5 Changes between 1.0.2g and 1.1.0 [xx XXX xxxx]
7 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
10 *) To enable users to have their own config files and build file templates,
11 Configure looks in the directory indicated by the environment variable
12 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
13 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
14 name and is used as is.
17 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
18 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
19 X509_CERT_FILE_CTX was removed.
22 *) "shared" builds are now the default. To create only static libraries use
23 the "no-shared" Configure option.
26 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
27 All of these option have not worked for some while and are fundamental
31 *) Make various cleanup routines no-ops and mark them as deprecated. Most
32 global cleanup functions are no longer required because they are handled
33 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
34 Explicitly de-initing can cause problems (e.g. where a library that uses
35 OpenSSL de-inits, but an application is still using it). The affected
36 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
37 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
38 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
42 *) --strict-warnings no longer enables runtime debugging options
43 such as REF_DEBUG. Instead, debug options are automatically
44 enabled with '--debug' builds.
45 [Andy Polyakov, Emilia Käsper]
47 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
48 have been moved out of the public header files. New functions for managing
49 these have been added.
52 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
53 objects have been moved out of the public header files. New
54 functions for managing these have been added.
57 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
58 have been moved out of the public header files. New functions for managing
59 these have been added.
62 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
63 moved out of the public header files. New functions for managing these
67 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
70 *) Removed the mk1mf build scripts.
73 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
74 it is always safe to #include a header now.
77 *) Removed the aged BC-32 config and all its supporting scripts
80 *) Removed support for Ultrix, Netware, and OS/2.
83 *) Add support for HKDF.
86 *) Add support for blake2b and blake2s
89 *) Added support for "pipelining". Ciphers that have the
90 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
91 encryptions/decryptions simultaneously. There are currently no built-in
92 ciphers with this property but the expectation is that engines will be able
93 to offer it to significantly improve throughput. Support has been extended
94 into libssl so that multiple records for a single connection can be
95 processed in one go (for >=TLS 1.1).
98 *) Added the AFALG engine. This is an async capable engine which is able to
99 offload work to the Linux kernel. In this initial version it only supports
100 AES128-CBC. The kernel must be version 4.1.0 or greater.
103 *) OpenSSL now uses a new threading API. It is no longer necessary to
104 set locking callbacks to use OpenSSL in a multi-threaded environment. There
105 are two supported threading models: pthreads and windows threads. It is
106 also possible to configure OpenSSL at compile time for "no-threads". The
107 old threading API should no longer be used. The functions have been
108 replaced with "no-op" compatibility macros.
109 [Alessandro Ghedini, Matt Caswell]
111 *) Modify behavior of ALPN to invoke callback after SNI/servername
112 callback, such that updates to the SSL_CTX affect ALPN.
115 *) Add SSL_CIPHER queries for authentication and key-exchange.
118 *) Changes to the DEFAULT cipherlist:
119 - Prefer (EC)DHE handshakes over plain RSA.
120 - Prefer AEAD ciphers over legacy ciphers.
121 - Prefer ECDSA over RSA when both certificates are available.
122 - Prefer TLSv1.2 ciphers/PRF.
123 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
127 *) Change the ECC default curve list to be this, in order: x25519,
128 secp256r1, secp521r1, secp384r1.
131 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
132 disabled by default. They can be re-enabled using the
133 enable-weak-ssl-ciphers option to Configure.
136 *) If the server has ALPN configured, but supports no protocols that the
137 client advertises, send a fatal "no_application_protocol" alert.
138 This behaviour is SHALL in RFC 7301, though it isn't universally
139 implemented by other servers.
142 *) Add X25519 support.
143 Integrate support for X25519 into EC library. This includes support
144 for public and private key encoding using the format documented in
145 draft-josefsson-pkix-newcurves-01: specifically X25519 uses the
146 OID from that draft, encodes public keys using little endian
147 format in the ECPoint structure and private keys using
148 little endian form in the privateKey field of the ECPrivateKey
149 structure. TLS support complies with draft-ietf-tls-rfc4492bis-06
152 Note: the current version supports key generation, public and
153 private key encoding and ECDH key agreement using the EC API.
154 Low level point operations such as EC_POINT_add(), EC_POINT_mul()
158 *) Deprecate SRP_VBASE_get_by_user.
159 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
160 In order to fix an unavoidable memory leak (CVE-2016-0798),
161 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
162 seed, even if the seed is configured.
164 Users should use SRP_VBASE_get1_by_user instead. Note that in
165 SRP_VBASE_get1_by_user, caller must free the returned value. Note
166 also that even though configuring the SRP seed attempts to hide
167 invalid usernames by continuing the handshake with fake
168 credentials, this behaviour is not constant time and no strong
169 guarantees are made that the handshake is indistinguishable from
170 that of a valid user.
173 *) Configuration change; it's now possible to build dynamic engines
174 without having to build shared libraries and vice versa. This
175 only applies to the engines in engines/, those in crypto/engine/
176 will always be built into libcrypto (i.e. "static").
178 Building dynamic engines is enabled by default; to disable, use
179 the configuration option "disable-dynamic-engine".
181 The only requirements for building dynamic engines are the
182 presence of the DSO module and building with position independent
183 code, so they will also automatically be disabled if configuring
184 with "disable-dso" or "disable-pic".
186 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
187 are also taken away from openssl/opensslconf.h, as they are
191 *) Configuration change; if there is a known flag to compile
192 position independent code, it will always be applied on the
193 libcrypto and libssl object files, and never on the application
194 object files. This means other libraries that use routines from
195 libcrypto / libssl can be made into shared libraries regardless
196 of how OpenSSL was configured.
198 If this isn't desirable, the configuration options "disable-pic"
199 or "no-pic" can be used to disable the use of PIC. This will
200 also disable building shared libraries and dynamic engines.
203 *) Removed JPAKE code. It was experimental and has no wide use.
206 *) The INSTALL_PREFIX Makefile variable has been renamed to
207 DESTDIR. That makes for less confusion on what this variable
208 is for. Also, the configuration option --install_prefix is
212 *) Heartbeat for TLS has been removed and is disabled by default
213 for DTLS; configure with enable-heartbeats. Code that uses the
214 old #define's might need to be updated.
215 [Emilia Käsper, Rich Salz]
217 *) Rename REF_CHECK to REF_DEBUG.
220 *) New "unified" build system
222 The "unified" build system is aimed to be a common system for all
223 platforms we support. With it comes new support for VMS.
225 This system builds supports building in a different directory tree
226 than the source tree. It produces one Makefile (for unix family
227 or lookalikes), or one descrip.mms (for VMS).
229 The source of information to make the Makefile / descrip.mms is
230 small files called 'build.info', holding the necessary
231 information for each directory with source to compile, and a
232 template in Configurations, like unix-Makefile.tmpl or
235 We rely heavily on the perl module Text::Template.
238 *) Added support for auto-initialisation and de-initialisation of the library.
239 OpenSSL no longer requires explicit init or deinit routines to be called,
240 except in certain circumstances. See the OPENSSL_init_crypto() and
241 OPENSSL_init_ssl() man pages for further information.
244 *) The arguments to the DTLSv1_listen function have changed. Specifically the
245 "peer" argument is now expected to be a BIO_ADDR object.
247 *) Rewrite of BIO networking library. The BIO library lacked consistent
248 support of IPv6, and adding it required some more extensive
249 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
250 which hold all types of addresses and chains of address information.
251 It also introduces a new API, with functions like BIO_socket,
252 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
253 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
254 have been adapted accordingly.
257 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
261 *) CRIME protection: disable compression by default, even if OpenSSL is
262 compiled with zlib enabled. Applications can still enable compression
263 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
264 using the SSL_CONF library to configure compression.
267 *) The signature of the session callback configured with
268 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
269 was explicitly marked as 'const unsigned char*' instead of
273 *) Always DPURIFY. Remove the use of uninitialized memory in the
274 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
277 *) Removed many obsolete configuration items, including
278 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
279 MD2_CHAR, MD2_INT, MD2_LONG
281 IDEA_SHORT, IDEA_LONG
282 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
283 [Rich Salz, with advice from Andy Polyakov]
285 *) Many BN internals have been moved to an internal header file.
286 [Rich Salz with help from Andy Polyakov]
288 *) Configuration and writing out the results from it has changed.
289 Files such as Makefile include/openssl/opensslconf.h and are now
290 produced through general templates, such as Makefile.in and
291 crypto/opensslconf.h.in and some help from the perl module
294 Also, the center of configuration information is no longer
295 Makefile. Instead, Configure produces a perl module in
296 configdata.pm which holds most of the config data (in the hash
297 table %config), the target data that comes from the target
298 configuration in one of the Configurations/*.conf files (in
302 *) To clarify their intended purposes, the Configure options
303 --prefix and --openssldir change their semantics, and become more
304 straightforward and less interdependent.
306 --prefix shall be used exclusively to give the location INSTALLTOP
307 where programs, scripts, libraries, include files and manuals are
308 going to be installed. The default is now /usr/local.
310 --openssldir shall be used exclusively to give the default
311 location OPENSSLDIR where certificates, private keys, CRLs are
312 managed. This is also where the default openssl.cnf gets
314 If the directory given with this option is a relative path, the
315 values of both the --prefix value and the --openssldir value will
316 be combined to become OPENSSLDIR.
317 The default for --openssldir is INSTALLTOP/ssl.
319 Anyone who uses --openssldir to specify where OpenSSL is to be
320 installed MUST change to use --prefix instead.
323 *) The GOST engine was out of date and therefore it has been removed. An up
324 to date GOST engine is now being maintained in an external repository.
325 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
326 support for GOST ciphersuites (these are only activated if a GOST engine
330 *) EGD is no longer supported by default; use enable-egd when
332 [Ben Kaduk and Rich Salz]
334 *) The distribution now has Makefile.in files, which are used to
335 create Makefile's when Configure is run. *Configure must be run
336 before trying to build now.*
339 *) The return value for SSL_CIPHER_description() for error conditions
343 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
345 Obtaining and performing DNSSEC validation of TLSA records is
346 the application's responsibility. The application provides
347 the TLSA records of its choice to OpenSSL, and these are then
348 used to authenticate the peer.
350 The TLSA records need not even come from DNS. They can, for
351 example, be used to implement local end-entity certificate or
352 trust-anchor "pinning", where the "pin" data takes the form
353 of TLSA records, which can augment or replace verification
354 based on the usual WebPKI public certification authorities.
357 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
358 continues to support deprecated interfaces in default builds.
359 However, applications are strongly advised to compile their
360 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
361 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
362 or the 1.1.0 releases.
364 In environments in which all applications have been ported to
365 not use any deprecated interfaces OpenSSL's Configure script
366 should be used with the --api=1.1.0 option to entirely remove
367 support for the deprecated features from the library and
368 unconditionally disable them in the installed headers.
369 Essentially the same effect can be achieved with the "no-deprecated"
370 argument to Configure, except that this will always restrict
371 the build to just the latest API, rather than a fixed API
374 As applications are ported to future revisions of the API,
375 they should update their compile-time OPENSSL_API_COMPAT define
376 accordingly, but in most cases should be able to continue to
377 compile with later releases.
379 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
380 0x10000000L and 0x00908000L, respectively. However those
381 versions did not support the OPENSSL_API_COMPAT feature, and
382 so applications are not typically tested for explicit support
383 of just the undeprecated features of either release.
386 *) Add support for setting the minimum and maximum supported protocol.
387 It can bet set via the SSL_set_min_proto_version() and
388 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
389 MaxProtcol. It's recommended to use the new APIs to disable
390 protocols instead of disabling individual protocols using
391 SSL_set_options() or SSL_CONF's Protocol. This change also
392 removes support for disabling TLS 1.2 in the OpenSSL TLS
393 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
396 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
399 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
400 and integrates ECDSA and ECDH functionality into EC. Implementations can
401 now redirect key generation and no longer need to convert to or from
404 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
405 include the ec.h header file instead.
408 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
409 ciphers who are no longer supported and drops support the ephemeral RSA key
410 exchange. The LOW ciphers currently doesn't have any ciphers in it.
413 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
414 opaque. For HMAC_CTX, the following constructors and destructors
417 HMAC_CTX *HMAC_CTX_new(void);
418 void HMAC_CTX_free(HMAC_CTX *ctx);
420 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
421 destroy such methods has been added. See EVP_MD_meth_new(3) and
422 EVP_CIPHER_meth_new(3) for documentation.
425 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
426 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
427 EVP_MD_CTX_reset() should be called instead to reinitialise
428 an already created structure.
429 2) For consistency with the majority of our object creators and
430 destructors, EVP_MD_CTX_(create|destroy) were renamed to
431 EVP_MD_CTX_(new|free). The old names are retained as macros
432 for deprecated builds.
435 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
436 cryptographic operations to be performed asynchronously as long as an
437 asynchronous capable engine is used. See the ASYNC_start_job() man page for
438 further details. Libssl has also had this capability integrated with the
439 introduction of the new mode SSL_MODE_ASYNC and associated error
440 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
441 pages. This work was developed in partnership with Intel Corp.
444 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
445 always enabled now. If you want to disable the support you should
446 exclude it using the list of supported ciphers. This also means that the
447 "-no_ecdhe" option has been removed from s_server.
450 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
451 SSL_{CTX_}set1_curves() which can set a list.
454 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
455 curve you want to support using SSL_{CTX_}set1_curves().
458 *) State machine rewrite. The state machine code has been significantly
459 refactored in order to remove much duplication of code and solve issues
460 with the old code (see ssl/statem/README for further details). This change
461 does have some associated API changes. Notably the SSL_state() function
462 has been removed and replaced by SSL_get_state which now returns an
463 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
464 altogether. The previous handshake states defined in ssl.h and ssl3.h have
468 *) All instances of the string "ssleay" in the public API were replaced
469 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
470 Some error codes related to internal RSA_eay API's were renamed.
473 *) The demo files in crypto/threads were moved to demo/threads.
476 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
478 [Matt Caswell, Rich Salz]
480 *) New ASN.1 embed macro.
482 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
483 structure is not allocated: it is part of the parent. That is instead of
491 This reduces memory fragmentation and make it impossible to accidentally
492 set a mandatory field to NULL.
494 This currently only works for some fields specifically a SEQUENCE, CHOICE,
495 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
496 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
500 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
503 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
504 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
505 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
506 DES and RC4 ciphersuites.
509 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
510 This changes the decoding behaviour for some invalid messages,
511 though the change is mostly in the more lenient direction, and
512 legacy behaviour is preserved as much as possible.
515 *) Fix no-stdio build.
516 [ David Woodhouse <David.Woodhouse@intel.com> and also
517 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
519 *) New testing framework
520 The testing framework has been largely rewritten and is now using
521 perl and the perl modules Test::Harness and an extended variant of
522 Test::More called OpenSSL::Test to do its work. All test scripts in
523 test/ have been rewritten into test recipes, and all direct calls to
524 executables in test/Makefile have become individual recipes using the
525 simplified testing OpenSSL::Test::Simple.
527 For documentation on our testing modules, do:
529 perldoc test/testlib/OpenSSL/Test/Simple.pm
530 perldoc test/testlib/OpenSSL/Test.pm
534 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
535 are used; the latter aborts on memory leaks (usually checked on exit).
536 Some undocumented "set malloc, etc., hooks" functions were removed
537 and others were changed. All are now documented.
540 *) In DSA_generate_parameters_ex, if the provided seed is too short,
542 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
544 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
545 from RFC4279, RFC4785, RFC5487, RFC5489.
547 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
548 original RSA_PSK patch.
551 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
552 era flag was never set throughout the codebase (only read). Also removed
553 SSL3_FLAGS_POP_BUFFER which was only used if
554 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
557 *) Changed the default name options in the "ca", "crl", "req" and "x509"
558 to be "oneline" instead of "compat".
561 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
562 not aware of clients that still exhibit this bug, and the workaround
563 hasn't been working properly for a while.
566 *) The return type of BIO_number_read() and BIO_number_written() as well as
567 the corresponding num_read and num_write members in the BIO structure has
568 changed from unsigned long to uint64_t. On platforms where an unsigned
569 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
573 *) Given the pervasive nature of TLS extensions it is inadvisable to run
574 OpenSSL without support for them. It also means that maintaining
575 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
576 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
579 *) Removed support for the two export grade static DH ciphersuites
580 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
581 were newly added (along with a number of other static DH ciphersuites) to
582 1.0.2. However the two export ones have *never* worked since they were
583 introduced. It seems strange in any case to be adding new export
584 ciphersuites, and given "logjam" it also does not seem correct to fix them.
587 *) Version negotiation has been rewritten. In particular SSLv23_method(),
588 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
589 and turned into macros which simply call the new preferred function names
590 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
591 should use the new names instead. Also as part of this change the ssl23.h
592 header file has been removed.
595 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
596 code and the associated standard is no longer considered fit-for-purpose.
599 *) RT2547 was closed. When generating a private key, try to make the
600 output file readable only by the owner. This behavior change might
601 be noticeable when interacting with other software.
603 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
607 *) Added HTTP GET support to the ocsp command.
610 *) Changed default digest for the dgst and enc commands from MD5 to
614 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
617 *) Added support for TLS extended master secret from
618 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
619 initial patch which was a great help during development.
622 *) All libssl internal structures have been removed from the public header
623 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
624 now redundant). Users should not attempt to access internal structures
625 directly. Instead they should use the provided API functions.
628 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
629 Access to deprecated functions can be re-enabled by running config with
630 "enable-deprecated". In addition applications wishing to use deprecated
631 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
632 will, by default, disable some transitive includes that previously existed
633 in the header files (e.g. ec.h will no longer, by default, include bn.h)
636 *) Added support for OCB mode. OpenSSL has been granted a patent license
637 compatible with the OpenSSL license for use of OCB. Details are available
638 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
639 for OCB can be removed by calling config with no-ocb.
642 *) SSLv2 support has been removed. It still supports receiving a SSLv2
643 compatible client hello.
646 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
647 done while fixing the error code for the key-too-small case.
648 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
650 *) CA.sh has been removmed; use CA.pl instead.
653 *) Removed old DES API.
656 *) Remove various unsupported platforms:
662 Sinix/ReliantUNIX RM400
667 16-bit platforms such as WIN16
670 *) Clean up OPENSSL_NO_xxx #define's
671 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
672 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
673 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
674 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
675 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
676 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
677 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
678 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
679 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
680 Remove MS_STATIC; it's a relic from platforms <32 bits.
683 *) Cleaned up dead code
684 Remove all but one '#ifdef undef' which is to be looked at.
687 *) Clean up calling of xxx_free routines.
688 Just like free(), fix most of the xxx_free routines to accept
689 NULL. Remove the non-null checks from callers. Save much code.
692 *) Add secure heap for storage of private keys (when possible).
693 Add BIO_s_secmem(), CBIGNUM, etc.
694 Contributed by Akamai Technologies under our Corporate CLA.
697 *) Experimental support for a new, fast, unbiased prime candidate generator,
698 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
699 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
701 *) New output format NSS in the sess_id command line tool. This allows
702 exporting the session id and the master key in NSS keylog format.
703 [Martin Kaiser <martin@kaiser.cx>]
705 *) Harmonize version and its documentation. -f flag is used to display
707 [mancha <mancha1@zoho.com>]
709 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
710 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
711 [mancha <mancha1@zoho.com>]
713 *) Fix some double frees. These are not thought to be exploitable.
714 [mancha <mancha1@zoho.com>]
716 *) A missing bounds check in the handling of the TLS heartbeat extension
717 can be used to reveal up to 64k of memory to a connected client or
720 Thanks for Neel Mehta of Google Security for discovering this bug and to
721 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
722 preparing the fix (CVE-2014-0160)
723 [Adam Langley, Bodo Moeller]
725 *) Fix for the attack described in the paper "Recovering OpenSSL
726 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
727 by Yuval Yarom and Naomi Benger. Details can be obtained from:
728 http://eprint.iacr.org/2014/140
730 Thanks to Yuval Yarom and Naomi Benger for discovering this
731 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
732 [Yuval Yarom and Naomi Benger]
734 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
735 this fixes a limitation in previous versions of OpenSSL.
738 *) Experimental encrypt-then-mac support.
740 Experimental support for encrypt then mac from
741 draft-gutmann-tls-encrypt-then-mac-02.txt
743 To enable it set the appropriate extension number (0x42 for the test
744 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
746 For non-compliant peers (i.e. just about everything) this should have no
749 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
753 *) Add EVP support for key wrapping algorithms, to avoid problems with
754 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
755 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
756 algorithms and include tests cases.
759 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
763 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
764 MGF1 digest and OAEP label.
767 *) Make openssl verify return errors.
768 [Chris Palmer <palmer@google.com> and Ben Laurie]
770 *) New function ASN1_TIME_diff to calculate the difference between two
771 ASN1_TIME structures or one structure and the current time.
774 *) Update fips_test_suite to support multiple command line options. New
775 test to induce all self test errors in sequence and check expected
779 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
780 sign or verify all in one operation.
783 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
784 test programs and fips_test_suite. Includes functionality to parse
785 the minimal script output of fipsalgest.pl directly.
788 *) Add authorisation parameter to FIPS_module_mode_set().
791 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
794 *) Use separate DRBG fields for internal and external flags. New function
795 FIPS_drbg_health_check() to perform on demand health checking. Add
796 generation tests to fips_test_suite with reduced health check interval to
797 demonstrate periodic health checking. Add "nodh" option to
798 fips_test_suite to skip very slow DH test.
801 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
805 *) More extensive health check for DRBG checking many more failure modes.
806 New function FIPS_selftest_drbg_all() to handle every possible DRBG
807 combination: call this in fips_test_suite.
810 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
811 and POST to handle Dual EC cases.
814 *) Add support for canonical generation of DSA parameter 'g'. See
817 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
818 POST to handle HMAC cases.
821 *) Add functions FIPS_module_version() and FIPS_module_version_text()
822 to return numerical and string versions of the FIPS module number.
825 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
826 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
827 outside the validated module in the FIPS capable OpenSSL.
830 *) Minor change to DRBG entropy callback semantics. In some cases
831 there is no multiple of the block length between min_len and
832 max_len. Allow the callback to return more than max_len bytes
833 of entropy but discard any extra: it is the callback's responsibility
834 to ensure that the extra data discarded does not impact the
835 requested amount of entropy.
838 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
839 information in FIPS186-3, SP800-57 and SP800-131A.
842 *) CCM support via EVP. Interface is very similar to GCM case except we
843 must supply all data in one chunk (i.e. no update, final) and the
844 message length must be supplied if AAD is used. Add algorithm test
848 *) Initial version of POST overhaul. Add POST callback to allow the status
849 of POST to be monitored and/or failures induced. Modify fips_test_suite
850 to use callback. Always run all selftests even if one fails.
853 *) XTS support including algorithm test driver in the fips_gcmtest program.
854 Note: this does increase the maximum key length from 32 to 64 bytes but
855 there should be no binary compatibility issues as existing applications
856 will never use XTS mode.
859 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
860 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
861 performs algorithm blocking for unapproved PRNG types. Also do not
862 set PRNG type in FIPS_mode_set(): leave this to the application.
863 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
864 the standard OpenSSL PRNG: set additional data to a date time vector.
867 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
868 This shouldn't present any incompatibility problems because applications
869 shouldn't be using these directly and any that are will need to rethink
870 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
873 *) Extensive self tests and health checking required by SP800-90 DRBG.
874 Remove strength parameter from FIPS_drbg_instantiate and always
875 instantiate at maximum supported strength.
878 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
881 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
884 *) New function DH_compute_key_padded() to compute a DH key and pad with
885 leading zeroes if needed: this complies with SP800-56A et al.
888 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
889 anything, incomplete, subject to change and largely untested at present.
892 *) Modify fipscanisteronly build option to only build the necessary object
893 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
896 *) Add experimental option FIPSSYMS to give all symbols in
897 fipscanister.o and FIPS or fips prefix. This will avoid
898 conflicts with future versions of OpenSSL. Add perl script
899 util/fipsas.pl to preprocess assembly language source files
900 and rename any affected symbols.
903 *) Add selftest checks and algorithm block of non-fips algorithms in
904 FIPS mode. Remove DES2 from selftests.
907 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
908 return internal method without any ENGINE dependencies. Add new
909 tiny fips sign and verify functions.
912 *) New build option no-ec2m to disable characteristic 2 code.
915 *) New build option "fipscanisteronly". This only builds fipscanister.o
916 and (currently) associated fips utilities. Uses the file Makefile.fips
917 instead of Makefile.org as the prototype.
920 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
921 Update fips_gcmtest to use IV generator.
924 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
925 setting output buffer to NULL. The *Final function must be
926 called although it will not retrieve any additional data. The tag
927 can be set or retrieved with a ctrl. The IV length is by default 12
928 bytes (96 bits) but can be set to an alternative value. If the IV
929 length exceeds the maximum IV length (currently 16 bytes) it cannot be
933 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
934 underlying do_cipher function handles all cipher semantics itself
935 including padding and finalisation. This is useful if (for example)
936 an ENGINE cipher handles block padding itself. The behaviour of
937 do_cipher is subtly changed if this flag is set: the return value
938 is the number of characters written to the output buffer (zero is
939 no longer an error code) or a negative error code. Also if the
940 input buffer is NULL and length 0 finalisation should be performed.
943 *) If a candidate issuer certificate is already part of the constructed
944 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
947 *) Improve forward-security support: add functions
949 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
950 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
952 for use by SSL/TLS servers; the callback function will be called whenever a
953 new session is created, and gets to decide whether the session may be
954 cached to make it resumable (return 0) or not (return 1). (As by the
955 SSL/TLS protocol specifications, the session_id sent by the server will be
956 empty to indicate that the session is not resumable; also, the server will
957 not generate RFC 4507 (RFC 5077) session tickets.)
959 A simple reasonable callback implementation is to return is_forward_secure.
960 This parameter will be set to 1 or 0 depending on the ciphersuite selected
961 by the SSL/TLS server library, indicating whether it can provide forward
963 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
965 *) New -verify_name option in command line utilities to set verification
969 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
970 Add CMAC pkey methods.
973 *) Experimental renegotiation in s_server -www mode. If the client
974 browses /reneg connection is renegotiated. If /renegcert it is
975 renegotiated requesting a certificate.
978 *) Add an "external" session cache for debugging purposes to s_server. This
979 should help trace issues which normally are only apparent in deployed
980 multi-process servers.
983 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
984 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
985 BIO_set_cipher() and some obscure PEM functions were changed so they
986 can now return an error. The RAND changes required a change to the
987 RAND_METHOD structure.
990 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
991 a gcc attribute to warn if the result of a function is ignored. This
992 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
993 whose return value is often ignored.
996 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
997 These allow SCTs (signed certificate timestamps) to be requested and
998 validated when establishing a connection.
999 [Rob Percival <robpercival@google.com>]
1001 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1003 *) Prevent padding oracle in AES-NI CBC MAC check
1005 A MITM attacker can use a padding oracle attack to decrypt traffic
1006 when the connection uses an AES CBC cipher and the server support
1009 This issue was introduced as part of the fix for Lucky 13 padding
1010 attack (CVE-2013-0169). The padding check was rewritten to be in
1011 constant time by making sure that always the same bytes are read and
1012 compared against either the MAC or padding bytes. But it no longer
1013 checked that there was enough data to have both the MAC and padding
1016 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1020 *) Fix EVP_EncodeUpdate overflow
1022 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1023 Base64 encoding of binary data. If an attacker is able to supply very large
1024 amounts of input data then a length check can overflow resulting in a heap
1027 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1028 the PEM_write_bio* family of functions. These are mainly used within the
1029 OpenSSL command line applications, so any application which processes data
1030 from an untrusted source and outputs it as a PEM file should be considered
1031 vulnerable to this issue. User applications that call these APIs directly
1032 with large amounts of untrusted data may also be vulnerable.
1034 This issue was reported by Guido Vranken.
1038 *) Fix EVP_EncryptUpdate overflow
1040 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1041 is able to supply very large amounts of input data after a previous call to
1042 EVP_EncryptUpdate() with a partial block then a length check can overflow
1043 resulting in a heap corruption. Following an analysis of all OpenSSL
1044 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1045 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1046 the first called function after an EVP_EncryptInit(), and therefore that
1047 specific call must be safe. The second form is where the length passed to
1048 EVP_EncryptUpdate() can be seen from the code to be some small value and
1049 therefore there is no possibility of an overflow. Since all instances are
1050 one of these two forms, it is believed that there can be no overflows in
1051 internal code due to this problem. It should be noted that
1052 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1053 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1054 of these calls have also been analysed too and it is believed there are no
1055 instances in internal usage where an overflow could occur.
1057 This issue was reported by Guido Vranken.
1061 *) Prevent ASN.1 BIO excessive memory allocation
1063 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1064 a short invalid encoding can cause allocation of large amounts of memory
1065 potentially consuming excessive resources or exhausting memory.
1067 Any application parsing untrusted data through d2i BIO functions is
1068 affected. The memory based functions such as d2i_X509() are *not* affected.
1069 Since the memory based functions are used by the TLS library, TLS
1070 applications are not affected.
1072 This issue was reported by Brian Carpenter.
1078 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1079 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1080 in arbitrary stack data being returned in the buffer.
1082 This issue was reported by Guido Vranken.
1086 *) Modify behavior of ALPN to invoke callback after SNI/servername
1087 callback, such that updates to the SSL_CTX affect ALPN.
1090 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1094 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1095 methods are enabled and ssl2 is disabled the methods return NULL.
1098 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1100 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1101 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1102 provide any "EXPORT" or "LOW" strength ciphers.
1105 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1106 is by default disabled at build-time. Builds that are not configured with
1107 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1108 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1109 will need to explicitly call either of:
1111 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1113 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1115 as appropriate. Even if either of those is used, or the application
1116 explicitly uses the version-specific SSLv2_method() or its client and
1117 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1118 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1119 ciphers, and SSLv2 56-bit DES are no longer available.
1123 *) Fix a double-free in DSA code
1125 A double free bug was discovered when OpenSSL parses malformed DSA private
1126 keys and could lead to a DoS attack or memory corruption for applications
1127 that receive DSA private keys from untrusted sources. This scenario is
1130 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1135 *) Disable SRP fake user seed to address a server memory leak.
1137 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1139 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1140 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1141 was changed to ignore the "fake user" SRP seed, even if the seed
1144 Users should use SRP_VBASE_get1_by_user instead. Note that in
1145 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1146 also that even though configuring the SRP seed attempts to hide
1147 invalid usernames by continuing the handshake with fake
1148 credentials, this behaviour is not constant time and no strong
1149 guarantees are made that the handshake is indistinguishable from
1150 that of a valid user.
1154 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1156 In the BN_hex2bn function the number of hex digits is calculated using an
1157 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1158 large values of |i| this can result in |bn_expand| not allocating any
1159 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1160 field as NULL leading to a subsequent NULL ptr deref. For very large values
1161 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1162 In this case memory is allocated to the internal BIGNUM data field, but it
1163 is insufficiently sized leading to heap corruption. A similar issue exists
1164 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1165 is ever called by user applications with very large untrusted hex/dec data.
1166 This is anticipated to be a rare occurrence.
1168 All OpenSSL internal usage of these functions use data that is not expected
1169 to be untrusted, e.g. config file data or application command line
1170 arguments. If user developed applications generate config file data based
1171 on untrusted data then it is possible that this could also lead to security
1172 consequences. This is also anticipated to be rare.
1174 This issue was reported to OpenSSL by Guido Vranken.
1178 *) Fix memory issues in BIO_*printf functions
1180 The internal |fmtstr| function used in processing a "%s" format string in
1181 the BIO_*printf functions could overflow while calculating the length of a
1182 string and cause an OOB read when printing very long strings.
1184 Additionally the internal |doapr_outch| function can attempt to write to an
1185 OOB memory location (at an offset from the NULL pointer) in the event of a
1186 memory allocation failure. In 1.0.2 and below this could be caused where
1187 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1188 could be in processing a very long "%s" format string. Memory leaks can
1191 The first issue may mask the second issue dependent on compiler behaviour.
1192 These problems could enable attacks where large amounts of untrusted data
1193 is passed to the BIO_*printf functions. If applications use these functions
1194 in this way then they could be vulnerable. OpenSSL itself uses these
1195 functions when printing out human-readable dumps of ASN.1 data. Therefore
1196 applications that print this data could be vulnerable if the data is from
1197 untrusted sources. OpenSSL command line applications could also be
1198 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1199 as command line arguments.
1201 Libssl is not considered directly vulnerable. Additionally certificates etc
1202 received via remote connections via libssl are also unlikely to be able to
1203 trigger these issues because of message size limits enforced within libssl.
1205 This issue was reported to OpenSSL Guido Vranken.
1209 *) Side channel attack on modular exponentiation
1211 A side-channel attack was found which makes use of cache-bank conflicts on
1212 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1213 of RSA keys. The ability to exploit this issue is limited as it relies on
1214 an attacker who has control of code in a thread running on the same
1215 hyper-threaded core as the victim thread which is performing decryptions.
1217 This issue was reported to OpenSSL by Yuval Yarom, The University of
1218 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1219 Nadia Heninger, University of Pennsylvania with more information at
1220 http://cachebleed.info.
1224 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1225 if no keysize is specified with default_bits. This fixes an
1226 omission in an earlier change that changed all RSA/DSA key generation
1227 apps to use 2048 bits by default.
1230 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1231 *) DH small subgroups
1233 Historically OpenSSL only ever generated DH parameters based on "safe"
1234 primes. More recently (in version 1.0.2) support was provided for
1235 generating X9.42 style parameter files such as those required for RFC 5114
1236 support. The primes used in such files may not be "safe". Where an
1237 application is using DH configured with parameters based on primes that are
1238 not "safe" then an attacker could use this fact to find a peer's private
1239 DH exponent. This attack requires that the attacker complete multiple
1240 handshakes in which the peer uses the same private DH exponent. For example
1241 this could be used to discover a TLS server's private DH exponent if it's
1242 reusing the private DH exponent or it's using a static DH ciphersuite.
1244 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1245 TLS. It is not on by default. If the option is not set then the server
1246 reuses the same private DH exponent for the life of the server process and
1247 would be vulnerable to this attack. It is believed that many popular
1248 applications do set this option and would therefore not be at risk.
1250 The fix for this issue adds an additional check where a "q" parameter is
1251 available (as is the case in X9.42 based parameters). This detects the
1252 only known attack, and is the only possible defense for static DH
1253 ciphersuites. This could have some performance impact.
1255 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1256 default and cannot be disabled. This could have some performance impact.
1258 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1262 *) SSLv2 doesn't block disabled ciphers
1264 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1265 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1266 been disabled, provided that the SSLv2 protocol was not also disabled via
1269 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1270 and Sebastian Schinzel.
1274 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1276 *) BN_mod_exp may produce incorrect results on x86_64
1278 There is a carry propagating bug in the x86_64 Montgomery squaring
1279 procedure. No EC algorithms are affected. Analysis suggests that attacks
1280 against RSA and DSA as a result of this defect would be very difficult to
1281 perform and are not believed likely. Attacks against DH are considered just
1282 feasible (although very difficult) because most of the work necessary to
1283 deduce information about a private key may be performed offline. The amount
1284 of resources required for such an attack would be very significant and
1285 likely only accessible to a limited number of attackers. An attacker would
1286 additionally need online access to an unpatched system using the target
1287 private key in a scenario with persistent DH parameters and a private
1288 key that is shared between multiple clients. For example this can occur by
1289 default in OpenSSL DHE based SSL/TLS ciphersuites.
1291 This issue was reported to OpenSSL by Hanno Böck.
1295 *) Certificate verify crash with missing PSS parameter
1297 The signature verification routines will crash with a NULL pointer
1298 dereference if presented with an ASN.1 signature using the RSA PSS
1299 algorithm and absent mask generation function parameter. Since these
1300 routines are used to verify certificate signature algorithms this can be
1301 used to crash any certificate verification operation and exploited in a
1302 DoS attack. Any application which performs certificate verification is
1303 vulnerable including OpenSSL clients and servers which enable client
1306 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1310 *) X509_ATTRIBUTE memory leak
1312 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1313 memory. This structure is used by the PKCS#7 and CMS routines so any
1314 application which reads PKCS#7 or CMS data from untrusted sources is
1315 affected. SSL/TLS is not affected.
1317 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1322 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1323 This changes the decoding behaviour for some invalid messages,
1324 though the change is mostly in the more lenient direction, and
1325 legacy behaviour is preserved as much as possible.
1328 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1330 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1332 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1334 *) Alternate chains certificate forgery
1336 During certificate verification, OpenSSL will attempt to find an
1337 alternative certificate chain if the first attempt to build such a chain
1338 fails. An error in the implementation of this logic can mean that an
1339 attacker could cause certain checks on untrusted certificates to be
1340 bypassed, such as the CA flag, enabling them to use a valid leaf
1341 certificate to act as a CA and "issue" an invalid certificate.
1343 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1347 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1349 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1350 incompatibility in the handling of HMAC. The previous ABI has now been
1354 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1356 *) Malformed ECParameters causes infinite loop
1358 When processing an ECParameters structure OpenSSL enters an infinite loop
1359 if the curve specified is over a specially malformed binary polynomial
1362 This can be used to perform denial of service against any
1363 system which processes public keys, certificate requests or
1364 certificates. This includes TLS clients and TLS servers with
1365 client authentication enabled.
1367 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1371 *) Exploitable out-of-bounds read in X509_cmp_time
1373 X509_cmp_time does not properly check the length of the ASN1_TIME
1374 string and can read a few bytes out of bounds. In addition,
1375 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1378 An attacker can use this to craft malformed certificates and CRLs of
1379 various sizes and potentially cause a segmentation fault, resulting in
1380 a DoS on applications that verify certificates or CRLs. TLS clients
1381 that verify CRLs are affected. TLS clients and servers with client
1382 authentication enabled may be affected if they use custom verification
1385 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1386 independently by Hanno Böck.
1390 *) PKCS7 crash with missing EnvelopedContent
1392 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1393 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1394 with missing content and trigger a NULL pointer dereference on parsing.
1396 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1397 structures from untrusted sources are affected. OpenSSL clients and
1398 servers are not affected.
1400 This issue was reported to OpenSSL by Michal Zalewski (Google).
1404 *) CMS verify infinite loop with unknown hash function
1406 When verifying a signedData message the CMS code can enter an infinite loop
1407 if presented with an unknown hash function OID. This can be used to perform
1408 denial of service against any system which verifies signedData messages using
1410 This issue was reported to OpenSSL by Johannes Bauer.
1414 *) Race condition handling NewSessionTicket
1416 If a NewSessionTicket is received by a multi-threaded client when attempting to
1417 reuse a previous ticket then a race condition can occur potentially leading to
1418 a double free of the ticket data.
1422 *) Only support 256-bit or stronger elliptic curves with the
1423 'ecdh_auto' setting (server) or by default (client). Of supported
1424 curves, prefer P-256 (both).
1427 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1429 *) ClientHello sigalgs DoS fix
1431 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1432 invalid signature algorithms extension a NULL pointer dereference will
1433 occur. This can be exploited in a DoS attack against the server.
1435 This issue was was reported to OpenSSL by David Ramos of Stanford
1438 [Stephen Henson and Matt Caswell]
1440 *) Multiblock corrupted pointer fix
1442 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1443 feature only applies on 64 bit x86 architecture platforms that support AES
1444 NI instructions. A defect in the implementation of "multiblock" can cause
1445 OpenSSL's internal write buffer to become incorrectly set to NULL when
1446 using non-blocking IO. Typically, when the user application is using a
1447 socket BIO for writing, this will only result in a failed connection.
1448 However if some other BIO is used then it is likely that a segmentation
1449 fault will be triggered, thus enabling a potential DoS attack.
1451 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1455 *) Segmentation fault in DTLSv1_listen fix
1457 The DTLSv1_listen function is intended to be stateless and processes the
1458 initial ClientHello from many peers. It is common for user code to loop
1459 over the call to DTLSv1_listen until a valid ClientHello is received with
1460 an associated cookie. A defect in the implementation of DTLSv1_listen means
1461 that state is preserved in the SSL object from one invocation to the next
1462 that can lead to a segmentation fault. Errors processing the initial
1463 ClientHello can trigger this scenario. An example of such an error could be
1464 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1467 This issue was reported to OpenSSL by Per Allansson.
1471 *) Segmentation fault in ASN1_TYPE_cmp fix
1473 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1474 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1475 certificate signature algorithm consistency this can be used to crash any
1476 certificate verification operation and exploited in a DoS attack. Any
1477 application which performs certificate verification is vulnerable including
1478 OpenSSL clients and servers which enable client authentication.
1482 *) Segmentation fault for invalid PSS parameters fix
1484 The signature verification routines will crash with a NULL pointer
1485 dereference if presented with an ASN.1 signature using the RSA PSS
1486 algorithm and invalid parameters. Since these routines are used to verify
1487 certificate signature algorithms this can be used to crash any
1488 certificate verification operation and exploited in a DoS attack. Any
1489 application which performs certificate verification is vulnerable including
1490 OpenSSL clients and servers which enable client authentication.
1492 This issue was was reported to OpenSSL by Brian Carpenter.
1496 *) ASN.1 structure reuse memory corruption fix
1498 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1499 memory corruption via an invalid write. Such reuse is and has been
1500 strongly discouraged and is believed to be rare.
1502 Applications that parse structures containing CHOICE or ANY DEFINED BY
1503 components may be affected. Certificate parsing (d2i_X509 and related
1504 functions) are however not affected. OpenSSL clients and servers are
1509 *) PKCS7 NULL pointer dereferences fix
1511 The PKCS#7 parsing code does not handle missing outer ContentInfo
1512 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1513 missing content and trigger a NULL pointer dereference on parsing.
1515 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1516 otherwise parse PKCS#7 structures from untrusted sources are
1517 affected. OpenSSL clients and servers are not affected.
1519 This issue was reported to OpenSSL by Michal Zalewski (Google).
1523 *) DoS via reachable assert in SSLv2 servers fix
1525 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1526 servers that both support SSLv2 and enable export cipher suites by sending
1527 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1529 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1530 (OpenSSL development team).
1534 *) Empty CKE with client auth and DHE fix
1536 If client auth is used then a server can seg fault in the event of a DHE
1537 ciphersuite being selected and a zero length ClientKeyExchange message
1538 being sent by the client. This could be exploited in a DoS attack.
1542 *) Handshake with unseeded PRNG fix
1544 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1545 with an unseeded PRNG. The conditions are:
1546 - The client is on a platform where the PRNG has not been seeded
1547 automatically, and the user has not seeded manually
1548 - A protocol specific client method version has been used (i.e. not
1549 SSL_client_methodv23)
1550 - A ciphersuite is used that does not require additional random data from
1551 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1553 If the handshake succeeds then the client random that has been used will
1554 have been generated from a PRNG with insufficient entropy and therefore the
1555 output may be predictable.
1557 For example using the following command with an unseeded openssl will
1558 succeed on an unpatched platform:
1560 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1564 *) Use After Free following d2i_ECPrivatekey error fix
1566 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1567 could cause a use after free condition. This, in turn, could cause a double
1568 free in several private key parsing functions (such as d2i_PrivateKey
1569 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1570 for applications that receive EC private keys from untrusted
1571 sources. This scenario is considered rare.
1573 This issue was discovered by the BoringSSL project and fixed in their
1578 *) X509_to_X509_REQ NULL pointer deref fix
1580 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1581 the certificate key is invalid. This function is rarely used in practice.
1583 This issue was discovered by Brian Carpenter.
1587 *) Removed the export ciphers from the DEFAULT ciphers
1590 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1592 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1593 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1594 So far those who have to target multiple platforms would compromise
1595 and argue that binary targeting say ARMv5 would still execute on
1596 ARMv8. "Universal" build resolves this compromise by providing
1597 near-optimal performance even on newer platforms.
1600 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1601 (other platforms pending).
1602 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1604 *) Add support for the SignedCertificateTimestampList certificate and
1605 OCSP response extensions from RFC6962.
1608 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1609 for corner cases. (Certain input points at infinity could lead to
1610 bogus results, with non-infinity inputs mapped to infinity too.)
1613 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1614 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1615 common cases are optimized and there still is room for further
1616 improvements. Vector Permutation AES for Altivec is also added.
1619 *) Add support for little-endian ppc64 Linux target.
1620 [Marcelo Cerri (IBM)]
1622 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1623 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1624 are optimized and there still is room for further improvements.
1625 Both 32- and 64-bit modes are supported.
1626 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1628 *) Improved ARMv7 NEON support.
1631 *) Support for SPARC Architecture 2011 crypto extensions, first
1632 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1633 SHA256/512, MD5, GHASH and modular exponentiation.
1634 [Andy Polyakov, David Miller]
1636 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1638 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1640 *) Support for new and upcoming Intel processors, including AVX2,
1641 BMI and SHA ISA extensions. This includes additional "stitched"
1642 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1645 This work was sponsored by Intel Corp.
1648 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1649 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1650 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1653 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1654 this fixes a limitation in previous versions of OpenSSL.
1657 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1658 MGF1 digest and OAEP label.
1661 *) Add EVP support for key wrapping algorithms, to avoid problems with
1662 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1663 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1664 algorithms and include tests cases.
1667 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1669 [Douglas E. Engert, Steve Henson]
1671 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1672 difference in days and seconds between two tm or ASN1_TIME structures.
1675 *) Add -rev test option to s_server to just reverse order of characters
1676 received by client and send back to server. Also prints an abbreviated
1677 summary of the connection parameters.
1680 *) New option -brief for s_client and s_server to print out a brief summary
1681 of connection parameters.
1684 *) Add callbacks for arbitrary TLS extensions.
1685 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1687 *) New option -crl_download in several openssl utilities to download CRLs
1688 from CRLDP extension in certificates.
1691 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1694 *) New function X509_CRL_diff to generate a delta CRL from the difference
1695 of two full CRLs. Add support to "crl" utility.
1698 *) New functions to set lookup_crls function and to retrieve
1699 X509_STORE from X509_STORE_CTX.
1702 *) Print out deprecated issuer and subject unique ID fields in
1706 *) Extend OCSP I/O functions so they can be used for simple general purpose
1707 HTTP as well as OCSP. New wrapper function which can be used to download
1708 CRLs using the OCSP API.
1711 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1714 *) SSL_CONF* functions. These provide a common framework for application
1715 configuration using configuration files or command lines.
1718 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1719 message callback and prints the results. Needs compile time option
1720 "enable-ssl-trace". New options to s_client and s_server to enable
1724 *) New ctrl and macro to retrieve supported points extensions.
1725 Print out extension in s_server and s_client.
1728 *) New functions to retrieve certificate signature and signature
1732 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1736 *) New Suite B modes for TLS code. These use and enforce the requirements
1737 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1738 only use Suite B curves. The Suite B modes can be set by using the
1739 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1742 *) New chain verification flags for Suite B levels of security. Check
1743 algorithms are acceptable when flags are set in X509_verify_cert.
1746 *) Make tls1_check_chain return a set of flags indicating checks passed
1747 by a certificate chain. Add additional tests to handle client
1748 certificates: checks for matching certificate type and issuer name
1752 *) If an attempt is made to use a signature algorithm not in the peer
1753 preference list abort the handshake. If client has no suitable
1754 signature algorithms in response to a certificate request do not
1755 use the certificate.
1758 *) If server EC tmp key is not in client preference list abort handshake.
1761 *) Add support for certificate stores in CERT structure. This makes it
1762 possible to have different stores per SSL structure or one store in
1763 the parent SSL_CTX. Include distinct stores for certificate chain
1764 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1765 to build and store a certificate chain in CERT structure: returing
1766 an error if the chain cannot be built: this will allow applications
1767 to test if a chain is correctly configured.
1769 Note: if the CERT based stores are not set then the parent SSL_CTX
1770 store is used to retain compatibility with existing behaviour.
1774 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1775 mask based on the current session, check mask when sending client
1776 hello and checking the requested ciphersuite.
1779 *) New ctrls to retrieve and set certificate types in a certificate
1780 request message. Print out received values in s_client. If certificate
1781 types is not set with custom values set sensible values based on
1782 supported signature algorithms.
1785 *) Support for distinct client and server supported signature algorithms.
1788 *) Add certificate callback. If set this is called whenever a certificate
1789 is required by client or server. An application can decide which
1790 certificate chain to present based on arbitrary criteria: for example
1791 supported signature algorithms. Add very simple example to s_server.
1792 This fixes many of the problems and restrictions of the existing client
1793 certificate callback: for example you can now clear an existing
1794 certificate and specify the whole chain.
1797 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1798 the certificate can be used for (if anything). Set valid_flags field
1799 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1800 to have similar checks in it.
1802 Add new "cert_flags" field to CERT structure and include a "strict mode".
1803 This enforces some TLS certificate requirements (such as only permitting
1804 certificate signature algorithms contained in the supported algorithms
1805 extension) which some implementations ignore: this option should be used
1806 with caution as it could cause interoperability issues.
1809 *) Update and tidy signature algorithm extension processing. Work out
1810 shared signature algorithms based on preferences and peer algorithms
1811 and print them out in s_client and s_server. Abort handshake if no
1812 shared signature algorithms.
1815 *) Add new functions to allow customised supported signature algorithms
1816 for SSL and SSL_CTX structures. Add options to s_client and s_server
1820 *) New function SSL_certs_clear() to delete all references to certificates
1821 from an SSL structure. Before this once a certificate had been added
1822 it couldn't be removed.
1825 *) Integrate hostname, email address and IP address checking with certificate
1826 verification. New verify options supporting checking in openssl utility.
1829 *) Fixes and wildcard matching support to hostname and email checking
1830 functions. Add manual page.
1831 [Florian Weimer (Red Hat Product Security Team)]
1833 *) New functions to check a hostname email or IP address against a
1834 certificate. Add options x509 utility to print results of checks against
1838 *) Fix OCSP checking.
1839 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
1841 *) Initial experimental support for explicitly trusted non-root CAs.
1842 OpenSSL still tries to build a complete chain to a root but if an
1843 intermediate CA has a trust setting included that is used. The first
1844 setting is used: whether to trust (e.g., -addtrust option to the x509
1848 *) Add -trusted_first option which attempts to find certificates in the
1849 trusted store even if an untrusted chain is also supplied.
1852 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
1853 platform support for Linux and Android.
1856 *) Support for linux-x32, ILP32 environment in x86_64 framework.
1859 *) Experimental multi-implementation support for FIPS capable OpenSSL.
1860 When in FIPS mode the approved implementations are used as normal,
1861 when not in FIPS mode the internal unapproved versions are used instead.
1862 This means that the FIPS capable OpenSSL isn't forced to use the
1863 (often lower performance) FIPS implementations outside FIPS mode.
1866 *) Transparently support X9.42 DH parameters when calling
1867 PEM_read_bio_DHparameters. This means existing applications can handle
1868 the new parameter format automatically.
1871 *) Initial experimental support for X9.42 DH parameter format: mainly
1872 to support use of 'q' parameter for RFC5114 parameters.
1875 *) Add DH parameters from RFC5114 including test data to dhtest.
1878 *) Support for automatic EC temporary key parameter selection. If enabled
1879 the most preferred EC parameters are automatically used instead of
1880 hardcoded fixed parameters. Now a server just has to call:
1881 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
1882 support ECDH and use the most appropriate parameters.
1885 *) Enhance and tidy EC curve and point format TLS extension code. Use
1886 static structures instead of allocation if default values are used.
1887 New ctrls to set curves we wish to support and to retrieve shared curves.
1888 Print out shared curves in s_server. New options to s_server and s_client
1889 to set list of supported curves.
1892 *) New ctrls to retrieve supported signature algorithms and
1893 supported curve values as an array of NIDs. Extend openssl utility
1894 to print out received values.
1897 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
1898 between NIDs and the more common NIST names such as "P-256". Enhance
1899 ecparam utility and ECC method to recognise the NIST names for curves.
1902 *) Enhance SSL/TLS certificate chain handling to support different
1903 chains for each certificate instead of one chain in the parent SSL_CTX.
1906 *) Support for fixed DH ciphersuite client authentication: where both
1907 server and client use DH certificates with common parameters.
1910 *) Support for fixed DH ciphersuites: those requiring DH server
1914 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
1916 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
1917 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
1918 X509_CINF_get_signature were reverted post internal team review.
1920 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
1922 *) Build fixes for the Windows and OpenVMS platforms
1923 [Matt Caswell and Richard Levitte]
1925 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
1927 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
1928 message can cause a segmentation fault in OpenSSL due to a NULL pointer
1929 dereference. This could lead to a Denial Of Service attack. Thanks to
1930 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
1934 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
1935 dtls1_buffer_record function under certain conditions. In particular this
1936 could occur if an attacker sent repeated DTLS records with the same
1937 sequence number but for the next epoch. The memory leak could be exploited
1938 by an attacker in a Denial of Service attack through memory exhaustion.
1939 Thanks to Chris Mueller for reporting this issue.
1943 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
1944 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
1945 method would be set to NULL which could later result in a NULL pointer
1946 dereference. Thanks to Frank Schmirler for reporting this issue.
1950 *) Abort handshake if server key exchange message is omitted for ephemeral
1953 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
1954 reporting this issue.
1958 *) Remove non-export ephemeral RSA code on client and server. This code
1959 violated the TLS standard by allowing the use of temporary RSA keys in
1960 non-export ciphersuites and could be used by a server to effectively
1961 downgrade the RSA key length used to a value smaller than the server
1962 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
1963 INRIA or reporting this issue.
1967 *) Fixed issue where DH client certificates are accepted without verification.
1968 An OpenSSL server will accept a DH certificate for client authentication
1969 without the certificate verify message. This effectively allows a client to
1970 authenticate without the use of a private key. This only affects servers
1971 which trust a client certificate authority which issues certificates
1972 containing DH keys: these are extremely rare and hardly ever encountered.
1973 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
1978 *) Ensure that the session ID context of an SSL is updated when its
1979 SSL_CTX is updated via SSL_set_SSL_CTX.
1981 The session ID context is typically set from the parent SSL_CTX,
1982 and can vary with the CTX.
1985 *) Fix various certificate fingerprint issues.
1987 By using non-DER or invalid encodings outside the signed portion of a
1988 certificate the fingerprint can be changed without breaking the signature.
1989 Although no details of the signed portion of the certificate can be changed
1990 this can cause problems with some applications: e.g. those using the
1991 certificate fingerprint for blacklists.
1993 1. Reject signatures with non zero unused bits.
1995 If the BIT STRING containing the signature has non zero unused bits reject
1996 the signature. All current signature algorithms require zero unused bits.
1998 2. Check certificate algorithm consistency.
2000 Check the AlgorithmIdentifier inside TBS matches the one in the
2001 certificate signature. NB: this will result in signature failure
2002 errors for some broken certificates.
2004 Thanks to Konrad Kraszewski from Google for reporting this issue.
2006 3. Check DSA/ECDSA signatures use DER.
2008 Reencode DSA/ECDSA signatures and compare with the original received
2009 signature. Return an error if there is a mismatch.
2011 This will reject various cases including garbage after signature
2012 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2013 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2014 (negative or with leading zeroes).
2016 Further analysis was conducted and fixes were developed by Stephen Henson
2017 of the OpenSSL core team.
2022 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2023 results on some platforms, including x86_64. This bug occurs at random
2024 with a very low probability, and is not known to be exploitable in any
2025 way, though its exact impact is difficult to determine. Thanks to Pieter
2026 Wuille (Blockstream) who reported this issue and also suggested an initial
2027 fix. Further analysis was conducted by the OpenSSL development team and
2028 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2029 the OpenSSL core team.
2033 *) Do not resume sessions on the server if the negotiated protocol
2034 version does not match the session's version. Resuming with a different
2035 version, while not strictly forbidden by the RFC, is of questionable
2036 sanity and breaks all known clients.
2037 [David Benjamin, Emilia Käsper]
2039 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2040 early CCS messages during renegotiation. (Note that because
2041 renegotiation is encrypted, this early CCS was not exploitable.)
2044 *) Tighten client-side session ticket handling during renegotiation:
2045 ensure that the client only accepts a session ticket if the server sends
2046 the extension anew in the ServerHello. Previously, a TLS client would
2047 reuse the old extension state and thus accept a session ticket if one was
2048 announced in the initial ServerHello.
2050 Similarly, ensure that the client requires a session ticket if one
2051 was advertised in the ServerHello. Previously, a TLS client would
2052 ignore a missing NewSessionTicket message.
2055 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2057 *) SRTP Memory Leak.
2059 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2060 sends a carefully crafted handshake message, to cause OpenSSL to fail
2061 to free up to 64k of memory causing a memory leak. This could be
2062 exploited in a Denial Of Service attack. This issue affects OpenSSL
2063 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2064 whether SRTP is used or configured. Implementations of OpenSSL that
2065 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2067 The fix was developed by the OpenSSL team.
2071 *) Session Ticket Memory Leak.
2073 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2074 integrity of that ticket is first verified. In the event of a session
2075 ticket integrity check failing, OpenSSL will fail to free memory
2076 causing a memory leak. By sending a large number of invalid session
2077 tickets an attacker could exploit this issue in a Denial Of Service
2082 *) Build option no-ssl3 is incomplete.
2084 When OpenSSL is configured with "no-ssl3" as a build option, servers
2085 could accept and complete a SSL 3.0 handshake, and clients could be
2086 configured to send them.
2088 [Akamai and the OpenSSL team]
2090 *) Add support for TLS_FALLBACK_SCSV.
2091 Client applications doing fallback retries should call
2092 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2094 [Adam Langley, Bodo Moeller]
2096 *) Add additional DigestInfo checks.
2098 Reencode DigestInto in DER and check against the original when
2099 verifying RSA signature: this will reject any improperly encoded
2100 DigestInfo structures.
2102 Note: this is a precautionary measure and no attacks are currently known.
2106 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2108 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2109 SRP code can be overrun an internal buffer. Add sanity check that
2110 g, A, B < N to SRP code.
2112 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2113 Group for discovering this issue.
2117 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2118 TLS 1.0 instead of higher protocol versions when the ClientHello message
2119 is badly fragmented. This allows a man-in-the-middle attacker to force a
2120 downgrade to TLS 1.0 even if both the server and the client support a
2121 higher protocol version, by modifying the client's TLS records.
2123 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2124 researching this issue.
2128 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2129 to a denial of service attack. A malicious server can crash the client
2130 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2131 ciphersuite and sending carefully crafted handshake messages.
2133 Thanks to Felix Gröbert (Google) for discovering and researching this
2138 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2139 to leak memory. This can be exploited through a Denial of Service attack.
2140 Thanks to Adam Langley for discovering and researching this issue.
2144 *) An attacker can force openssl to consume large amounts of memory whilst
2145 processing DTLS handshake messages. This can be exploited through a
2146 Denial of Service attack.
2147 Thanks to Adam Langley for discovering and researching this issue.
2151 *) An attacker can force an error condition which causes openssl to crash
2152 whilst processing DTLS packets due to memory being freed twice. This
2153 can be exploited through a Denial of Service attack.
2154 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2159 *) If a multithreaded client connects to a malicious server using a resumed
2160 session and the server sends an ec point format extension it could write
2161 up to 255 bytes to freed memory.
2163 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2168 *) A malicious server can crash an OpenSSL client with a null pointer
2169 dereference (read) by specifying an SRP ciphersuite even though it was not
2170 properly negotiated with the client. This can be exploited through a
2171 Denial of Service attack.
2173 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2174 discovering and researching this issue.
2178 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2179 X509_name_oneline, X509_name_print_ex et al. to leak some information
2180 from the stack. Applications may be affected if they echo pretty printing
2181 output to the attacker.
2183 Thanks to Ivan Fratric (Google) for discovering this issue.
2185 [Emilia Käsper, and Steve Henson]
2187 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2188 for corner cases. (Certain input points at infinity could lead to
2189 bogus results, with non-infinity inputs mapped to infinity too.)
2192 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2194 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2195 handshake can force the use of weak keying material in OpenSSL
2196 SSL/TLS clients and servers.
2198 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2199 researching this issue. (CVE-2014-0224)
2200 [KIKUCHI Masashi, Steve Henson]
2202 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2203 OpenSSL DTLS client the code can be made to recurse eventually crashing
2206 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2208 [Imre Rad, Steve Henson]
2210 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2211 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2212 client or server. This is potentially exploitable to run arbitrary
2213 code on a vulnerable client or server.
2215 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2216 [Jüri Aedla, Steve Henson]
2218 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2219 are subject to a denial of service attack.
2221 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2222 this issue. (CVE-2014-3470)
2223 [Felix Gröbert, Ivan Fratric, Steve Henson]
2225 *) Harmonize version and its documentation. -f flag is used to display
2227 [mancha <mancha1@zoho.com>]
2229 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2230 in i2d_ECPrivateKey.
2231 [mancha <mancha1@zoho.com>]
2233 *) Fix some double frees. These are not thought to be exploitable.
2234 [mancha <mancha1@zoho.com>]
2236 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2238 *) A missing bounds check in the handling of the TLS heartbeat extension
2239 can be used to reveal up to 64k of memory to a connected client or
2242 Thanks for Neel Mehta of Google Security for discovering this bug and to
2243 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2244 preparing the fix (CVE-2014-0160)
2245 [Adam Langley, Bodo Moeller]
2247 *) Fix for the attack described in the paper "Recovering OpenSSL
2248 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2249 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2250 http://eprint.iacr.org/2014/140
2252 Thanks to Yuval Yarom and Naomi Benger for discovering this
2253 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2254 [Yuval Yarom and Naomi Benger]
2256 *) TLS pad extension: draft-agl-tls-padding-03
2258 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2259 TLS client Hello record length value would otherwise be > 255 and
2260 less that 512 pad with a dummy extension containing zeroes so it
2261 is at least 512 bytes long.
2263 [Adam Langley, Steve Henson]
2265 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2267 *) Fix for TLS record tampering bug. A carefully crafted invalid
2268 handshake could crash OpenSSL with a NULL pointer exception.
2269 Thanks to Anton Johansson for reporting this issues.
2272 *) Keep original DTLS digest and encryption contexts in retransmission
2273 structures so we can use the previous session parameters if they need
2274 to be resent. (CVE-2013-6450)
2277 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2278 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2279 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2280 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2281 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2282 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2283 [Rob Stradling, Adam Langley]
2285 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2287 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2288 supporting platforms or when small records were transferred.
2289 [Andy Polyakov, Steve Henson]
2291 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2293 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2295 This addresses the flaw in CBC record processing discovered by
2296 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2297 at: http://www.isg.rhul.ac.uk/tls/
2299 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2300 Security Group at Royal Holloway, University of London
2301 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2302 Emilia Käsper for the initial patch.
2304 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2306 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2307 ciphersuites which can be exploited in a denial of service attack.
2308 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2309 and detecting this bug and to Wolfgang Ettlinger
2310 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2314 *) Return an error when checking OCSP signatures when key is NULL.
2315 This fixes a DoS attack. (CVE-2013-0166)
2318 *) Make openssl verify return errors.
2319 [Chris Palmer <palmer@google.com> and Ben Laurie]
2321 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2322 the right response is stapled. Also change SSL_get_certificate()
2323 so it returns the certificate actually sent.
2324 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2325 [Rob Stradling <rob.stradling@comodo.com>]
2327 *) Fix possible deadlock when decoding public keys.
2330 *) Don't use TLS 1.0 record version number in initial client hello
2334 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2336 *) Sanity check record length before skipping explicit IV in TLS
2337 1.2, 1.1 and DTLS to fix DoS attack.
2339 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2340 fuzzing as a service testing platform.
2344 *) Initialise tkeylen properly when encrypting CMS messages.
2345 Thanks to Solar Designer of Openwall for reporting this issue.
2348 *) In FIPS mode don't try to use composite ciphers as they are not
2352 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2354 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2355 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2356 mean any application compiled against OpenSSL 1.0.0 headers setting
2357 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2358 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2359 0x10000000L Any application which was previously compiled against
2360 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2361 will need to be recompiled as a result. Letting be results in
2362 inability to disable specifically TLS 1.1 and in client context,
2363 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2366 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2367 disable just protocol X, but all protocols above X *if* there are
2368 protocols *below* X still enabled. In more practical terms it means
2369 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2370 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2371 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2375 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2377 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2378 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2379 in CRYPTO_realloc_clean.
2381 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2382 issue and to Adam Langley <agl@chromium.org> for fixing it.
2384 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2386 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2389 *) Workarounds for some broken servers that "hang" if a client hello
2390 record length exceeds 255 bytes.
2392 1. Do not use record version number > TLS 1.0 in initial client
2393 hello: some (but not all) hanging servers will now work.
2394 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2395 the number of ciphers sent in the client hello. This should be
2396 set to an even number, such as 50, for example by passing:
2397 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2398 Most broken servers should now work.
2399 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2400 TLS 1.2 client support entirely.
2403 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2406 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2408 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2409 STRING form instead of a DigestInfo.
2412 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2413 and the RSA_sign/RSA_verify functions. This was made more apparent when
2414 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2415 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2416 the correct format in RSA_verify so both forms transparently work.
2419 *) Some servers which support TLS 1.0 can choke if we initially indicate
2420 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2421 encrypted premaster secret. As a workaround use the maximum permitted
2422 client version in client hello, this should keep such servers happy
2423 and still work with previous versions of OpenSSL.
2426 *) Add support for TLS/DTLS heartbeats.
2427 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2429 *) Add support for SCTP.
2430 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2432 *) Improved PRNG seeding for VOS.
2433 [Paul Green <Paul.Green@stratus.com>]
2435 *) Extensive assembler packs updates, most notably:
2437 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2438 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2439 - x86_64: bit-sliced AES implementation;
2440 - ARM: NEON support, contemporary platforms optimizations;
2441 - s390x: z196 support;
2442 - *: GHASH and GF(2^m) multiplication implementations;
2446 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2447 (removal of unnecessary code)
2448 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2450 *) Add TLS key material exporter from RFC 5705.
2453 *) Add DTLS-SRTP negotiation from RFC 5764.
2456 *) Add Next Protocol Negotiation,
2457 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2458 disabled with a no-npn flag to config or Configure. Code donated
2460 [Adam Langley <agl@google.com> and Ben Laurie]
2462 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2463 NIST-P256, NIST-P521, with constant-time single point multiplication on
2464 typical inputs. Compiler support for the nonstandard type __uint128_t is
2465 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2466 Code made available under Apache License version 2.0.
2468 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2469 line to include this in your build of OpenSSL, and run "make depend" (or
2470 "make update"). This enables the following EC_METHODs:
2472 EC_GFp_nistp224_method()
2473 EC_GFp_nistp256_method()
2474 EC_GFp_nistp521_method()
2476 EC_GROUP_new_by_curve_name() will automatically use these (while
2477 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2479 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2481 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2482 all platforms. Move ssize_t definition from e_os.h to the public
2483 header file e_os2.h as it now appears in public header file cms.h
2486 *) New -sigopt option to the ca, req and x509 utilities. Additional
2487 signature parameters can be passed using this option and in
2491 *) Add RSA PSS signing function. This will generate and set the
2492 appropriate AlgorithmIdentifiers for PSS based on those in the
2493 corresponding EVP_MD_CTX structure. No application support yet.
2496 *) Support for companion algorithm specific ASN1 signing routines.
2497 New function ASN1_item_sign_ctx() signs a pre-initialised
2498 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2499 the appropriate parameters.
2502 *) Add new algorithm specific ASN1 verification initialisation function
2503 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2504 handling will be the same no matter what EVP_PKEY_METHOD is used.
2505 Add a PSS handler to support verification of PSS signatures: checked
2506 against a number of sample certificates.
2509 *) Add signature printing for PSS. Add PSS OIDs.
2510 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2512 *) Add algorithm specific signature printing. An individual ASN1 method
2513 can now print out signatures instead of the standard hex dump.
2515 More complex signatures (e.g. PSS) can print out more meaningful
2516 information. Include DSA version that prints out the signature
2520 *) Password based recipient info support for CMS library: implementing
2524 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2525 neatly separates the code into cipher and PBE sections and is required
2526 for some algorithms that split PBES2 into separate pieces (such as
2527 password based CMS).
2530 *) Session-handling fixes:
2531 - Fix handling of connections that are resuming with a session ID,
2532 but also support Session Tickets.
2533 - Fix a bug that suppressed issuing of a new ticket if the client
2534 presented a ticket with an expired session.
2535 - Try to set the ticket lifetime hint to something reasonable.
2536 - Make tickets shorter by excluding irrelevant information.
2537 - On the client side, don't ignore renewed tickets.
2538 [Adam Langley, Bodo Moeller (Google)]
2540 *) Fix PSK session representation.
2543 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2545 This work was sponsored by Intel.
2548 *) Add GCM support to TLS library. Some custom code is needed to split
2549 the IV between the fixed (from PRF) and explicit (from TLS record)
2550 portions. This adds all GCM ciphersuites supported by RFC5288 and
2551 RFC5289. Generalise some AES* cipherstrings to include GCM and
2552 add a special AESGCM string for GCM only.
2555 *) Expand range of ctrls for AES GCM. Permit setting invocation
2556 field on decrypt and retrieval of invocation field only on encrypt.
2559 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2560 As required by RFC5289 these ciphersuites cannot be used if for
2561 versions of TLS earlier than 1.2.
2564 *) For FIPS capable OpenSSL interpret a NULL default public key method
2565 as unset and return the appropriate default but do *not* set the default.
2566 This means we can return the appropriate method in applications that
2567 switch between FIPS and non-FIPS modes.
2570 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2571 ENGINE is used then we cannot handle that in the FIPS module so we
2572 keep original code iff non-FIPS operations are allowed.
2575 *) Add -attime option to openssl utilities.
2576 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2578 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2581 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2582 FIPS EC methods unconditionally for now.
2585 *) New build option no-ec2m to disable characteristic 2 code.
2588 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2589 all cases can be covered as some introduce binary incompatibilities.
2592 *) Redirect RSA operations to FIPS module including keygen,
2593 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2596 *) Add similar low level API blocking to ciphers.
2599 *) Low level digest APIs are not approved in FIPS mode: any attempt
2600 to use these will cause a fatal error. Applications that *really* want
2601 to use them can use the private_* version instead.
2604 *) Redirect cipher operations to FIPS module for FIPS builds.
2607 *) Redirect digest operations to FIPS module for FIPS builds.
2610 *) Update build system to add "fips" flag which will link in fipscanister.o
2611 for static and shared library builds embedding a signature if needed.
2614 *) Output TLS supported curves in preference order instead of numerical
2615 order. This is currently hardcoded for the highest order curves first.
2616 This should be configurable so applications can judge speed vs strength.
2619 *) Add TLS v1.2 server support for client authentication.
2622 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2626 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2627 FIPS modules versions.
2630 *) Add TLS v1.2 client side support for client authentication. Keep cache
2631 of handshake records longer as we don't know the hash algorithm to use
2632 until after the certificate request message is received.
2635 *) Initial TLS v1.2 client support. Add a default signature algorithms
2636 extension including all the algorithms we support. Parse new signature
2637 format in client key exchange. Relax some ECC signing restrictions for
2638 TLS v1.2 as indicated in RFC5246.
2641 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2642 to new signature format when needed using client digest preference.
2643 All server ciphersuites should now work correctly in TLS v1.2. No client
2644 support yet and no support for client certificates.
2647 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2648 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2649 ciphersuites. At present only RSA key exchange ciphersuites work with
2650 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2651 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2652 and version checking.
2655 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2656 with this defined it will not be affected by any changes to ssl internal
2657 structures. Add several utility functions to allow openssl application
2658 to work with OPENSSL_NO_SSL_INTERN defined.
2662 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2664 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2667 *) Permit abbreviated handshakes when renegotiating using the function
2668 SSL_renegotiate_abbreviated().
2669 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2671 *) Add call to ENGINE_register_all_complete() to
2672 ENGINE_load_builtin_engines(), so some implementations get used
2673 automatically instead of needing explicit application support.
2676 *) Add support for TLS key exporter as described in RFC5705.
2677 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2679 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2680 a few changes are required:
2682 Add SSL_OP_NO_TLSv1_1 flag.
2683 Add TLSv1_1 methods.
2684 Update version checking logic to handle version 1.1.
2685 Add explicit IV handling (ported from DTLS code).
2686 Add command line options to s_client/s_server.
2689 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2691 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2692 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2693 content decryption and always return the same error. Note: this attack
2694 needs on average 2^20 messages so it only affects automated senders. The
2695 old behaviour can be reenabled in the CMS code by setting the
2696 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2697 an MMA defence is not necessary.
2698 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2699 this issue. (CVE-2012-0884)
2702 *) Fix CVE-2011-4619: make sure we really are receiving a
2703 client hello before rejecting multiple SGC restarts. Thanks to
2704 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2707 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2709 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2710 Thanks to Antonio Martin, Enterprise Secure Access Research and
2711 Development, Cisco Systems, Inc. for discovering this bug and
2712 preparing a fix. (CVE-2012-0050)
2715 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2717 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2718 of the Vaudenay padding oracle attack on CBC mode encryption
2719 which enables an efficient plaintext recovery attack against
2720 the OpenSSL implementation of DTLS. Their attack exploits timing
2721 differences arising during decryption processing. A research
2722 paper describing this attack can be found at:
2723 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2724 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2725 Security Group at Royal Holloway, University of London
2726 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2727 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2728 for preparing the fix. (CVE-2011-4108)
2729 [Robin Seggelmann, Michael Tuexen]
2731 *) Clear bytes used for block padding of SSL 3.0 records.
2733 [Adam Langley (Google)]
2735 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2736 Kadianakis <desnacked@gmail.com> for discovering this issue and
2737 Adam Langley for preparing the fix. (CVE-2011-4619)
2738 [Adam Langley (Google)]
2740 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2741 [Andrey Kulikov <amdeich@gmail.com>]
2743 *) Prevent malformed RFC3779 data triggering an assertion failure.
2744 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2745 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2746 [Rob Austein <sra@hactrn.net>]
2748 *) Improved PRNG seeding for VOS.
2749 [Paul Green <Paul.Green@stratus.com>]
2751 *) Fix ssl_ciph.c set-up race.
2752 [Adam Langley (Google)]
2754 *) Fix spurious failures in ecdsatest.c.
2755 [Emilia Käsper (Google)]
2757 *) Fix the BIO_f_buffer() implementation (which was mixing different
2758 interpretations of the '..._len' fields).
2759 [Adam Langley (Google)]
2761 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2762 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2763 threads won't reuse the same blinding coefficients.
2765 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2766 lock to call BN_BLINDING_invert_ex, and avoids one use of
2767 BN_BLINDING_update for each BN_BLINDING structure (previously,
2768 the last update always remained unused).
2769 [Emilia Käsper (Google)]
2771 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2772 [Bob Buckholz (Google)]
2774 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2776 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2777 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2778 [Kaspar Brand <ossl@velox.ch>]
2780 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2781 for multi-threaded use of ECDH. (CVE-2011-3210)
2782 [Adam Langley (Google)]
2784 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2787 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2788 signature public key algorithm by using OID xref utilities instead.
2789 Before this you could only use some ECC ciphersuites with SHA1 only.
2792 *) Add protection against ECDSA timing attacks as mentioned in the paper
2793 by Billy Bob Brumley and Nicola Tuveri, see:
2795 http://eprint.iacr.org/2011/232.pdf
2797 [Billy Bob Brumley and Nicola Tuveri]
2799 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2801 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2802 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2804 *) Fix bug in string printing code: if *any* escaping is enabled we must
2805 escape the escape character (backslash) or the resulting string is
2809 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2811 *) Disable code workaround for ancient and obsolete Netscape browsers
2812 and servers: an attacker can use it in a ciphersuite downgrade attack.
2813 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2816 *) Fixed J-PAKE implementation error, originally discovered by
2817 Sebastien Martini, further info and confirmation from Stefan
2818 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2821 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
2823 *) Fix extension code to avoid race conditions which can result in a buffer
2824 overrun vulnerability: resumed sessions must not be modified as they can
2825 be shared by multiple threads. CVE-2010-3864
2828 *) Fix WIN32 build system to correctly link an ENGINE directory into
2832 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
2834 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
2836 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
2838 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
2840 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
2841 context. The operation can be customised via the ctrl mechanism in
2842 case ENGINEs want to include additional functionality.
2845 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
2848 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
2849 output hashes compatible with older versions of OpenSSL.
2850 [Willy Weisz <weisz@vcpc.univie.ac.at>]
2852 *) Fix compression algorithm handling: if resuming a session use the
2853 compression algorithm of the resumed session instead of determining
2854 it from client hello again. Don't allow server to change algorithm.
2857 *) Add load_crls() function to apps tidying load_certs() too. Add option
2858 to verify utility to allow additional CRLs to be included.
2861 *) Update OCSP request code to permit adding custom headers to the request:
2862 some responders need this.
2865 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
2867 [Julia Lawall <julia@diku.dk>]
2869 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
2870 needlessly dereferenced structures, used obsolete functions and
2871 didn't handle all updated verify codes correctly.
2874 *) Disable MD2 in the default configuration.
2877 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
2878 indicate the initial BIO being pushed or popped. This makes it possible
2879 to determine whether the BIO is the one explicitly called or as a result
2880 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
2881 it handles reference counts correctly and doesn't zero out the I/O bio
2882 when it is not being explicitly popped. WARNING: applications which
2883 included workarounds for the old buggy behaviour will need to be modified
2884 or they could free up already freed BIOs.
2887 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
2888 renaming to all platforms (within the 0.9.8 branch, this was
2889 done conditionally on Netware platforms to avoid a name clash).
2890 [Guenter <lists@gknw.net>]
2892 *) Add ECDHE and PSK support to DTLS.
2893 [Michael Tuexen <tuexen@fh-muenster.de>]
2895 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
2899 *) Add "missing" function EVP_MD_flags() (without this the only way to
2900 retrieve a digest flags is by accessing the structure directly. Update
2901 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
2902 or cipher is registered as in the "from" argument. Print out all
2903 registered digests in the dgst usage message instead of manually
2904 attempting to work them out.
2907 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
2908 this allows the use of compression and extensions. Change default cipher
2909 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
2910 by default unless an application cipher string requests it.
2913 *) Alter match criteria in PKCS12_parse(). It used to try to use local
2914 key ids to find matching certificates and keys but some PKCS#12 files
2915 don't follow the (somewhat unwritten) rules and this strategy fails.
2916 Now just gather all certificates together and the first private key
2917 then look for the first certificate that matches the key.
2920 *) Support use of registered digest and cipher names for dgst and cipher
2921 commands instead of having to add each one as a special case. So now
2928 openssl dgst -sha256 foo
2930 and this works for ENGINE based algorithms too.
2934 *) Update Gost ENGINE to support parameter files.
2935 [Victor B. Wagner <vitus@cryptocom.ru>]
2937 *) Support GeneralizedTime in ca utility.
2938 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
2940 *) Enhance the hash format used for certificate directory links. The new
2941 form uses the canonical encoding (meaning equivalent names will work
2942 even if they aren't identical) and uses SHA1 instead of MD5. This form
2943 is incompatible with the older format and as a result c_rehash should
2944 be used to rebuild symbolic links.
2947 *) Make PKCS#8 the default write format for private keys, replacing the
2948 traditional format. This form is standardised, more secure and doesn't
2949 include an implicit MD5 dependency.
2952 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
2953 committed to OpenSSL should pass this lot as a minimum.
2956 *) Add session ticket override functionality for use by EAP-FAST.
2957 [Jouni Malinen <j@w1.fi>]
2959 *) Modify HMAC functions to return a value. Since these can be implemented
2960 in an ENGINE errors can occur.
2963 *) Type-checked OBJ_bsearch_ex.
2966 *) Type-checked OBJ_bsearch. Also some constification necessitated
2967 by type-checking. Still to come: TXT_DB, bsearch(?),
2968 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
2972 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
2973 seconds to a tm structure directly, instead of going through OS
2974 specific date routines. This avoids any issues with OS routines such
2975 as the year 2038 bug. New *_adj() functions for ASN1 time structures
2976 and X509_time_adj_ex() to cover the extended range. The existing
2977 X509_time_adj() is still usable and will no longer have any date issues.
2980 *) Delta CRL support. New use deltas option which will attempt to locate
2981 and search any appropriate delta CRLs available.
2983 This work was sponsored by Google.
2986 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
2987 code and add additional score elements. Validate alternate CRL paths
2988 as part of the CRL checking and indicate a new error "CRL path validation
2989 error" in this case. Applications wanting additional details can use
2990 the verify callback and check the new "parent" field. If this is not
2991 NULL CRL path validation is taking place. Existing applications wont
2992 see this because it requires extended CRL support which is off by
2995 This work was sponsored by Google.
2998 *) Support for freshest CRL extension.
3000 This work was sponsored by Google.
3003 *) Initial indirect CRL support. Currently only supported in the CRLs
3004 passed directly and not via lookup. Process certificate issuer
3005 CRL entry extension and lookup CRL entries by bother issuer name
3006 and serial number. Check and process CRL issuer entry in IDP extension.
3008 This work was sponsored by Google.
3011 *) Add support for distinct certificate and CRL paths. The CRL issuer
3012 certificate is validated separately in this case. Only enabled if
3013 an extended CRL support flag is set: this flag will enable additional
3014 CRL functionality in future.
3016 This work was sponsored by Google.
3019 *) Add support for policy mappings extension.
3021 This work was sponsored by Google.
3024 *) Fixes to pathlength constraint, self issued certificate handling,
3025 policy processing to align with RFC3280 and PKITS tests.
3027 This work was sponsored by Google.
3030 *) Support for name constraints certificate extension. DN, email, DNS
3031 and URI types are currently supported.
3033 This work was sponsored by Google.
3036 *) To cater for systems that provide a pointer-based thread ID rather
3037 than numeric, deprecate the current numeric thread ID mechanism and
3038 replace it with a structure and associated callback type. This
3039 mechanism allows a numeric "hash" to be extracted from a thread ID in
3040 either case, and on platforms where pointers are larger than 'long',
3041 mixing is done to help ensure the numeric 'hash' is usable even if it
3042 can't be guaranteed unique. The default mechanism is to use "&errno"
3043 as a pointer-based thread ID to distinguish between threads.
3045 Applications that want to provide their own thread IDs should now use
3046 CRYPTO_THREADID_set_callback() to register a callback that will call
3047 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3049 Note that ERR_remove_state() is now deprecated, because it is tied
3050 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3051 to free the current thread's error state should be replaced by
3052 ERR_remove_thread_state(NULL).
3054 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3055 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3056 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3057 application was previously providing a numeric thread callback that
3058 was inappropriate for distinguishing threads, then uniqueness might
3059 have been obtained with &errno that happened immediately in the
3060 intermediate development versions of OpenSSL; this is no longer the
3061 case, the numeric thread callback will now override the automatic use
3063 [Geoff Thorpe, with help from Bodo Moeller]
3065 *) Initial support for different CRL issuing certificates. This covers a
3066 simple case where the self issued certificates in the chain exist and
3067 the real CRL issuer is higher in the existing chain.
3069 This work was sponsored by Google.
3072 *) Removed effectively defunct crypto/store from the build.
3075 *) Revamp of STACK to provide stronger type-checking. Still to come:
3076 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3077 ASN1_STRING, CONF_VALUE.
3080 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3081 RAM on SSL connections. This option can save about 34k per idle SSL.
3084 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3085 STACK, TXT_DB, bsearch, qsort.
3088 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3089 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3090 support for data, signedData, compressedData, digestedData and
3091 encryptedData, envelopedData types included. Scripts to check against
3092 RFC4134 examples draft and interop and consistency checks of many
3093 content types and variants.
3096 *) Add options to enc utility to support use of zlib compression BIO.
3099 *) Extend mk1mf to support importing of options and assembly language
3100 files from Configure script, currently only included in VC-WIN32.
3101 The assembly language rules can now optionally generate the source
3102 files from the associated perl scripts.
3105 *) Implement remaining functionality needed to support GOST ciphersuites.
3106 Interop testing has been performed using CryptoPro implementations.
3107 [Victor B. Wagner <vitus@cryptocom.ru>]
3109 *) s390x assembler pack.
3112 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3116 *) Implement Opaque PRF Input TLS extension as specified in
3117 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3118 official specification yet and no extension type assignment by
3119 IANA exists, this extension (for now) will have to be explicitly
3120 enabled when building OpenSSL by providing the extension number
3121 to use. For example, specify an option
3123 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3125 to the "config" or "Configure" script to enable the extension,
3126 assuming extension number 0x9527 (which is a completely arbitrary
3127 and unofficial assignment based on the MD5 hash of the Internet
3128 Draft). Note that by doing so, you potentially lose
3129 interoperability with other TLS implementations since these might
3130 be using the same extension number for other purposes.
3132 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3133 opaque PRF input value to use in the handshake. This will create
3134 an interal copy of the length-'len' string at 'src', and will
3135 return non-zero for success.
3137 To get more control and flexibility, provide a callback function
3140 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3141 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3145 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3148 Callback function 'cb' will be called in handshakes, and is
3149 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3150 Argument 'arg' is for application purposes (the value as given to
3151 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3152 be provided to the callback function). The callback function
3153 has to return non-zero to report success: usually 1 to use opaque
3154 PRF input just if possible, or 2 to enforce use of the opaque PRF
3155 input. In the latter case, the library will abort the handshake
3156 if opaque PRF input is not successfully negotiated.
3158 Arguments 'peerinput' and 'len' given to the callback function
3159 will always be NULL and 0 in the case of a client. A server will
3160 see the client's opaque PRF input through these variables if
3161 available (NULL and 0 otherwise). Note that if the server
3162 provides an opaque PRF input, the length must be the same as the
3163 length of the client's opaque PRF input.
3165 Note that the callback function will only be called when creating
3166 a new session (session resumption can resume whatever was
3167 previously negotiated), and will not be called in SSL 2.0
3168 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3169 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3170 for applications that need to enforce opaque PRF input.
3174 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3177 [Victor B. Wagner <vitus@cryptocom.ru>]
3179 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3180 RFC4507bis. The encrypted ticket format is an encrypted encoded
3181 SSL_SESSION structure, that way new session features are automatically
3184 If a client application caches session in an SSL_SESSION structure
3185 support is transparent because tickets are now stored in the encoded
3188 The SSL_CTX structure automatically generates keys for ticket
3189 protection in servers so again support should be possible
3190 with no application modification.
3192 If a client or server wishes to disable RFC4507 support then the option
3193 SSL_OP_NO_TICKET can be set.
3195 Add a TLS extension debugging callback to allow the contents of any client
3196 or server extensions to be examined.
3198 This work was sponsored by Google.
3201 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3202 OpenSSL should now compile cleanly on gcc 4.2
3203 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3205 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3206 support including streaming MAC support: this is required for GOST
3207 ciphersuite support.
3208 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3210 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3211 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3212 to output in BER and PEM format.
3215 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3216 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3217 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3218 ENGINE support for HMAC keys which are unextractable. New -mac and
3219 -macopt options to dgst utility.
3222 *) New option -sigopt to dgst utility. Update dgst to use
3223 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3224 alternative signing parameters such as X9.31 or PSS in the dgst
3228 *) Change ssl_cipher_apply_rule(), the internal function that does
3229 the work each time a ciphersuite string requests enabling
3230 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3231 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3232 the order of disabled ciphersuites such that those ciphersuites
3233 that most recently went from enabled to disabled not only stay
3234 in order with respect to each other, but also have higher priority
3235 than other disabled ciphersuites the next time ciphersuites are
3238 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3239 the same ciphersuites as with "HIGH" alone, but in a specific
3240 order where the PSK ciphersuites come first (since they are the
3241 most recently disabled ciphersuites when "HIGH" is parsed).
3243 Also, change ssl_create_cipher_list() (using this new
3244 funcionality) such that between otherwise identical
3245 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3249 *) Change ssl_create_cipher_list() so that it automatically
3250 arranges the ciphersuites in reasonable order before starting
3251 to process the rule string. Thus, the definition for "DEFAULT"
3252 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3253 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3254 This makes it much easier to arrive at a reasonable default order
3255 in applications for which anonymous ciphers are OK (meaning
3256 that you can't actually use DEFAULT).
3257 [Bodo Moeller; suggested by Victor Duchovni]
3259 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3260 processing) into multiple integers instead of setting
3261 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3262 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3263 (These masks as well as the individual bit definitions are hidden
3264 away into the non-exported interface ssl/ssl_locl.h, so this
3265 change to the definition of the SSL_CIPHER structure shouldn't
3266 affect applications.) This give us more bits for each of these
3267 categories, so there is no longer a need to coagulate AES128 and
3268 AES256 into a single algorithm bit, and to coagulate Camellia128
3269 and Camellia256 into a single algorithm bit, which has led to all
3272 Thus, among other things, the kludge introduced in 0.9.7m and
3273 0.9.8e for masking out AES256 independently of AES128 or masking
3274 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3276 With the change, we also introduce new ciphersuite aliases that
3277 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3281 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3282 Use the leftmost N bytes of the signature input if the input is
3283 larger than the prime q (with N being the size in bytes of q).
3286 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3287 it yet and it is largely untested.
3290 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3293 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3294 some compilers (gcc 4.2 and later) reject their use. Safestack is
3295 reimplemented. Update ASN1 to avoid use of legacy functions.
3298 *) Win32/64 targets are linked with Winsock2.
3301 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3302 to external functions. This can be used to increase CRL handling
3303 efficiency especially when CRLs are very large by (for example) storing
3304 the CRL revoked certificates in a database.
3307 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3308 new CRLs added to a directory can be used. New command line option
3309 -verify_return_error to s_client and s_server. This causes real errors
3310 to be returned by the verify callback instead of carrying on no matter
3311 what. This reflects the way a "real world" verify callback would behave.
3314 *) GOST engine, supporting several GOST algorithms and public key formats.
3315 Kindly donated by Cryptocom.
3318 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3319 partitioned by DP are handled but no indirect CRL or reason partitioning
3320 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3321 selected via a scoring technique which handles IDP and AKID in CRLs.
3324 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3325 will ultimately be used for all verify operations: this will remove the
3326 X509_STORE dependency on certificate verification and allow alternative
3327 lookup methods. X509_STORE based implementations of these two callbacks.
3330 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3331 Modify get_crl() to find a valid (unexpired) CRL if possible.
3334 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3335 this would be called X509_CRL_cmp() but that name is already used by
3336 a function that just compares CRL issuer names. Cache several CRL
3337 extensions in X509_CRL structure and cache CRLDP in X509.
3340 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3341 this maps equivalent X509_NAME structures into a consistent structure.
3342 Name comparison can then be performed rapidly using memcmp().
3345 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3349 *) Allow digests to supply their own micalg string for S/MIME type using
3350 the ctrl EVP_MD_CTRL_MICALG.
3353 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3354 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3355 ctrl. It can then customise the structure before and/or after signing
3359 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3360 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3361 to free up any added signature OIDs.
3364 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3365 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3366 digest and cipher tables. New options added to openssl utility:
3367 list-message-digest-algorithms and list-cipher-algorithms.
3370 *) Change the array representation of binary polynomials: the list
3371 of degrees of non-zero coefficients is now terminated with -1.
3372 Previously it was terminated with 0, which was also part of the
3373 value; thus, the array representation was not applicable to
3374 polynomials where t^0 has coefficient zero. This change makes
3375 the array representation useful in a more general context.
3378 *) Various modifications and fixes to SSL/TLS cipher string
3379 handling. For ECC, the code now distinguishes between fixed ECDH
3380 with RSA certificates on the one hand and with ECDSA certificates
3381 on the other hand, since these are separate ciphersuites. The
3382 unused code for Fortezza ciphersuites has been removed.
3384 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3385 (not "ECDHE"). For consistency with the code for DH
3386 certificates, use of ECDH certificates is now considered ECDH
3387 authentication, not RSA or ECDSA authentication (the latter is
3388 merely the CA's signing algorithm and not actively used in the
3391 The temporary ciphersuite alias "ECCdraft" is no longer
3392 available, and ECC ciphersuites are no longer excluded from "ALL"
3393 and "DEFAULT". The following aliases now exist for RFC 4492
3394 ciphersuites, most of these by analogy with the DH case:
3396 kECDHr - ECDH cert, signed with RSA
3397 kECDHe - ECDH cert, signed with ECDSA
3398 kECDH - ECDH cert (signed with either RSA or ECDSA)
3399 kEECDH - ephemeral ECDH
3400 ECDH - ECDH cert or ephemeral ECDH
3406 AECDH - anonymous ECDH
3407 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3411 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3412 Use correct micalg parameters depending on digest(s) in signed message.
3415 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3416 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3419 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3420 an engine to register a method. Add ENGINE lookups for methods and
3421 functional reference processing.
3424 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3425 EVP_{Sign,Verify}* which allow an application to customise the signature
3429 *) New -resign option to smime utility. This adds one or more signers
3430 to an existing PKCS#7 signedData structure. Also -md option to use an
3431 alternative message digest algorithm for signing.
3434 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3435 create PKCS7 structures containing multiple signers. Update smime
3436 application to support multiple signers.
3439 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3443 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3444 Reorganize PBE internals to lookup from a static table using NIDs,
3445 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3446 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3447 PRF which will be automatically used with PBES2.
3450 *) Replace the algorithm specific calls to generate keys in "req" with the
3454 *) Update PKCS#7 enveloped data routines to use new API. This is now
3455 supported by any public key method supporting the encrypt operation. A
3456 ctrl is added to allow the public key algorithm to examine or modify
3457 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3461 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3462 a default digest type to use. In most cases this will be SHA1 but some
3463 algorithms (such as GOST) need to specify an alternative digest. The
3464 return value indicates how strong the preference is 1 means optional and
3465 2 is mandatory (that is it is the only supported type). Modify
3466 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3467 use the default md. Update openssl utilities to use the default digest
3468 type for signing if it is not explicitly indicated.
3471 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3472 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3473 signing method from the key type. This effectively removes the link
3474 between digests and public key types.
3477 *) Add an OID cross reference table and utility functions. Its purpose is to
3478 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3479 rsaEncryption. This will allow some of the algorithm specific hackery
3480 needed to use the correct OID to be removed.
3483 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3484 structures for PKCS7_sign(). They are now set up by the relevant public
3488 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3491 *) Add support for key derivation (agreement) in the API, DH method and
3495 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3496 public and private key formats. As a side effect these add additional
3497 command line functionality not previously available: DSA signatures can be
3498 generated and verified using pkeyutl and DH key support and generation in
3503 [Oliver Tappe <zooey@hirschkaefer.de>]
3505 *) New make target "install_html_docs" installs HTML renditions of the
3507 [Oliver Tappe <zooey@hirschkaefer.de>]
3509 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3510 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3511 support key and parameter generation and add initial key generation
3512 functionality for RSA.
3515 *) Add functions for main EVP_PKEY_method operations. The undocumented
3516 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3517 EVP_PKEY_{encrypt,decrypt}_old.
3520 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3521 key API, doesn't do much yet.
3524 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3525 public key algorithms. New option to openssl utility:
3526 "list-public-key-algorithms" to print out info.
3529 *) Implement the Supported Elliptic Curves Extension for
3530 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3533 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3534 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3537 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3538 utilities such as rsa, dsa, dsaparam etc except they process any key
3542 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3543 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3544 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3548 *) Initial support for pluggable public key ASN1.
3549 De-spaghettify the public key ASN1 handling. Move public and private
3550 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3551 algorithm specific handling to a single module within the relevant
3552 algorithm directory. Add functions to allow (near) opaque processing
3553 of public and private key structures.
3556 *) Implement the Supported Point Formats Extension for
3557 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3560 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3561 for the psk identity [hint] and the psk callback functions to the
3562 SSL_SESSION, SSL and SSL_CTX structure.
3565 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3569 SSL_CTX_use_psk_identity_hint
3570 SSL_get_psk_identity_hint
3571 SSL_get_psk_identity
3572 SSL_use_psk_identity_hint
3574 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3576 *) Add RFC 3161 compliant time stamp request creation, response generation
3577 and response verification functionality.
3578 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3580 *) Add initial support for TLS extensions, specifically for the server_name
3581 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3582 have new members for a host name. The SSL data structure has an
3583 additional member SSL_CTX *initial_ctx so that new sessions can be
3584 stored in that context to allow for session resumption, even after the
3585 SSL has been switched to a new SSL_CTX in reaction to a client's
3586 server_name extension.
3588 New functions (subject to change):
3590 SSL_get_servername()
3591 SSL_get_servername_type()
3594 New CTRL codes and macros (subject to change):
3596 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3597 - SSL_CTX_set_tlsext_servername_callback()
3598 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3599 - SSL_CTX_set_tlsext_servername_arg()
3600 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3602 openssl s_client has a new '-servername ...' option.
3604 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3605 '-key2 ...', '-servername_fatal' (subject to change). This allows
3606 testing the HostName extension for a specific single host name ('-cert'
3607 and '-key' remain fallbacks for handshakes without HostName
3608 negotiation). If the unrecognized_name alert has to be sent, this by
3609 default is a warning; it becomes fatal with the '-servername_fatal'
3612 [Peter Sylvester, Remy Allais, Christophe Renou]
3614 *) Whirlpool hash implementation is added.
3617 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3618 bn(64,32). Because of instruction set limitations it doesn't have
3619 any negative impact on performance. This was done mostly in order
3620 to make it possible to share assembler modules, such as bn_mul_mont
3621 implementations, between 32- and 64-bit builds without hassle.
3624 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3625 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3629 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3630 dedicated Montgomery multiplication procedure, is introduced.
3631 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3632 "64-bit" performance on certain 32-bit targets.
3635 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3636 in SSL structures. New SSL ctrl to set maximum send fragment size.
3637 Save memory by seeting the I/O buffer sizes dynamically instead of
3638 using the maximum available value.
3641 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3642 in addition to the text details.
3645 *) Very, very preliminary EXPERIMENTAL support for printing of general
3646 ASN1 structures. This currently produces rather ugly output and doesn't
3647 handle several customised structures at all.
3650 *) Integrated support for PVK file format and some related formats such
3651 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
3652 these in the 'rsa' and 'dsa' utilities.
3655 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
3658 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
3659 place for the (very old) "NETSCAPE" format certificates which are now
3660 handled using new ASN1 code equivalents.
3663 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
3664 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
3665 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
3668 *) Modify CRL distribution points extension code to print out previously
3669 unsupported fields. Enhance extension setting code to allow setting of
3673 *) Add print and set support for Issuing Distribution Point CRL extension.
3676 *) Change 'Configure' script to enable Camellia by default.
3679 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
3681 *) When rejecting SSL/TLS records due to an incorrect version number, never
3682 update s->server with a new major version number. As of
3683 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
3684 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
3685 the previous behavior could result in a read attempt at NULL when
3686 receiving specific incorrect SSL/TLS records once record payload
3687 protection is active. (CVE-2010-0740)
3688 [Bodo Moeller, Adam Langley <agl@chromium.org>]
3690 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
3691 could be crashed if the relevant tables were not present (e.g. chrooted).
3692 [Tomas Hoger <thoger@redhat.com>]
3694 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
3696 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
3697 [Martin Olsson, Neel Mehta]
3699 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
3700 accommodate for stack sorting, always a write lock!).
3703 *) On some versions of WIN32 Heap32Next is very slow. This can cause
3704 excessive delays in the RAND_poll(): over a minute. As a workaround
3705 include a time check in the inner Heap32Next loop too.
3708 *) The code that handled flushing of data in SSL/TLS originally used the
3709 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
3710 the problem outlined in PR#1949. The fix suggested there however can
3711 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
3712 of Apache). So instead simplify the code to flush unconditionally.
3713 This should be fine since flushing with no data to flush is a no op.
3716 *) Handle TLS versions 2.0 and later properly and correctly use the
3717 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
3718 off ancient servers have a habit of sticking around for a while...
3721 *) Modify compression code so it frees up structures without using the
3722 ex_data callbacks. This works around a problem where some applications
3723 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when