
1228 All OpenSSL internal usage of these functions use data that is not expected
1229 to be untrusted, e.g. config file data or application command line
1230 arguments. If user developed applications generate config file data based
1231 on untrusted data then it is possible that this could also lead to security
1232 consequences. This is also anticipated to be rare.
1234 This issue was reported to OpenSSL by Guido Vranken.
1238 *) Fix memory issues in BIO_*printf functions
1240 The internal |fmtstr| function used in processing a "%s" format string in
1241 the BIO_*printf functions could overflow while calculating the length of a
1242 string and cause an OOB read when printing very long strings.
1244 Additionally the internal |doapr_outch| function can attempt to write to an
1245 OOB memory location (at an offset from the NULL pointer) in the event of a
1246 memory allocation failure. In 1.0.2 and below this could be caused where
1247 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1248 could be in processing a very long "%s" format string. Memory leaks can
1251 The first issue may mask the second issue dependent on compiler behaviour.
1252 These problems could enable attacks where large amounts of untrusted data
1253 is passed to the BIO_*printf functions. If applications use these functions
1254 in this way then they could be vulnerable. OpenSSL itself uses these
1255 functions when printing out human-readable dumps of ASN.1 data. Therefore
1256 applications that print this data could be vulnerable if the data is from
1257 untrusted sources. OpenSSL command line applications could also be
1258 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1259 as command line arguments.
1261 Libssl is not considered directly vulnerable. Additionally certificates etc
1262 received via remote connections via libssl are also unlikely to be able to
1263 trigger these issues because of message size limits enforced within libssl.
1265 This issue was reported to OpenSSL Guido Vranken.
1269 *) Side channel attack on modular exponentiation
1271 A side-channel attack was found which makes use of cache-bank conflicts on
1272 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1273 of RSA keys. The ability to exploit this issue is limited as it relies on
1274 an attacker who has control of code in a thread running on the same
1275 hyper-threaded core as the victim thread which is performing decryptions.
1277 This issue was reported to OpenSSL by Yuval Yarom, The University of
1278 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1279 Nadia Heninger, University of Pennsylvania with more information at
1280 http://cachebleed.info.
1284 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1285 if no keysize is specified with default_bits. This fixes an
1286 omission in an earlier change that changed all RSA/DSA key generation
1287 apps to use 2048 bits by default.
1290 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1291 *) DH small subgroups
1293 Historically OpenSSL only ever generated DH parameters based on "safe"
1294 primes. More recently (in version 1.0.2) support was provided for
1295 generating X9.42 style parameter files such as those required for RFC 5114
1296 support. The primes used in such files may not be "safe". Where an
1297 application is using DH configured with parameters based on primes that are
1298 not "safe" then an attacker could use this fact to find a peer's private
1299 DH exponent. This attack requires that the attacker complete multiple
1300 handshakes in which the peer uses the same private DH exponent. For example
1301 this could be used to discover a TLS server's private DH exponent if it's
1302 reusing the private DH exponent or it's using a static DH ciphersuite.
1304 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1305 TLS. It is not on by default. If the option is not set then the server
1306 reuses the same private DH exponent for the life of the server process and
1307 would be vulnerable to this attack. It is believed that many popular
1308 applications do set this option and would therefore not be at risk.
1310 The fix for this issue adds an additional check where a "q" parameter is
1311 available (as is the case in X9.42 based parameters). This detects the
1312 only known attack, and is the only possible defense for static DH
1313 ciphersuites. This could have some performance impact.
1315 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1316 default and cannot be disabled. This could have some performance impact.
1318 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1322 *) SSLv2 doesn't block disabled ciphers
1324 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1325 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1326 been disabled, provided that the SSLv2 protocol was not also disabled via
1329 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1330 and Sebastian Schinzel.
1334 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1336 *) BN_mod_exp may produce incorrect results on x86_64
1338 There is a carry propagating bug in the x86_64 Montgomery squaring
1339 procedure. No EC algorithms are affected. Analysis suggests that attacks
1340 against RSA and DSA as a result of this defect would be very difficult to
1341 perform and are not believed likely. Attacks against DH are considered just
1342 feasible (although very difficult) because most of the work necessary to
1343 deduce information about a private key may be performed offline. The amount
1344 of resources required for such an attack would be very significant and
1345 likely only accessible to a limited number of attackers. An attacker would
1346 additionally need online access to an unpatched system using the target
1347 private key in a scenario with persistent DH parameters and a private
1348 key that is shared between multiple clients. For example this can occur by
1349 default in OpenSSL DHE based SSL/TLS ciphersuites.
1351 This issue was reported to OpenSSL by Hanno Böck.
1355 *) Certificate verify crash with missing PSS parameter
1357 The signature verification routines will crash with a NULL pointer
1358 dereference if presented with an ASN.1 signature using the RSA PSS
1359 algorithm and absent mask generation function parameter. Since these
1360 routines are used to verify certificate signature algorithms this can be
1361 used to crash any certificate verification operation and exploited in a
1362 DoS attack. Any application which performs certificate verification is
1363 vulnerable including OpenSSL clients and servers which enable client
1366 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1370 *) X509_ATTRIBUTE memory leak
1372 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1373 memory. This structure is used by the PKCS#7 and CMS routines so any
1374 application which reads PKCS#7 or CMS data from untrusted sources is
1375 affected. SSL/TLS is not affected.
1377 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1382 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1383 This changes the decoding behaviour for some invalid messages,
1384 though the change is mostly in the more lenient direction, and
1385 legacy behaviour is preserved as much as possible.
1388 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1390 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1392 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1394 *) Alternate chains certificate forgery
1396 During certificate verification, OpenSSL will attempt to find an
1397 alternative certificate chain if the first attempt to build such a chain
1398 fails. An error in the implementation of this logic can mean that an
1399 attacker could cause certain checks on untrusted certificates to be
1400 bypassed, such as the CA flag, enabling them to use a valid leaf
1401 certificate to act as a CA and "issue" an invalid certificate.
1403 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1407 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1409 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1410 incompatibility in the handling of HMAC. The previous ABI has now been
1414 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1416 *) Malformed ECParameters causes infinite loop
1418 When processing an ECParameters structure OpenSSL enters an infinite loop
1419 if the curve specified is over a specially malformed binary polynomial
1422 This can be used to perform denial of service against any
1423 system which processes public keys, certificate requests or
1424 certificates. This includes TLS clients and TLS servers with
1425 client authentication enabled.
1427 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1431 *) Exploitable out-of-bounds read in X509_cmp_time
1433 X509_cmp_time does not properly check the length of the ASN1_TIME
1434 string and can read a few bytes out of bounds. In addition,
1435 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1438 An attacker can use this to craft malformed certificates and CRLs of
1439 various sizes and potentially cause a segmentation fault, resulting in
1440 a DoS on applications that verify certificates or CRLs. TLS clients
1441 that verify CRLs are affected. TLS clients and servers with client
1442 authentication enabled may be affected if they use custom verification
1445 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1446 independently by Hanno Böck.
1450 *) PKCS7 crash with missing EnvelopedContent
1452 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1453 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1454 with missing content and trigger a NULL pointer dereference on parsing.
1456 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1457 structures from untrusted sources are affected. OpenSSL clients and
1458 servers are not affected.
1460 This issue was reported to OpenSSL by Michal Zalewski (Google).
1464 *) CMS verify infinite loop with unknown hash function
1466 When verifying a signedData message the CMS code can enter an infinite loop
1467 if presented with an unknown hash function OID. This can be used to perform
1468 denial of service against any system which verifies signedData messages using
1470 This issue was reported to OpenSSL by Johannes Bauer.
1474 *) Race condition handling NewSessionTicket
1476 If a NewSessionTicket is received by a multi-threaded client when attempting to
1477 reuse a previous ticket then a race condition can occur potentially leading to
1478 a double free of the ticket data.
1482 *) Only support 256-bit or stronger elliptic curves with the
1483 'ecdh_auto' setting (server) or by default (client). Of supported
1484 curves, prefer P-256 (both).
1487 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1489 *) ClientHello sigalgs DoS fix
1491 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1492 invalid signature algorithms extension a NULL pointer dereference will
1493 occur. This can be exploited in a DoS attack against the server.
1495 This issue was was reported to OpenSSL by David Ramos of Stanford
1498 [Stephen Henson and Matt Caswell]
1500 *) Multiblock corrupted pointer fix
1502 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1503 feature only applies on 64 bit x86 architecture platforms that support AES
1504 NI instructions. A defect in the implementation of "multiblock" can cause
1505 OpenSSL's internal write buffer to become incorrectly set to NULL when
1506 using non-blocking IO. Typically, when the user application is using a
1507 socket BIO for writing, this will only result in a failed connection.
1508 However if some other BIO is used then it is likely that a segmentation
1509 fault will be triggered, thus enabling a potential DoS attack.
1511 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1515 *) Segmentation fault in DTLSv1_listen fix
1517 The DTLSv1_listen function is intended to be stateless and processes the
1518 initial ClientHello from many peers. It is common for user code to loop
1519 over the call to DTLSv1_listen until a valid ClientHello is received with
1520 an associated cookie. A defect in the implementation of DTLSv1_listen means
1521 that state is preserved in the SSL object from one invocation to the next
1522 that can lead to a segmentation fault. Errors processing the initial
1523 ClientHello can trigger this scenario. An example of such an error could be
1524 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1527 This issue was reported to OpenSSL by Per Allansson.
1531 *) Segmentation fault in ASN1_TYPE_cmp fix
1533 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1534 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1535 certificate signature algorithm consistency this can be used to crash any
1536 certificate verification operation and exploited in a DoS attack. Any
1537 application which performs certificate verification is vulnerable including
1538 OpenSSL clients and servers which enable client authentication.
1542 *) Segmentation fault for invalid PSS parameters fix
1544 The signature verification routines will crash with a NULL pointer
1545 dereference if presented with an ASN.1 signature using the RSA PSS
1546 algorithm and invalid parameters. Since these routines are used to verify
1547 certificate signature algorithms this can be used to crash any
1548 certificate verification operation and exploited in a DoS attack. Any
1549 application which performs certificate verification is vulnerable including
1550 OpenSSL clients and servers which enable client authentication.
1552 This issue was was reported to OpenSSL by Brian Carpenter.
1556 *) ASN.1 structure reuse memory corruption fix
1558 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1559 memory corruption via an invalid write. Such reuse is and has been
1560 strongly discouraged and is believed to be rare.
1562 Applications that parse structures containing CHOICE or ANY DEFINED BY
1563 components may be affected. Certificate parsing (d2i_X509 and related
1564 functions) are however not affected. OpenSSL clients and servers are
1569 *) PKCS7 NULL pointer dereferences fix
1571 The PKCS#7 parsing code does not handle missing outer ContentInfo
1572 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1573 missing content and trigger a NULL pointer dereference on parsing.
1575 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1576 otherwise parse PKCS#7 structures from untrusted sources are
1577 affected. OpenSSL clients and servers are not affected.
1579 This issue was reported to OpenSSL by Michal Zalewski (Google).
1583 *) DoS via reachable assert in SSLv2 servers fix
1585 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1586 servers that both support SSLv2 and enable export cipher suites by sending
1587 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1589 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1590 (OpenSSL development team).
1594 *) Empty CKE with client auth and DHE fix
1596 If client auth is used then a server can seg fault in the event of a DHE
1597 ciphersuite being selected and a zero length ClientKeyExchange message
1598 being sent by the client. This could be exploited in a DoS attack.
1602 *) Handshake with unseeded PRNG fix
1604 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1605 with an unseeded PRNG. The conditions are:
1606 - The client is on a platform where the PRNG has not been seeded
1607 automatically, and the user has not seeded manually
1608 - A protocol specific client method version has been used (i.e. not
1609 SSL_client_methodv23)
1610 - A ciphersuite is used that does not require additional random data from
1611 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1613 If the handshake succeeds then the client random that has been used will
1614 have been generated from a PRNG with insufficient entropy and therefore the
1615 output may be predictable.
1617 For example using the following command with an unseeded openssl will
1618 succeed on an unpatched platform:
1620 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1624 *) Use After Free following d2i_ECPrivatekey error fix
1626 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1627 could cause a use after free condition. This, in turn, could cause a double
1628 free in several private key parsing functions (such as d2i_PrivateKey
1629 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1630 for applications that receive EC private keys from untrusted
1631 sources. This scenario is considered rare.
1633 This issue was discovered by the BoringSSL project and fixed in their
1638 *) X509_to_X509_REQ NULL pointer deref fix
1640 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1641 the certificate key is invalid. This function is rarely used in practice.
1643 This issue was discovered by Brian Carpenter.
1647 *) Removed the export ciphers from the DEFAULT ciphers
1650 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1652 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1653 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1654 So far those who have to target multiple platforms would compromise
1655 and argue that binary targeting say ARMv5 would still execute on
1656 ARMv8. "Universal" build resolves this compromise by providing
1657 near-optimal performance even on newer platforms.
1660 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1661 (other platforms pending).
1662 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1664 *) Add support for the SignedCertificateTimestampList certificate and
1665 OCSP response extensions from RFC6962.
1668 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1669 for corner cases. (Certain input points at infinity could lead to
1670 bogus results, with non-infinity inputs mapped to infinity too.)
1673 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1674 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1675 common cases are optimized and there still is room for further
1676 improvements. Vector Permutation AES for Altivec is also added.
1679 *) Add support for little-endian ppc64 Linux target.
1680 [Marcelo Cerri (IBM)]
1682 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1683 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1684 are optimized and there still is room for further improvements.
1685 Both 32- and 64-bit modes are supported.
1686 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1688 *) Improved ARMv7 NEON support.
1691 *) Support for SPARC Architecture 2011 crypto extensions, first
1692 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1693 SHA256/512, MD5, GHASH and modular exponentiation.
1694 [Andy Polyakov, David Miller]
1696 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1698 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1700 *) Support for new and upcoming Intel processors, including AVX2,
1701 BMI and SHA ISA extensions. This includes additional "stitched"
1702 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1705 This work was sponsored by Intel Corp.
1708 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1709 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1710 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1713 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1714 this fixes a limitation in previous versions of OpenSSL.
1717 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1718 MGF1 digest and OAEP label.
1721 *) Add EVP support for key wrapping algorithms, to avoid problems with
1722 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1723 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1724 algorithms and include tests cases.
1727 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1729 [Douglas E. Engert, Steve Henson]
1731 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1732 difference in days and seconds between two tm or ASN1_TIME structures.
1735 *) Add -rev test option to s_server to just reverse order of characters
1736 received by client and send back to server. Also prints an abbreviated
1737 summary of the connection parameters.
1740 *) New option -brief for s_client and s_server to print out a brief summary
1741 of connection parameters.
1744 *) Add callbacks for arbitrary TLS extensions.
1745 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1747 *) New option -crl_download in several openssl utilities to download CRLs
1748 from CRLDP extension in certificates.
1751 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1754 *) New function X509_CRL_diff to generate a delta CRL from the difference
1755 of two full CRLs. Add support to "crl" utility.
1758 *) New functions to set lookup_crls function and to retrieve
1759 X509_STORE from X509_STORE_CTX.
1762 *) Print out deprecated issuer and subject unique ID fields in
1766 *) Extend OCSP I/O functions so they can be used for simple general purpose
1767 HTTP as well as OCSP. New wrapper function which can be used to download
1768 CRLs using the OCSP API.
1771 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1774 *) SSL_CONF* functions. These provide a common framework for application
1775 configuration using configuration files or command lines.
1778 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1779 message callback and prints the results. Needs compile time option
1780 "enable-ssl-trace". New options to s_client and s_server to enable
1784 *) New ctrl and macro to retrieve supported points extensions.
1785 Print out extension in s_server and s_client.
1788 *) New functions to retrieve certificate signature and signature
1792 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1796 *) New Suite B modes for TLS code. These use and enforce the requirements
1797 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1798 only use Suite B curves. The Suite B modes can be set by using the
1799 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1802 *) New chain verification flags for Suite B levels of security. Check
1803 algorithms are acceptable when flags are set in X509_verify_cert.
1806 *) Make tls1_check_chain return a set of flags indicating checks passed
1807 by a certificate chain. Add additional tests to handle client
1808 certificates: checks for matching certificate type and issuer name
1812 *) If an attempt is made to use a signature algorithm not in the peer
1813 preference list abort the handshake. If client has no suitable
1814 signature algorithms in response to a certificate request do not
1815 use the certificate.
1818 *) If server EC tmp key is not in client preference list abort handshake.
1821 *) Add support for certificate stores in CERT structure. This makes it
1822 possible to have different stores per SSL structure or one store in
1823 the parent SSL_CTX. Include distinct stores for certificate chain
1824 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1825 to build and store a certificate chain in CERT structure: returing
1826 an error if the chain cannot be built: this will allow applications
1827 to test if a chain is correctly configured.
1829 Note: if the CERT based stores are not set then the parent SSL_CTX
1830 store is used to retain compatibility with existing behaviour.
1834 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1835 mask based on the current session, check mask when sending client
1836 hello and checking the requested ciphersuite.
1839 *) New ctrls to retrieve and set certificate types in a certificate
1840 request message. Print out received values in s_client. If certificate
1841 types is not set with custom values set sensible values based on
1842 supported signature algorithms.
1845 *) Support for distinct client and server supported signature algorithms.
1848 *) Add certificate callback. If set this is called whenever a certificate
1849 is required by client or server. An application can decide which
1850 certificate chain to present based on arbitrary criteria: for example
1851 supported signature algorithms. Add very simple example to s_server.
1852 This fixes many of the problems and restrictions of the existing client
1853 certificate callback: for example you can now clear an existing
1854 certificate and specify the whole chain.
1857 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1858 the certificate can be used for (if anything). Set valid_flags field
1859 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1860 to have similar checks in it.
1862 Add new "cert_flags" field to CERT structure and include a "strict mode".
1863 This enforces some TLS certificate requirements (such as only permitting
1864 certificate signature algorithms contained in the supported algorithms
1865 extension) which some implementations ignore: this option should be used
1866 with caution as it could cause interoperability issues.
1869 *) Update and tidy signature algorithm extension processing. Work out
1870 shared signature algorithms based on preferences and peer algorithms
1871 and print them out in s_client and s_server. Abort handshake if no
1872 shared signature algorithms.
1875 *) Add new functions to allow customised supported signature algorithms
1876 for SSL and SSL_CTX structures. Add options to s_client and s_server
1880 *) New function SSL_certs_clear() to delete all references to certificates
1881 from an SSL structure. Before this once a certificate had been added
1882 it couldn't be removed.
1885 *) Integrate hostname, email address and IP address checking with certificate
1886 verification. New verify options supporting checking in openssl utility.
1889 *) Fixes and wildcard matching support to hostname and email checking
1890 functions. Add manual page.
1891 [Florian Weimer (Red Hat Product Security Team)]
1893 *) New functions to check a hostname email or IP address against a
1894 certificate. Add options x509 utility to print results of checks against
1898 *) Fix OCSP checking.
1899 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1901 *) Initial experimental support for explicitly trusted non-root CAs.
1902 OpenSSL still tries to build a complete chain to a root but if an
1903 intermediate CA has a trust setting included that is used. The first
1904 setting is used: whether to trust (e.g., -addtrust option to the x509
1908 *) Add -trusted_first option which attempts to find certificates in the
1909 trusted store even if an untrusted chain is also supplied.
1912 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1913 platform support for Linux and Android.
1916 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1919 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1920 When in FIPS mode the approved implementations are used as normal,
1921 when not in FIPS mode the internal unapproved versions are used instead.
1922 This means that the FIPS capable OpenSSL isn't forced to use the
1923 (often lower performance) FIPS implementations outside FIPS mode.
1926 *) Transparently support X9.42 DH parameters when calling
1927 PEM_read_bio_DHparameters. This means existing applications can handle
1928 the new parameter format automatically.
1931 *) Initial experimental support for X9.42 DH parameter format: mainly
1932 to support use of 'q' parameter for RFC5114 parameters.
1935 *) Add DH parameters from RFC5114 including test data to dhtest.
1938 *) Support for automatic EC temporary key parameter selection. If enabled
1939 the most preferred EC parameters are automatically used instead of
1940 hardcoded fixed parameters. Now a server just has to call:
1941 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1942 support ECDH and use the most appropriate parameters.
1945 *) Enhance and tidy EC curve and point format TLS extension code. Use
1946 static structures instead of allocation if default values are used.
1947 New ctrls to set curves we wish to support and to retrieve shared curves.
1948 Print out shared curves in s_server. New options to s_server and s_client
1949 to set list of supported curves.
1952 *) New ctrls to retrieve supported signature algorithms and
1953 supported curve values as an array of NIDs. Extend openssl utility
1954 to print out received values.
1957 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1958 between NIDs and the more common NIST names such as "P-256". Enhance
1959 ecparam utility and ECC method to recognise the NIST names for curves.
1962 *) Enhance SSL/TLS certificate chain handling to support different
1963 chains for each certificate instead of one chain in the parent SSL_CTX.
1966 *) Support for fixed DH ciphersuite client authentication: where both
1967 server and client use DH certificates with common parameters.
1970 *) Support for fixed DH ciphersuites: those requiring DH server
1974 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1976 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1977 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1978 X509_CINF_get_signature were reverted post internal team review.
1980 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1982 *) Build fixes for the Windows and OpenVMS platforms
1983 [Matt Caswell and Richard Levitte]
1985 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1987 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1988 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1989 dereference. This could lead to a Denial Of Service attack. Thanks to
1990 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1994 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1995 dtls1_buffer_record function under certain conditions. In particular this
1996 could occur if an attacker sent repeated DTLS records with the same
1997 sequence number but for the next epoch. The memory leak could be exploited
1998 by an attacker in a Denial of Service attack through memory exhaustion.
1999 Thanks to Chris Mueller for reporting this issue.
2003 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2004 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2005 method would be set to NULL which could later result in a NULL pointer
2006 dereference. Thanks to Frank Schmirler for reporting this issue.
2010 *) Abort handshake if server key exchange message is omitted for ephemeral
2013 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2014 reporting this issue.
2018 *) Remove non-export ephemeral RSA code on client and server. This code
2019 violated the TLS standard by allowing the use of temporary RSA keys in
2020 non-export ciphersuites and could be used by a server to effectively
2021 downgrade the RSA key length used to a value smaller than the server
2022 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2023 INRIA or reporting this issue.
2027 *) Fixed issue where DH client certificates are accepted without verification.
2028 An OpenSSL server will accept a DH certificate for client authentication
2029 without the certificate verify message. This effectively allows a client to
2030 authenticate without the use of a private key. This only affects servers
2031 which trust a client certificate authority which issues certificates
2032 containing DH keys: these are extremely rare and hardly ever encountered.
2033 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2038 *) Ensure that the session ID context of an SSL is updated when its
2039 SSL_CTX is updated via SSL_set_SSL_CTX.
2041 The session ID context is typically set from the parent SSL_CTX,
2042 and can vary with the CTX.
2045 *) Fix various certificate fingerprint issues.
2047 By using non-DER or invalid encodings outside the signed portion of a
2048 certificate the fingerprint can be changed without breaking the signature.
2049 Although no details of the signed portion of the certificate can be changed
2050 this can cause problems with some applications: e.g. those using the
2051 certificate fingerprint for blacklists.
2053 1. Reject signatures with non zero unused bits.
2055 If the BIT STRING containing the signature has non zero unused bits reject
2056 the signature. All current signature algorithms require zero unused bits.
2058 2. Check certificate algorithm consistency.
2060 Check the AlgorithmIdentifier inside TBS matches the one in the
2061 certificate signature. NB: this will result in signature failure
2062 errors for some broken certificates.
2064 Thanks to Konrad Kraszewski from Google for reporting this issue.
2066 3. Check DSA/ECDSA signatures use DER.
2068 Reencode DSA/ECDSA signatures and compare with the original received
2069 signature. Return an error if there is a mismatch.
2071 This will reject various cases including garbage after signature
2072 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2073 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2074 (negative or with leading zeroes).
2076 Further analysis was conducted and fixes were developed by Stephen Henson
2077 of the OpenSSL core team.
2082 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2083 results on some platforms, including x86_64. This bug occurs at random
2084 with a very low probability, and is not known to be exploitable in any
2085 way, though its exact impact is difficult to determine. Thanks to Pieter
2086 Wuille (Blockstream) who reported this issue and also suggested an initial
2087 fix. Further analysis was conducted by the OpenSSL development team and
2088 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2089 the OpenSSL core team.
2093 *) Do not resume sessions on the server if the negotiated protocol
2094 version does not match the session's version. Resuming with a different
2095 version, while not strictly forbidden by the RFC, is of questionable
2096 sanity and breaks all known clients.
2097 [David Benjamin, Emilia Käsper]
2099 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2100 early CCS messages during renegotiation. (Note that because
2101 renegotiation is encrypted, this early CCS was not exploitable.)
2104 *) Tighten client-side session ticket handling during renegotiation:
2105 ensure that the client only accepts a session ticket if the server sends
2106 the extension anew in the ServerHello. Previously, a TLS client would
2107 reuse the old extension state and thus accept a session ticket if one was
2108 announced in the initial ServerHello.
2110 Similarly, ensure that the client requires a session ticket if one
2111 was advertised in the ServerHello. Previously, a TLS client would
2112 ignore a missing NewSessionTicket message.
2115 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2117 *) SRTP Memory Leak.
2119 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2120 sends a carefully crafted handshake message, to cause OpenSSL to fail
2121 to free up to 64k of memory causing a memory leak. This could be
2122 exploited in a Denial Of Service attack. This issue affects OpenSSL
2123 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2124 whether SRTP is used or configured. Implementations of OpenSSL that
2125 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2127 The fix was developed by the OpenSSL team.
2131 *) Session Ticket Memory Leak.
2133 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2134 integrity of that ticket is first verified. In the event of a session
2135 ticket integrity check failing, OpenSSL will fail to free memory
2136 causing a memory leak. By sending a large number of invalid session
2137 tickets an attacker could exploit this issue in a Denial Of Service
2142 *) Build option no-ssl3 is incomplete.
2144 When OpenSSL is configured with "no-ssl3" as a build option, servers
2145 could accept and complete a SSL 3.0 handshake, and clients could be
2146 configured to send them.
2148 [Akamai and the OpenSSL team]
2150 *) Add support for TLS_FALLBACK_SCSV.
2151 Client applications doing fallback retries should call
2152 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2154 [Adam Langley, Bodo Moeller]
2156 *) Add additional DigestInfo checks.
2158 Reencode DigestInto in DER and check against the original when
2159 verifying RSA signature: this will reject any improperly encoded
2160 DigestInfo structures.
2162 Note: this is a precautionary measure and no attacks are currently known.
2166 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2168 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2169 SRP code can be overrun an internal buffer. Add sanity check that
2170 g, A, B < N to SRP code.
2172 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2173 Group for discovering this issue.
2177 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2178 TLS 1.0 instead of higher protocol versions when the ClientHello message
2179 is badly fragmented. This allows a man-in-the-middle attacker to force a
2180 downgrade to TLS 1.0 even if both the server and the client support a
2181 higher protocol version, by modifying the client's TLS records.
2183 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2184 researching this issue.
2188 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2189 to a denial of service attack. A malicious server can crash the client
2190 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2191 ciphersuite and sending carefully crafted handshake messages.
2193 Thanks to Felix Gröbert (Google) for discovering and researching this
2198 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2199 to leak memory. This can be exploited through a Denial of Service attack.
2200 Thanks to Adam Langley for discovering and researching this issue.
2204 *) An attacker can force openssl to consume large amounts of memory whilst
2205 processing DTLS handshake messages. This can be exploited through a
2206 Denial of Service attack.
2207 Thanks to Adam Langley for discovering and researching this issue.
2211 *) An attacker can force an error condition which causes openssl to crash
2212 whilst processing DTLS packets due to memory being freed twice. This
2213 can be exploited through a Denial of Service attack.
2214 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2219 *) If a multithreaded client connects to a malicious server using a resumed
2220 session and the server sends an ec point format extension it could write
2221 up to 255 bytes to freed memory.
2223 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2228 *) A malicious server can crash an OpenSSL client with a null pointer
2229 dereference (read) by specifying an SRP ciphersuite even though it was not
2230 properly negotiated with the client. This can be exploited through a
2231 Denial of Service attack.
2233 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2234 discovering and researching this issue.
2238 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2239 X509_name_oneline, X509_name_print_ex et al. to leak some information
2240 from the stack. Applications may be affected if they echo pretty printing
2241 output to the attacker.
2243 Thanks to Ivan Fratric (Google) for discovering this issue.
2245 [Emilia Käsper, and Steve Henson]
2247 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2248 for corner cases. (Certain input points at infinity could lead to
2249 bogus results, with non-infinity inputs mapped to infinity too.)
2252 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2254 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2255 handshake can force the use of weak keying material in OpenSSL
2256 SSL/TLS clients and servers.
2258 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2259 researching this issue. (CVE-2014-0224)
2260 [KIKUCHI Masashi, Steve Henson]
2262 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2263 OpenSSL DTLS client the code can be made to recurse eventually crashing
2266 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2268 [Imre Rad, Steve Henson]
2270 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2271 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2272 client or server. This is potentially exploitable to run arbitrary
2273 code on a vulnerable client or server.
2275 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2276 [Jüri Aedla, Steve Henson]
2278 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2279 are subject to a denial of service attack.
2281 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2282 this issue. (CVE-2014-3470)
2283 [Felix Gröbert, Ivan Fratric, Steve Henson]
2285 *) Harmonize version and its documentation. -f flag is used to display
2287 [mancha <mancha1@zoho.com>]
2289 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2290 in i2d_ECPrivateKey.
2291 [mancha <mancha1@zoho.com>]
2293 *) Fix some double frees. These are not thought to be exploitable.
2294 [mancha <mancha1@zoho.com>]
2296 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2298 *) A missing bounds check in the handling of the TLS heartbeat extension
2299 can be used to reveal up to 64k of memory to a connected client or
2302 Thanks for Neel Mehta of Google Security for discovering this bug and to
2303 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2304 preparing the fix (CVE-2014-0160)
2305 [Adam Langley, Bodo Moeller]
2307 *) Fix for the attack described in the paper "Recovering OpenSSL
2308 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2309 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2310 http://eprint.iacr.org/2014/140
2312 Thanks to Yuval Yarom and Naomi Benger for discovering this
2313 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2314 [Yuval Yarom and Naomi Benger]
2316 *) TLS pad extension: draft-agl-tls-padding-03
2318 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2319 TLS client Hello record length value would otherwise be > 255 and
2320 less that 512 pad with a dummy extension containing zeroes so it
2321 is at least 512 bytes long.
2323 [Adam Langley, Steve Henson]
2325 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2327 *) Fix for TLS record tampering bug. A carefully crafted invalid
2328 handshake could crash OpenSSL with a NULL pointer exception.
2329 Thanks to Anton Johansson for reporting this issues.
2332 *) Keep original DTLS digest and encryption contexts in retransmission
2333 structures so we can use the previous session parameters if they need
2334 to be resent. (CVE-2013-6450)
2337 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2338 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2339 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2340 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2341 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2342 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2343 [Rob Stradling, Adam Langley]
2345 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2347 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2348 supporting platforms or when small records were transferred.
2349 [Andy Polyakov, Steve Henson]
2351 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2353 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2355 This addresses the flaw in CBC record processing discovered by
2356 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2357 at: http://www.isg.rhul.ac.uk/tls/
2359 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2360 Security Group at Royal Holloway, University of London
2361 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2362 Emilia Käsper for the initial patch.
2364 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2366 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2367 ciphersuites which can be exploited in a denial of service attack.
2368 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2369 and detecting this bug and to Wolfgang Ettlinger
2370 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2374 *) Return an error when checking OCSP signatures when key is NULL.
2375 This fixes a DoS attack. (CVE-2013-0166)
2378 *) Make openssl verify return errors.
2379 [Chris Palmer <palmer@google.com> and Ben Laurie]
2381 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2382 the right response is stapled. Also change SSL_get_certificate()
2383 so it returns the certificate actually sent.
2384 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2385 [Rob Stradling <rob.stradling@comodo.com>]
2387 *) Fix possible deadlock when decoding public keys.
2390 *) Don't use TLS 1.0 record version number in initial client hello
2394 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2396 *) Sanity check record length before skipping explicit IV in TLS
2397 1.2, 1.1 and DTLS to fix DoS attack.
2399 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2400 fuzzing as a service testing platform.
2404 *) Initialise tkeylen properly when encrypting CMS messages.
2405 Thanks to Solar Designer of Openwall for reporting this issue.
2408 *) In FIPS mode don't try to use composite ciphers as they are not
2412 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2414 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2415 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2416 mean any application compiled against OpenSSL 1.0.0 headers setting
2417 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2418 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2419 0x10000000L Any application which was previously compiled against
2420 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2421 will need to be recompiled as a result. Letting be results in
2422 inability to disable specifically TLS 1.1 and in client context,
2423 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2426 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2427 disable just protocol X, but all protocols above X *if* there are
2428 protocols *below* X still enabled. In more practical terms it means
2429 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2430 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2431 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2435 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2437 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2438 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2439 in CRYPTO_realloc_clean.
2441 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2442 issue and to Adam Langley <agl@chromium.org> for fixing it.
2444 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2446 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2449 *) Workarounds for some broken servers that "hang" if a client hello
2450 record length exceeds 255 bytes.
2452 1. Do not use record version number > TLS 1.0 in initial client
2453 hello: some (but not all) hanging servers will now work.
2454 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2455 the number of ciphers sent in the client hello. This should be
2456 set to an even number, such as 50, for example by passing:
2457 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2458 Most broken servers should now work.
2459 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2460 TLS 1.2 client support entirely.
2463 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2466 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2468 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2469 STRING form instead of a DigestInfo.
2472 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2473 and the RSA_sign/RSA_verify functions. This was made more apparent when
2474 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2475 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2476 the correct format in RSA_verify so both forms transparently work.
2479 *) Some servers which support TLS 1.0 can choke if we initially indicate
2480 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2481 encrypted premaster secret. As a workaround use the maximum permitted
2482 client version in client hello, this should keep such servers happy
2483 and still work with previous versions of OpenSSL.
2486 *) Add support for TLS/DTLS heartbeats.
2487 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2489 *) Add support for SCTP.
2490 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2492 *) Improved PRNG seeding for VOS.
2493 [Paul Green <Paul.Green@stratus.com>]
2495 *) Extensive assembler packs updates, most notably:
2497 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2498 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2499 - x86_64: bit-sliced AES implementation;
2500 - ARM: NEON support, contemporary platforms optimizations;
2501 - s390x: z196 support;
2502 - *: GHASH and GF(2^m) multiplication implementations;
2506 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2507 (removal of unnecessary code)
2508 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2510 *) Add TLS key material exporter from RFC 5705.
2513 *) Add DTLS-SRTP negotiation from RFC 5764.
2516 *) Add Next Protocol Negotiation,
2517 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2518 disabled with a no-npn flag to config or Configure. Code donated
2520 [Adam Langley <agl@google.com> and Ben Laurie]
2522 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2523 NIST-P256, NIST-P521, with constant-time single point multiplication on
2524 typical inputs. Compiler support for the nonstandard type __uint128_t is
2525 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2526 Code made available under Apache License version 2.0.
2528 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2529 line to include this in your build of OpenSSL, and run "make depend" (or
2530 "make update"). This enables the following EC_METHODs:
2532 EC_GFp_nistp224_method()
2533 EC_GFp_nistp256_method()
2534 EC_GFp_nistp521_method()
2536 EC_GROUP_new_by_curve_name() will automatically use these (while
2537 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2539 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2541 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2542 all platforms. Move ssize_t definition from e_os.h to the public
2543 header file e_os2.h as it now appears in public header file cms.h
2546 *) New -sigopt option to the ca, req and x509 utilities. Additional
2547 signature parameters can be passed using this option and in
2551 *) Add RSA PSS signing function. This will generate and set the
2552 appropriate AlgorithmIdentifiers for PSS based on those in the
2553 corresponding EVP_MD_CTX structure. No application support yet.
2556 *) Support for companion algorithm specific ASN1 signing routines.
2557 New function ASN1_item_sign_ctx() signs a pre-initialised
2558 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2559 the appropriate parameters.
2562 *) Add new algorithm specific ASN1 verification initialisation function
2563 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2564 handling will be the same no matter what EVP_PKEY_METHOD is used.
2565 Add a PSS handler to support verification of PSS signatures: checked
2566 against a number of sample certificates.
2569 *) Add signature printing for PSS. Add PSS OIDs.
2570 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2572 *) Add algorithm specific signature printing. An individual ASN1 method
2573 can now print out signatures instead of the standard hex dump.
2575 More complex signatures (e.g. PSS) can print out more meaningful
2576 information. Include DSA version that prints out the signature
2580 *) Password based recipient info support for CMS library: implementing
2584 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2585 neatly separates the code into cipher and PBE sections and is required
2586 for some algorithms that split PBES2 into separate pieces (such as
2587 password based CMS).
2590 *) Session-handling fixes:
2591 - Fix handling of connections that are resuming with a session ID,
2592 but also support Session Tickets.
2593 - Fix a bug that suppressed issuing of a new ticket if the client
2594 presented a ticket with an expired session.
2595 - Try to set the ticket lifetime hint to something reasonable.
2596 - Make tickets shorter by excluding irrelevant information.
2597 - On the client side, don't ignore renewed tickets.
2598 [Adam Langley, Bodo Moeller (Google)]
2600 *) Fix PSK session representation.
2603 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2605 This work was sponsored by Intel.
2608 *) Add GCM support to TLS library. Some custom code is needed to split
2609 the IV between the fixed (from PRF) and explicit (from TLS record)
2610 portions. This adds all GCM ciphersuites supported by RFC5288 and
2611 RFC5289. Generalise some AES* cipherstrings to include GCM and
2612 add a special AESGCM string for GCM only.
2615 *) Expand range of ctrls for AES GCM. Permit setting invocation
2616 field on decrypt and retrieval of invocation field only on encrypt.
2619 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2620 As required by RFC5289 these ciphersuites cannot be used if for
2621 versions of TLS earlier than 1.2.
2624 *) For FIPS capable OpenSSL interpret a NULL default public key method
2625 as unset and return the appropriate default but do *not* set the default.
2626 This means we can return the appropriate method in applications that
2627 switch between FIPS and non-FIPS modes.
2630 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2631 ENGINE is used then we cannot handle that in the FIPS module so we
2632 keep original code iff non-FIPS operations are allowed.
2635 *) Add -attime option to openssl utilities.
2636 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2638 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2641 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2642 FIPS EC methods unconditionally for now.
2645 *) New build option no-ec2m to disable characteristic 2 code.
2648 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2649 all cases can be covered as some introduce binary incompatibilities.
2652 *) Redirect RSA operations to FIPS module including keygen,
2653 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2656 *) Add similar low level API blocking to ciphers.
2659 *) Low level digest APIs are not approved in FIPS mode: any attempt
2660 to use these will cause a fatal error. Applications that *really* want
2661 to use them can use the private_* version instead.
2664 *) Redirect cipher operations to FIPS module for FIPS builds.
2667 *) Redirect digest operations to FIPS module for FIPS builds.
2670 *) Update build system to add "fips" flag which will link in fipscanister.o
2671 for static and shared library builds embedding a signature if needed.
2674 *) Output TLS supported curves in preference order instead of numerical
2675 order. This is currently hardcoded for the highest order curves first.
2676 This should be configurable so applications can judge speed vs strength.
2679 *) Add TLS v1.2 server support for client authentication.
2682 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2686 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2687 FIPS modules versions.
2690 *) Add TLS v1.2 client side support for client authentication. Keep cache
2691 of handshake records longer as we don't know the hash algorithm to use
2692 until after the certificate request message is received.
2695 *) Initial TLS v1.2 client support. Add a default signature algorithms
2696 extension including all the algorithms we support. Parse new signature
2697 format in client key exchange. Relax some ECC signing restrictions for
2698 TLS v1.2 as indicated in RFC5246.
2701 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2702 to new signature format when needed using client digest preference.
2703 All server ciphersuites should now work correctly in TLS v1.2. No client
2704 support yet and no support for client certificates.
2707 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2708 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2709 ciphersuites. At present only RSA key exchange ciphersuites work with
2710 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2711 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2712 and version checking.
2715 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2716 with this defined it will not be affected by any changes to ssl internal
2717 structures. Add several utility functions to allow openssl application
2718 to work with OPENSSL_NO_SSL_INTERN defined.
2722 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2724 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2727 *) Permit abbreviated handshakes when renegotiating using the function
2728 SSL_renegotiate_abbreviated().
2729 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2731 *) Add call to ENGINE_register_all_complete() to
2732 ENGINE_load_builtin_engines(), so some implementations get used
2733 automatically instead of needing explicit application support.
2736 *) Add support for TLS key exporter as described in RFC5705.
2737 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2739 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2740 a few changes are required:
2742 Add SSL_OP_NO_TLSv1_1 flag.
2743 Add TLSv1_1 methods.
2744 Update version checking logic to handle version 1.1.
2745 Add explicit IV handling (ported from DTLS code).
2746 Add command line options to s_client/s_server.
2749 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2751 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2752 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2753 content decryption and always return the same error. Note: this attack
2754 needs on average 2^20 messages so it only affects automated senders. The
2755 old behaviour can be reenabled in the CMS code by setting the
2756 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2757 an MMA defence is not necessary.
2758 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2759 this issue. (CVE-2012-0884)
2762 *) Fix CVE-2011-4619: make sure we really are receiving a
2763 client hello before rejecting multiple SGC restarts. Thanks to
2764 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2767 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2769 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2770 Thanks to Antonio Martin, Enterprise Secure Access Research and
2771 Development, Cisco Systems, Inc. for discovering this bug and
2772 preparing a fix. (CVE-2012-0050)
2775 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2777 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2778 of the Vaudenay padding oracle attack on CBC mode encryption
2779 which enables an efficient plaintext recovery attack against
2780 the OpenSSL implementation of DTLS. Their attack exploits timing
2781 differences arising during decryption processing. A research
2782 paper describing this attack can be found at:
2783 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2784 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2785 Security Group at Royal Holloway, University of London
2786 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2787 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2788 for preparing the fix. (CVE-2011-4108)
2789 [Robin Seggelmann, Michael Tuexen]
2791 *) Clear bytes used for block padding of SSL 3.0 records.
2793 [Adam Langley (Google)]
2795 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2796 Kadianakis <desnacked@gmail.com> for discovering this issue and
2797 Adam Langley for preparing the fix. (CVE-2011-4619)
2798 [Adam Langley (Google)]
2800 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2801 [Andrey Kulikov <amdeich@gmail.com>]
2803 *) Prevent malformed RFC3779 data triggering an assertion failure.
2804 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2805 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2806 [Rob Austein <sra@hactrn.net>]
2808 *) Improved PRNG seeding for VOS.
2809 [Paul Green <Paul.Green@stratus.com>]
2811 *) Fix ssl_ciph.c set-up race.
2812 [Adam Langley (Google)]
2814 *) Fix spurious failures in ecdsatest.c.
2815 [Emilia Käsper (Google)]
2817 *) Fix the BIO_f_buffer() implementation (which was mixing different
2818 interpretations of the '..._len' fields).
2819 [Adam Langley (Google)]
2821 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2822 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2823 threads won't reuse the same blinding coefficients.
2825 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2826 lock to call BN_BLINDING_invert_ex, and avoids one use of
2827 BN_BLINDING_update for each BN_BLINDING structure (previously,
2828 the last update always remained unused).
2829 [Emilia Käsper (Google)]
2831 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2832 [Bob Buckholz (Google)]
2834 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2836 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2837 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2838 [Kaspar Brand <ossl@velox.ch>]
2840 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2841 for multi-threaded use of ECDH. (CVE-2011-3210)
2842 [Adam Langley (Google)]
2844 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2847 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2848 signature public key algorithm by using OID xref utilities instead.
2849 Before this you could only use some ECC ciphersuites with SHA1 only.
2852 *) Add protection against ECDSA timing attacks as mentioned in the paper
2853 by Billy Bob Brumley and Nicola Tuveri, see:
2855 http://eprint.iacr.org/2011/232.pdf
2857 [Billy Bob Brumley and Nicola Tuveri]
2859 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2861 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2862 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2864 *) Fix bug in string printing code: if *any* escaping is enabled we must
2865 escape the escape character (backslash) or the resulting string is
2869 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2871 *) Disable code workaround for ancient and obsolete Netscape browsers
2872 and servers: an attacker can use it in a ciphersuite downgrade attack.
2873 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2876 *) Fixed J-PAKE implementation error, originally discovered by
2877 Sebastien Martini, further info and confirmation from Stefan
2878 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2881 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2883 *) Fix extension code to avoid race conditions which can result in a buffer
2884 overrun vulnerability: resumed sessions must not be modified as they can
2885 be shared by multiple threads. CVE-2010-3864
2888 *) Fix WIN32 build system to correctly link an ENGINE directory into
2892 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2894 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2896 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2898 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2900 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2901 context. The operation can be customised via the ctrl mechanism in
2902 case ENGINEs want to include additional functionality.
2905 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2908 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2909 output hashes compatible with older versions of OpenSSL.
2910 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2912 *) Fix compression algorithm handling: if resuming a session use the
2913 compression algorithm of the resumed session instead of determining
2914 it from client hello again. Don't allow server to change algorithm.
2917 *) Add load_crls() function to apps tidying load_certs() too. Add option
2918 to verify utility to allow additional CRLs to be included.
2921 *) Update OCSP request code to permit adding custom headers to the request:
2922 some responders need this.
2925 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2927 [Julia Lawall <julia@diku.dk>]
2929 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2930 needlessly dereferenced structures, used obsolete functions and
2931 didn't handle all updated verify codes correctly.
2934 *) Disable MD2 in the default configuration.
2937 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2938 indicate the initial BIO being pushed or popped. This makes it possible
2939 to determine whether the BIO is the one explicitly called or as a result
2940 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2941 it handles reference counts correctly and doesn't zero out the I/O bio
2942 when it is not being explicitly popped. WARNING: applications which
2943 included workarounds for the old buggy behaviour will need to be modified
2944 or they could free up already freed BIOs.
2947 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2948 renaming to all platforms (within the 0.9.8 branch, this was
2949 done conditionally on Netware platforms to avoid a name clash).
2950 [Guenter <lists@gknw.net>]
2952 *) Add ECDHE and PSK support to DTLS.
2953 [Michael Tuexen <tuexen@fh-muenster.de>]
2955 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2959 *) Add "missing" function EVP_MD_flags() (without this the only way to
2960 retrieve a digest flags is by accessing the structure directly. Update
2961 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2962 or cipher is registered as in the "from" argument. Print out all
2963 registered digests in the dgst usage message instead of manually
2964 attempting to work them out.
2967 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2968 this allows the use of compression and extensions. Change default cipher
2969 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2970 by default unless an application cipher string requests it.
2973 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2974 key ids to find matching certificates and keys but some PKCS#12 files
2975 don't follow the (somewhat unwritten) rules and this strategy fails.
2976 Now just gather all certificates together and the first private key
2977 then look for the first certificate that matches the key.
2980 *) Support use of registered digest and cipher names for dgst and cipher
2981 commands instead of having to add each one as a special case. So now
2988 openssl dgst -sha256 foo
2990 and this works for ENGINE based algorithms too.
2994 *) Update Gost ENGINE to support parameter files.
2995 [Victor B. Wagner <vitus@cryptocom.ru>]
2997 *) Support GeneralizedTime in ca utility.
2998 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3000 *) Enhance the hash format used for certificate directory links. The new
3001 form uses the canonical encoding (meaning equivalent names will work
3002 even if they aren't identical) and uses SHA1 instead of MD5. This form
3003 is incompatible with the older format and as a result c_rehash should
3004 be used to rebuild symbolic links.
3007 *) Make PKCS#8 the default write format for private keys, replacing the
3008 traditional format. This form is standardised, more secure and doesn't
3009 include an implicit MD5 dependency.
3012 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3013 committed to OpenSSL should pass this lot as a minimum.
3016 *) Add session ticket override functionality for use by EAP-FAST.
3017 [Jouni Malinen <j@w1.fi>]
3019 *) Modify HMAC functions to return a value. Since these can be implemented
3020 in an ENGINE errors can occur.
3023 *) Type-checked OBJ_bsearch_ex.
3026 *) Type-checked OBJ_bsearch. Also some constification necessitated
3027 by type-checking. Still to come: TXT_DB, bsearch(?),
3028 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3032 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3033 seconds to a tm structure directly, instead of going through OS
3034 specific date routines. This avoids any issues with OS routines such
3035 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3036 and X509_time_adj_ex() to cover the extended range. The existing
3037 X509_time_adj() is still usable and will no longer have any date issues.
3040 *) Delta CRL support. New use deltas option which will attempt to locate
3041 and search any appropriate delta CRLs available.
3043 This work was sponsored by Google.
3046 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3047 code and add additional score elements. Validate alternate CRL paths
3048 as part of the CRL checking and indicate a new error "CRL path validation
3049 error" in this case. Applications wanting additional details can use
3050 the verify callback and check the new "parent" field. If this is not
3051 NULL CRL path validation is taking place. Existing applications wont
3052 see this because it requires extended CRL support which is off by
3055 This work was sponsored by Google.
3058 *) Support for freshest CRL extension.
3060 This work was sponsored by Google.
3063 *) Initial indirect CRL support. Currently only supported in the CRLs
3064 passed directly and not via lookup. Process certificate issuer
3065 CRL entry extension and lookup CRL entries by bother issuer name
3066 and serial number. Check and process CRL issuer entry in IDP extension.
3068 This work was sponsored by Google.
3071 *) Add support for distinct certificate and CRL paths. The CRL issuer
3072 certificate is validated separately in this case. Only enabled if
3073 an extended CRL support flag is set: this flag will enable additional
3074 CRL functionality in future.
3076 This work was sponsored by Google.
3079 *) Add support for policy mappings extension.
3081 This work was sponsored by Google.
3084 *) Fixes to pathlength constraint, self issued certificate handling,
3085 policy processing to align with RFC3280 and PKITS tests.
3087 This work was sponsored by Google.
3090 *) Support for name constraints certificate extension. DN, email, DNS
3091 and URI types are currently supported.
3093 This work was sponsored by Google.
3096 *) To cater for systems that provide a pointer-based thread ID rather
3097 than numeric, deprecate the current numeric thread ID mechanism and
3098 replace it with a structure and associated callback type. This
3099 mechanism allows a numeric "hash" to be extracted from a thread ID in
3100 either case, and on platforms where pointers are larger than 'long',
3101 mixing is done to help ensure the numeric 'hash' is usable even if it
3102 can't be guaranteed unique. The default mechanism is to use "&errno"
3103 as a pointer-based thread ID to distinguish between threads.
3105 Applications that want to provide their own thread IDs should now use
3106 CRYPTO_THREADID_set_callback() to register a callback that will call
3107 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3109 Note that ERR_remove_state() is now deprecated, because it is tied
3110 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3111 to free the current thread's error state should be replaced by
3112 ERR_remove_thread_state(NULL).
3114 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3115 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3116 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3117 application was previously providing a numeric thread callback that
3118 was inappropriate for distinguishing threads, then uniqueness might
3119 have been obtained with &errno that happened immediately in the
3120 intermediate development versions of OpenSSL; this is no longer the
3121 case, the numeric thread callback will now override the automatic use
3123 [Geoff Thorpe, with help from Bodo Moeller]
3125 *) Initial support for different CRL issuing certificates. This covers a
3126 simple case where the self issued certificates in the chain exist and
3127 the real CRL issuer is higher in the existing chain.
3129 This work was sponsored by Google.
3132 *) Removed effectively defunct crypto/store from the build.
3135 *) Revamp of STACK to provide stronger type-checking. Still to come:
3136 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3137 ASN1_STRING, CONF_VALUE.
3140 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3141 RAM on SSL connections. This option can save about 34k per idle SSL.
3144 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3145 STACK, TXT_DB, bsearch, qsort.
3148 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3149 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3150 support for data, signedData, compressedData, digestedData and
3151 encryptedData, envelopedData types included. Scripts to check against
3152 RFC4134 examples draft and interop and consistency checks of many
3153 content types and variants.
3156 *) Add options to enc utility to support use of zlib compression BIO.
3159 *) Extend mk1mf to support importing of options and assembly language
3160 files from Configure script, currently only included in VC-WIN32.
3161 The assembly language rules can now optionally generate the source
3162 files from the associated perl scripts.
3165 *) Implement remaining functionality needed to support GOST ciphersuites.
3166 Interop testing has been performed using CryptoPro implementations.
3167 [Victor B. Wagner <vitus@cryptocom.ru>]
3169 *) s390x assembler pack.
3172 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3176 *) Implement Opaque PRF Input TLS extension as specified in
3177 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3178 official specification yet and no extension type assignment by
3179 IANA exists, this extension (for now) will have to be explicitly
3180 enabled when building OpenSSL by providing the extension number
3181 to use. For example, specify an option
3183 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3185 to the "config" or "Configure" script to enable the extension,
3186 assuming extension number 0x9527 (which is a completely arbitrary
3187 and unofficial assignment based on the MD5 hash of the Internet
3188 Draft). Note that by doing so, you potentially lose
3189 interoperability with other TLS implementations since these might
3190 be using the same extension number for other purposes.
3192 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3193 opaque PRF input value to use in the handshake. This will create
3194 an interal copy of the length-'len' string at 'src', and will
3195 return non-zero for success.
3197 To get more control and flexibility, provide a callback function
3200 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3201 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3205 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3208 Callback function 'cb' will be called in handshakes, and is
3209 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3210 Argument 'arg' is for application purposes (the value as given to
3211 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3212 be provided to the callback function). The callback function
3213 has to return non-zero to report success: usually 1 to use opaque
3214 PRF input just if possible, or 2 to enforce use of the opaque PRF
3215 input. In the latter case, the library will abort the handshake
3216 if opaque PRF input is not successfully negotiated.
3218 Arguments 'peerinput' and 'len' given to the callback function
3219 will always be NULL and 0 in the case of a client. A server will
3220 see the client's opaque PRF input through these variables if
3221 available (NULL and 0 otherwise). Note that if the server
3222 provides an opaque PRF input, the length must be the same as the
3223 length of the client's opaque PRF input.
3225 Note that the callback function will only be called when creating
3226 a new session (session resumption can resume whatever was
3227 previously negotiated), and will not be called in SSL 2.0
3228 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3229 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3230 for applications that need to enforce opaque PRF input.
3234 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3237 [Victor B. Wagner <vitus@cryptocom.ru>]
3239 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3240 RFC4507bis. The encrypted ticket format is an encrypted encoded
3241 SSL_SESSION structure, that way new session features are automatically
3244 If a client application caches session in an SSL_SESSION structure
3245 support is transparent because tickets are now stored in the encoded
3248 The SSL_CTX structure automatically generates keys for ticket
3249 protection in servers so again support should be possible
3250 with no application modification.
3252 If a client or server wishes to disable RFC4507 support then the option
3253 SSL_OP_NO_TICKET can be set.
3255 Add a TLS extension debugging callback to allow the contents of any client
3256 or server extensions to be examined.
3258 This work was sponsored by Google.
3261 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3262 OpenSSL should now compile cleanly on gcc 4.2
3263 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3265 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3266 support including streaming MAC support: this is required for GOST
3267 ciphersuite support.
3268 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3270 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3271 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3272 to output in BER and PEM format.
3275 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3276 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3277 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3278 ENGINE support for HMAC keys which are unextractable. New -mac and
3279 -macopt options to dgst utility.
3282 *) New option -sigopt to dgst utility. Update dgst to use
3283 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3284 alternative signing parameters such as X9.31 or PSS in the dgst
3288 *) Change ssl_cipher_apply_rule(), the internal function that does
3289 the work each time a ciphersuite string requests enabling
3290 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3291 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3292 the order of disabled ciphersuites such that those ciphersuites
3293 that most recently went from enabled to disabled not only stay
3294 in order with respect to each other, but also have higher priority
3295 than other disabled ciphersuites the next time ciphersuites are
3298 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3299 the same ciphersuites as with "HIGH" alone, but in a specific
3300 order where the PSK ciphersuites come first (since they are the
3301 most recently disabled ciphersuites when "HIGH" is parsed).
3303 Also, change ssl_create_cipher_list() (using this new
3304 funcionality) such that between otherwise identical
3305 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3309 *) Change ssl_create_cipher_list() so that it automatically
3310 arranges the ciphersuites in reasonable order before starting
3311 to process the rule string. Thus, the definition for "DEFAULT"
3312 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3313 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3314 This makes it much easier to arrive at a reasonable default order
3315 in applications for which anonymous ciphers are OK (meaning
3316 that you can't actually use DEFAULT).
3317 [Bodo Moeller; suggested by Victor Duchovni]
3319 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3320 processing) into multiple integers instead of setting
3321 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3322 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3323 (These masks as well as the individual bit definitions are hidden
3324 away into the non-exported interface ssl/ssl_locl.h, so this
3325 change to the definition of the SSL_CIPHER structure shouldn't
3326 affect applications.) This give us more bits for each of these
3327 categories, so there is no longer a need to coagulate AES128 and
3328 AES256 into a single algorithm bit, and to coagulate Camellia128
3329 and Camellia256 into a single algorithm bit, which has led to all
3332 Thus, among other things, the kludge introduced in 0.9.7m and
3333 0.9.8e for masking out AES256 independently of AES128 or masking
3334 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3336 With the change, we also introduce new ciphersuite aliases that
3337 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3341 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3342 Use the leftmost N bytes of the signature input if the input is
3343 larger than the prime q (with N being the size in bytes of q).
3346 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3347 it yet and it is largely untested.
3350 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3353 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3354 some compilers (gcc 4.2 and later) reject their use. Safestack is
3355 reimplemented. Update ASN1 to avoid use of legacy functions.
3358 *) Win32/64 targets are linked with Winsock2.
3361 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3362 to external functions. This can be used to increase CRL handling
3363 efficiency especially when CRLs are very large by (for example) storing
3364 the CRL revoked certificates in a database.
3367 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3368 new CRLs added to a directory can be used. New command line option
3369 -verify_return_error to s_client and s_server. This causes real errors
3370 to be returned by the verify callback instead of carrying on no matter
3371 what. This reflects the way a "real world" verify callback would behave.
3374 *) GOST engine, supporting several GOST algorithms and public key formats.
3375 Kindly donated by Cryptocom.
3378 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3379 partitioned by DP are handled but no indirect CRL or reason partitioning
3380 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3381 selected via a scoring technique which handles IDP and AKID in CRLs.
3384 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3385 will ultimately be used for all verify operations: this will remove the
3386 X509_STORE dependency on certificate verification and allow alternative
3387 lookup methods. X509_STORE based implementations of these two callbacks.
3390 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3391 Modify get_crl() to find a valid (unexpired) CRL if possible.
3394 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3395 this would be called X509_CRL_cmp() but that name is already used by
3396 a function that just compares CRL issuer names. Cache several CRL
3397 extensions in X509_CRL structure and cache CRLDP in X509.
3400 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3401 this maps equivalent X509_NAME structures into a consistent structure.
3402 Name comparison can then be performed rapidly using memcmp().
3405 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3409 *) Allow digests to supply their own micalg string for S/MIME type using
3410 the ctrl EVP_MD_CTRL_MICALG.
3413 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3414 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3415 ctrl. It can then customise the structure before and/or after signing
3419 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3420 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3421 to free up any added signature OIDs.
3424 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3425 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3426 digest and cipher tables. New options added to openssl utility:
3427 list-message-digest-algorithms and list-cipher-algorithms.
3430 *) Change the array representation of binary polynomials: the list
3431 of degrees of non-zero coefficients is now terminated with -1.
3432 Previously it was terminated with 0, which was also part of the
3433 value; thus, the array representation was not applicable to
3434 polynomials where t^0 has coefficient zero. This change makes
3435 the array representation useful in a more general context.
3438 *) Various modifications and fixes to SSL/TLS cipher string
3439 handling. For ECC, the code now distinguishes between fixed ECDH
3440 with RSA certificates on the one hand and with ECDSA certificates
3441 on the other hand, since these are separate ciphersuites. The
3442 unused code for Fortezza ciphersuites has been removed.
3444 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3445 (not "ECDHE"). For consistency with the code for DH
3446 certificates, use of ECDH certificates is now considered ECDH
3447 authentication, not RSA or ECDSA authentication (the latter is
3448 merely the CA's signing algorithm and not actively used in the
3451 The temporary ciphersuite alias "ECCdraft" is no longer
3452 available, and ECC ciphersuites are no longer excluded from "ALL"
3453 and "DEFAULT". The following aliases now exist for RFC 4492
3454 ciphersuites, most of these by analogy with the DH case:
3456 kECDHr - ECDH cert, signed with RSA
3457 kECDHe - ECDH cert, signed with ECDSA
3458 kECDH - ECDH cert (signed with either RSA or ECDSA)
3459 kEECDH - ephemeral ECDH
3460 ECDH - ECDH cert or ephemeral ECDH
3466 AECDH - anonymous ECDH
3467 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3471 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3472 Use correct micalg parameters depending on digest(s) in signed message.
3475 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3476 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3479 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3480 an engine to register a method. Add ENGINE lookups for methods and
3481 functional reference processing.
3484 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3485 EVP_{Sign,Verify}* which allow an application to customise the signature
3489 *) New -resign option to smime utility. This adds one or more signers
3490 to an existing PKCS#7 signedData structure. Also -md option to use an
3491 alternative message digest algorithm for signing.
3494 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3495 create PKCS7 structures containing multiple signers. Update smime
3496 application to support multiple signers.
3499 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3503 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3504 Reorganize PBE internals to lookup from a static table using NIDs,
3505 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3506 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3507 PRF which will be automatically used with PBES2.
3510 *) Replace the algorithm specific calls to generate keys in "req" with the
3514 *) Update PKCS#7 enveloped data routines to use new API. This is now
3515 supported by any public key method supporting the encrypt operation. A
3516 ctrl is added to allow the public key algorithm to examine or modify
3517 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3521 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3522 a default digest type to use. In most cases this will be SHA1 but some
3523 algorithms (such as GOST) need to specify an alternative digest. The
3524 return value indicates how strong the preference is 1 means optional and
3525 2 is mandatory (that is it is the only supported type). Modify
3526 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3527 use the default md. Update openssl utilities to use the default digest
3528 type for signing if it is not explicitly indicated.
3531 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3532 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3533 signing method from the key type. This effectively removes the link
3534 between digests and public key types.
3537 *) Add an OID cross reference table and utility functions. Its purpose is to
3538 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3539 rsaEncryption. This will allow some of the algorithm specific hackery
3540 needed to use the correct OID to be removed.
3543 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3544 structures for PKCS7_sign(). They are now set up by the relevant public
3548 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3551 *) Add support for key derivation (agreement) in the API, DH method and
3555 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3556 public and private key formats. As a side effect these add additional
3557 command line functionality not previously available: DSA signatures can be
3558 generated and verified using pkeyutl and DH key support and generation in
3563 [Oliver Tappe <zooey@hirschkaefer.de>]
3565 *) New make target "install_html_docs" installs HTML renditions of the
3567 [Oliver Tappe <zooey@hirschkaefer.de>]
3569 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3570 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3571 support key and parameter generation and add initial key generation
3572 functionality for RSA.
3575 *) Add functions for main EVP_PKEY_method operations. The undocumented
3576 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3577 EVP_PKEY_{encrypt,decrypt}_old.
3580 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3581 key API, doesn't do much yet.
3584 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3585 public key algorithms. New option to openssl utility:
3586 "list-public-key-algorithms" to print out info.
3589 *) Implement the Supported Elliptic Curves Extension for
3590 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3593 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3594 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3597 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3598 utilities such as rsa, dsa, dsaparam etc except they process any key
3602 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3603 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3604 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3608 *) Initial support for pluggable public key ASN1.
3609 De-spaghettify the public key ASN1 handling. Move public and private
3610 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3611 algorithm specific handling to a single module within the relevant
3612 algorithm directory. Add functions to allow (near) opaque processing
3613 of public and private key structures.
3616 *) Implement the Supported Point Formats Extension for
3617 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3620 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3621 for the psk identity [hint] and the psk callback functions to the
3622 SSL_SESSION, SSL and SSL_CTX structure.
3625 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3629 SSL_CTX_use_psk_identity_hint
3630 SSL_get_psk_identity_hint
3631 SSL_get_psk_identity
3632 SSL_use_psk_identity_hint
3634 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3636 *) Add RFC 3161 compliant time stamp request creation, response generation
3637 and response verification functionality.
3638 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3640 *) Add initial support for TLS extensions, specifically for the server_name
3641 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3642 have new members for a host name. The SSL data structure has an
3643 additional member SSL_CTX *initial_ctx so that new sessions can be
3644 stored in that context to allow for session resumption, even after the
3645 SSL has been switched to a new SSL_CTX in reaction to a client's
3646 server_name extension.
3648 New functions (subject to change):
3650 SSL_get_servername()
3651 SSL_get_servername_type()
3654 New CTRL codes and macros (subject to change):
3656 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3657 - SSL_CTX_set_tlsext_servername_callback()
3658 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3659 - SSL_CTX_set_tlsext_servername_arg()
3660 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3662 openssl s_client has a new '-servername ...' option.
3664 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3665 '-key2 ...', '-servername_fatal' (subject to change). This allows
3666 testing the HostName extension for a specific single host name ('-cert'
3667 and '-key' remain fallbacks for handshakes without HostName
3668 negotiation). If the unrecognized_name alert has to be sent, this by
3669 default is a warning; it becomes fatal with the '-servername_fatal'
3672 [Peter Sylvester, Remy Allais, Christophe Renou]
3674 *) Whirlpool hash implementation is added.
3677 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3678 bn(64,32). Because of instruction set limitations it doesn't have
3679 any negative impact on performance. This was done mostly in order
3680 to make it possible to share assembler modules, such as bn_mul_mont
3681 implementations, between 32- and 64-bit builds without hassle.
3684 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3685 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3689 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3690 dedicated Montgomery multiplication procedure, is introduced.
3691 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3692 "64-bit" performance on certain 32-bit targets.
3695 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3696 in SSL structures. New SSL ctrl to set maximum send fragment size.
3697 Save memory by seeting the I/O buffer sizes dynamically instead of
3698 using the maximum available value.
3701 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3702 in addition to the text details.
3705 *) Very, very preliminary EXPERIMENTAL support for printing of general
3706 ASN1 structures. This currently produces rather ugly output and doesn't
3707 handle several customised structures at all.
3710 *) Integrated support for PVK file format and some related formats such
3711 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3712 these in the 'rsa' and 'dsa' utilities.
3715 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3718 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3719 place for the (very old) "NETSCAPE" format certificates which are now
3720 handled using new ASN1 code equivalents.
3723 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3724 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3725 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3728 *) Modify CRL distribution points extension code to print out previously
3729 unsupported fields. Enhance extension setting code to allow setting of
3733 *) Add print and set support for Issuing Distribution Point CRL extension.
3736 *) Change 'Configure' script to enable Camellia by default.
3739 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3741 *) When rejecting SSL/TLS records due to an incorrect version number, never
3742 update s->server with a new major version number. As of
3743 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3744 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3745 the previous behavior could result in a read attempt at NULL when
3746 receiving specific incorrect SSL/TLS records once record payload
3747 protection is active. (CVE-2010-0740)
3748 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3750 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3751 could be crashed if the relevant tables were not present (e.g. chrooted).
3752 [Tomas Hoger <thoger@redhat.com>]
3754 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3756 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3757 [Martin Olsson, Neel Mehta]
3759 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3760 accommodate for stack sorting, always a write lock!).
3763 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3764 excessive delays in the RAND_poll(): over a minute. As a workaround
3765 include a time check in the inner Heap32Next loop too.
3768 *) The code that handled flushing of data in SSL/TLS originally used the
3769 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3770 the problem outlined in PR#1949. The fix suggested there however can
3771 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3772 of Apache). So instead simplify the code to flush unconditionally.
3773 This should be fine since flushing with no data to flush is a no op.
3776 *) Handle TLS versions 2.0 and later properly and correctly use the
3777 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3778 off ancient servers have a habit of sticking around for a while...
3781 *) Modify compression code so it frees up structures without using the
3782 ex_data callbacks. This works around a problem where some applications
3783 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
3784 restarting) then use compression (e.g. SSL with compression) later.
3785 This results in significant per-connection memory leaks and
3786 has caused some security issues including CVE-2008-1678 and
3790 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
3791 change when encrypting or decrypting.
3794 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
3795 connect and renegotiate with servers which do not support RI.
3796 Until RI is more widely deployed this option is enabled by default.
3799 *) Add "missing" ssl ctrls to clear options and mode.
3802 *) If client attempts to renegotiate and doesn't support RI respond with
3803 a no_renegotiation alert as required by RFC5746. Some renegotiating
3804 TLS clients will continue a connection gracefully when they receive
3805 the alert. Unfortunately OpenSSL mishandled this alert and would hang
3806 waiting for a server hello which it will never receive. Now we treat a
3807 received no_renegotiation alert as a fatal error. This is because
3808 applications requesting a renegotiation might well expect it to succeed
3809 and would have no code in place to handle the server denying it so the
3810 only safe thing to do is to terminate the connection.
3813 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
3814 peer supports secure renegotiation and 0 otherwise. Print out peer
3815 renegotiation support in s_client/s_server.
3818 *) Replace the highly broken and deprecated SPKAC certification method with
3819 the updated NID creation version. This should correctly handle UTF8.
3822 *) Implement RFC5746. Re-enable renegotiation but require the extension
3823 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
3824 turns out to be a bad idea. It has been replaced by
3825 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
3826 SSL_CTX_set_options(). This is really not recommended unless you
3827 know what you are doing.
3828 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
3830 *) Fixes to stateless session resumption handling. Use initial_ctx when
3831 issuing and attempting to decrypt tickets in case it has changed during
3832 servername handling. Use a non-zero length session ID when attempting
3833 stateless session resumption: this makes it possible to determine if
3834 a resumption has occurred immediately after receiving server hello
3835 (several places in OpenSSL subtly assume this) instead of later in
3839 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
3840 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
3841 fixes for a few places where the return code is not checked
3843 [Julia Lawall <julia@diku.dk>]
3845 *) Add --strict-warnings option to Configure script to include devteam
3846 warnings in other configurations.
3849 *) Add support for --libdir option and LIBDIR variable in makefiles. This
3850 makes it possible to install openssl libraries in locations which
3851 have names other than "lib", for example "/usr/lib64" which some
3853 [Steve Henson, based on patch from Jeremy Utley]
3855 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
3856 X690 8.9.12 and can produce some misleading textual output of OIDs.
3857 [Steve Henson, reported by Dan Kaminsky]
3859 *) Delete MD2 from algorithm tables. This follows the recommendation in
3860 several standards that it is not used in new applications due to
3861 several cryptographic weaknesses. For binary compatibility reasons
3862 the MD2 API is still compiled in by default.
3865 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
3869 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
3870 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
3872 [Guenter <lists@gknw.net>]
3874 *) Fix the server certificate chain building code to use X509_verify_cert(),
3875 it used to have an ad-hoc builder which was unable to cope with anything
3876 other than a simple chain.
3877 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
3879 *) Don't check self signed certificate signatures in X509_verify_cert()
3880 by default (a flag can override this): it just wastes time without
3881 adding any security. As a useful side effect self signed root CAs
3882 with non-FIPS digests are now usable in FIPS mode.
3885 *) In dtls1_process_out_of_seq_message() the check if the current message
3886 is already buffered was missing. For every new message was memory
3887 allocated, allowing an attacker to perform an denial of service attack
3888 with sending out of seq handshake messages until there is no memory
3889 left. Additionally every future messege was buffered, even if the
3890 sequence number made no sense and would be part of another handshake.
3891 So only messages with sequence numbers less than 10 in advance will be
3892 buffered. (CVE-2009-1378)
3893 [Robin Seggelmann, discovered by Daniel Mentz]
3895 *) Records are buffered if they arrive with a future epoch to be
3896 processed after finishing the corresponding handshake. There is
3897 currently no limitation to this buffer allowing an attacker to perform
3898 a DOS attack with sending records with future epochs until there is no
3899 memory left. This patch adds the pqueue_size() function to determine
3900 the size of a buffer and limits the record buffer to 100 entries.
3902 [Robin Seggelmann, discovered by Daniel Mentz]
3904 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
3905 parent structure is freed. (CVE-2009-1379)
3908 *) Handle non-blocking I/O properly in SSL_shutdown() call.
3909 [Darryl Miles <darryl-mailinglists@netbauds.net>]
3912 [Ilya O. <vrghost@gmail.com>]
3914 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
3916 *) Disable renegotiation completely - this fixes a severe security
3917 problem (CVE-2009-3555) at the cost of breaking all
3918 renegotiation. Renegotiation can be re-enabled by setting
3919 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
3920 run-time. This is really not recommended unless you know what
3924 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
3926 *) Don't set val to NULL when freeing up structures, it is freed up by
3927 underlying code. If sizeof(void *) > sizeof(long) this can result in
3928 zeroing past the valid field. (CVE-2009-0789)
3929 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
3931 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
3932 checked correctly. This would allow some invalid signed attributes to
3933 appear to verify correctly. (CVE-2009-0591)
3934 [Ivan Nestlerode <inestlerode@us.ibm.com>]
3936 *) Reject UniversalString and BMPString types with invalid lengths. This
3937 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
3938 a legal length. (CVE-2009-0590)
3941 *) Set S/MIME signing as the default purpose rather than setting it
3942 unconditionally. This allows applications to override it at the store
3946 *) Permit restricted recursion of ASN1 strings. This is needed in practice
3947 to handle some structures.
3950 *) Improve efficiency of mem_gets: don't search whole buffer each time
3952 [Jeremy Shapiro <jnshapir@us.ibm.com>]
3954 *) New -hex option for openssl rand.
3957 *) Print out UTF8String and NumericString when parsing ASN1.
3960 *) Support NumericString type for name components.
3963 *) Allow CC in the environment to override the automatically chosen
3964 compiler. Note that nothing is done to ensure flags work with the
3968 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
3970 *) Properly check EVP_VerifyFinal() and similar return values
3972 [Ben Laurie, Bodo Moeller, Google Security Team]
3974 *) Enable TLS extensions by default.
3977 *) Allow the CHIL engine to be loaded, whether the application is
3978 multithreaded or not. (This does not release the developer from the
3979 obligation to set up the dynamic locking callbacks.)
3980 [Sander Temme <sander@temme.net>]
3982 *) Use correct exit code if there is an error in dgst command.
3983 [Steve Henson; problem pointed out by Roland Dirlewanger]
3985 *) Tweak Configure so that you need to say "experimental-jpake" to enable
3986 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
3989 *) Add experimental JPAKE support, including demo authentication in
3990 s_client and s_server.
3993 *) Set the comparison function in v3_addr_canonize().
3994 [Rob Austein <sra@hactrn.net>]
3996 *) Add support for XMPP STARTTLS in s_client.
3997 [Philip Paeps <philip@freebsd.org>]
3999 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
4000 to ensure that even with this option, only ciphersuites in the
4001 server's preference list will be accepted. (Note that the option
4002 applies only when resuming a session, so the earlier behavior was
4003 just about the algorithm choice for symmetric cryptography.)
4006 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
4008 *) Fix NULL pointer dereference if a DTLS server received
4009 ChangeCipherSpec as first record (CVE-2009-1386).
4012 *) Fix a state transition in s3_srvr.c and d1_srvr.c
4013 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
4016 *) The fix in 0.9.8c that supposedly got rid of unsafe
4017 double-checked locking was incomplete for RSA blinding,
4018 addressing just one layer of what turns out to have been
4019 doubly unsafe triple-checked locking.
4021 So now fix this for real by retiring the MONT_HELPER macro
4022 in crypto/rsa/rsa_eay.c.
4024 [Bodo Moeller; problem pointed out by Marius Schilder]
4026 *) Various precautionary measures:
4028 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
4030 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
4031 (NB: This would require knowledge of the secret session ticket key
4032 to exploit, in which case you'd be SOL either way.)
4034 - Change bn_nist.c so that it will properly handle input BIGNUMs
4035 outside the expected range.
4037 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
4040 [Neel Mehta, Bodo Moeller]
4042 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
4043 the load fails. Useful for distros.
4044 [Ben Laurie and the FreeBSD team]
4046 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
4049 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
4052 *) Expand ENGINE to support engine supplied SSL client certificate functions.
4054 This work was sponsored by Logica.
4057 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
4058 keystores. Support for SSL/TLS client authentication too.
4059 Not compiled unless enable-capieng specified to Configure.
4061 This work was sponsored by Logica.
4064 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
4065 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
4066 attribute creation routines such as certifcate requests and PKCS#12
4070 Changes between 0.9.8g and 0.9.8h [28 May 2008]
4072 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
4073 handshake which could lead to a cilent crash as found using the
4074 Codenomicon TLS test suite (CVE-2008-1672)
4075 [Steve Henson, Mark Cox]
4077 *) Fix double free in TLS server name extensions which could lead to
4078 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
4081 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
4083 Clear the error queue to ensure that error entries left from
4084 older function calls do not interfere with the correct operation.
4085 [Lutz Jaenicke, Erik de Castro Lopo]
4087 *) Remove root CA certificates of commercial CAs:
4089 The OpenSSL project does not recommend any specific CA and does not
4090 have any policy with respect to including or excluding any CA.
4091 Therefore it does not make any sense to ship an arbitrary selection
4092 of root CA certificates with the OpenSSL software.
4095 *) RSA OAEP patches to fix two separate invalid memory reads.
4096 The first one involves inputs when 'lzero' is greater than
4097 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
4098 before the beginning of from). The second one involves inputs where
4099 the 'db' section contains nothing but zeroes (there is a one-byte
4100 invalid read after the end of 'db').
4101 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4103 *) Partial backport from 0.9.9-dev:
4105 Introduce bn_mul_mont (dedicated Montgomery multiplication
4106 procedure) as a candidate for BIGNUM assembler implementation.
4107 While 0.9.9-dev uses assembler for various architectures, only
4108 x86_64 is available by default here in the 0.9.8 branch, and
4109 32-bit x86 is available through a compile-time setting.
4111 To try the 32-bit x86 assembler implementation, use Configure
4112 option "enable-montasm" (which exists only for this backport).
4114 As "enable-montasm" for 32-bit x86 disclaims code stability
4115 anyway, in this constellation we activate additional code
4116 backported from 0.9.9-dev for further performance improvements,
4117 namely BN_from_montgomery_word. (To enable this otherwise,
4118 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
4120 [Andy Polyakov (backport partially by Bodo Moeller)]
4122 *) Add TLS session ticket callback. This allows an application to set
4123 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
4124 values. This is useful for key rollover for example where several key
4125 sets may exist with different names.
4128 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
4129 This was broken until now in 0.9.8 releases, such that the only way
4130 a registered ENGINE could be used (assuming it initialises
4131 successfully on the host) was to explicitly set it as the default
4132 for the relevant algorithms. This is in contradiction with 0.9.7
4133 behaviour and the documentation. With this fix, when an ENGINE is
4134 registered into a given algorithm's table of implementations, the
4135 'uptodate' flag is reset so that auto-discovery will be used next
4136 time a new context for that algorithm attempts to select an
4138 [Ian Lister (tweaked by Geoff Thorpe)]
4140 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
4141 implemention in the following ways:
4143 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
4146 Lack of BER streaming support means one pass streaming processing is
4147 only supported if data is detached: setting the streaming flag is
4148 ignored for embedded content.
4150 CMS support is disabled by default and must be explicitly enabled
4151 with the enable-cms configuration option.
4154 *) Update the GMP engine glue to do direct copies between BIGNUM and
4155 mpz_t when openssl and GMP use the same limb size. Otherwise the
4156 existing "conversion via a text string export" trick is still used.
4157 [Paul Sheer <paulsheer@gmail.com>]
4159 *) Zlib compression BIO. This is a filter BIO which compressed and
4160 uncompresses any data passed through it.
4163 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
4164 RFC3394 compatible AES key wrapping.
4167 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
4168 sets string data without copying. X509_ALGOR_set0() and
4169 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
4170 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
4171 from an X509_ATTRIBUTE structure optionally checking it occurs only
4172 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
4176 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
4177 to get the expected BN_FLG_CONSTTIME behavior.
4178 [Bodo Moeller (Google)]
4182 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
4183 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
4184 - added some more tests to do_tests.pl
4185 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
4186 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
4187 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
4188 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
4189 - various changes to netware.pl to enable gcc-cross builds on Win32
4191 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
4192 - various changes to fix missing prototype warnings
4193 - fixed x86nasm.pl to create correct asm files for NASM COFF output
4194 - added AES, WHIRLPOOL and CPUID assembler code to build files
4195 - added missing AES assembler make rules to mk1mf.pl
4196 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
4197 [Guenter Knauf <eflash@gmx.net>]
4199 *) Implement certificate status request TLS extension defined in RFC3546.
4200 A client can set the appropriate parameters and receive the encoded
4201 OCSP response via a callback. A server can query the supplied parameters
4202 and set the encoded OCSP response in the callback. Add simplified examples
4203 to s_client and s_server.
4206 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
4208 *) Fix various bugs:
4209 + Binary incompatibility of ssl_ctx_st structure
4210 + DTLS interoperation with non-compliant servers
4211 + Don't call get_session_cb() without proposed session
4212 + Fix ia64 assembler code
4213 [Andy Polyakov, Steve Henson]
4215 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
4217 *) DTLS Handshake overhaul. There were longstanding issues with
4218 OpenSSL DTLS implementation, which were making it impossible for
4219 RFC 4347 compliant client to communicate with OpenSSL server.
4220 Unfortunately just fixing these incompatibilities would "cut off"
4221 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
4222 server keeps tolerating non RFC compliant syntax. The opposite is
4223 not true, 0.9.8f client can not communicate with earlier server.
4224 This update even addresses CVE-2007-4995.
4227 *) Changes to avoid need for function casts in OpenSSL: some compilers
4228 (gcc 4.2 and later) reject their use.
4229 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
4232 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4233 RFC4507bis. The encrypted ticket format is an encrypted encoded
4234 SSL_SESSION structure, that way new session features are automatically
4237 If a client application caches session in an SSL_SESSION structure
4238 support is transparent because tickets are now stored in the encoded
4241 The SSL_CTX structure automatically generates keys for ticket
4242 protection in servers so again support should be possible
4243 with no application modification.
4245 If a client or server wishes to disable RFC4507 support then the option
4246 SSL_OP_NO_TICKET can be set.
4248 Add a TLS extension debugging callback to allow the contents of any client
4249 or server extensions to be examined.
4251 This work was sponsored by Google.
4254 *) Add initial support for TLS extensions, specifically for the server_name
4255 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4256 have new members for a host name. The SSL data structure has an
4257 additional member SSL_CTX *initial_ctx so that new sessions can be
4258 stored in that context to allow for session resumption, even after the
4259 SSL has been switched to a new SSL_CTX in reaction to a client's
4260 server_name extension.
4262 New functions (subject to change):
4264 SSL_get_servername()
4265 SSL_get_servername_type()
4268 New CTRL codes and macros (subject to change):
4270 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4271 - SSL_CTX_set_tlsext_servername_callback()
4272 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4273 - SSL_CTX_set_tlsext_servername_arg()
4274 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4276 openssl s_client has a new '-servername ...' option.
4278 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4279 '-key2 ...', '-servername_fatal' (subject to change). This allows
4280 testing the HostName extension for a specific single host name ('-cert'
4281 and '-key' remain fallbacks for handshakes without HostName
4282 negotiation). If the unrecognized_name alert has to be sent, this by
4283 default is a warning; it becomes fatal with the '-servername_fatal'
4286 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
4288 *) Add AES and SSE2 assembly language support to VC++ build.
4291 *) Mitigate attack on final subtraction in Montgomery reduction.
4294 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
4295 (which previously caused an internal error).
4298 *) Squeeze another 10% out of IGE mode when in != out.
4301 *) AES IGE mode speedup.
4302 [Dean Gaudet (Google)]
4304 *) Add the Korean symmetric 128-bit cipher SEED (see
4305 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
4306 add SEED ciphersuites from RFC 4162:
4308 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
4309 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
4310 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
4311 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
4313 To minimize changes between patchlevels in the OpenSSL 0.9.8
4314 series, SEED remains excluded from compilation unless OpenSSL
4315 is configured with 'enable-seed'.
4316 [KISA, Bodo Moeller]
4318 *) Mitigate branch prediction attacks, which can be practical if a
4319 single processor is shared, allowing a spy process to extract
4320 information. For detailed background information, see
4321 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
4322 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
4323 and Necessary Software Countermeasures"). The core of the change
4324 are new versions BN_div_no_branch() and
4325 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
4326 respectively, which are slower, but avoid the security-relevant
4327 conditional branches. These are automatically called by BN_div()
4328 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
4329 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
4330 remove a conditional branch.
4332 BN_FLG_CONSTTIME is the new name for the previous
4333 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
4334 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
4335 in the exponent causes BN_mod_exp_mont() to use the alternative
4336 implementation in BN_mod_exp_mont_consttime().) The old name
4337 remains as a deprecated alias.
4339 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
4340 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
4341 constant-time implementations for more than just exponentiation.
4342 Here too the old name is kept as a deprecated alias.
4344 BN_BLINDING_new() will now use BN_dup() for the modulus so that
4345 the BN_BLINDING structure gets an independent copy of the
4346 modulus. This means that the previous "BIGNUM *m" argument to
4347 BN_BLINDING_new() and to BN_BLINDING_create_param() now
4348 essentially becomes "const BIGNUM *m", although we can't actually
4349 change this in the header file before 0.9.9. It allows
4350 RSA_setup_blinding() to use BN_with_flags() on the modulus to
4351 enable BN_FLG_CONSTTIME.
4353 [Matthew D Wood (Intel Corp)]
4355 *) In the SSL/TLS server implementation, be strict about session ID
4356 context matching (which matters if an application uses a single
4357 external cache for different purposes). Previously,
4358 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
4359 set. This did ensure strict client verification, but meant that,
4360 with applications using a single external cache for quite
4361 different requirements, clients could circumvent ciphersuite
4362 restrictions for a given session ID context by starting a session
4363 in a different context.
4366 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4367 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4368 authentication-only ciphersuites.
4371 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
4372 not complete and could lead to a possible single byte overflow
4373 (CVE-2007-5135) [Ben Laurie]
4375 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
4377 *) Since AES128 and AES256 (and similarly Camellia128 and
4378 Camellia256) share a single mask bit in the logic of
4379 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4380 kludge to work properly if AES128 is available and AES256 isn't
4381 (or if Camellia128 is available and Camellia256 isn't).
4384 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
4385 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
4386 When a point or a seed is encoded in a BIT STRING, we need to
4387 prevent the removal of trailing zero bits to get the proper DER
4388 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
4389 of a NamedBitList, for which trailing 0 bits need to be removed.)
4392 *) Have SSL/TLS server implementation tolerate "mismatched" record
4393 protocol version while receiving ClientHello even if the
4394 ClientHello is fragmented. (The server can't insist on the
4395 particular protocol version it has chosen before the ServerHello
4396 message has informed the client about his choice.)
4399 *) Add RFC 3779 support.
4400 [Rob Austein for ARIN, Ben Laurie]
4402 *) Load error codes if they are not already present instead of using a
4403 static variable. This allows them to be cleanly unloaded and reloaded.
4404 Improve header file function name parsing.
4407 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
4408 or CAPABILITY handshake as required by RFCs.
4411 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
4413 *) Introduce limits to prevent malicious keys being able to
4414 cause a denial of service. (CVE-2006-2940)
4415 [Steve Henson, Bodo Moeller]
4417 *) Fix ASN.1 parsing of certain invalid structures that can result
projects / openssl.git / blob