5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Remove openssl.spec. Packaging files belong with the packagers.
10 *) Automatic Darwin/OSX configuration has had a refresh, it will now
11 recognise x86_64 architectures automatically. You can still decide
12 to build for a different bitness with the environment variable
13 KERNEL_BITS (can be 32 or 64), for example:
15 KERNEL_BITS=32 ./config
19 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
20 256 bit AES and HMAC with SHA256.
23 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
26 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
29 *) To enable users to have their own config files and build file templates,
30 Configure looks in the directory indicated by the environment variable
31 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
32 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
33 name and is used as is.
36 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
37 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
38 X509_CERT_FILE_CTX was removed.
41 *) "shared" builds are now the default. To create only static libraries use
42 the "no-shared" Configure option.
45 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
46 All of these option have not worked for some while and are fundamental
50 *) Make various cleanup routines no-ops and mark them as deprecated. Most
51 global cleanup functions are no longer required because they are handled
52 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
53 Explicitly de-initing can cause problems (e.g. where a library that uses
54 OpenSSL de-inits, but an application is still using it). The affected
55 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
56 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
57 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
61 *) --strict-warnings no longer enables runtime debugging options
62 such as REF_DEBUG. Instead, debug options are automatically
63 enabled with '--debug' builds.
64 [Andy Polyakov, Emilia Käsper]
66 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
67 have been moved out of the public header files. New functions for managing
68 these have been added.
71 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
72 objects have been moved out of the public header files. New
73 functions for managing these have been added.
76 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
77 have been moved out of the public header files. New functions for managing
78 these have been added.
81 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
82 moved out of the public header files. New functions for managing these
86 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
89 *) Removed the mk1mf build scripts.
92 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
93 it is always safe to #include a header now.
96 *) Removed the aged BC-32 config and all its supporting scripts
99 *) Removed support for Ultrix, Netware, and OS/2.
102 *) Add support for HKDF.
105 *) Add support for blake2b and blake2s
108 *) Added support for "pipelining". Ciphers that have the
109 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
110 encryptions/decryptions simultaneously. There are currently no built-in
111 ciphers with this property but the expectation is that engines will be able
112 to offer it to significantly improve throughput. Support has been extended
113 into libssl so that multiple records for a single connection can be
114 processed in one go (for >=TLS 1.1).
117 *) Added the AFALG engine. This is an async capable engine which is able to
118 offload work to the Linux kernel. In this initial version it only supports
119 AES128-CBC. The kernel must be version 4.1.0 or greater.
122 *) OpenSSL now uses a new threading API. It is no longer necessary to
123 set locking callbacks to use OpenSSL in a multi-threaded environment. There
124 are two supported threading models: pthreads and windows threads. It is
125 also possible to configure OpenSSL at compile time for "no-threads". The
126 old threading API should no longer be used. The functions have been
127 replaced with "no-op" compatibility macros.
128 [Alessandro Ghedini, Matt Caswell]
130 *) Modify behavior of ALPN to invoke callback after SNI/servername
131 callback, such that updates to the SSL_CTX affect ALPN.
134 *) Add SSL_CIPHER queries for authentication and key-exchange.
137 *) Changes to the DEFAULT cipherlist:
138 - Prefer (EC)DHE handshakes over plain RSA.
139 - Prefer AEAD ciphers over legacy ciphers.
140 - Prefer ECDSA over RSA when both certificates are available.
141 - Prefer TLSv1.2 ciphers/PRF.
142 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
146 *) Change the ECC default curve list to be this, in order: x25519,
147 secp256r1, secp521r1, secp384r1.
150 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
151 disabled by default. They can be re-enabled using the
152 enable-weak-ssl-ciphers option to Configure.
155 *) If the server has ALPN configured, but supports no protocols that the
156 client advertises, send a fatal "no_application_protocol" alert.
157 This behaviour is SHALL in RFC 7301, though it isn't universally
158 implemented by other servers.
161 *) Add X25519 support.
162 Integrate support for X25519 into EC library. This includes support
163 for public and private key encoding using the format documented in
164 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
165 OID from that draft, encodes public keys using little endian
166 format in the ECPoint structure and private keys using
167 little endian form in the privateKey field of the ECPrivateKey
168 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
171 Note: the current version supports key generation, public and
172 private key encoding and ECDH key agreement using the EC API.
173 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
177 *) Deprecate SRP_VBASE_get_by_user.
178 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
179 In order to fix an unavoidable memory leak (CVE-2016-0798),
180 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
181 seed, even if the seed is configured.
183 Users should use SRP_VBASE_get1_by_user instead. Note that in
184 SRP_VBASE_get1_by_user, caller must free the returned value. Note
185 also that even though configuring the SRP seed attempts to hide
186 invalid usernames by continuing the handshake with fake
187 credentials, this behaviour is not constant time and no strong
188 guarantees are made that the handshake is indistinguishable from
189 that of a valid user.
192 *) Configuration change; it's now possible to build dynamic engines
193 without having to build shared libraries and vice versa. This
194 only applies to the engines in engines/, those in crypto/engine/
195 will always be built into libcrypto (i.e. "static").
197 Building dynamic engines is enabled by default; to disable, use
198 the configuration option "disable-dynamic-engine".
200 The only requirements for building dynamic engines are the
201 presence of the DSO module and building with position independent
202 code, so they will also automatically be disabled if configuring
203 with "disable-dso" or "disable-pic".
205 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
206 are also taken away from openssl/opensslconf.h, as they are
210 *) Configuration change; if there is a known flag to compile
211 position independent code, it will always be applied on the
212 libcrypto and libssl object files, and never on the application
213 object files. This means other libraries that use routines from
214 libcrypto / libssl can be made into shared libraries regardless
215 of how OpenSSL was configured.
217 If this isn't desirable, the configuration options "disable-pic"
218 or "no-pic" can be used to disable the use of PIC. This will
219 also disable building shared libraries and dynamic engines.
222 *) Removed JPAKE code. It was experimental and has no wide use.
225 *) The INSTALL_PREFIX Makefile variable has been renamed to
226 DESTDIR. That makes for less confusion on what this variable
227 is for. Also, the configuration option --install_prefix is
231 *) Heartbeat for TLS has been removed and is disabled by default
232 for DTLS; configure with enable-heartbeats. Code that uses the
233 old #define's might need to be updated.
234 [Emilia Käsper, Rich Salz]
236 *) Rename REF_CHECK to REF_DEBUG.
239 *) New "unified" build system
241 The "unified" build system is aimed to be a common system for all
242 platforms we support. With it comes new support for VMS.
244 This system builds supports building in a different directory tree
245 than the source tree. It produces one Makefile (for unix family
246 or lookalikes), or one descrip.mms (for VMS).
248 The source of information to make the Makefile / descrip.mms is
249 small files called 'build.info', holding the necessary
250 information for each directory with source to compile, and a
251 template in Configurations, like unix-Makefile.tmpl or
254 We rely heavily on the perl module Text::Template.
257 *) Added support for auto-initialisation and de-initialisation of the library.
258 OpenSSL no longer requires explicit init or deinit routines to be called,
259 except in certain circumstances. See the OPENSSL_init_crypto() and
260 OPENSSL_init_ssl() man pages for further information.
263 *) The arguments to the DTLSv1_listen function have changed. Specifically the
264 "peer" argument is now expected to be a BIO_ADDR object.
266 *) Rewrite of BIO networking library. The BIO library lacked consistent
267 support of IPv6, and adding it required some more extensive
268 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
269 which hold all types of addresses and chains of address information.
270 It also introduces a new API, with functions like BIO_socket,
271 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
272 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
273 have been adapted accordingly.
276 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
280 *) CRIME protection: disable compression by default, even if OpenSSL is
281 compiled with zlib enabled. Applications can still enable compression
282 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
283 using the SSL_CONF library to configure compression.
286 *) The signature of the session callback configured with
287 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
288 was explicitly marked as 'const unsigned char*' instead of
292 *) Always DPURIFY. Remove the use of uninitialized memory in the
293 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
296 *) Removed many obsolete configuration items, including
297 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
298 MD2_CHAR, MD2_INT, MD2_LONG
300 IDEA_SHORT, IDEA_LONG
301 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
302 [Rich Salz, with advice from Andy Polyakov]
304 *) Many BN internals have been moved to an internal header file.
305 [Rich Salz with help from Andy Polyakov]
307 *) Configuration and writing out the results from it has changed.
308 Files such as Makefile include/openssl/opensslconf.h and are now
309 produced through general templates, such as Makefile.in and
310 crypto/opensslconf.h.in and some help from the perl module
313 Also, the center of configuration information is no longer
314 Makefile. Instead, Configure produces a perl module in
315 configdata.pm which holds most of the config data (in the hash
316 table %config), the target data that comes from the target
317 configuration in one of the Configurations/*.conf files (in
321 *) To clarify their intended purposes, the Configure options
322 --prefix and --openssldir change their semantics, and become more
323 straightforward and less interdependent.
325 --prefix shall be used exclusively to give the location INSTALLTOP
326 where programs, scripts, libraries, include files and manuals are
327 going to be installed. The default is now /usr/local.
329 --openssldir shall be used exclusively to give the default
330 location OPENSSLDIR where certificates, private keys, CRLs are
331 managed. This is also where the default openssl.cnf gets
333 If the directory given with this option is a relative path, the
334 values of both the --prefix value and the --openssldir value will
335 be combined to become OPENSSLDIR.
336 The default for --openssldir is INSTALLTOP/ssl.
338 Anyone who uses --openssldir to specify where OpenSSL is to be
339 installed MUST change to use --prefix instead.
342 *) The GOST engine was out of date and therefore it has been removed. An up
343 to date GOST engine is now being maintained in an external repository.
344 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
345 support for GOST ciphersuites (these are only activated if a GOST engine
349 *) EGD is no longer supported by default; use enable-egd when
351 [Ben Kaduk and Rich Salz]
353 *) The distribution now has Makefile.in files, which are used to
354 create Makefile's when Configure is run. *Configure must be run
355 before trying to build now.*
358 *) The return value for SSL_CIPHER_description() for error conditions
362 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
364 Obtaining and performing DNSSEC validation of TLSA records is
365 the application's responsibility. The application provides
366 the TLSA records of its choice to OpenSSL, and these are then
367 used to authenticate the peer.
369 The TLSA records need not even come from DNS. They can, for
370 example, be used to implement local end-entity certificate or
371 trust-anchor "pinning", where the "pin" data takes the form
372 of TLSA records, which can augment or replace verification
373 based on the usual WebPKI public certification authorities.
376 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
377 continues to support deprecated interfaces in default builds.
378 However, applications are strongly advised to compile their
379 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
380 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
381 or the 1.1.0 releases.
383 In environments in which all applications have been ported to
384 not use any deprecated interfaces OpenSSL's Configure script
385 should be used with the --api=1.1.0 option to entirely remove
386 support for the deprecated features from the library and
387 unconditionally disable them in the installed headers.
388 Essentially the same effect can be achieved with the "no-deprecated"
389 argument to Configure, except that this will always restrict
390 the build to just the latest API, rather than a fixed API
393 As applications are ported to future revisions of the API,
394 they should update their compile-time OPENSSL_API_COMPAT define
395 accordingly, but in most cases should be able to continue to
396 compile with later releases.
398 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
399 0x10000000L and 0x00908000L, respectively. However those
400 versions did not support the OPENSSL_API_COMPAT feature, and
401 so applications are not typically tested for explicit support
402 of just the undeprecated features of either release.
405 *) Add support for setting the minimum and maximum supported protocol.
406 It can bet set via the SSL_set_min_proto_version() and
407 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
408 MaxProtcol. It's recommended to use the new APIs to disable
409 protocols instead of disabling individual protocols using
410 SSL_set_options() or SSL_CONF's Protocol. This change also
411 removes support for disabling TLS 1.2 in the OpenSSL TLS
412 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
415 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
418 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
419 and integrates ECDSA and ECDH functionality into EC. Implementations can
420 now redirect key generation and no longer need to convert to or from
423 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
424 include the ec.h header file instead.
427 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
428 ciphers who are no longer supported and drops support the ephemeral RSA key
429 exchange. The LOW ciphers currently doesn't have any ciphers in it.
432 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
433 opaque. For HMAC_CTX, the following constructors and destructors
436 HMAC_CTX *HMAC_CTX_new(void);
437 void HMAC_CTX_free(HMAC_CTX *ctx);
439 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
440 destroy such methods has been added. See EVP_MD_meth_new(3) and
441 EVP_CIPHER_meth_new(3) for documentation.
444 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
445 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
446 EVP_MD_CTX_reset() should be called instead to reinitialise
447 an already created structure.
448 2) For consistency with the majority of our object creators and
449 destructors, EVP_MD_CTX_(create|destroy) were renamed to
450 EVP_MD_CTX_(new|free). The old names are retained as macros
451 for deprecated builds.
454 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
455 cryptographic operations to be performed asynchronously as long as an
456 asynchronous capable engine is used. See the ASYNC_start_job() man page for
457 further details. Libssl has also had this capability integrated with the
458 introduction of the new mode SSL_MODE_ASYNC and associated error
459 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
460 pages. This work was developed in partnership with Intel Corp.
463 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
464 always enabled now. If you want to disable the support you should
465 exclude it using the list of supported ciphers. This also means that the
466 "-no_ecdhe" option has been removed from s_server.
469 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
470 SSL_{CTX_}set1_curves() which can set a list.
473 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
474 curve you want to support using SSL_{CTX_}set1_curves().
477 *) State machine rewrite. The state machine code has been significantly
478 refactored in order to remove much duplication of code and solve issues
479 with the old code (see ssl/statem/README for further details). This change
480 does have some associated API changes. Notably the SSL_state() function
481 has been removed and replaced by SSL_get_state which now returns an
482 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
483 altogether. The previous handshake states defined in ssl.h and ssl3.h have
487 *) All instances of the string "ssleay" in the public API were replaced
488 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
489 Some error codes related to internal RSA_eay API's were renamed.
492 *) The demo files in crypto/threads were moved to demo/threads.
495 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
497 [Matt Caswell, Rich Salz]
499 *) New ASN.1 embed macro.
501 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
502 structure is not allocated: it is part of the parent. That is instead of
510 This reduces memory fragmentation and make it impossible to accidentally
511 set a mandatory field to NULL.
513 This currently only works for some fields specifically a SEQUENCE, CHOICE,
514 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
515 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
519 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
522 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
523 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
524 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
525 DES and RC4 ciphersuites.
528 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
529 This changes the decoding behaviour for some invalid messages,
530 though the change is mostly in the more lenient direction, and
531 legacy behaviour is preserved as much as possible.
534 *) Fix no-stdio build.
535 [ David Woodhouse <David.Woodhouse@intel.com> and also
536 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
538 *) New testing framework
539 The testing framework has been largely rewritten and is now using
540 perl and the perl modules Test::Harness and an extended variant of
541 Test::More called OpenSSL::Test to do its work. All test scripts in
542 test/ have been rewritten into test recipes, and all direct calls to
543 executables in test/Makefile have become individual recipes using the
544 simplified testing OpenSSL::Test::Simple.
546 For documentation on our testing modules, do:
548 perldoc test/testlib/OpenSSL/Test/Simple.pm
549 perldoc test/testlib/OpenSSL/Test.pm
553 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
554 are used; the latter aborts on memory leaks (usually checked on exit).
555 Some undocumented "set malloc, etc., hooks" functions were removed
556 and others were changed. All are now documented.
559 *) In DSA_generate_parameters_ex, if the provided seed is too short,
561 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
563 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
564 from RFC4279, RFC4785, RFC5487, RFC5489.
566 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
567 original RSA_PSK patch.
570 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
571 era flag was never set throughout the codebase (only read). Also removed
572 SSL3_FLAGS_POP_BUFFER which was only used if
573 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
576 *) Changed the default name options in the "ca", "crl", "req" and "x509"
577 to be "oneline" instead of "compat".
580 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
581 not aware of clients that still exhibit this bug, and the workaround
582 hasn't been working properly for a while.
585 *) The return type of BIO_number_read() and BIO_number_written() as well as
586 the corresponding num_read and num_write members in the BIO structure has
587 changed from unsigned long to uint64_t. On platforms where an unsigned
588 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
592 *) Given the pervasive nature of TLS extensions it is inadvisable to run
593 OpenSSL without support for them. It also means that maintaining
594 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
595 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
598 *) Removed support for the two export grade static DH ciphersuites
599 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
600 were newly added (along with a number of other static DH ciphersuites) to
601 1.0.2. However the two export ones have *never* worked since they were
602 introduced. It seems strange in any case to be adding new export
603 ciphersuites, and given "logjam" it also does not seem correct to fix them.
606 *) Version negotiation has been rewritten. In particular SSLv23_method(),
607 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
608 and turned into macros which simply call the new preferred function names
609 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
610 should use the new names instead. Also as part of this change the ssl23.h
611 header file has been removed.
614 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
615 code and the associated standard is no longer considered fit-for-purpose.
618 *) RT2547 was closed. When generating a private key, try to make the
619 output file readable only by the owner. This behavior change might
620 be noticeable when interacting with other software.
622 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
626 *) Added HTTP GET support to the ocsp command.
629 *) Changed default digest for the dgst and enc commands from MD5 to
633 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
636 *) Added support for TLS extended master secret from
637 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
638 initial patch which was a great help during development.
641 *) All libssl internal structures have been removed from the public header
642 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
643 now redundant). Users should not attempt to access internal structures
644 directly. Instead they should use the provided API functions.
647 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
648 Access to deprecated functions can be re-enabled by running config with
649 "enable-deprecated". In addition applications wishing to use deprecated
650 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
651 will, by default, disable some transitive includes that previously existed
652 in the header files (e.g. ec.h will no longer, by default, include bn.h)
655 *) Added support for OCB mode. OpenSSL has been granted a patent license
656 compatible with the OpenSSL license for use of OCB. Details are available
657 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
658 for OCB can be removed by calling config with no-ocb.
661 *) SSLv2 support has been removed. It still supports receiving a SSLv2
662 compatible client hello.
665 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
666 done while fixing the error code for the key-too-small case.
667 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
669 *) CA.sh has been removmed; use CA.pl instead.
672 *) Removed old DES API.
675 *) Remove various unsupported platforms:
681 Sinix/ReliantUNIX RM400
686 16-bit platforms such as WIN16
689 *) Clean up OPENSSL_NO_xxx #define's
690 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
691 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
692 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
693 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
694 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
695 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
696 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
697 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
698 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
699 Remove MS_STATIC; it's a relic from platforms <32 bits.
702 *) Cleaned up dead code
703 Remove all but one '#ifdef undef' which is to be looked at.
706 *) Clean up calling of xxx_free routines.
707 Just like free(), fix most of the xxx_free routines to accept
708 NULL. Remove the non-null checks from callers. Save much code.
711 *) Add secure heap for storage of private keys (when possible).
712 Add BIO_s_secmem(), CBIGNUM, etc.
713 Contributed by Akamai Technologies under our Corporate CLA.
716 *) Experimental support for a new, fast, unbiased prime candidate generator,
717 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
718 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
720 *) New output format NSS in the sess_id command line tool. This allows
721 exporting the session id and the master key in NSS keylog format.
722 [Martin Kaiser <martin@kaiser.cx>]
724 *) Harmonize version and its documentation. -f flag is used to display
726 [mancha <mancha1@zoho.com>]
728 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
729 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
730 [mancha <mancha1@zoho.com>]
732 *) Fix some double frees. These are not thought to be exploitable.
733 [mancha <mancha1@zoho.com>]
735 *) A missing bounds check in the handling of the TLS heartbeat extension
736 can be used to reveal up to 64k of memory to a connected client or
739 Thanks for Neel Mehta of Google Security for discovering this bug and to
740 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
741 preparing the fix (CVE-2014-0160)
742 [Adam Langley, Bodo Moeller]
744 *) Fix for the attack described in the paper "Recovering OpenSSL
745 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
746 by Yuval Yarom and Naomi Benger. Details can be obtained from:
747 http://eprint.iacr.org/2014/140
749 Thanks to Yuval Yarom and Naomi Benger for discovering this
750 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
751 [Yuval Yarom and Naomi Benger]
753 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
754 this fixes a limitation in previous versions of OpenSSL.
757 *) Experimental encrypt-then-mac support.
759 Experimental support for encrypt then mac from
760 draft-gutmann-tls-encrypt-then-mac-02.txt
762 To enable it set the appropriate extension number (0x42 for the test
763 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
765 For non-compliant peers (i.e. just about everything) this should have no
768 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
772 *) Add EVP support for key wrapping algorithms, to avoid problems with
773 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
774 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
775 algorithms and include tests cases.
778 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
782 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
783 MGF1 digest and OAEP label.
786 *) Make openssl verify return errors.
787 [Chris Palmer <palmer@google.com> and Ben Laurie]
789 *) New function ASN1_TIME_diff to calculate the difference between two
790 ASN1_TIME structures or one structure and the current time.
793 *) Update fips_test_suite to support multiple command line options. New
794 test to induce all self test errors in sequence and check expected
798 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
799 sign or verify all in one operation.
802 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
803 test programs and fips_test_suite. Includes functionality to parse
804 the minimal script output of fipsalgest.pl directly.
807 *) Add authorisation parameter to FIPS_module_mode_set().
810 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
813 *) Use separate DRBG fields for internal and external flags. New function
814 FIPS_drbg_health_check() to perform on demand health checking. Add
815 generation tests to fips_test_suite with reduced health check interval to
816 demonstrate periodic health checking. Add "nodh" option to
817 fips_test_suite to skip very slow DH test.
820 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
824 *) More extensive health check for DRBG checking many more failure modes.
825 New function FIPS_selftest_drbg_all() to handle every possible DRBG
826 combination: call this in fips_test_suite.
829 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
830 and POST to handle Dual EC cases.
833 *) Add support for canonical generation of DSA parameter 'g'. See
836 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
837 POST to handle HMAC cases.
840 *) Add functions FIPS_module_version() and FIPS_module_version_text()
841 to return numerical and string versions of the FIPS module number.
844 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
845 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
846 outside the validated module in the FIPS capable OpenSSL.
849 *) Minor change to DRBG entropy callback semantics. In some cases
850 there is no multiple of the block length between min_len and
851 max_len. Allow the callback to return more than max_len bytes
852 of entropy but discard any extra: it is the callback's responsibility
853 to ensure that the extra data discarded does not impact the
854 requested amount of entropy.
857 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
858 information in FIPS186-3, SP800-57 and SP800-131A.
861 *) CCM support via EVP. Interface is very similar to GCM case except we
862 must supply all data in one chunk (i.e. no update, final) and the
863 message length must be supplied if AAD is used. Add algorithm test
867 *) Initial version of POST overhaul. Add POST callback to allow the status
868 of POST to be monitored and/or failures induced. Modify fips_test_suite
869 to use callback. Always run all selftests even if one fails.
872 *) XTS support including algorithm test driver in the fips_gcmtest program.
873 Note: this does increase the maximum key length from 32 to 64 bytes but
874 there should be no binary compatibility issues as existing applications
875 will never use XTS mode.
878 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
879 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
880 performs algorithm blocking for unapproved PRNG types. Also do not
881 set PRNG type in FIPS_mode_set(): leave this to the application.
882 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
883 the standard OpenSSL PRNG: set additional data to a date time vector.
886 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
887 This shouldn't present any incompatibility problems because applications
888 shouldn't be using these directly and any that are will need to rethink
889 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
892 *) Extensive self tests and health checking required by SP800-90 DRBG.
893 Remove strength parameter from FIPS_drbg_instantiate and always
894 instantiate at maximum supported strength.
897 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
900 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
903 *) New function DH_compute_key_padded() to compute a DH key and pad with
904 leading zeroes if needed: this complies with SP800-56A et al.
907 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
908 anything, incomplete, subject to change and largely untested at present.
911 *) Modify fipscanisteronly build option to only build the necessary object
912 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
915 *) Add experimental option FIPSSYMS to give all symbols in
916 fipscanister.o and FIPS or fips prefix. This will avoid
917 conflicts with future versions of OpenSSL. Add perl script
918 util/fipsas.pl to preprocess assembly language source files
919 and rename any affected symbols.
922 *) Add selftest checks and algorithm block of non-fips algorithms in
923 FIPS mode. Remove DES2 from selftests.
926 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
927 return internal method without any ENGINE dependencies. Add new
928 tiny fips sign and verify functions.
931 *) New build option no-ec2m to disable characteristic 2 code.
934 *) New build option "fipscanisteronly". This only builds fipscanister.o
935 and (currently) associated fips utilities. Uses the file Makefile.fips
936 instead of Makefile.org as the prototype.
939 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
940 Update fips_gcmtest to use IV generator.
943 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
944 setting output buffer to NULL. The *Final function must be
945 called although it will not retrieve any additional data. The tag
946 can be set or retrieved with a ctrl. The IV length is by default 12
947 bytes (96 bits) but can be set to an alternative value. If the IV
948 length exceeds the maximum IV length (currently 16 bytes) it cannot be
952 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
953 underlying do_cipher function handles all cipher semantics itself
954 including padding and finalisation. This is useful if (for example)
955 an ENGINE cipher handles block padding itself. The behaviour of
956 do_cipher is subtly changed if this flag is set: the return value
957 is the number of characters written to the output buffer (zero is
958 no longer an error code) or a negative error code. Also if the
959 input buffer is NULL and length 0 finalisation should be performed.
962 *) If a candidate issuer certificate is already part of the constructed
963 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
966 *) Improve forward-security support: add functions
968 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
969 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
971 for use by SSL/TLS servers; the callback function will be called whenever a
972 new session is created, and gets to decide whether the session may be
973 cached to make it resumable (return 0) or not (return 1). (As by the
974 SSL/TLS protocol specifications, the session_id sent by the server will be
975 empty to indicate that the session is not resumable; also, the server will
976 not generate RFC 4507 (RFC 5077) session tickets.)
978 A simple reasonable callback implementation is to return is_forward_secure.
979 This parameter will be set to 1 or 0 depending on the ciphersuite selected
980 by the SSL/TLS server library, indicating whether it can provide forward
982 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
984 *) New -verify_name option in command line utilities to set verification
988 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
989 Add CMAC pkey methods.
992 *) Experimental renegotiation in s_server -www mode. If the client
993 browses /reneg connection is renegotiated. If /renegcert it is
994 renegotiated requesting a certificate.
997 *) Add an "external" session cache for debugging purposes to s_server. This
998 should help trace issues which normally are only apparent in deployed
999 multi-process servers.
1002 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1003 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1004 BIO_set_cipher() and some obscure PEM functions were changed so they
1005 can now return an error. The RAND changes required a change to the
1006 RAND_METHOD structure.
1009 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1010 a gcc attribute to warn if the result of a function is ignored. This
1011 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1012 whose return value is often ignored.
1015 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1016 These allow SCTs (signed certificate timestamps) to be requested and
1017 validated when establishing a connection.
1018 [Rob Percival <robpercival@google.com>]
1020 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1022 *) Prevent padding oracle in AES-NI CBC MAC check
1024 A MITM attacker can use a padding oracle attack to decrypt traffic
1025 when the connection uses an AES CBC cipher and the server support
1028 This issue was introduced as part of the fix for Lucky 13 padding
1029 attack (CVE-2013-0169). The padding check was rewritten to be in
1030 constant time by making sure that always the same bytes are read and
1031 compared against either the MAC or padding bytes. But it no longer
1032 checked that there was enough data to have both the MAC and padding
1035 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1039 *) Fix EVP_EncodeUpdate overflow
1041 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1042 Base64 encoding of binary data. If an attacker is able to supply very large
1043 amounts of input data then a length check can overflow resulting in a heap
1046 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1047 the PEM_write_bio* family of functions. These are mainly used within the
1048 OpenSSL command line applications, so any application which processes data
1049 from an untrusted source and outputs it as a PEM file should be considered
1050 vulnerable to this issue. User applications that call these APIs directly
1051 with large amounts of untrusted data may also be vulnerable.
1053 This issue was reported by Guido Vranken.
1057 *) Fix EVP_EncryptUpdate overflow
1059 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1060 is able to supply very large amounts of input data after a previous call to
1061 EVP_EncryptUpdate() with a partial block then a length check can overflow
1062 resulting in a heap corruption. Following an analysis of all OpenSSL
1063 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1064 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1065 the first called function after an EVP_EncryptInit(), and therefore that
1066 specific call must be safe. The second form is where the length passed to
1067 EVP_EncryptUpdate() can be seen from the code to be some small value and
1068 therefore there is no possibility of an overflow. Since all instances are
1069 one of these two forms, it is believed that there can be no overflows in
1070 internal code due to this problem. It should be noted that
1071 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1072 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1073 of these calls have also been analysed too and it is believed there are no
1074 instances in internal usage where an overflow could occur.
1076 This issue was reported by Guido Vranken.
1080 *) Prevent ASN.1 BIO excessive memory allocation
1082 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1083 a short invalid encoding can cause allocation of large amounts of memory
1084 potentially consuming excessive resources or exhausting memory.
1086 Any application parsing untrusted data through d2i BIO functions is
1087 affected. The memory based functions such as d2i_X509() are *not* affected.
1088 Since the memory based functions are used by the TLS library, TLS
1089 applications are not affected.
1091 This issue was reported by Brian Carpenter.
1097 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1098 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1099 in arbitrary stack data being returned in the buffer.
1101 This issue was reported by Guido Vranken.
1105 *) Modify behavior of ALPN to invoke callback after SNI/servername
1106 callback, such that updates to the SSL_CTX affect ALPN.
1109 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1113 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1114 methods are enabled and ssl2 is disabled the methods return NULL.
1117 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1119 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1120 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1121 provide any "EXPORT" or "LOW" strength ciphers.
1124 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1125 is by default disabled at build-time. Builds that are not configured with
1126 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1127 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1128 will need to explicitly call either of:
1130 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1132 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1134 as appropriate. Even if either of those is used, or the application
1135 explicitly uses the version-specific SSLv2_method() or its client and
1136 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1137 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1138 ciphers, and SSLv2 56-bit DES are no longer available.
1142 *) Fix a double-free in DSA code
1144 A double free bug was discovered when OpenSSL parses malformed DSA private
1145 keys and could lead to a DoS attack or memory corruption for applications
1146 that receive DSA private keys from untrusted sources. This scenario is
1149 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1154 *) Disable SRP fake user seed to address a server memory leak.
1156 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1158 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1159 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1160 was changed to ignore the "fake user" SRP seed, even if the seed
1163 Users should use SRP_VBASE_get1_by_user instead. Note that in
1164 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1165 also that even though configuring the SRP seed attempts to hide
1166 invalid usernames by continuing the handshake with fake
1167 credentials, this behaviour is not constant time and no strong
1168 guarantees are made that the handshake is indistinguishable from
1169 that of a valid user.
1173 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1175 In the BN_hex2bn function the number of hex digits is calculated using an
1176 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1177 large values of |i| this can result in |bn_expand| not allocating any
1178 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1179 field as NULL leading to a subsequent NULL ptr deref. For very large values
1180 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1181 In this case memory is allocated to the internal BIGNUM data field, but it
1182 is insufficiently sized leading to heap corruption. A similar issue exists
1183 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1184 is ever called by user applications with very large untrusted hex/dec data.
1185 This is anticipated to be a rare occurrence.
1187 All OpenSSL internal usage of these functions use data that is not expected
1188 to be untrusted, e.g. config file data or application command line
1189 arguments. If user developed applications generate config file data based
1190 on untrusted data then it is possible that this could also lead to security
1191 consequences. This is also anticipated to be rare.
1193 This issue was reported to OpenSSL by Guido Vranken.
1197 *) Fix memory issues in BIO_*printf functions
1199 The internal |fmtstr| function used in processing a "%s" format string in
1200 the BIO_*printf functions could overflow while calculating the length of a
1201 string and cause an OOB read when printing very long strings.
1203 Additionally the internal |doapr_outch| function can attempt to write to an
1204 OOB memory location (at an offset from the NULL pointer) in the event of a
1205 memory allocation failure. In 1.0.2 and below this could be caused where
1206 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1207 could be in processing a very long "%s" format string. Memory leaks can
1210 The first issue may mask the second issue dependent on compiler behaviour.
1211 These problems could enable attacks where large amounts of untrusted data
1212 is passed to the BIO_*printf functions. If applications use these functions
1213 in this way then they could be vulnerable. OpenSSL itself uses these
1214 functions when printing out human-readable dumps of ASN.1 data. Therefore
1215 applications that print this data could be vulnerable if the data is from
1216 untrusted sources. OpenSSL command line applications could also be
1217 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1218 as command line arguments.
1220 Libssl is not considered directly vulnerable. Additionally certificates etc
1221 received via remote connections via libssl are also unlikely to be able to
1222 trigger these issues because of message size limits enforced within libssl.
1224 This issue was reported to OpenSSL Guido Vranken.
1228 *) Side channel attack on modular exponentiation
1230 A side-channel attack was found which makes use of cache-bank conflicts on
1231 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1232 of RSA keys. The ability to exploit this issue is limited as it relies on
1233 an attacker who has control of code in a thread running on the same
1234 hyper-threaded core as the victim thread which is performing decryptions.
1236 This issue was reported to OpenSSL by Yuval Yarom, The University of
1237 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1238 Nadia Heninger, University of Pennsylvania with more information at
1239 http://cachebleed.info.
1243 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1244 if no keysize is specified with default_bits. This fixes an
1245 omission in an earlier change that changed all RSA/DSA key generation
1246 apps to use 2048 bits by default.
1249 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1250 *) DH small subgroups
1252 Historically OpenSSL only ever generated DH parameters based on "safe"
1253 primes. More recently (in version 1.0.2) support was provided for
1254 generating X9.42 style parameter files such as those required for RFC 5114
1255 support. The primes used in such files may not be "safe". Where an
1256 application is using DH configured with parameters based on primes that are
1257 not "safe" then an attacker could use this fact to find a peer's private
1258 DH exponent. This attack requires that the attacker complete multiple
1259 handshakes in which the peer uses the same private DH exponent. For example
1260 this could be used to discover a TLS server's private DH exponent if it's
1261 reusing the private DH exponent or it's using a static DH ciphersuite.
1263 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1264 TLS. It is not on by default. If the option is not set then the server
1265 reuses the same private DH exponent for the life of the server process and
1266 would be vulnerable to this attack. It is believed that many popular
1267 applications do set this option and would therefore not be at risk.
1269 The fix for this issue adds an additional check where a "q" parameter is
1270 available (as is the case in X9.42 based parameters). This detects the
1271 only known attack, and is the only possible defense for static DH
1272 ciphersuites. This could have some performance impact.
1274 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1275 default and cannot be disabled. This could have some performance impact.
1277 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1281 *) SSLv2 doesn't block disabled ciphers
1283 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1284 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1285 been disabled, provided that the SSLv2 protocol was not also disabled via
1288 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1289 and Sebastian Schinzel.
1293 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1295 *) BN_mod_exp may produce incorrect results on x86_64
1297 There is a carry propagating bug in the x86_64 Montgomery squaring
1298 procedure. No EC algorithms are affected. Analysis suggests that attacks
1299 against RSA and DSA as a result of this defect would be very difficult to
1300 perform and are not believed likely. Attacks against DH are considered just
1301 feasible (although very difficult) because most of the work necessary to
1302 deduce information about a private key may be performed offline. The amount
1303 of resources required for such an attack would be very significant and
1304 likely only accessible to a limited number of attackers. An attacker would
1305 additionally need online access to an unpatched system using the target
1306 private key in a scenario with persistent DH parameters and a private
1307 key that is shared between multiple clients. For example this can occur by
1308 default in OpenSSL DHE based SSL/TLS ciphersuites.
1310 This issue was reported to OpenSSL by Hanno Böck.
1314 *) Certificate verify crash with missing PSS parameter
1316 The signature verification routines will crash with a NULL pointer
1317 dereference if presented with an ASN.1 signature using the RSA PSS
1318 algorithm and absent mask generation function parameter. Since these
1319 routines are used to verify certificate signature algorithms this can be
1320 used to crash any certificate verification operation and exploited in a
1321 DoS attack. Any application which performs certificate verification is
1322 vulnerable including OpenSSL clients and servers which enable client
1325 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1329 *) X509_ATTRIBUTE memory leak
1331 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1332 memory. This structure is used by the PKCS#7 and CMS routines so any
1333 application which reads PKCS#7 or CMS data from untrusted sources is
1334 affected. SSL/TLS is not affected.
1336 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1341 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1342 This changes the decoding behaviour for some invalid messages,
1343 though the change is mostly in the more lenient direction, and
1344 legacy behaviour is preserved as much as possible.
1347 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1349 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1351 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1353 *) Alternate chains certificate forgery
1355 During certificate verification, OpenSSL will attempt to find an
1356 alternative certificate chain if the first attempt to build such a chain
1357 fails. An error in the implementation of this logic can mean that an
1358 attacker could cause certain checks on untrusted certificates to be
1359 bypassed, such as the CA flag, enabling them to use a valid leaf
1360 certificate to act as a CA and "issue" an invalid certificate.
1362 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1366 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1368 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1369 incompatibility in the handling of HMAC. The previous ABI has now been
1373 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1375 *) Malformed ECParameters causes infinite loop
1377 When processing an ECParameters structure OpenSSL enters an infinite loop
1378 if the curve specified is over a specially malformed binary polynomial
1381 This can be used to perform denial of service against any
1382 system which processes public keys, certificate requests or
1383 certificates. This includes TLS clients and TLS servers with
1384 client authentication enabled.
1386 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1390 *) Exploitable out-of-bounds read in X509_cmp_time
1392 X509_cmp_time does not properly check the length of the ASN1_TIME
1393 string and can read a few bytes out of bounds. In addition,
1394 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1397 An attacker can use this to craft malformed certificates and CRLs of
1398 various sizes and potentially cause a segmentation fault, resulting in
1399 a DoS on applications that verify certificates or CRLs. TLS clients
1400 that verify CRLs are affected. TLS clients and servers with client
1401 authentication enabled may be affected if they use custom verification
1404 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1405 independently by Hanno Böck.
1409 *) PKCS7 crash with missing EnvelopedContent
1411 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1412 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1413 with missing content and trigger a NULL pointer dereference on parsing.
1415 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1416 structures from untrusted sources are affected. OpenSSL clients and
1417 servers are not affected.
1419 This issue was reported to OpenSSL by Michal Zalewski (Google).
1423 *) CMS verify infinite loop with unknown hash function
1425 When verifying a signedData message the CMS code can enter an infinite loop
1426 if presented with an unknown hash function OID. This can be used to perform
1427 denial of service against any system which verifies signedData messages using
1429 This issue was reported to OpenSSL by Johannes Bauer.
1433 *) Race condition handling NewSessionTicket
1435 If a NewSessionTicket is received by a multi-threaded client when attempting to
1436 reuse a previous ticket then a race condition can occur potentially leading to
1437 a double free of the ticket data.
1441 *) Only support 256-bit or stronger elliptic curves with the
1442 'ecdh_auto' setting (server) or by default (client). Of supported
1443 curves, prefer P-256 (both).
1446 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1448 *) ClientHello sigalgs DoS fix
1450 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1451 invalid signature algorithms extension a NULL pointer dereference will
1452 occur. This can be exploited in a DoS attack against the server.
1454 This issue was was reported to OpenSSL by David Ramos of Stanford
1457 [Stephen Henson and Matt Caswell]
1459 *) Multiblock corrupted pointer fix
1461 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1462 feature only applies on 64 bit x86 architecture platforms that support AES
1463 NI instructions. A defect in the implementation of "multiblock" can cause
1464 OpenSSL's internal write buffer to become incorrectly set to NULL when
1465 using non-blocking IO. Typically, when the user application is using a
1466 socket BIO for writing, this will only result in a failed connection.
1467 However if some other BIO is used then it is likely that a segmentation
1468 fault will be triggered, thus enabling a potential DoS attack.
1470 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1474 *) Segmentation fault in DTLSv1_listen fix
1476 The DTLSv1_listen function is intended to be stateless and processes the
1477 initial ClientHello from many peers. It is common for user code to loop
1478 over the call to DTLSv1_listen until a valid ClientHello is received with
1479 an associated cookie. A defect in the implementation of DTLSv1_listen means
1480 that state is preserved in the SSL object from one invocation to the next
1481 that can lead to a segmentation fault. Errors processing the initial
1482 ClientHello can trigger this scenario. An example of such an error could be
1483 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1486 This issue was reported to OpenSSL by Per Allansson.
1490 *) Segmentation fault in ASN1_TYPE_cmp fix
1492 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1493 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1494 certificate signature algorithm consistency this can be used to crash any
1495 certificate verification operation and exploited in a DoS attack. Any
1496 application which performs certificate verification is vulnerable including
1497 OpenSSL clients and servers which enable client authentication.
1501 *) Segmentation fault for invalid PSS parameters fix
1503 The signature verification routines will crash with a NULL pointer
1504 dereference if presented with an ASN.1 signature using the RSA PSS
1505 algorithm and invalid parameters. Since these routines are used to verify
1506 certificate signature algorithms this can be used to crash any
1507 certificate verification operation and exploited in a DoS attack. Any
1508 application which performs certificate verification is vulnerable including
1509 OpenSSL clients and servers which enable client authentication.
1511 This issue was was reported to OpenSSL by Brian Carpenter.
1515 *) ASN.1 structure reuse memory corruption fix
1517 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1518 memory corruption via an invalid write. Such reuse is and has been
1519 strongly discouraged and is believed to be rare.
1521 Applications that parse structures containing CHOICE or ANY DEFINED BY
1522 components may be affected. Certificate parsing (d2i_X509 and related
1523 functions) are however not affected. OpenSSL clients and servers are
1528 *) PKCS7 NULL pointer dereferences fix
1530 The PKCS#7 parsing code does not handle missing outer ContentInfo
1531 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1532 missing content and trigger a NULL pointer dereference on parsing.
1534 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1535 otherwise parse PKCS#7 structures from untrusted sources are
1536 affected. OpenSSL clients and servers are not affected.
1538 This issue was reported to OpenSSL by Michal Zalewski (Google).
1542 *) DoS via reachable assert in SSLv2 servers fix
1544 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1545 servers that both support SSLv2 and enable export cipher suites by sending
1546 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1548 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1549 (OpenSSL development team).
1553 *) Empty CKE with client auth and DHE fix
1555 If client auth is used then a server can seg fault in the event of a DHE
1556 ciphersuite being selected and a zero length ClientKeyExchange message
1557 being sent by the client. This could be exploited in a DoS attack.
1561 *) Handshake with unseeded PRNG fix
1563 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1564 with an unseeded PRNG. The conditions are:
1565 - The client is on a platform where the PRNG has not been seeded
1566 automatically, and the user has not seeded manually
1567 - A protocol specific client method version has been used (i.e. not
1568 SSL_client_methodv23)
1569 - A ciphersuite is used that does not require additional random data from
1570 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1572 If the handshake succeeds then the client random that has been used will
1573 have been generated from a PRNG with insufficient entropy and therefore the
1574 output may be predictable.
1576 For example using the following command with an unseeded openssl will
1577 succeed on an unpatched platform:
1579 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1583 *) Use After Free following d2i_ECPrivatekey error fix
1585 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1586 could cause a use after free condition. This, in turn, could cause a double
1587 free in several private key parsing functions (such as d2i_PrivateKey
1588 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1589 for applications that receive EC private keys from untrusted
1590 sources. This scenario is considered rare.
1592 This issue was discovered by the BoringSSL project and fixed in their
1597 *) X509_to_X509_REQ NULL pointer deref fix
1599 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1600 the certificate key is invalid. This function is rarely used in practice.
1602 This issue was discovered by Brian Carpenter.
1606 *) Removed the export ciphers from the DEFAULT ciphers
1609 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1611 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1612 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1613 So far those who have to target multiple platforms would compromise
1614 and argue that binary targeting say ARMv5 would still execute on
1615 ARMv8. "Universal" build resolves this compromise by providing
1616 near-optimal performance even on newer platforms.
1619 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1620 (other platforms pending).
1621 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1623 *) Add support for the SignedCertificateTimestampList certificate and
1624 OCSP response extensions from RFC6962.
1627 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1628 for corner cases. (Certain input points at infinity could lead to
1629 bogus results, with non-infinity inputs mapped to infinity too.)
1632 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1633 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1634 common cases are optimized and there still is room for further
1635 improvements. Vector Permutation AES for Altivec is also added.
1638 *) Add support for little-endian ppc64 Linux target.
1639 [Marcelo Cerri (IBM)]
1641 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1642 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1643 are optimized and there still is room for further improvements.
1644 Both 32- and 64-bit modes are supported.
1645 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1647 *) Improved ARMv7 NEON support.
1650 *) Support for SPARC Architecture 2011 crypto extensions, first
1651 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1652 SHA256/512, MD5, GHASH and modular exponentiation.
1653 [Andy Polyakov, David Miller]
1655 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1657 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1659 *) Support for new and upcoming Intel processors, including AVX2,
1660 BMI and SHA ISA extensions. This includes additional "stitched"
1661 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1664 This work was sponsored by Intel Corp.
1667 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1668 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1669 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1672 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1673 this fixes a limitation in previous versions of OpenSSL.
1676 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1677 MGF1 digest and OAEP label.
1680 *) Add EVP support for key wrapping algorithms, to avoid problems with
1681 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1682 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1683 algorithms and include tests cases.
1686 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1688 [Douglas E. Engert, Steve Henson]
1690 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1691 difference in days and seconds between two tm or ASN1_TIME structures.
1694 *) Add -rev test option to s_server to just reverse order of characters
1695 received by client and send back to server. Also prints an abbreviated
1696 summary of the connection parameters.
1699 *) New option -brief for s_client and s_server to print out a brief summary
1700 of connection parameters.
1703 *) Add callbacks for arbitrary TLS extensions.
1704 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1706 *) New option -crl_download in several openssl utilities to download CRLs
1707 from CRLDP extension in certificates.
1710 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1713 *) New function X509_CRL_diff to generate a delta CRL from the difference
1714 of two full CRLs. Add support to "crl" utility.
1717 *) New functions to set lookup_crls function and to retrieve
1718 X509_STORE from X509_STORE_CTX.
1721 *) Print out deprecated issuer and subject unique ID fields in
1725 *) Extend OCSP I/O functions so they can be used for simple general purpose
1726 HTTP as well as OCSP. New wrapper function which can be used to download
1727 CRLs using the OCSP API.
1730 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1733 *) SSL_CONF* functions. These provide a common framework for application
1734 configuration using configuration files or command lines.
1737 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1738 message callback and prints the results. Needs compile time option
1739 "enable-ssl-trace". New options to s_client and s_server to enable
1743 *) New ctrl and macro to retrieve supported points extensions.
1744 Print out extension in s_server and s_client.
1747 *) New functions to retrieve certificate signature and signature
1751 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1755 *) New Suite B modes for TLS code. These use and enforce the requirements
1756 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1757 only use Suite B curves. The Suite B modes can be set by using the
1758 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1761 *) New chain verification flags for Suite B levels of security. Check
1762 algorithms are acceptable when flags are set in X509_verify_cert.
1765 *) Make tls1_check_chain return a set of flags indicating checks passed
1766 by a certificate chain. Add additional tests to handle client
1767 certificates: checks for matching certificate type and issuer name
1771 *) If an attempt is made to use a signature algorithm not in the peer
1772 preference list abort the handshake. If client has no suitable
1773 signature algorithms in response to a certificate request do not
1774 use the certificate.
1777 *) If server EC tmp key is not in client preference list abort handshake.
1780 *) Add support for certificate stores in CERT structure. This makes it
1781 possible to have different stores per SSL structure or one store in
1782 the parent SSL_CTX. Include distinct stores for certificate chain
1783 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1784 to build and store a certificate chain in CERT structure: returing
1785 an error if the chain cannot be built: this will allow applications
1786 to test if a chain is correctly configured.
1788 Note: if the CERT based stores are not set then the parent SSL_CTX
1789 store is used to retain compatibility with existing behaviour.
1793 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1794 mask based on the current session, check mask when sending client
1795 hello and checking the requested ciphersuite.
1798 *) New ctrls to retrieve and set certificate types in a certificate
1799 request message. Print out received values in s_client. If certificate
1800 types is not set with custom values set sensible values based on
1801 supported signature algorithms.
1804 *) Support for distinct client and server supported signature algorithms.
1807 *) Add certificate callback. If set this is called whenever a certificate
1808 is required by client or server. An application can decide which
1809 certificate chain to present based on arbitrary criteria: for example
1810 supported signature algorithms. Add very simple example to s_server.
1811 This fixes many of the problems and restrictions of the existing client
1812 certificate callback: for example you can now clear an existing
1813 certificate and specify the whole chain.
1816 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1817 the certificate can be used for (if anything). Set valid_flags field
1818 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1819 to have similar checks in it.
1821 Add new "cert_flags" field to CERT structure and include a "strict mode".
1822 This enforces some TLS certificate requirements (such as only permitting
1823 certificate signature algorithms contained in the supported algorithms
1824 extension) which some implementations ignore: this option should be used
1825 with caution as it could cause interoperability issues.
1828 *) Update and tidy signature algorithm extension processing. Work out
1829 shared signature algorithms based on preferences and peer algorithms
1830 and print them out in s_client and s_server. Abort handshake if no
1831 shared signature algorithms.
1834 *) Add new functions to allow customised supported signature algorithms
1835 for SSL and SSL_CTX structures. Add options to s_client and s_server
1839 *) New function SSL_certs_clear() to delete all references to certificates
1840 from an SSL structure. Before this once a certificate had been added
1841 it couldn't be removed.
1844 *) Integrate hostname, email address and IP address checking with certificate
1845 verification. New verify options supporting checking in openssl utility.
1848 *) Fixes and wildcard matching support to hostname and email checking
1849 functions. Add manual page.
1850 [Florian Weimer (Red Hat Product Security Team)]
1852 *) New functions to check a hostname email or IP address against a
1853 certificate. Add options x509 utility to print results of checks against
1857 *) Fix OCSP checking.
1858 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1860 *) Initial experimental support for explicitly trusted non-root CAs.
1861 OpenSSL still tries to build a complete chain to a root but if an
1862 intermediate CA has a trust setting included that is used. The first
1863 setting is used: whether to trust (e.g., -addtrust option to the x509
1867 *) Add -trusted_first option which attempts to find certificates in the
1868 trusted store even if an untrusted chain is also supplied.
1871 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1872 platform support for Linux and Android.
1875 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1878 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1879 When in FIPS mode the approved implementations are used as normal,
1880 when not in FIPS mode the internal unapproved versions are used instead.
1881 This means that the FIPS capable OpenSSL isn't forced to use the
1882 (often lower performance) FIPS implementations outside FIPS mode.
1885 *) Transparently support X9.42 DH parameters when calling
1886 PEM_read_bio_DHparameters. This means existing applications can handle
1887 the new parameter format automatically.
1890 *) Initial experimental support for X9.42 DH parameter format: mainly
1891 to support use of 'q' parameter for RFC5114 parameters.
1894 *) Add DH parameters from RFC5114 including test data to dhtest.
1897 *) Support for automatic EC temporary key parameter selection. If enabled
1898 the most preferred EC parameters are automatically used instead of
1899 hardcoded fixed parameters. Now a server just has to call:
1900 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1901 support ECDH and use the most appropriate parameters.
1904 *) Enhance and tidy EC curve and point format TLS extension code. Use
1905 static structures instead of allocation if default values are used.
1906 New ctrls to set curves we wish to support and to retrieve shared curves.
1907 Print out shared curves in s_server. New options to s_server and s_client
1908 to set list of supported curves.
1911 *) New ctrls to retrieve supported signature algorithms and
1912 supported curve values as an array of NIDs. Extend openssl utility
1913 to print out received values.
1916 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1917 between NIDs and the more common NIST names such as "P-256". Enhance
1918 ecparam utility and ECC method to recognise the NIST names for curves.
1921 *) Enhance SSL/TLS certificate chain handling to support different
1922 chains for each certificate instead of one chain in the parent SSL_CTX.
1925 *) Support for fixed DH ciphersuite client authentication: where both
1926 server and client use DH certificates with common parameters.
1929 *) Support for fixed DH ciphersuites: those requiring DH server
1933 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1935 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1936 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1937 X509_CINF_get_signature were reverted post internal team review.
1939 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1941 *) Build fixes for the Windows and OpenVMS platforms
1942 [Matt Caswell and Richard Levitte]
1944 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1946 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1947 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1948 dereference. This could lead to a Denial Of Service attack. Thanks to
1949 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1953 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1954 dtls1_buffer_record function under certain conditions. In particular this
1955 could occur if an attacker sent repeated DTLS records with the same
1956 sequence number but for the next epoch. The memory leak could be exploited
1957 by an attacker in a Denial of Service attack through memory exhaustion.
1958 Thanks to Chris Mueller for reporting this issue.
1962 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1963 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1964 method would be set to NULL which could later result in a NULL pointer
1965 dereference. Thanks to Frank Schmirler for reporting this issue.
1969 *) Abort handshake if server key exchange message is omitted for ephemeral
1972 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1973 reporting this issue.
1977 *) Remove non-export ephemeral RSA code on client and server. This code
1978 violated the TLS standard by allowing the use of temporary RSA keys in
1979 non-export ciphersuites and could be used by a server to effectively
1980 downgrade the RSA key length used to a value smaller than the server
1981 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1982 INRIA or reporting this issue.
1986 *) Fixed issue where DH client certificates are accepted without verification.
1987 An OpenSSL server will accept a DH certificate for client authentication
1988 without the certificate verify message. This effectively allows a client to
1989 authenticate without the use of a private key. This only affects servers
1990 which trust a client certificate authority which issues certificates
1991 containing DH keys: these are extremely rare and hardly ever encountered.
1992 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1997 *) Ensure that the session ID context of an SSL is updated when its
1998 SSL_CTX is updated via SSL_set_SSL_CTX.
2000 The session ID context is typically set from the parent SSL_CTX,
2001 and can vary with the CTX.
2004 *) Fix various certificate fingerprint issues.
2006 By using non-DER or invalid encodings outside the signed portion of a
2007 certificate the fingerprint can be changed without breaking the signature.
2008 Although no details of the signed portion of the certificate can be changed
2009 this can cause problems with some applications: e.g. those using the
2010 certificate fingerprint for blacklists.
2012 1. Reject signatures with non zero unused bits.
2014 If the BIT STRING containing the signature has non zero unused bits reject
2015 the signature. All current signature algorithms require zero unused bits.
2017 2. Check certificate algorithm consistency.
2019 Check the AlgorithmIdentifier inside TBS matches the one in the
2020 certificate signature. NB: this will result in signature failure
2021 errors for some broken certificates.
2023 Thanks to Konrad Kraszewski from Google for reporting this issue.
2025 3. Check DSA/ECDSA signatures use DER.
2027 Reencode DSA/ECDSA signatures and compare with the original received
2028 signature. Return an error if there is a mismatch.
2030 This will reject various cases including garbage after signature
2031 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2032 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2033 (negative or with leading zeroes).
2035 Further analysis was conducted and fixes were developed by Stephen Henson
2036 of the OpenSSL core team.
2041 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2042 results on some platforms, including x86_64. This bug occurs at random
2043 with a very low probability, and is not known to be exploitable in any
2044 way, though its exact impact is difficult to determine. Thanks to Pieter
2045 Wuille (Blockstream) who reported this issue and also suggested an initial
2046 fix. Further analysis was conducted by the OpenSSL development team and
2047 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2048 the OpenSSL core team.
2052 *) Do not resume sessions on the server if the negotiated protocol
2053 version does not match the session's version. Resuming with a different
2054 version, while not strictly forbidden by the RFC, is of questionable
2055 sanity and breaks all known clients.
2056 [David Benjamin, Emilia Käsper]
2058 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2059 early CCS messages during renegotiation. (Note that because
2060 renegotiation is encrypted, this early CCS was not exploitable.)
2063 *) Tighten client-side session ticket handling during renegotiation:
2064 ensure that the client only accepts a session ticket if the server sends
2065 the extension anew in the ServerHello. Previously, a TLS client would
2066 reuse the old extension state and thus accept a session ticket if one was
2067 announced in the initial ServerHello.
2069 Similarly, ensure that the client requires a session ticket if one
2070 was advertised in the ServerHello. Previously, a TLS client would
2071 ignore a missing NewSessionTicket message.
2074 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2076 *) SRTP Memory Leak.
2078 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2079 sends a carefully crafted handshake message, to cause OpenSSL to fail
2080 to free up to 64k of memory causing a memory leak. This could be
2081 exploited in a Denial Of Service attack. This issue affects OpenSSL
2082 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2083 whether SRTP is used or configured. Implementations of OpenSSL that
2084 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2086 The fix was developed by the OpenSSL team.
2090 *) Session Ticket Memory Leak.
2092 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2093 integrity of that ticket is first verified. In the event of a session
2094 ticket integrity check failing, OpenSSL will fail to free memory
2095 causing a memory leak. By sending a large number of invalid session
2096 tickets an attacker could exploit this issue in a Denial Of Service
2101 *) Build option no-ssl3 is incomplete.
2103 When OpenSSL is configured with "no-ssl3" as a build option, servers
2104 could accept and complete a SSL 3.0 handshake, and clients could be
2105 configured to send them.
2107 [Akamai and the OpenSSL team]
2109 *) Add support for TLS_FALLBACK_SCSV.
2110 Client applications doing fallback retries should call
2111 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2113 [Adam Langley, Bodo Moeller]
2115 *) Add additional DigestInfo checks.
2117 Reencode DigestInto in DER and check against the original when
2118 verifying RSA signature: this will reject any improperly encoded
2119 DigestInfo structures.
2121 Note: this is a precautionary measure and no attacks are currently known.
2125 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2127 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2128 SRP code can be overrun an internal buffer. Add sanity check that
2129 g, A, B < N to SRP code.
2131 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2132 Group for discovering this issue.
2136 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2137 TLS 1.0 instead of higher protocol versions when the ClientHello message
2138 is badly fragmented. This allows a man-in-the-middle attacker to force a
2139 downgrade to TLS 1.0 even if both the server and the client support a
2140 higher protocol version, by modifying the client's TLS records.
2142 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2143 researching this issue.
2147 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2148 to a denial of service attack. A malicious server can crash the client
2149 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2150 ciphersuite and sending carefully crafted handshake messages.
2152 Thanks to Felix Gröbert (Google) for discovering and researching this
2157 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2158 to leak memory. This can be exploited through a Denial of Service attack.
2159 Thanks to Adam Langley for discovering and researching this issue.
2163 *) An attacker can force openssl to consume large amounts of memory whilst
2164 processing DTLS handshake messages. This can be exploited through a
2165 Denial of Service attack.
2166 Thanks to Adam Langley for discovering and researching this issue.
2170 *) An attacker can force an error condition which causes openssl to crash
2171 whilst processing DTLS packets due to memory being freed twice. This
2172 can be exploited through a Denial of Service attack.
2173 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2178 *) If a multithreaded client connects to a malicious server using a resumed
2179 session and the server sends an ec point format extension it could write
2180 up to 255 bytes to freed memory.
2182 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2187 *) A malicious server can crash an OpenSSL client with a null pointer
2188 dereference (read) by specifying an SRP ciphersuite even though it was not
2189 properly negotiated with the client. This can be exploited through a
2190 Denial of Service attack.
2192 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2193 discovering and researching this issue.
2197 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2198 X509_name_oneline, X509_name_print_ex et al. to leak some information
2199 from the stack. Applications may be affected if they echo pretty printing
2200 output to the attacker.
2202 Thanks to Ivan Fratric (Google) for discovering this issue.
2204 [Emilia Käsper, and Steve Henson]
2206 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2207 for corner cases. (Certain input points at infinity could lead to
2208 bogus results, with non-infinity inputs mapped to infinity too.)
2211 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2213 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2214 handshake can force the use of weak keying material in OpenSSL
2215 SSL/TLS clients and servers.
2217 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2218 researching this issue. (CVE-2014-0224)
2219 [KIKUCHI Masashi, Steve Henson]
2221 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2222 OpenSSL DTLS client the code can be made to recurse eventually crashing
2225 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2227 [Imre Rad, Steve Henson]
2229 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2230 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2231 client or server. This is potentially exploitable to run arbitrary
2232 code on a vulnerable client or server.
2234 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2235 [Jüri Aedla, Steve Henson]
2237 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2238 are subject to a denial of service attack.
2240 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2241 this issue. (CVE-2014-3470)
2242 [Felix Gröbert, Ivan Fratric, Steve Henson]
2244 *) Harmonize version and its documentation. -f flag is used to display
2246 [mancha <mancha1@zoho.com>]
2248 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2249 in i2d_ECPrivateKey.
2250 [mancha <mancha1@zoho.com>]
2252 *) Fix some double frees. These are not thought to be exploitable.
2253 [mancha <mancha1@zoho.com>]
2255 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2257 *) A missing bounds check in the handling of the TLS heartbeat extension
2258 can be used to reveal up to 64k of memory to a connected client or
2261 Thanks for Neel Mehta of Google Security for discovering this bug and to
2262 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2263 preparing the fix (CVE-2014-0160)
2264 [Adam Langley, Bodo Moeller]
2266 *) Fix for the attack described in the paper "Recovering OpenSSL
2267 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2268 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2269 http://eprint.iacr.org/2014/140
2271 Thanks to Yuval Yarom and Naomi Benger for discovering this
2272 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2273 [Yuval Yarom and Naomi Benger]
2275 *) TLS pad extension: draft-agl-tls-padding-03
2277 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2278 TLS client Hello record length value would otherwise be > 255 and
2279 less that 512 pad with a dummy extension containing zeroes so it
2280 is at least 512 bytes long.
2282 [Adam Langley, Steve Henson]
2284 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2286 *) Fix for TLS record tampering bug. A carefully crafted invalid
2287 handshake could crash OpenSSL with a NULL pointer exception.
2288 Thanks to Anton Johansson for reporting this issues.
2291 *) Keep original DTLS digest and encryption contexts in retransmission
2292 structures so we can use the previous session parameters if they need
2293 to be resent. (CVE-2013-6450)
2296 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2297 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2298 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2299 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2300 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2301 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2302 [Rob Stradling, Adam Langley]
2304 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2306 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2307 supporting platforms or when small records were transferred.
2308 [Andy Polyakov, Steve Henson]
2310 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2312 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2314 This addresses the flaw in CBC record processing discovered by
2315 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2316 at: http://www.isg.rhul.ac.uk/tls/
2318 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2319 Security Group at Royal Holloway, University of London
2320 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2321 Emilia Käsper for the initial patch.
2323 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2325 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2326 ciphersuites which can be exploited in a denial of service attack.
2327 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2328 and detecting this bug and to Wolfgang Ettlinger
2329 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2333 *) Return an error when checking OCSP signatures when key is NULL.
2334 This fixes a DoS attack. (CVE-2013-0166)
2337 *) Make openssl verify return errors.
2338 [Chris Palmer <palmer@google.com> and Ben Laurie]
2340 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2341 the right response is stapled. Also change SSL_get_certificate()
2342 so it returns the certificate actually sent.
2343 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2344 [Rob Stradling <rob.stradling@comodo.com>]
2346 *) Fix possible deadlock when decoding public keys.
2349 *) Don't use TLS 1.0 record version number in initial client hello
2353 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2355 *) Sanity check record length before skipping explicit IV in TLS
2356 1.2, 1.1 and DTLS to fix DoS attack.
2358 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2359 fuzzing as a service testing platform.
2363 *) Initialise tkeylen properly when encrypting CMS messages.
2364 Thanks to Solar Designer of Openwall for reporting this issue.
2367 *) In FIPS mode don't try to use composite ciphers as they are not
2371 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2373 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2374 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2375 mean any application compiled against OpenSSL 1.0.0 headers setting
2376 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2377 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2378 0x10000000L Any application which was previously compiled against
2379 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2380 will need to be recompiled as a result. Letting be results in
2381 inability to disable specifically TLS 1.1 and in client context,
2382 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2385 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2386 disable just protocol X, but all protocols above X *if* there are
2387 protocols *below* X still enabled. In more practical terms it means
2388 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2389 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2390 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2394 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2396 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2397 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2398 in CRYPTO_realloc_clean.
2400 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2401 issue and to Adam Langley <agl@chromium.org> for fixing it.
2403 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2405 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2408 *) Workarounds for some broken servers that "hang" if a client hello
2409 record length exceeds 255 bytes.
2411 1. Do not use record version number > TLS 1.0 in initial client
2412 hello: some (but not all) hanging servers will now work.
2413 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2414 the number of ciphers sent in the client hello. This should be
2415 set to an even number, such as 50, for example by passing:
2416 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2417 Most broken servers should now work.
2418 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2419 TLS 1.2 client support entirely.
2422 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2425 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2427 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2428 STRING form instead of a DigestInfo.
2431 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2432 and the RSA_sign/RSA_verify functions. This was made more apparent when
2433 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2434 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2435 the correct format in RSA_verify so both forms transparently work.
2438 *) Some servers which support TLS 1.0 can choke if we initially indicate
2439 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2440 encrypted premaster secret. As a workaround use the maximum permitted
2441 client version in client hello, this should keep such servers happy
2442 and still work with previous versions of OpenSSL.
2445 *) Add support for TLS/DTLS heartbeats.
2446 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2448 *) Add support for SCTP.
2449 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2451 *) Improved PRNG seeding for VOS.
2452 [Paul Green <Paul.Green@stratus.com>]
2454 *) Extensive assembler packs updates, most notably:
2456 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2457 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2458 - x86_64: bit-sliced AES implementation;
2459 - ARM: NEON support, contemporary platforms optimizations;
2460 - s390x: z196 support;
2461 - *: GHASH and GF(2^m) multiplication implementations;
2465 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2466 (removal of unnecessary code)
2467 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2469 *) Add TLS key material exporter from RFC 5705.
2472 *) Add DTLS-SRTP negotiation from RFC 5764.
2475 *) Add Next Protocol Negotiation,
2476 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2477 disabled with a no-npn flag to config or Configure. Code donated
2479 [Adam Langley <agl@google.com> and Ben Laurie]
2481 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2482 NIST-P256, NIST-P521, with constant-time single point multiplication on
2483 typical inputs. Compiler support for the nonstandard type __uint128_t is
2484 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2485 Code made available under Apache License version 2.0.
2487 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2488 line to include this in your build of OpenSSL, and run "make depend" (or
2489 "make update"). This enables the following EC_METHODs:
2491 EC_GFp_nistp224_method()
2492 EC_GFp_nistp256_method()
2493 EC_GFp_nistp521_method()
2495 EC_GROUP_new_by_curve_name() will automatically use these (while
2496 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2498 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2500 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2501 all platforms. Move ssize_t definition from e_os.h to the public
2502 header file e_os2.h as it now appears in public header file cms.h
2505 *) New -sigopt option to the ca, req and x509 utilities. Additional
2506 signature parameters can be passed using this option and in
2510 *) Add RSA PSS signing function. This will generate and set the
2511 appropriate AlgorithmIdentifiers for PSS based on those in the
2512 corresponding EVP_MD_CTX structure. No application support yet.
2515 *) Support for companion algorithm specific ASN1 signing routines.
2516 New function ASN1_item_sign_ctx() signs a pre-initialised
2517 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2518 the appropriate parameters.
2521 *) Add new algorithm specific ASN1 verification initialisation function
2522 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2523 handling will be the same no matter what EVP_PKEY_METHOD is used.
2524 Add a PSS handler to support verification of PSS signatures: checked
2525 against a number of sample certificates.
2528 *) Add signature printing for PSS. Add PSS OIDs.
2529 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2531 *) Add algorithm specific signature printing. An individual ASN1 method
2532 can now print out signatures instead of the standard hex dump.
2534 More complex signatures (e.g. PSS) can print out more meaningful
2535 information. Include DSA version that prints out the signature
2539 *) Password based recipient info support for CMS library: implementing
2543 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2544 neatly separates the code into cipher and PBE sections and is required
2545 for some algorithms that split PBES2 into separate pieces (such as
2546 password based CMS).
2549 *) Session-handling fixes:
2550 - Fix handling of connections that are resuming with a session ID,
2551 but also support Session Tickets.
2552 - Fix a bug that suppressed issuing of a new ticket if the client
2553 presented a ticket with an expired session.
2554 - Try to set the ticket lifetime hint to something reasonable.
2555 - Make tickets shorter by excluding irrelevant information.
2556 - On the client side, don't ignore renewed tickets.
2557 [Adam Langley, Bodo Moeller (Google)]
2559 *) Fix PSK session representation.
2562 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2564 This work was sponsored by Intel.
2567 *) Add GCM support to TLS library. Some custom code is needed to split
2568 the IV between the fixed (from PRF) and explicit (from TLS record)
2569 portions. This adds all GCM ciphersuites supported by RFC5288 and
2570 RFC5289. Generalise some AES* cipherstrings to include GCM and
2571 add a special AESGCM string for GCM only.
2574 *) Expand range of ctrls for AES GCM. Permit setting invocation
2575 field on decrypt and retrieval of invocation field only on encrypt.
2578 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2579 As required by RFC5289 these ciphersuites cannot be used if for
2580 versions of TLS earlier than 1.2.
2583 *) For FIPS capable OpenSSL interpret a NULL default public key method
2584 as unset and return the appropriate default but do *not* set the default.
2585 This means we can return the appropriate method in applications that
2586 switch between FIPS and non-FIPS modes.
2589 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2590 ENGINE is used then we cannot handle that in the FIPS module so we
2591 keep original code iff non-FIPS operations are allowed.
2594 *) Add -attime option to openssl utilities.
2595 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2597 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2600 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2601 FIPS EC methods unconditionally for now.
2604 *) New build option no-ec2m to disable characteristic 2 code.
2607 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2608 all cases can be covered as some introduce binary incompatibilities.
2611 *) Redirect RSA operations to FIPS module including keygen,
2612 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2615 *) Add similar low level API blocking to ciphers.
2618 *) Low level digest APIs are not approved in FIPS mode: any attempt
2619 to use these will cause a fatal error. Applications that *really* want
2620 to use them can use the private_* version instead.
2623 *) Redirect cipher operations to FIPS module for FIPS builds.
2626 *) Redirect digest operations to FIPS module for FIPS builds.
2629 *) Update build system to add "fips" flag which will link in fipscanister.o
2630 for static and shared library builds embedding a signature if needed.
2633 *) Output TLS supported curves in preference order instead of numerical
2634 order. This is currently hardcoded for the highest order curves first.
2635 This should be configurable so applications can judge speed vs strength.
2638 *) Add TLS v1.2 server support for client authentication.
2641 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2645 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2646 FIPS modules versions.
2649 *) Add TLS v1.2 client side support for client authentication. Keep cache
2650 of handshake records longer as we don't know the hash algorithm to use
2651 until after the certificate request message is received.
2654 *) Initial TLS v1.2 client support. Add a default signature algorithms
2655 extension including all the algorithms we support. Parse new signature
2656 format in client key exchange. Relax some ECC signing restrictions for
2657 TLS v1.2 as indicated in RFC5246.
2660 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2661 to new signature format when needed using client digest preference.
2662 All server ciphersuites should now work correctly in TLS v1.2. No client
2663 support yet and no support for client certificates.
2666 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2667 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2668 ciphersuites. At present only RSA key exchange ciphersuites work with
2669 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2670 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2671 and version checking.
2674 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2675 with this defined it will not be affected by any changes to ssl internal
2676 structures. Add several utility functions to allow openssl application
2677 to work with OPENSSL_NO_SSL_INTERN defined.
2681 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2683 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2686 *) Permit abbreviated handshakes when renegotiating using the function
2687 SSL_renegotiate_abbreviated().
2688 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2690 *) Add call to ENGINE_register_all_complete() to
2691 ENGINE_load_builtin_engines(), so some implementations get used
2692 automatically instead of needing explicit application support.
2695 *) Add support for TLS key exporter as described in RFC5705.
2696 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2698 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2699 a few changes are required:
2701 Add SSL_OP_NO_TLSv1_1 flag.
2702 Add TLSv1_1 methods.
2703 Update version checking logic to handle version 1.1.
2704 Add explicit IV handling (ported from DTLS code).
2705 Add command line options to s_client/s_server.
2708 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2710 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2711 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2712 content decryption and always return the same error. Note: this attack
2713 needs on average 2^20 messages so it only affects automated senders. The
2714 old behaviour can be reenabled in the CMS code by setting the
2715 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2716 an MMA defence is not necessary.
2717 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2718 this issue. (CVE-2012-0884)
2721 *) Fix CVE-2011-4619: make sure we really are receiving a
2722 client hello before rejecting multiple SGC restarts. Thanks to
2723 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2726 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2728 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2729 Thanks to Antonio Martin, Enterprise Secure Access Research and
2730 Development, Cisco Systems, Inc. for discovering this bug and
2731 preparing a fix. (CVE-2012-0050)
2734 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2736 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2737 of the Vaudenay padding oracle attack on CBC mode encryption
2738 which enables an efficient plaintext recovery attack against
2739 the OpenSSL implementation of DTLS. Their attack exploits timing
2740 differences arising during decryption processing. A research
2741 paper describing this attack can be found at:
2742 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2743 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2744 Security Group at Royal Holloway, University of London
2745 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2746 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2747 for preparing the fix. (CVE-2011-4108)
2748 [Robin Seggelmann, Michael Tuexen]
2750 *) Clear bytes used for block padding of SSL 3.0 records.
2752 [Adam Langley (Google)]
2754 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2755 Kadianakis <desnacked@gmail.com> for discovering this issue and
2756 Adam Langley for preparing the fix. (CVE-2011-4619)
2757 [Adam Langley (Google)]
2759 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2760 [Andrey Kulikov <amdeich@gmail.com>]
2762 *) Prevent malformed RFC3779 data triggering an assertion failure.
2763 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2764 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2765 [Rob Austein <sra@hactrn.net>]
2767 *) Improved PRNG seeding for VOS.
2768 [Paul Green <Paul.Green@stratus.com>]
2770 *) Fix ssl_ciph.c set-up race.
2771 [Adam Langley (Google)]
2773 *) Fix spurious failures in ecdsatest.c.
2774 [Emilia Käsper (Google)]
2776 *) Fix the BIO_f_buffer() implementation (which was mixing different
2777 interpretations of the '..._len' fields).
2778 [Adam Langley (Google)]
2780 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2781 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2782 threads won't reuse the same blinding coefficients.
2784 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2785 lock to call BN_BLINDING_invert_ex, and avoids one use of
2786 BN_BLINDING_update for each BN_BLINDING structure (previously,
2787 the last update always remained unused).
2788 [Emilia Käsper (Google)]
2790 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2791 [Bob Buckholz (Google)]
2793 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2795 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2796 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2797 [Kaspar Brand <ossl@velox.ch>]
2799 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2800 for multi-threaded use of ECDH. (CVE-2011-3210)
2801 [Adam Langley (Google)]
2803 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2806 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2807 signature public key algorithm by using OID xref utilities instead.
2808 Before this you could only use some ECC ciphersuites with SHA1 only.
2811 *) Add protection against ECDSA timing attacks as mentioned in the paper
2812 by Billy Bob Brumley and Nicola Tuveri, see:
2814 http://eprint.iacr.org/2011/232.pdf
2816 [Billy Bob Brumley and Nicola Tuveri]
2818 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2820 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2821 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2823 *) Fix bug in string printing code: if *any* escaping is enabled we must
2824 escape the escape character (backslash) or the resulting string is
2828 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2830 *) Disable code workaround for ancient and obsolete Netscape browsers
2831 and servers: an attacker can use it in a ciphersuite downgrade attack.
2832 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2835 *) Fixed J-PAKE implementation error, originally discovered by
2836 Sebastien Martini, further info and confirmation from Stefan
2837 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2840 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2842 *) Fix extension code to avoid race conditions which can result in a buffer
2843 overrun vulnerability: resumed sessions must not be modified as they can
2844 be shared by multiple threads. CVE-2010-3864
2847 *) Fix WIN32 build system to correctly link an ENGINE directory into
2851 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2853 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2855 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2857 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2859 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2860 context. The operation can be customised via the ctrl mechanism in
2861 case ENGINEs want to include additional functionality.
2864 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2867 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2868 output hashes compatible with older versions of OpenSSL.
2869 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2871 *) Fix compression algorithm handling: if resuming a session use the
2872 compression algorithm of the resumed session instead of determining
2873 it from client hello again. Don't allow server to change algorithm.
2876 *) Add load_crls() function to apps tidying load_certs() too. Add option
2877 to verify utility to allow additional CRLs to be included.
2880 *) Update OCSP request code to permit adding custom headers to the request:
2881 some responders need this.
2884 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2886 [Julia Lawall <julia@diku.dk>]
2888 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2889 needlessly dereferenced structures, used obsolete functions and
2890 didn't handle all updated verify codes correctly.
2893 *) Disable MD2 in the default configuration.
2896 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2897 indicate the initial BIO being pushed or popped. This makes it possible
2898 to determine whether the BIO is the one explicitly called or as a result
2899 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2900 it handles reference counts correctly and doesn't zero out the I/O bio
2901 when it is not being explicitly popped. WARNING: applications which
2902 included workarounds for the old buggy behaviour will need to be modified
2903 or they could free up already freed BIOs.
2906 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2907 renaming to all platforms (within the 0.9.8 branch, this was
2908 done conditionally on Netware platforms to avoid a name clash).
2909 [Guenter <lists@gknw.net>]
2911 *) Add ECDHE and PSK support to DTLS.
2912 [Michael Tuexen <tuexen@fh-muenster.de>]
2914 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2918 *) Add "missing" function EVP_MD_flags() (without this the only way to
2919 retrieve a digest flags is by accessing the structure directly. Update
2920 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2921 or cipher is registered as in the "from" argument. Print out all
2922 registered digests in the dgst usage message instead of manually
2923 attempting to work them out.
2926 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2927 this allows the use of compression and extensions. Change default cipher
2928 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2929 by default unless an application cipher string requests it.
2932 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2933 key ids to find matching certificates and keys but some PKCS#12 files
2934 don't follow the (somewhat unwritten) rules and this strategy fails.
2935 Now just gather all certificates together and the first private key
2936 then look for the first certificate that matches the key.
2939 *) Support use of registered digest and cipher names for dgst and cipher
2940 commands instead of having to add each one as a special case. So now
2947 openssl dgst -sha256 foo
2949 and this works for ENGINE based algorithms too.
2953 *) Update Gost ENGINE to support parameter files.
2954 [Victor B. Wagner <vitus@cryptocom.ru>]
2956 *) Support GeneralizedTime in ca utility.
2957 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2959 *) Enhance the hash format used for certificate directory links. The new
2960 form uses the canonical encoding (meaning equivalent names will work
2961 even if they aren't identical) and uses SHA1 instead of MD5. This form
2962 is incompatible with the older format and as a result c_rehash should
2963 be used to rebuild symbolic links.
2966 *) Make PKCS#8 the default write format for private keys, replacing the
2967 traditional format. This form is standardised, more secure and doesn't
2968 include an implicit MD5 dependency.
2971 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2972 committed to OpenSSL should pass this lot as a minimum.
2975 *) Add session ticket override functionality for use by EAP-FAST.
2976 [Jouni Malinen <j@w1.fi>]
2978 *) Modify HMAC functions to return a value. Since these can be implemented
2979 in an ENGINE errors can occur.
2982 *) Type-checked OBJ_bsearch_ex.
2985 *) Type-checked OBJ_bsearch. Also some constification necessitated
2986 by type-checking. Still to come: TXT_DB, bsearch(?),
2987 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2991 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2992 seconds to a tm structure directly, instead of going through OS
2993 specific date routines. This avoids any issues with OS routines such
2994 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2995 and X509_time_adj_ex() to cover the extended range. The existing
2996 X509_time_adj() is still usable and will no longer have any date issues.
2999 *) Delta CRL support. New use deltas option which will attempt to locate
3000 and search any appropriate delta CRLs available.
3002 This work was sponsored by Google.
3005 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3006 code and add additional score elements. Validate alternate CRL paths
3007 as part of the CRL checking and indicate a new error "CRL path validation
3008 error" in this case. Applications wanting additional details can use
3009 the verify callback and check the new "parent" field. If this is not
3010 NULL CRL path validation is taking place. Existing applications wont
3011 see this because it requires extended CRL support which is off by
3014 This work was sponsored by Google.
3017 *) Support for freshest CRL extension.
3019 This work was sponsored by Google.
3022 *) Initial indirect CRL support. Currently only supported in the CRLs
3023 passed directly and not via lookup. Process certificate issuer
3024 CRL entry extension and lookup CRL entries by bother issuer name
3025 and serial number. Check and process CRL issuer entry in IDP extension.
3027 This work was sponsored by Google.
3030 *) Add support for distinct certificate and CRL paths. The CRL issuer
3031 certificate is validated separately in this case. Only enabled if
3032 an extended CRL support flag is set: this flag will enable additional
3033 CRL functionality in future.
3035 This work was sponsored by Google.
3038 *) Add support for policy mappings extension.
3040 This work was sponsored by Google.
3043 *) Fixes to pathlength constraint, self issued certificate handling,
3044 policy processing to align with RFC3280 and PKITS tests.
3046 This work was sponsored by Google.
3049 *) Support for name constraints certificate extension. DN, email, DNS
3050 and URI types are currently supported.
3052 This work was sponsored by Google.
3055 *) To cater for systems that provide a pointer-based thread ID rather
3056 than numeric, deprecate the current numeric thread ID mechanism and
3057 replace it with a structure and associated callback type. This
3058 mechanism allows a numeric "hash" to be extracted from a thread ID in
3059 either case, and on platforms where pointers are larger than 'long',
3060 mixing is done to help ensure the numeric 'hash' is usable even if it
3061 can't be guaranteed unique. The default mechanism is to use "&errno"
3062 as a pointer-based thread ID to distinguish between threads.
3064 Applications that want to provide their own thread IDs should now use
3065 CRYPTO_THREADID_set_callback() to register a callback that will call
3066 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3068 Note that ERR_remove_state() is now deprecated, because it is tied
3069 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3070 to free the current thread's error state should be replaced by
3071 ERR_remove_thread_state(NULL).
3073 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3074 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3075 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3076 application was previously providing a numeric thread callback that
3077 was inappropriate for distinguishing threads, then uniqueness might
3078 have been obtained with &errno that happened immediately in the
3079 intermediate development versions of OpenSSL; this is no longer the
3080 case, the numeric thread callback will now override the automatic use
3082 [Geoff Thorpe, with help from Bodo Moeller]
3084 *) Initial support for different CRL issuing certificates. This covers a
3085 simple case where the self issued certificates in the chain exist and
3086 the real CRL issuer is higher in the existing chain.
3088 This work was sponsored by Google.
3091 *) Removed effectively defunct crypto/store from the build.
3094 *) Revamp of STACK to provide stronger type-checking. Still to come:
3095 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3096 ASN1_STRING, CONF_VALUE.
3099 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3100 RAM on SSL connections. This option can save about 34k per idle SSL.
3103 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3104 STACK, TXT_DB, bsearch, qsort.
3107 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3108 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3109 support for data, signedData, compressedData, digestedData and
3110 encryptedData, envelopedData types included. Scripts to check against
3111 RFC4134 examples draft and interop and consistency checks of many
3112 content types and variants.
3115 *) Add options to enc utility to support use of zlib compression BIO.
3118 *) Extend mk1mf to support importing of options and assembly language
3119 files from Configure script, currently only included in VC-WIN32.
3120 The assembly language rules can now optionally generate the source
3121 files from the associated perl scripts.
3124 *) Implement remaining functionality needed to support GOST ciphersuites.
3125 Interop testing has been performed using CryptoPro implementations.
3126 [Victor B. Wagner <vitus@cryptocom.ru>]
3128 *) s390x assembler pack.
3131 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3135 *) Implement Opaque PRF Input TLS extension as specified in
3136 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3137 official specification yet and no extension type assignment by
3138 IANA exists, this extension (for now) will have to be explicitly
3139 enabled when building OpenSSL by providing the extension number
3140 to use. For example, specify an option
3142 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3144 to the "config" or "Configure" script to enable the extension,
3145 assuming extension number 0x9527 (which is a completely arbitrary
3146 and unofficial assignment based on the MD5 hash of the Internet
3147 Draft). Note that by doing so, you potentially lose
3148 interoperability with other TLS implementations since these might
3149 be using the same extension number for other purposes.
3151 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3152 opaque PRF input value to use in the handshake. This will create
3153 an interal copy of the length-'len' string at 'src', and will
3154 return non-zero for success.
3156 To get more control and flexibility, provide a callback function
3159 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3160 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3164 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3167 Callback function 'cb' will be called in handshakes, and is
3168 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3169 Argument 'arg' is for application purposes (the value as given to
3170 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3171 be provided to the callback function). The callback function
3172 has to return non-zero to report success: usually 1 to use opaque
3173 PRF input just if possible, or 2 to enforce use of the opaque PRF
3174 input. In the latter case, the library will abort the handshake
3175 if opaque PRF input is not successfully negotiated.
3177 Arguments 'peerinput' and 'len' given to the callback function
3178 will always be NULL and 0 in the case of a client. A server will
3179 see the client's opaque PRF input through these variables if
3180 available (NULL and 0 otherwise). Note that if the server
3181 provides an opaque PRF input, the length must be the same as the
3182 length of the client's opaque PRF input.
3184 Note that the callback function will only be called when creating
3185 a new session (session resumption can resume whatever was
3186 previously negotiated), and will not be called in SSL 2.0
3187 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3188 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3189 for applications that need to enforce opaque PRF input.
3193 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3196 [Victor B. Wagner <vitus@cryptocom.ru>]
3198 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3199 RFC4507bis. The encrypted ticket format is an encrypted encoded
3200 SSL_SESSION structure, that way new session features are automatically
3203 If a client application caches session in an SSL_SESSION structure
3204 support is transparent because tickets are now stored in the encoded
3207 The SSL_CTX structure automatically generates keys for ticket
3208 protection in servers so again support should be possible
3209 with no application modification.
3211 If a client or server wishes to disable RFC4507 support then the option
3212 SSL_OP_NO_TICKET can be set.
3214 Add a TLS extension debugging callback to allow the contents of any client
3215 or server extensions to be examined.
3217 This work was sponsored by Google.
3220 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3221 OpenSSL should now compile cleanly on gcc 4.2
3222 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3224 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3225 support including streaming MAC support: this is required for GOST
3226 ciphersuite support.
3227 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3229 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3230 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3231 to output in BER and PEM format.
3234 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3235 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3236 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3237 ENGINE support for HMAC keys which are unextractable. New -mac and
3238 -macopt options to dgst utility.
3241 *) New option -sigopt to dgst utility. Update dgst to use
3242 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3243 alternative signing parameters such as X9.31 or PSS in the dgst
3247 *) Change ssl_cipher_apply_rule(), the internal function that does
3248 the work each time a ciphersuite string requests enabling
3249 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3250 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3251 the order of disabled ciphersuites such that those ciphersuites
3252 that most recently went from enabled to disabled not only stay
3253 in order with respect to each other, but also have higher priority
3254 than other disabled ciphersuites the next time ciphersuites are
3257 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3258 the same ciphersuites as with "HIGH" alone, but in a specific
3259 order where the PSK ciphersuites come first (since they are the
3260 most recently disabled ciphersuites when "HIGH" is parsed).
3262 Also, change ssl_create_cipher_list() (using this new
3263 funcionality) such that between otherwise identical
3264 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3268 *) Change ssl_create_cipher_list() so that it automatically
3269 arranges the ciphersuites in reasonable order before starting
3270 to process the rule string. Thus, the definition for "DEFAULT"
3271 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3272 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3273 This makes it much easier to arrive at a reasonable default order
3274 in applications for which anonymous ciphers are OK (meaning
3275 that you can't actually use DEFAULT).
3276 [Bodo Moeller; suggested by Victor Duchovni]
3278 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3279 processing) into multiple integers instead of setting
3280 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3281 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3282 (These masks as well as the individual bit definitions are hidden
3283 away into the non-exported interface ssl/ssl_locl.h, so this
3284 change to the definition of the SSL_CIPHER structure shouldn't
3285 affect applications.) This give us more bits for each of these
3286 categories, so there is no longer a need to coagulate AES128 and
3287 AES256 into a single algorithm bit, and to coagulate Camellia128
3288 and Camellia256 into a single algorithm bit, which has led to all
3291 Thus, among other things, the kludge introduced in 0.9.7m and
3292 0.9.8e for masking out AES256 independently of AES128 or masking
3293 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3295 With the change, we also introduce new ciphersuite aliases that
3296 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3300 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3301 Use the leftmost N bytes of the signature input if the input is
3302 larger than the prime q (with N being the size in bytes of q).
3305 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3306 it yet and it is largely untested.
3309 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3312 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3313 some compilers (gcc 4.2 and later) reject their use. Safestack is
3314 reimplemented. Update ASN1 to avoid use of legacy functions.
3317 *) Win32/64 targets are linked with Winsock2.
3320 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3321 to external functions. This can be used to increase CRL handling
3322 efficiency especially when CRLs are very large by (for example) storing
3323 the CRL revoked certificates in a database.
3326 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3327 new CRLs added to a directory can be used. New command line option
3328 -verify_return_error to s_client and s_server. This causes real errors
3329 to be returned by the verify callback instead of carrying on no matter
3330 what. This reflects the way a "real world" verify callback would behave.
3333 *) GOST engine, supporting several GOST algorithms and public key formats.
3334 Kindly donated by Cryptocom.
3337 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3338 partitioned by DP are handled but no indirect CRL or reason partitioning
3339 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3340 selected via a scoring technique which handles IDP and AKID in CRLs.
3343 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3344 will ultimately be used for all verify operations: this will remove the
3345 X509_STORE dependency on certificate verification and allow alternative
3346 lookup methods. X509_STORE based implementations of these two callbacks.
3349 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3350 Modify get_crl() to find a valid (unexpired) CRL if possible.
3353 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3354 this would be called X509_CRL_cmp() but that name is already used by
3355 a function that just compares CRL issuer names. Cache several CRL
3356 extensions in X509_CRL structure and cache CRLDP in X509.
3359 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3360 this maps equivalent X509_NAME structures into a consistent structure.
3361 Name comparison can then be performed rapidly using memcmp().
3364 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3368 *) Allow digests to supply their own micalg string for S/MIME type using
3369 the ctrl EVP_MD_CTRL_MICALG.
3372 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3373 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3374 ctrl. It can then customise the structure before and/or after signing
3378 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3379 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3380 to free up any added signature OIDs.
3383 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3384 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3385 digest and cipher tables. New options added to openssl utility:
3386 list-message-digest-algorithms and list-cipher-algorithms.
3389 *) Change the array representation of binary polynomials: the list
3390 of degrees of non-zero coefficients is now terminated with -1.
3391 Previously it was terminated with 0, which was also part of the
3392 value; thus, the array representation was not applicable to
3393 polynomials where t^0 has coefficient zero. This change makes
3394 the array representation useful in a more general context.
3397 *) Various modifications and fixes to SSL/TLS cipher string
3398 handling. For ECC, the code now distinguishes between fixed ECDH
3399 with RSA certificates on the one hand and with ECDSA certificates
3400 on the other hand, since these are separate ciphersuites. The
3401 unused code for Fortezza ciphersuites has been removed.
3403 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3404 (not "ECDHE"). For consistency with the code for DH
3405 certificates, use of ECDH certificates is now considered ECDH
3406 authentication, not RSA or ECDSA authentication (the latter is
3407 merely the CA's signing algorithm and not actively used in the
3410 The temporary ciphersuite alias "ECCdraft" is no longer
3411 available, and ECC ciphersuites are no longer excluded from "ALL"
3412 and "DEFAULT". The following aliases now exist for RFC 4492
3413 ciphersuites, most of these by analogy with the DH case:
3415 kECDHr - ECDH cert, signed with RSA
3416 kECDHe - ECDH cert, signed with ECDSA
3417 kECDH - ECDH cert (signed with either RSA or ECDSA)
3418 kEECDH - ephemeral ECDH
3419 ECDH - ECDH cert or ephemeral ECDH
3425 AECDH - anonymous ECDH
3426 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3430 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3431 Use correct micalg parameters depending on digest(s) in signed message.
3434 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3435 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3438 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3439 an engine to register a method. Add ENGINE lookups for methods and
3440 functional reference processing.
3443 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3444 EVP_{Sign,Verify}* which allow an application to customise the signature
3448 *) New -resign option to smime utility. This adds one or more signers
3449 to an existing PKCS#7 signedData structure. Also -md option to use an
3450 alternative message digest algorithm for signing.
3453 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3454 create PKCS7 structures containing multiple signers. Update smime
3455 application to support multiple signers.
3458 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3462 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3463 Reorganize PBE internals to lookup from a static table using NIDs,
3464 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3465 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3466 PRF which will be automatically used with PBES2.
3469 *) Replace the algorithm specific calls to generate keys in "req" with the
3473 *) Update PKCS#7 enveloped data routines to use new API. This is now
3474 supported by any public key method supporting the encrypt operation. A
3475 ctrl is added to allow the public key algorithm to examine or modify
3476 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3480 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3481 a default digest type to use. In most cases this will be SHA1 but some
3482 algorithms (such as GOST) need to specify an alternative digest. The
3483 return value indicates how strong the preference is 1 means optional and
3484 2 is mandatory (that is it is the only supported type). Modify
3485 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3486 use the default md. Update openssl utilities to use the default digest
3487 type for signing if it is not explicitly indicated.
3490 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3491 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3492 signing method from the key type. This effectively removes the link
3493 between digests and public key types.
3496 *) Add an OID cross reference table and utility functions. Its purpose is to
3497 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3498 rsaEncryption. This will allow some of the algorithm specific hackery
3499 needed to use the correct OID to be removed.
3502 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3503 structures for PKCS7_sign(). They are now set up by the relevant public
3507 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3510 *) Add support for key derivation (agreement) in the API, DH method and
3514 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3515 public and private key formats. As a side effect these add additional
3516 command line functionality not previously available: DSA signatures can be
3517 generated and verified using pkeyutl and DH key support and generation in
3522 [Oliver Tappe <zooey@hirschkaefer.de>]
3524 *) New make target "install_html_docs" installs HTML renditions of the
3526 [Oliver Tappe <zooey@hirschkaefer.de>]
3528 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3529 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3530 support key and parameter generation and add initial key generation
3531 functionality for RSA.
3534 *) Add functions for main EVP_PKEY_method operations. The undocumented
3535 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3536 EVP_PKEY_{encrypt,decrypt}_old.
3539 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3540 key API, doesn't do much yet.
3543 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3544 public key algorithms. New option to openssl utility:
3545 "list-public-key-algorithms" to print out info.
3548 *) Implement the Supported Elliptic Curves Extension for
3549 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3552 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3553 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3556 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3557 utilities such as rsa, dsa, dsaparam etc except they process any key
3561 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3562 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3563 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3567 *) Initial support for pluggable public key ASN1.
3568 De-spaghettify the public key ASN1 handling. Move public and private
3569 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3570 algorithm specific handling to a single module within the relevant
3571 algorithm directory. Add functions to allow (near) opaque processing
3572 of public and private key structures.
3575 *) Implement the Supported Point Formats Extension for
3576 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3579 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3580 for the psk identity [hint] and the psk callback functions to the
3581 SSL_SESSION, SSL and SSL_CTX structure.
3584 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3588 SSL_CTX_use_psk_identity_hint
3589 SSL_get_psk_identity_hint
3590 SSL_get_psk_identity
3591 SSL_use_psk_identity_hint
3593 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3595 *) Add RFC 3161 compliant time stamp request creation, response generation
3596 and response verification functionality.
3597 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3599 *) Add initial support for TLS extensions, specifically for the server_name
3600 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3601 have new members for a host name. The SSL data structure has an
3602 additional member SSL_CTX *initial_ctx so that new sessions can be
3603 stored in that context to allow for session resumption, even after the
3604 SSL has been switched to a new SSL_CTX in reaction to a client's
3605 server_name extension.
3607 New functions (subject to change):
3609 SSL_get_servername()
3610 SSL_get_servername_type()
3613 New CTRL codes and macros (subject to change):
3615 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3616 - SSL_CTX_set_tlsext_servername_callback()
3617 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3618 - SSL_CTX_set_tlsext_servername_arg()
3619 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3621 openssl s_client has a new '-servername ...' option.
3623 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3624 '-key2 ...', '-servername_fatal' (subject to change). This allows
3625 testing the HostName extension for a specific single host name ('-cert'
3626 and '-key' remain fallbacks for handshakes without HostName
3627 negotiation). If the unrecognized_name alert has to be sent, this by
3628 default is a warning; it becomes fatal with the '-servername_fatal'
3631 [Peter Sylvester, Remy Allais, Christophe Renou]
3633 *) Whirlpool hash implementation is added.
3636 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3637 bn(64,32). Because of instruction set limitations it doesn't have
3638 any negative impact on performance. This was done mostly in order
3639 to make it possible to share assembler modules, such as bn_mul_mont
3640 implementations, between 32- and 64-bit builds without hassle.
3643 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3644 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3648 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3649 dedicated Montgomery multiplication procedure, is introduced.
3650 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3651 "64-bit" performance on certain 32-bit targets.
3654 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3655 in SSL structures. New SSL ctrl to set maximum send fragment size.
3656 Save memory by seeting the I/O buffer sizes dynamically instead of
3657 using the maximum available value.
3660 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3661 in addition to the text details.
3664 *) Very, very preliminary EXPERIMENTAL support for printing of general
3665 ASN1 structures. This currently produces rather ugly output and doesn't
3666 handle several customised structures at all.
3669 *) Integrated support for PVK file format and some related formats such
3670 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3671 these in the 'rsa' and 'dsa' utilities.
3674 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3677 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3678 place for the (very old) "NETSCAPE" format certificates which are now
3679 handled using new ASN1 code equivalents.
3682 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3683 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3684 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3687 *) Modify CRL distribution points extension code to print out previously
3688 unsupported fields. Enhance extension setting code to allow setting of
3692 *) Add print and set support for Issuing Distribution Point CRL extension.
3695 *) Change 'Configure' script to enable Camellia by default.
3698 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3700 *) When rejecting SSL/TLS records due to an incorrect version number, never
3701 update s->server with a new major version number. As of
3702 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3703 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3704 the previous behavior could result in a read attempt at NULL when
3705 receiving specific incorrect SSL/TLS records once record payload
3706 protection is active. (CVE-2010-0740)
3707 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3709 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3710 could be crashed if the relevant tables were not present (e.g. chrooted).
3711 [Tomas Hoger <thoger@redhat.com>]
3713 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3715 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3716 [Martin Olsson, Neel Mehta]
3718 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3719 accommodate for stack sorting, always a write lock!).
3722 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3723 excessive delays in the RAND_poll(): over a minute. As a workaround
3724 include a time check in the inner Heap32Next loop too.
projects / openssl.git / blob