5 Changes between 1.1.0a and 1.1.1 [xx XXX xxxx]
7 *) MDC2 is now disabled by default.
10 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
11 using the algorithm defined in
12 https://www.akkadia.org/drepper/SHA-crypt.txt
15 Changes between 1.1.0b and 1.1.0c [xx XXX xxxx]
17 *) Removed automatic addition of RPATH in shared libraries and executables,
18 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
21 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
23 *) Fix Use After Free for large message sizes
25 The patch applied to address CVE-2016-6307 resulted in an issue where if a
26 message larger than approx 16k is received then the underlying buffer to
27 store the incoming message is reallocated and moved. Unfortunately a
28 dangling pointer to the old location is left which results in an attempt to
29 write to the previously freed location. This is likely to result in a
30 crash, however it could potentially lead to execution of arbitrary code.
32 This issue only affects OpenSSL 1.1.0a.
34 This issue was reported to OpenSSL by Robert Święcki.
38 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
40 *) OCSP Status Request extension unbounded memory growth
42 A malicious client can send an excessively large OCSP Status Request
43 extension. If that client continually requests renegotiation, sending a
44 large OCSP Status Request extension each time, then there will be unbounded
45 memory growth on the server. This will eventually lead to a Denial Of
46 Service attack through memory exhaustion. Servers with a default
47 configuration are vulnerable even if they do not support OCSP. Builds using
48 the "no-ocsp" build time option are not affected.
50 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
54 *) SSL_peek() hang on empty record
56 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
57 sends an empty record. This could be exploited by a malicious peer in a
58 Denial Of Service attack.
60 This issue was reported to OpenSSL by Alex Gaynor.
64 *) Excessive allocation of memory in tls_get_message_header() and
65 dtls1_preprocess_fragment()
67 A (D)TLS message includes 3 bytes for its length in the header for the
68 message. This would allow for messages up to 16Mb in length. Messages of
69 this length are excessive and OpenSSL includes a check to ensure that a
70 peer is sending reasonably sized messages in order to avoid too much memory
71 being consumed to service a connection. A flaw in the logic of version
72 1.1.0 means that memory for the message is allocated too early, prior to
73 the excessive message length check. Due to way memory is allocated in
74 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
75 to service a connection. This could lead to a Denial of Service through
76 memory exhaustion. However, the excessive message length check still takes
77 place, and this would cause the connection to immediately fail. Assuming
78 that the application calls SSL_free() on the failed conneciton in a timely
79 manner then the 21Mb of allocated memory will then be immediately freed
80 again. Therefore the excessive memory allocation will be transitory in
81 nature. This then means that there is only a security impact if:
83 1) The application does not call SSL_free() in a timely manner in the event
84 that the connection fails
86 2) The application is working in a constrained environment where there is
87 very little free memory
89 3) The attacker initiates multiple connection attempts such that there are
90 multiple connections in a state where memory has been allocated for the
91 connection; SSL_free() has not yet been called; and there is insufficient
92 memory to service the multiple requests.
94 Except in the instance of (1) above any Denial Of Service is likely to be
95 transitory because as soon as the connection fails the memory is
96 subsequently freed again in the SSL_free() call. However there is an
97 increased risk during this period of application crashes due to the lack of
98 memory - which would then mean a more serious Denial of Service.
100 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
101 (CVE-2016-6307 and CVE-2016-6308)
104 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
105 had to be removed. Primary reason is that vendor assembler can't
106 assemble our modules with -KPIC flag. As result it, assembly
107 support, was not even available as option. But its lack means
108 lack of side-channel resistant code, which is incompatible with
109 security by todays standards. Fortunately gcc is readily available
110 prepackaged option, which we firmly point at...
113 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
115 *) Windows command-line tool supports UTF-8 opt-in option for arguments
116 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
117 (to any value) allows Windows user to access PKCS#12 file generated
118 with Windows CryptoAPI and protected with non-ASCII password, as well
119 as files generated under UTF-8 locale on Linux also protected with
123 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
124 have been disabled by default and removed from DEFAULT, just like RC4.
125 See the RC4 item below to re-enable both.
128 *) The method for finding the storage location for the Windows RAND seed file
129 has changed. First we check %RANDFILE%. If that is not set then we check
130 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
131 all else fails we fall back to C:\.
134 *) The EVP_EncryptUpdate() function has had its return type changed from void
135 to int. A return of 0 indicates and error while a return of 1 indicates
139 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
140 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
141 off the constant time implementation for RSA, DSA and DH have been made
142 no-ops and deprecated.
145 *) Windows RAND implementation was simplified to only get entropy by
146 calling CryptGenRandom(). Various other RAND-related tickets
148 [Joseph Wylie Yandle, Rich Salz]
150 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
151 and OPENSSL_LH_, respectively. The old names are available
152 with API compatibility. They new names are now completely documented.
155 *) Unify TYPE_up_ref(obj) methods signature.
156 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
157 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
158 int (instead of void) like all others TYPE_up_ref() methods.
159 So now these methods also check the return value of CRYPTO_atomic_add(),
160 and the validity of object reference counter.
161 [fdasilvayy@gmail.com]
163 *) With Windows Visual Studio builds, the .pdb files are installed
164 alongside the installed libraries and executables. For a static
165 library installation, ossl_static.pdb is the associate compiler
166 generated .pdb file to be used when linking programs.
169 *) Remove openssl.spec. Packaging files belong with the packagers.
172 *) Automatic Darwin/OSX configuration has had a refresh, it will now
173 recognise x86_64 architectures automatically. You can still decide
174 to build for a different bitness with the environment variable
175 KERNEL_BITS (can be 32 or 64), for example:
177 KERNEL_BITS=32 ./config
181 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
182 256 bit AES and HMAC with SHA256.
185 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
188 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
191 *) To enable users to have their own config files and build file templates,
192 Configure looks in the directory indicated by the environment variable
193 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
194 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
195 name and is used as is.
198 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
199 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
200 X509_CERT_FILE_CTX was removed.
203 *) "shared" builds are now the default. To create only static libraries use
204 the "no-shared" Configure option.
207 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
208 All of these option have not worked for some while and are fundamental
212 *) Make various cleanup routines no-ops and mark them as deprecated. Most
213 global cleanup functions are no longer required because they are handled
214 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
215 Explicitly de-initing can cause problems (e.g. where a library that uses
216 OpenSSL de-inits, but an application is still using it). The affected
217 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
218 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
219 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
223 *) --strict-warnings no longer enables runtime debugging options
224 such as REF_DEBUG. Instead, debug options are automatically
225 enabled with '--debug' builds.
226 [Andy Polyakov, Emilia Käsper]
228 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
229 have been moved out of the public header files. New functions for managing
230 these have been added.
233 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
234 objects have been moved out of the public header files. New
235 functions for managing these have been added.
238 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
239 have been moved out of the public header files. New functions for managing
240 these have been added.
243 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
244 moved out of the public header files. New functions for managing these
248 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
251 *) Removed the mk1mf build scripts.
254 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
255 it is always safe to #include a header now.
258 *) Removed the aged BC-32 config and all its supporting scripts
261 *) Removed support for Ultrix, Netware, and OS/2.
264 *) Add support for HKDF.
267 *) Add support for blake2b and blake2s
270 *) Added support for "pipelining". Ciphers that have the
271 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
272 encryptions/decryptions simultaneously. There are currently no built-in
273 ciphers with this property but the expectation is that engines will be able
274 to offer it to significantly improve throughput. Support has been extended
275 into libssl so that multiple records for a single connection can be
276 processed in one go (for >=TLS 1.1).
279 *) Added the AFALG engine. This is an async capable engine which is able to
280 offload work to the Linux kernel. In this initial version it only supports
281 AES128-CBC. The kernel must be version 4.1.0 or greater.
284 *) OpenSSL now uses a new threading API. It is no longer necessary to
285 set locking callbacks to use OpenSSL in a multi-threaded environment. There
286 are two supported threading models: pthreads and windows threads. It is
287 also possible to configure OpenSSL at compile time for "no-threads". The
288 old threading API should no longer be used. The functions have been
289 replaced with "no-op" compatibility macros.
290 [Alessandro Ghedini, Matt Caswell]
292 *) Modify behavior of ALPN to invoke callback after SNI/servername
293 callback, such that updates to the SSL_CTX affect ALPN.
296 *) Add SSL_CIPHER queries for authentication and key-exchange.
299 *) Changes to the DEFAULT cipherlist:
300 - Prefer (EC)DHE handshakes over plain RSA.
301 - Prefer AEAD ciphers over legacy ciphers.
302 - Prefer ECDSA over RSA when both certificates are available.
303 - Prefer TLSv1.2 ciphers/PRF.
304 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
308 *) Change the ECC default curve list to be this, in order: x25519,
309 secp256r1, secp521r1, secp384r1.
312 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
313 disabled by default. They can be re-enabled using the
314 enable-weak-ssl-ciphers option to Configure.
317 *) If the server has ALPN configured, but supports no protocols that the
318 client advertises, send a fatal "no_application_protocol" alert.
319 This behaviour is SHALL in RFC 7301, though it isn't universally
320 implemented by other servers.
323 *) Add X25519 support.
324 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
325 for public and private key encoding using the format documented in
326 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
327 key generation and key derivation.
329 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
333 *) Deprecate SRP_VBASE_get_by_user.
334 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
335 In order to fix an unavoidable memory leak (CVE-2016-0798),
336 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
337 seed, even if the seed is configured.
339 Users should use SRP_VBASE_get1_by_user instead. Note that in
340 SRP_VBASE_get1_by_user, caller must free the returned value. Note
341 also that even though configuring the SRP seed attempts to hide
342 invalid usernames by continuing the handshake with fake
343 credentials, this behaviour is not constant time and no strong
344 guarantees are made that the handshake is indistinguishable from
345 that of a valid user.
348 *) Configuration change; it's now possible to build dynamic engines
349 without having to build shared libraries and vice versa. This
350 only applies to the engines in engines/, those in crypto/engine/
351 will always be built into libcrypto (i.e. "static").
353 Building dynamic engines is enabled by default; to disable, use
354 the configuration option "disable-dynamic-engine".
356 The only requirements for building dynamic engines are the
357 presence of the DSO module and building with position independent
358 code, so they will also automatically be disabled if configuring
359 with "disable-dso" or "disable-pic".
361 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
362 are also taken away from openssl/opensslconf.h, as they are
366 *) Configuration change; if there is a known flag to compile
367 position independent code, it will always be applied on the
368 libcrypto and libssl object files, and never on the application
369 object files. This means other libraries that use routines from
370 libcrypto / libssl can be made into shared libraries regardless
371 of how OpenSSL was configured.
373 If this isn't desirable, the configuration options "disable-pic"
374 or "no-pic" can be used to disable the use of PIC. This will
375 also disable building shared libraries and dynamic engines.
378 *) Removed JPAKE code. It was experimental and has no wide use.
381 *) The INSTALL_PREFIX Makefile variable has been renamed to
382 DESTDIR. That makes for less confusion on what this variable
383 is for. Also, the configuration option --install_prefix is
387 *) Heartbeat for TLS has been removed and is disabled by default
388 for DTLS; configure with enable-heartbeats. Code that uses the
389 old #define's might need to be updated.
390 [Emilia Käsper, Rich Salz]
392 *) Rename REF_CHECK to REF_DEBUG.
395 *) New "unified" build system
397 The "unified" build system is aimed to be a common system for all
398 platforms we support. With it comes new support for VMS.
400 This system builds supports building in a different directory tree
401 than the source tree. It produces one Makefile (for unix family
402 or lookalikes), or one descrip.mms (for VMS).
404 The source of information to make the Makefile / descrip.mms is
405 small files called 'build.info', holding the necessary
406 information for each directory with source to compile, and a
407 template in Configurations, like unix-Makefile.tmpl or
410 With this change, the library names were also renamed on Windows
411 and on VMS. They now have names that are closer to the standard
412 on Unix, and include the major version number, and in certain
413 cases, the architecture they are built for. See "Notes on shared
414 libraries" in INSTALL.
416 We rely heavily on the perl module Text::Template.
419 *) Added support for auto-initialisation and de-initialisation of the library.
420 OpenSSL no longer requires explicit init or deinit routines to be called,
421 except in certain circumstances. See the OPENSSL_init_crypto() and
422 OPENSSL_init_ssl() man pages for further information.
425 *) The arguments to the DTLSv1_listen function have changed. Specifically the
426 "peer" argument is now expected to be a BIO_ADDR object.
428 *) Rewrite of BIO networking library. The BIO library lacked consistent
429 support of IPv6, and adding it required some more extensive
430 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
431 which hold all types of addresses and chains of address information.
432 It also introduces a new API, with functions like BIO_socket,
433 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
434 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
435 have been adapted accordingly.
438 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
442 *) CRIME protection: disable compression by default, even if OpenSSL is
443 compiled with zlib enabled. Applications can still enable compression
444 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
445 using the SSL_CONF library to configure compression.
448 *) The signature of the session callback configured with
449 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
450 was explicitly marked as 'const unsigned char*' instead of
454 *) Always DPURIFY. Remove the use of uninitialized memory in the
455 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
458 *) Removed many obsolete configuration items, including
459 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
460 MD2_CHAR, MD2_INT, MD2_LONG
462 IDEA_SHORT, IDEA_LONG
463 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
464 [Rich Salz, with advice from Andy Polyakov]
466 *) Many BN internals have been moved to an internal header file.
467 [Rich Salz with help from Andy Polyakov]
469 *) Configuration and writing out the results from it has changed.
470 Files such as Makefile include/openssl/opensslconf.h and are now
471 produced through general templates, such as Makefile.in and
472 crypto/opensslconf.h.in and some help from the perl module
475 Also, the center of configuration information is no longer
476 Makefile. Instead, Configure produces a perl module in
477 configdata.pm which holds most of the config data (in the hash
478 table %config), the target data that comes from the target
479 configuration in one of the Configurations/*.conf files (in
483 *) To clarify their intended purposes, the Configure options
484 --prefix and --openssldir change their semantics, and become more
485 straightforward and less interdependent.
487 --prefix shall be used exclusively to give the location INSTALLTOP
488 where programs, scripts, libraries, include files and manuals are
489 going to be installed. The default is now /usr/local.
491 --openssldir shall be used exclusively to give the default
492 location OPENSSLDIR where certificates, private keys, CRLs are
493 managed. This is also where the default openssl.cnf gets
495 If the directory given with this option is a relative path, the
496 values of both the --prefix value and the --openssldir value will
497 be combined to become OPENSSLDIR.
498 The default for --openssldir is INSTALLTOP/ssl.
500 Anyone who uses --openssldir to specify where OpenSSL is to be
501 installed MUST change to use --prefix instead.
504 *) The GOST engine was out of date and therefore it has been removed. An up
505 to date GOST engine is now being maintained in an external repository.
506 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
507 support for GOST ciphersuites (these are only activated if a GOST engine
511 *) EGD is no longer supported by default; use enable-egd when
513 [Ben Kaduk and Rich Salz]
515 *) The distribution now has Makefile.in files, which are used to
516 create Makefile's when Configure is run. *Configure must be run
517 before trying to build now.*
520 *) The return value for SSL_CIPHER_description() for error conditions
524 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
526 Obtaining and performing DNSSEC validation of TLSA records is
527 the application's responsibility. The application provides
528 the TLSA records of its choice to OpenSSL, and these are then
529 used to authenticate the peer.
531 The TLSA records need not even come from DNS. They can, for
532 example, be used to implement local end-entity certificate or
533 trust-anchor "pinning", where the "pin" data takes the form
534 of TLSA records, which can augment or replace verification
535 based on the usual WebPKI public certification authorities.
538 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
539 continues to support deprecated interfaces in default builds.
540 However, applications are strongly advised to compile their
541 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
542 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
543 or the 1.1.0 releases.
545 In environments in which all applications have been ported to
546 not use any deprecated interfaces OpenSSL's Configure script
547 should be used with the --api=1.1.0 option to entirely remove
548 support for the deprecated features from the library and
549 unconditionally disable them in the installed headers.
550 Essentially the same effect can be achieved with the "no-deprecated"
551 argument to Configure, except that this will always restrict
552 the build to just the latest API, rather than a fixed API
555 As applications are ported to future revisions of the API,
556 they should update their compile-time OPENSSL_API_COMPAT define
557 accordingly, but in most cases should be able to continue to
558 compile with later releases.
560 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
561 0x10000000L and 0x00908000L, respectively. However those
562 versions did not support the OPENSSL_API_COMPAT feature, and
563 so applications are not typically tested for explicit support
564 of just the undeprecated features of either release.
567 *) Add support for setting the minimum and maximum supported protocol.
568 It can bet set via the SSL_set_min_proto_version() and
569 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
570 MaxProtcol. It's recommended to use the new APIs to disable
571 protocols instead of disabling individual protocols using
572 SSL_set_options() or SSL_CONF's Protocol. This change also
573 removes support for disabling TLS 1.2 in the OpenSSL TLS
574 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
577 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
580 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
581 and integrates ECDSA and ECDH functionality into EC. Implementations can
582 now redirect key generation and no longer need to convert to or from
585 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
586 include the ec.h header file instead.
589 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
590 ciphers who are no longer supported and drops support the ephemeral RSA key
591 exchange. The LOW ciphers currently doesn't have any ciphers in it.
594 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
595 opaque. For HMAC_CTX, the following constructors and destructors
598 HMAC_CTX *HMAC_CTX_new(void);
599 void HMAC_CTX_free(HMAC_CTX *ctx);
601 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
602 destroy such methods has been added. See EVP_MD_meth_new(3) and
603 EVP_CIPHER_meth_new(3) for documentation.
606 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
607 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
608 EVP_MD_CTX_reset() should be called instead to reinitialise
609 an already created structure.
610 2) For consistency with the majority of our object creators and
611 destructors, EVP_MD_CTX_(create|destroy) were renamed to
612 EVP_MD_CTX_(new|free). The old names are retained as macros
613 for deprecated builds.
616 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
617 cryptographic operations to be performed asynchronously as long as an
618 asynchronous capable engine is used. See the ASYNC_start_job() man page for
619 further details. Libssl has also had this capability integrated with the
620 introduction of the new mode SSL_MODE_ASYNC and associated error
621 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
622 pages. This work was developed in partnership with Intel Corp.
625 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
626 always enabled now. If you want to disable the support you should
627 exclude it using the list of supported ciphers. This also means that the
628 "-no_ecdhe" option has been removed from s_server.
631 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
632 SSL_{CTX_}set1_curves() which can set a list.
635 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
636 curve you want to support using SSL_{CTX_}set1_curves().
639 *) State machine rewrite. The state machine code has been significantly
640 refactored in order to remove much duplication of code and solve issues
641 with the old code (see ssl/statem/README for further details). This change
642 does have some associated API changes. Notably the SSL_state() function
643 has been removed and replaced by SSL_get_state which now returns an
644 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
645 altogether. The previous handshake states defined in ssl.h and ssl3.h have
649 *) All instances of the string "ssleay" in the public API were replaced
650 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
651 Some error codes related to internal RSA_eay API's were renamed.
654 *) The demo files in crypto/threads were moved to demo/threads.
657 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
659 [Matt Caswell, Rich Salz]
661 *) New ASN.1 embed macro.
663 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
664 structure is not allocated: it is part of the parent. That is instead of
672 This reduces memory fragmentation and make it impossible to accidentally
673 set a mandatory field to NULL.
675 This currently only works for some fields specifically a SEQUENCE, CHOICE,
676 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
677 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
681 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
684 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
685 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
686 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
687 DES and RC4 ciphersuites.
690 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
691 This changes the decoding behaviour for some invalid messages,
692 though the change is mostly in the more lenient direction, and
693 legacy behaviour is preserved as much as possible.
696 *) Fix no-stdio build.
697 [ David Woodhouse <David.Woodhouse@intel.com> and also
698 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
700 *) New testing framework
701 The testing framework has been largely rewritten and is now using
702 perl and the perl modules Test::Harness and an extended variant of
703 Test::More called OpenSSL::Test to do its work. All test scripts in
704 test/ have been rewritten into test recipes, and all direct calls to
705 executables in test/Makefile have become individual recipes using the
706 simplified testing OpenSSL::Test::Simple.
708 For documentation on our testing modules, do:
710 perldoc test/testlib/OpenSSL/Test/Simple.pm
711 perldoc test/testlib/OpenSSL/Test.pm
715 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
716 are used; the latter aborts on memory leaks (usually checked on exit).
717 Some undocumented "set malloc, etc., hooks" functions were removed
718 and others were changed. All are now documented.
721 *) In DSA_generate_parameters_ex, if the provided seed is too short,
723 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
725 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
726 from RFC4279, RFC4785, RFC5487, RFC5489.
728 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
729 original RSA_PSK patch.
732 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
733 era flag was never set throughout the codebase (only read). Also removed
734 SSL3_FLAGS_POP_BUFFER which was only used if
735 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
738 *) Changed the default name options in the "ca", "crl", "req" and "x509"
739 to be "oneline" instead of "compat".
742 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
743 not aware of clients that still exhibit this bug, and the workaround
744 hasn't been working properly for a while.
747 *) The return type of BIO_number_read() and BIO_number_written() as well as
748 the corresponding num_read and num_write members in the BIO structure has
749 changed from unsigned long to uint64_t. On platforms where an unsigned
750 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
754 *) Given the pervasive nature of TLS extensions it is inadvisable to run
755 OpenSSL without support for them. It also means that maintaining
756 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
757 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
760 *) Removed support for the two export grade static DH ciphersuites
761 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
762 were newly added (along with a number of other static DH ciphersuites) to
763 1.0.2. However the two export ones have *never* worked since they were
764 introduced. It seems strange in any case to be adding new export
765 ciphersuites, and given "logjam" it also does not seem correct to fix them.
768 *) Version negotiation has been rewritten. In particular SSLv23_method(),
769 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
770 and turned into macros which simply call the new preferred function names
771 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
772 should use the new names instead. Also as part of this change the ssl23.h
773 header file has been removed.
776 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
777 code and the associated standard is no longer considered fit-for-purpose.
780 *) RT2547 was closed. When generating a private key, try to make the
781 output file readable only by the owner. This behavior change might
782 be noticeable when interacting with other software.
784 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
788 *) Added HTTP GET support to the ocsp command.
791 *) Changed default digest for the dgst and enc commands from MD5 to
795 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
798 *) Added support for TLS extended master secret from
799 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
800 initial patch which was a great help during development.
803 *) All libssl internal structures have been removed from the public header
804 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
805 now redundant). Users should not attempt to access internal structures
806 directly. Instead they should use the provided API functions.
809 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
810 Access to deprecated functions can be re-enabled by running config with
811 "enable-deprecated". In addition applications wishing to use deprecated
812 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
813 will, by default, disable some transitive includes that previously existed
814 in the header files (e.g. ec.h will no longer, by default, include bn.h)
817 *) Added support for OCB mode. OpenSSL has been granted a patent license
818 compatible with the OpenSSL license for use of OCB. Details are available
819 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
820 for OCB can be removed by calling config with no-ocb.
823 *) SSLv2 support has been removed. It still supports receiving a SSLv2
824 compatible client hello.
827 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
828 done while fixing the error code for the key-too-small case.
829 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
831 *) CA.sh has been removmed; use CA.pl instead.
834 *) Removed old DES API.
837 *) Remove various unsupported platforms:
843 Sinix/ReliantUNIX RM400
848 16-bit platforms such as WIN16
851 *) Clean up OPENSSL_NO_xxx #define's
852 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
853 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
854 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
855 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
856 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
857 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
858 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
859 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
860 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
861 Remove MS_STATIC; it's a relic from platforms <32 bits.
864 *) Cleaned up dead code
865 Remove all but one '#ifdef undef' which is to be looked at.
868 *) Clean up calling of xxx_free routines.
869 Just like free(), fix most of the xxx_free routines to accept
870 NULL. Remove the non-null checks from callers. Save much code.
873 *) Add secure heap for storage of private keys (when possible).
874 Add BIO_s_secmem(), CBIGNUM, etc.
875 Contributed by Akamai Technologies under our Corporate CLA.
878 *) Experimental support for a new, fast, unbiased prime candidate generator,
879 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
880 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
882 *) New output format NSS in the sess_id command line tool. This allows
883 exporting the session id and the master key in NSS keylog format.
884 [Martin Kaiser <martin@kaiser.cx>]
886 *) Harmonize version and its documentation. -f flag is used to display
888 [mancha <mancha1@zoho.com>]
890 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
891 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
892 [mancha <mancha1@zoho.com>]
894 *) Fix some double frees. These are not thought to be exploitable.
895 [mancha <mancha1@zoho.com>]
897 *) A missing bounds check in the handling of the TLS heartbeat extension
898 can be used to reveal up to 64k of memory to a connected client or
901 Thanks for Neel Mehta of Google Security for discovering this bug and to
902 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
903 preparing the fix (CVE-2014-0160)
904 [Adam Langley, Bodo Moeller]
906 *) Fix for the attack described in the paper "Recovering OpenSSL
907 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
908 by Yuval Yarom and Naomi Benger. Details can be obtained from:
909 http://eprint.iacr.org/2014/140
911 Thanks to Yuval Yarom and Naomi Benger for discovering this
912 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
913 [Yuval Yarom and Naomi Benger]
915 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
916 this fixes a limitation in previous versions of OpenSSL.
919 *) Experimental encrypt-then-mac support.
921 Experimental support for encrypt then mac from
922 draft-gutmann-tls-encrypt-then-mac-02.txt
924 To enable it set the appropriate extension number (0x42 for the test
925 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
927 For non-compliant peers (i.e. just about everything) this should have no
930 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
934 *) Add EVP support for key wrapping algorithms, to avoid problems with
935 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
936 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
937 algorithms and include tests cases.
940 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
944 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
945 MGF1 digest and OAEP label.
948 *) Make openssl verify return errors.
949 [Chris Palmer <palmer@google.com> and Ben Laurie]
951 *) New function ASN1_TIME_diff to calculate the difference between two
952 ASN1_TIME structures or one structure and the current time.
955 *) Update fips_test_suite to support multiple command line options. New
956 test to induce all self test errors in sequence and check expected
960 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
961 sign or verify all in one operation.
964 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
965 test programs and fips_test_suite. Includes functionality to parse
966 the minimal script output of fipsalgest.pl directly.
969 *) Add authorisation parameter to FIPS_module_mode_set().
972 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
975 *) Use separate DRBG fields for internal and external flags. New function
976 FIPS_drbg_health_check() to perform on demand health checking. Add
977 generation tests to fips_test_suite with reduced health check interval to
978 demonstrate periodic health checking. Add "nodh" option to
979 fips_test_suite to skip very slow DH test.
982 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
986 *) More extensive health check for DRBG checking many more failure modes.
987 New function FIPS_selftest_drbg_all() to handle every possible DRBG
988 combination: call this in fips_test_suite.
991 *) Add support for canonical generation of DSA parameter 'g'. See
994 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
995 POST to handle HMAC cases.
998 *) Add functions FIPS_module_version() and FIPS_module_version_text()
999 to return numerical and string versions of the FIPS module number.
1002 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1003 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1004 outside the validated module in the FIPS capable OpenSSL.
1007 *) Minor change to DRBG entropy callback semantics. In some cases
1008 there is no multiple of the block length between min_len and
1009 max_len. Allow the callback to return more than max_len bytes
1010 of entropy but discard any extra: it is the callback's responsibility
1011 to ensure that the extra data discarded does not impact the
1012 requested amount of entropy.
1015 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1016 information in FIPS186-3, SP800-57 and SP800-131A.
1019 *) CCM support via EVP. Interface is very similar to GCM case except we
1020 must supply all data in one chunk (i.e. no update, final) and the
1021 message length must be supplied if AAD is used. Add algorithm test
1025 *) Initial version of POST overhaul. Add POST callback to allow the status
1026 of POST to be monitored and/or failures induced. Modify fips_test_suite
1027 to use callback. Always run all selftests even if one fails.
1030 *) XTS support including algorithm test driver in the fips_gcmtest program.
1031 Note: this does increase the maximum key length from 32 to 64 bytes but
1032 there should be no binary compatibility issues as existing applications
1033 will never use XTS mode.
1036 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1037 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1038 performs algorithm blocking for unapproved PRNG types. Also do not
1039 set PRNG type in FIPS_mode_set(): leave this to the application.
1040 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1041 the standard OpenSSL PRNG: set additional data to a date time vector.
1044 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1045 This shouldn't present any incompatibility problems because applications
1046 shouldn't be using these directly and any that are will need to rethink
1047 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1050 *) Extensive self tests and health checking required by SP800-90 DRBG.
1051 Remove strength parameter from FIPS_drbg_instantiate and always
1052 instantiate at maximum supported strength.
1055 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1058 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1061 *) New function DH_compute_key_padded() to compute a DH key and pad with
1062 leading zeroes if needed: this complies with SP800-56A et al.
1065 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1066 anything, incomplete, subject to change and largely untested at present.
1069 *) Modify fipscanisteronly build option to only build the necessary object
1070 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1073 *) Add experimental option FIPSSYMS to give all symbols in
1074 fipscanister.o and FIPS or fips prefix. This will avoid
1075 conflicts with future versions of OpenSSL. Add perl script
1076 util/fipsas.pl to preprocess assembly language source files
1077 and rename any affected symbols.
1080 *) Add selftest checks and algorithm block of non-fips algorithms in
1081 FIPS mode. Remove DES2 from selftests.
1084 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1085 return internal method without any ENGINE dependencies. Add new
1086 tiny fips sign and verify functions.
1089 *) New build option no-ec2m to disable characteristic 2 code.
1092 *) New build option "fipscanisteronly". This only builds fipscanister.o
1093 and (currently) associated fips utilities. Uses the file Makefile.fips
1094 instead of Makefile.org as the prototype.
1097 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1098 Update fips_gcmtest to use IV generator.
1101 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1102 setting output buffer to NULL. The *Final function must be
1103 called although it will not retrieve any additional data. The tag
1104 can be set or retrieved with a ctrl. The IV length is by default 12
1105 bytes (96 bits) but can be set to an alternative value. If the IV
1106 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1110 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1111 underlying do_cipher function handles all cipher semantics itself
1112 including padding and finalisation. This is useful if (for example)
1113 an ENGINE cipher handles block padding itself. The behaviour of
1114 do_cipher is subtly changed if this flag is set: the return value
1115 is the number of characters written to the output buffer (zero is
1116 no longer an error code) or a negative error code. Also if the
1117 input buffer is NULL and length 0 finalisation should be performed.
1120 *) If a candidate issuer certificate is already part of the constructed
1121 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1124 *) Improve forward-security support: add functions
1126 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1127 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1129 for use by SSL/TLS servers; the callback function will be called whenever a
1130 new session is created, and gets to decide whether the session may be
1131 cached to make it resumable (return 0) or not (return 1). (As by the
1132 SSL/TLS protocol specifications, the session_id sent by the server will be
1133 empty to indicate that the session is not resumable; also, the server will
1134 not generate RFC 4507 (RFC 5077) session tickets.)
1136 A simple reasonable callback implementation is to return is_forward_secure.
1137 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1138 by the SSL/TLS server library, indicating whether it can provide forward
1140 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1142 *) New -verify_name option in command line utilities to set verification
1146 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1147 Add CMAC pkey methods.
1150 *) Experimental renegotiation in s_server -www mode. If the client
1151 browses /reneg connection is renegotiated. If /renegcert it is
1152 renegotiated requesting a certificate.
1155 *) Add an "external" session cache for debugging purposes to s_server. This
1156 should help trace issues which normally are only apparent in deployed
1157 multi-process servers.
1160 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1161 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1162 BIO_set_cipher() and some obscure PEM functions were changed so they
1163 can now return an error. The RAND changes required a change to the
1164 RAND_METHOD structure.
1167 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1168 a gcc attribute to warn if the result of a function is ignored. This
1169 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1170 whose return value is often ignored.
1173 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1174 These allow SCTs (signed certificate timestamps) to be requested and
1175 validated when establishing a connection.
1176 [Rob Percival <robpercival@google.com>]
1178 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1180 *) Prevent padding oracle in AES-NI CBC MAC check
1182 A MITM attacker can use a padding oracle attack to decrypt traffic
1183 when the connection uses an AES CBC cipher and the server support
1186 This issue was introduced as part of the fix for Lucky 13 padding
1187 attack (CVE-2013-0169). The padding check was rewritten to be in
1188 constant time by making sure that always the same bytes are read and
1189 compared against either the MAC or padding bytes. But it no longer
1190 checked that there was enough data to have both the MAC and padding
1193 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1197 *) Fix EVP_EncodeUpdate overflow
1199 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1200 Base64 encoding of binary data. If an attacker is able to supply very large
1201 amounts of input data then a length check can overflow resulting in a heap
1204 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1205 the PEM_write_bio* family of functions. These are mainly used within the
1206 OpenSSL command line applications, so any application which processes data
1207 from an untrusted source and outputs it as a PEM file should be considered
1208 vulnerable to this issue. User applications that call these APIs directly
1209 with large amounts of untrusted data may also be vulnerable.
1211 This issue was reported by Guido Vranken.
1215 *) Fix EVP_EncryptUpdate overflow
1217 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1218 is able to supply very large amounts of input data after a previous call to
1219 EVP_EncryptUpdate() with a partial block then a length check can overflow
1220 resulting in a heap corruption. Following an analysis of all OpenSSL
1221 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1222 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1223 the first called function after an EVP_EncryptInit(), and therefore that
1224 specific call must be safe. The second form is where the length passed to
1225 EVP_EncryptUpdate() can be seen from the code to be some small value and
1226 therefore there is no possibility of an overflow. Since all instances are
1227 one of these two forms, it is believed that there can be no overflows in
1228 internal code due to this problem. It should be noted that
1229 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1230 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1231 of these calls have also been analysed too and it is believed there are no
1232 instances in internal usage where an overflow could occur.
1234 This issue was reported by Guido Vranken.
1238 *) Prevent ASN.1 BIO excessive memory allocation
1240 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1241 a short invalid encoding can cause allocation of large amounts of memory
1242 potentially consuming excessive resources or exhausting memory.
1244 Any application parsing untrusted data through d2i BIO functions is
1245 affected. The memory based functions such as d2i_X509() are *not* affected.
1246 Since the memory based functions are used by the TLS library, TLS
1247 applications are not affected.
1249 This issue was reported by Brian Carpenter.
1255 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1256 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1257 in arbitrary stack data being returned in the buffer.
1259 This issue was reported by Guido Vranken.
1263 *) Modify behavior of ALPN to invoke callback after SNI/servername
1264 callback, such that updates to the SSL_CTX affect ALPN.
1267 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1271 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1272 methods are enabled and ssl2 is disabled the methods return NULL.
1275 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1277 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1278 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1279 provide any "EXPORT" or "LOW" strength ciphers.
1282 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1283 is by default disabled at build-time. Builds that are not configured with
1284 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1285 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1286 will need to explicitly call either of:
1288 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1290 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1292 as appropriate. Even if either of those is used, or the application
1293 explicitly uses the version-specific SSLv2_method() or its client and
1294 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1295 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1296 ciphers, and SSLv2 56-bit DES are no longer available.
1300 *) Fix a double-free in DSA code
1302 A double free bug was discovered when OpenSSL parses malformed DSA private
1303 keys and could lead to a DoS attack or memory corruption for applications
1304 that receive DSA private keys from untrusted sources. This scenario is
1307 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1312 *) Disable SRP fake user seed to address a server memory leak.
1314 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1316 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1317 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1318 was changed to ignore the "fake user" SRP seed, even if the seed
1321 Users should use SRP_VBASE_get1_by_user instead. Note that in
1322 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1323 also that even though configuring the SRP seed attempts to hide
1324 invalid usernames by continuing the handshake with fake
1325 credentials, this behaviour is not constant time and no strong
1326 guarantees are made that the handshake is indistinguishable from
1327 that of a valid user.
1331 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1333 In the BN_hex2bn function the number of hex digits is calculated using an
1334 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1335 large values of |i| this can result in |bn_expand| not allocating any
1336 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1337 field as NULL leading to a subsequent NULL ptr deref. For very large values
1338 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1339 In this case memory is allocated to the internal BIGNUM data field, but it
1340 is insufficiently sized leading to heap corruption. A similar issue exists
1341 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1342 is ever called by user applications with very large untrusted hex/dec data.
1343 This is anticipated to be a rare occurrence.
1345 All OpenSSL internal usage of these functions use data that is not expected
1346 to be untrusted, e.g. config file data or application command line
1347 arguments. If user developed applications generate config file data based
1348 on untrusted data then it is possible that this could also lead to security
1349 consequences. This is also anticipated to be rare.
1351 This issue was reported to OpenSSL by Guido Vranken.
1355 *) Fix memory issues in BIO_*printf functions
1357 The internal |fmtstr| function used in processing a "%s" format string in
1358 the BIO_*printf functions could overflow while calculating the length of a
1359 string and cause an OOB read when printing very long strings.
1361 Additionally the internal |doapr_outch| function can attempt to write to an
1362 OOB memory location (at an offset from the NULL pointer) in the event of a
1363 memory allocation failure. In 1.0.2 and below this could be caused where
1364 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1365 could be in processing a very long "%s" format string. Memory leaks can
1368 The first issue may mask the second issue dependent on compiler behaviour.
1369 These problems could enable attacks where large amounts of untrusted data
1370 is passed to the BIO_*printf functions. If applications use these functions
1371 in this way then they could be vulnerable. OpenSSL itself uses these
1372 functions when printing out human-readable dumps of ASN.1 data. Therefore
1373 applications that print this data could be vulnerable if the data is from
1374 untrusted sources. OpenSSL command line applications could also be
1375 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1376 as command line arguments.
1378 Libssl is not considered directly vulnerable. Additionally certificates etc
1379 received via remote connections via libssl are also unlikely to be able to
1380 trigger these issues because of message size limits enforced within libssl.
1382 This issue was reported to OpenSSL Guido Vranken.
1386 *) Side channel attack on modular exponentiation
1388 A side-channel attack was found which makes use of cache-bank conflicts on
1389 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1390 of RSA keys. The ability to exploit this issue is limited as it relies on
1391 an attacker who has control of code in a thread running on the same
1392 hyper-threaded core as the victim thread which is performing decryptions.
1394 This issue was reported to OpenSSL by Yuval Yarom, The University of
1395 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1396 Nadia Heninger, University of Pennsylvania with more information at
1397 http://cachebleed.info.
1401 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1402 if no keysize is specified with default_bits. This fixes an
1403 omission in an earlier change that changed all RSA/DSA key generation
1404 apps to use 2048 bits by default.
1407 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1408 *) DH small subgroups
1410 Historically OpenSSL only ever generated DH parameters based on "safe"
1411 primes. More recently (in version 1.0.2) support was provided for
1412 generating X9.42 style parameter files such as those required for RFC 5114
1413 support. The primes used in such files may not be "safe". Where an
1414 application is using DH configured with parameters based on primes that are
1415 not "safe" then an attacker could use this fact to find a peer's private
1416 DH exponent. This attack requires that the attacker complete multiple
1417 handshakes in which the peer uses the same private DH exponent. For example
1418 this could be used to discover a TLS server's private DH exponent if it's
1419 reusing the private DH exponent or it's using a static DH ciphersuite.
1421 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1422 TLS. It is not on by default. If the option is not set then the server
1423 reuses the same private DH exponent for the life of the server process and
1424 would be vulnerable to this attack. It is believed that many popular
1425 applications do set this option and would therefore not be at risk.
1427 The fix for this issue adds an additional check where a "q" parameter is
1428 available (as is the case in X9.42 based parameters). This detects the
1429 only known attack, and is the only possible defense for static DH
1430 ciphersuites. This could have some performance impact.
1432 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1433 default and cannot be disabled. This could have some performance impact.
1435 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1439 *) SSLv2 doesn't block disabled ciphers
1441 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1442 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1443 been disabled, provided that the SSLv2 protocol was not also disabled via
1446 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1447 and Sebastian Schinzel.
1451 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1453 *) BN_mod_exp may produce incorrect results on x86_64
1455 There is a carry propagating bug in the x86_64 Montgomery squaring
1456 procedure. No EC algorithms are affected. Analysis suggests that attacks
1457 against RSA and DSA as a result of this defect would be very difficult to
1458 perform and are not believed likely. Attacks against DH are considered just
1459 feasible (although very difficult) because most of the work necessary to
1460 deduce information about a private key may be performed offline. The amount
1461 of resources required for such an attack would be very significant and
1462 likely only accessible to a limited number of attackers. An attacker would
1463 additionally need online access to an unpatched system using the target
1464 private key in a scenario with persistent DH parameters and a private
1465 key that is shared between multiple clients. For example this can occur by
1466 default in OpenSSL DHE based SSL/TLS ciphersuites.
1468 This issue was reported to OpenSSL by Hanno Böck.
1472 *) Certificate verify crash with missing PSS parameter
1474 The signature verification routines will crash with a NULL pointer
1475 dereference if presented with an ASN.1 signature using the RSA PSS
1476 algorithm and absent mask generation function parameter. Since these
1477 routines are used to verify certificate signature algorithms this can be
1478 used to crash any certificate verification operation and exploited in a
1479 DoS attack. Any application which performs certificate verification is
1480 vulnerable including OpenSSL clients and servers which enable client
1483 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1487 *) X509_ATTRIBUTE memory leak
1489 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1490 memory. This structure is used by the PKCS#7 and CMS routines so any
1491 application which reads PKCS#7 or CMS data from untrusted sources is
1492 affected. SSL/TLS is not affected.
1494 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1499 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1500 This changes the decoding behaviour for some invalid messages,
1501 though the change is mostly in the more lenient direction, and
1502 legacy behaviour is preserved as much as possible.
1505 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1507 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1509 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1511 *) Alternate chains certificate forgery
1513 During certificate verification, OpenSSL will attempt to find an
1514 alternative certificate chain if the first attempt to build such a chain
1515 fails. An error in the implementation of this logic can mean that an
1516 attacker could cause certain checks on untrusted certificates to be
1517 bypassed, such as the CA flag, enabling them to use a valid leaf
1518 certificate to act as a CA and "issue" an invalid certificate.
1520 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1524 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1526 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1527 incompatibility in the handling of HMAC. The previous ABI has now been
1531 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1533 *) Malformed ECParameters causes infinite loop
1535 When processing an ECParameters structure OpenSSL enters an infinite loop
1536 if the curve specified is over a specially malformed binary polynomial
1539 This can be used to perform denial of service against any
1540 system which processes public keys, certificate requests or
1541 certificates. This includes TLS clients and TLS servers with
1542 client authentication enabled.
1544 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1548 *) Exploitable out-of-bounds read in X509_cmp_time
1550 X509_cmp_time does not properly check the length of the ASN1_TIME
1551 string and can read a few bytes out of bounds. In addition,
1552 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1555 An attacker can use this to craft malformed certificates and CRLs of
1556 various sizes and potentially cause a segmentation fault, resulting in
1557 a DoS on applications that verify certificates or CRLs. TLS clients
1558 that verify CRLs are affected. TLS clients and servers with client
1559 authentication enabled may be affected if they use custom verification
1562 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1563 independently by Hanno Böck.
1567 *) PKCS7 crash with missing EnvelopedContent
1569 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1570 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1571 with missing content and trigger a NULL pointer dereference on parsing.
1573 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1574 structures from untrusted sources are affected. OpenSSL clients and
1575 servers are not affected.
1577 This issue was reported to OpenSSL by Michal Zalewski (Google).
1581 *) CMS verify infinite loop with unknown hash function
1583 When verifying a signedData message the CMS code can enter an infinite loop
1584 if presented with an unknown hash function OID. This can be used to perform
1585 denial of service against any system which verifies signedData messages using
1587 This issue was reported to OpenSSL by Johannes Bauer.
1591 *) Race condition handling NewSessionTicket
1593 If a NewSessionTicket is received by a multi-threaded client when attempting to
1594 reuse a previous ticket then a race condition can occur potentially leading to
1595 a double free of the ticket data.
1599 *) Only support 256-bit or stronger elliptic curves with the
1600 'ecdh_auto' setting (server) or by default (client). Of supported
1601 curves, prefer P-256 (both).
1604 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1606 *) ClientHello sigalgs DoS fix
1608 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1609 invalid signature algorithms extension a NULL pointer dereference will
1610 occur. This can be exploited in a DoS attack against the server.
1612 This issue was was reported to OpenSSL by David Ramos of Stanford
1615 [Stephen Henson and Matt Caswell]
1617 *) Multiblock corrupted pointer fix
1619 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1620 feature only applies on 64 bit x86 architecture platforms that support AES
1621 NI instructions. A defect in the implementation of "multiblock" can cause
1622 OpenSSL's internal write buffer to become incorrectly set to NULL when
1623 using non-blocking IO. Typically, when the user application is using a
1624 socket BIO for writing, this will only result in a failed connection.
1625 However if some other BIO is used then it is likely that a segmentation
1626 fault will be triggered, thus enabling a potential DoS attack.
1628 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1632 *) Segmentation fault in DTLSv1_listen fix
1634 The DTLSv1_listen function is intended to be stateless and processes the
1635 initial ClientHello from many peers. It is common for user code to loop
1636 over the call to DTLSv1_listen until a valid ClientHello is received with
1637 an associated cookie. A defect in the implementation of DTLSv1_listen means
1638 that state is preserved in the SSL object from one invocation to the next
1639 that can lead to a segmentation fault. Errors processing the initial
1640 ClientHello can trigger this scenario. An example of such an error could be
1641 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1644 This issue was reported to OpenSSL by Per Allansson.
1648 *) Segmentation fault in ASN1_TYPE_cmp fix
1650 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1651 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1652 certificate signature algorithm consistency this can be used to crash any
1653 certificate verification operation and exploited in a DoS attack. Any
1654 application which performs certificate verification is vulnerable including
1655 OpenSSL clients and servers which enable client authentication.
1659 *) Segmentation fault for invalid PSS parameters fix
1661 The signature verification routines will crash with a NULL pointer
1662 dereference if presented with an ASN.1 signature using the RSA PSS
1663 algorithm and invalid parameters. Since these routines are used to verify
1664 certificate signature algorithms this can be used to crash any
1665 certificate verification operation and exploited in a DoS attack. Any
1666 application which performs certificate verification is vulnerable including
1667 OpenSSL clients and servers which enable client authentication.
1669 This issue was was reported to OpenSSL by Brian Carpenter.
1673 *) ASN.1 structure reuse memory corruption fix
1675 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1676 memory corruption via an invalid write. Such reuse is and has been
1677 strongly discouraged and is believed to be rare.
1679 Applications that parse structures containing CHOICE or ANY DEFINED BY
1680 components may be affected. Certificate parsing (d2i_X509 and related
1681 functions) are however not affected. OpenSSL clients and servers are
1686 *) PKCS7 NULL pointer dereferences fix
1688 The PKCS#7 parsing code does not handle missing outer ContentInfo
1689 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1690 missing content and trigger a NULL pointer dereference on parsing.
1692 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1693 otherwise parse PKCS#7 structures from untrusted sources are
1694 affected. OpenSSL clients and servers are not affected.
1696 This issue was reported to OpenSSL by Michal Zalewski (Google).
1700 *) DoS via reachable assert in SSLv2 servers fix
1702 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1703 servers that both support SSLv2 and enable export cipher suites by sending
1704 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1706 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1707 (OpenSSL development team).
1711 *) Empty CKE with client auth and DHE fix
1713 If client auth is used then a server can seg fault in the event of a DHE
1714 ciphersuite being selected and a zero length ClientKeyExchange message
1715 being sent by the client. This could be exploited in a DoS attack.
1719 *) Handshake with unseeded PRNG fix
1721 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1722 with an unseeded PRNG. The conditions are:
1723 - The client is on a platform where the PRNG has not been seeded
1724 automatically, and the user has not seeded manually
1725 - A protocol specific client method version has been used (i.e. not
1726 SSL_client_methodv23)
1727 - A ciphersuite is used that does not require additional random data from
1728 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1730 If the handshake succeeds then the client random that has been used will
1731 have been generated from a PRNG with insufficient entropy and therefore the
1732 output may be predictable.
1734 For example using the following command with an unseeded openssl will
1735 succeed on an unpatched platform:
1737 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1741 *) Use After Free following d2i_ECPrivatekey error fix
1743 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1744 could cause a use after free condition. This, in turn, could cause a double
1745 free in several private key parsing functions (such as d2i_PrivateKey
1746 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1747 for applications that receive EC private keys from untrusted
1748 sources. This scenario is considered rare.
1750 This issue was discovered by the BoringSSL project and fixed in their
1755 *) X509_to_X509_REQ NULL pointer deref fix
1757 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1758 the certificate key is invalid. This function is rarely used in practice.
1760 This issue was discovered by Brian Carpenter.
1764 *) Removed the export ciphers from the DEFAULT ciphers
1767 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1769 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1770 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1771 So far those who have to target multiple platforms would compromise
1772 and argue that binary targeting say ARMv5 would still execute on
1773 ARMv8. "Universal" build resolves this compromise by providing
1774 near-optimal performance even on newer platforms.
1777 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1778 (other platforms pending).
1779 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1781 *) Add support for the SignedCertificateTimestampList certificate and
1782 OCSP response extensions from RFC6962.
1785 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1786 for corner cases. (Certain input points at infinity could lead to
1787 bogus results, with non-infinity inputs mapped to infinity too.)
1790 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1791 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1792 common cases are optimized and there still is room for further
1793 improvements. Vector Permutation AES for Altivec is also added.
1796 *) Add support for little-endian ppc64 Linux target.
1797 [Marcelo Cerri (IBM)]
1799 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1800 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1801 are optimized and there still is room for further improvements.
1802 Both 32- and 64-bit modes are supported.
1803 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1805 *) Improved ARMv7 NEON support.
1808 *) Support for SPARC Architecture 2011 crypto extensions, first
1809 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1810 SHA256/512, MD5, GHASH and modular exponentiation.
1811 [Andy Polyakov, David Miller]
1813 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1815 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1817 *) Support for new and upcoming Intel processors, including AVX2,
1818 BMI and SHA ISA extensions. This includes additional "stitched"
1819 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1822 This work was sponsored by Intel Corp.
1825 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1826 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1827 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1830 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1831 this fixes a limitation in previous versions of OpenSSL.
1834 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1835 MGF1 digest and OAEP label.
1838 *) Add EVP support for key wrapping algorithms, to avoid problems with
1839 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1840 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1841 algorithms and include tests cases.
1844 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1846 [Douglas E. Engert, Steve Henson]
1848 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1849 difference in days and seconds between two tm or ASN1_TIME structures.
1852 *) Add -rev test option to s_server to just reverse order of characters
1853 received by client and send back to server. Also prints an abbreviated
1854 summary of the connection parameters.
1857 *) New option -brief for s_client and s_server to print out a brief summary
1858 of connection parameters.
1861 *) Add callbacks for arbitrary TLS extensions.
1862 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1864 *) New option -crl_download in several openssl utilities to download CRLs
1865 from CRLDP extension in certificates.
1868 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1871 *) New function X509_CRL_diff to generate a delta CRL from the difference
1872 of two full CRLs. Add support to "crl" utility.
1875 *) New functions to set lookup_crls function and to retrieve
1876 X509_STORE from X509_STORE_CTX.
1879 *) Print out deprecated issuer and subject unique ID fields in
1883 *) Extend OCSP I/O functions so they can be used for simple general purpose
1884 HTTP as well as OCSP. New wrapper function which can be used to download
1885 CRLs using the OCSP API.
1888 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1891 *) SSL_CONF* functions. These provide a common framework for application
1892 configuration using configuration files or command lines.
1895 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1896 message callback and prints the results. Needs compile time option
1897 "enable-ssl-trace". New options to s_client and s_server to enable
1901 *) New ctrl and macro to retrieve supported points extensions.
1902 Print out extension in s_server and s_client.
1905 *) New functions to retrieve certificate signature and signature
1909 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1913 *) New Suite B modes for TLS code. These use and enforce the requirements
1914 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1915 only use Suite B curves. The Suite B modes can be set by using the
1916 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1919 *) New chain verification flags for Suite B levels of security. Check
1920 algorithms are acceptable when flags are set in X509_verify_cert.
1923 *) Make tls1_check_chain return a set of flags indicating checks passed
1924 by a certificate chain. Add additional tests to handle client
1925 certificates: checks for matching certificate type and issuer name
1929 *) If an attempt is made to use a signature algorithm not in the peer
1930 preference list abort the handshake. If client has no suitable
1931 signature algorithms in response to a certificate request do not
1932 use the certificate.
1935 *) If server EC tmp key is not in client preference list abort handshake.
1938 *) Add support for certificate stores in CERT structure. This makes it
1939 possible to have different stores per SSL structure or one store in
1940 the parent SSL_CTX. Include distinct stores for certificate chain
1941 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1942 to build and store a certificate chain in CERT structure: returning
1943 an error if the chain cannot be built: this will allow applications
1944 to test if a chain is correctly configured.
1946 Note: if the CERT based stores are not set then the parent SSL_CTX
1947 store is used to retain compatibility with existing behaviour.
1951 *) New function ssl_set_client_disabled to set a ciphersuite disabled
1952 mask based on the current session, check mask when sending client
1953 hello and checking the requested ciphersuite.
1956 *) New ctrls to retrieve and set certificate types in a certificate
1957 request message. Print out received values in s_client. If certificate
1958 types is not set with custom values set sensible values based on
1959 supported signature algorithms.
1962 *) Support for distinct client and server supported signature algorithms.
1965 *) Add certificate callback. If set this is called whenever a certificate
1966 is required by client or server. An application can decide which
1967 certificate chain to present based on arbitrary criteria: for example
1968 supported signature algorithms. Add very simple example to s_server.
1969 This fixes many of the problems and restrictions of the existing client
1970 certificate callback: for example you can now clear an existing
1971 certificate and specify the whole chain.
1974 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
1975 the certificate can be used for (if anything). Set valid_flags field
1976 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
1977 to have similar checks in it.
1979 Add new "cert_flags" field to CERT structure and include a "strict mode".
1980 This enforces some TLS certificate requirements (such as only permitting
1981 certificate signature algorithms contained in the supported algorithms
1982 extension) which some implementations ignore: this option should be used
1983 with caution as it could cause interoperability issues.
1986 *) Update and tidy signature algorithm extension processing. Work out
1987 shared signature algorithms based on preferences and peer algorithms
1988 and print them out in s_client and s_server. Abort handshake if no
1989 shared signature algorithms.
1992 *) Add new functions to allow customised supported signature algorithms
1993 for SSL and SSL_CTX structures. Add options to s_client and s_server
1997 *) New function SSL_certs_clear() to delete all references to certificates
1998 from an SSL structure. Before this once a certificate had been added
1999 it couldn't be removed.
2002 *) Integrate hostname, email address and IP address checking with certificate
2003 verification. New verify options supporting checking in openssl utility.
2006 *) Fixes and wildcard matching support to hostname and email checking
2007 functions. Add manual page.
2008 [Florian Weimer (Red Hat Product Security Team)]
2010 *) New functions to check a hostname email or IP address against a
2011 certificate. Add options x509 utility to print results of checks against
2015 *) Fix OCSP checking.
2016 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2018 *) Initial experimental support for explicitly trusted non-root CAs.
2019 OpenSSL still tries to build a complete chain to a root but if an
2020 intermediate CA has a trust setting included that is used. The first
2021 setting is used: whether to trust (e.g., -addtrust option to the x509
2025 *) Add -trusted_first option which attempts to find certificates in the
2026 trusted store even if an untrusted chain is also supplied.
2029 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2030 platform support for Linux and Android.
2033 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2036 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2037 When in FIPS mode the approved implementations are used as normal,
2038 when not in FIPS mode the internal unapproved versions are used instead.
2039 This means that the FIPS capable OpenSSL isn't forced to use the
2040 (often lower performance) FIPS implementations outside FIPS mode.
2043 *) Transparently support X9.42 DH parameters when calling
2044 PEM_read_bio_DHparameters. This means existing applications can handle
2045 the new parameter format automatically.
2048 *) Initial experimental support for X9.42 DH parameter format: mainly
2049 to support use of 'q' parameter for RFC5114 parameters.
2052 *) Add DH parameters from RFC5114 including test data to dhtest.
2055 *) Support for automatic EC temporary key parameter selection. If enabled
2056 the most preferred EC parameters are automatically used instead of
2057 hardcoded fixed parameters. Now a server just has to call:
2058 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2059 support ECDH and use the most appropriate parameters.
2062 *) Enhance and tidy EC curve and point format TLS extension code. Use
2063 static structures instead of allocation if default values are used.
2064 New ctrls to set curves we wish to support and to retrieve shared curves.
2065 Print out shared curves in s_server. New options to s_server and s_client
2066 to set list of supported curves.
2069 *) New ctrls to retrieve supported signature algorithms and
2070 supported curve values as an array of NIDs. Extend openssl utility
2071 to print out received values.
2074 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2075 between NIDs and the more common NIST names such as "P-256". Enhance
2076 ecparam utility and ECC method to recognise the NIST names for curves.
2079 *) Enhance SSL/TLS certificate chain handling to support different
2080 chains for each certificate instead of one chain in the parent SSL_CTX.
2083 *) Support for fixed DH ciphersuite client authentication: where both
2084 server and client use DH certificates with common parameters.
2087 *) Support for fixed DH ciphersuites: those requiring DH server
2091 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2093 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2094 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2095 X509_CINF_get_signature were reverted post internal team review.
2097 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2099 *) Build fixes for the Windows and OpenVMS platforms
2100 [Matt Caswell and Richard Levitte]
2102 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2104 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2105 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2106 dereference. This could lead to a Denial Of Service attack. Thanks to
2107 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2111 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2112 dtls1_buffer_record function under certain conditions. In particular this
2113 could occur if an attacker sent repeated DTLS records with the same
2114 sequence number but for the next epoch. The memory leak could be exploited
2115 by an attacker in a Denial of Service attack through memory exhaustion.
2116 Thanks to Chris Mueller for reporting this issue.
2120 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2121 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2122 method would be set to NULL which could later result in a NULL pointer
2123 dereference. Thanks to Frank Schmirler for reporting this issue.
2127 *) Abort handshake if server key exchange message is omitted for ephemeral
2130 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2131 reporting this issue.
2135 *) Remove non-export ephemeral RSA code on client and server. This code
2136 violated the TLS standard by allowing the use of temporary RSA keys in
2137 non-export ciphersuites and could be used by a server to effectively
2138 downgrade the RSA key length used to a value smaller than the server
2139 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2140 INRIA or reporting this issue.
2144 *) Fixed issue where DH client certificates are accepted without verification.
2145 An OpenSSL server will accept a DH certificate for client authentication
2146 without the certificate verify message. This effectively allows a client to
2147 authenticate without the use of a private key. This only affects servers
2148 which trust a client certificate authority which issues certificates
2149 containing DH keys: these are extremely rare and hardly ever encountered.
2150 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2155 *) Ensure that the session ID context of an SSL is updated when its
2156 SSL_CTX is updated via SSL_set_SSL_CTX.
2158 The session ID context is typically set from the parent SSL_CTX,
2159 and can vary with the CTX.
2162 *) Fix various certificate fingerprint issues.
2164 By using non-DER or invalid encodings outside the signed portion of a
2165 certificate the fingerprint can be changed without breaking the signature.
2166 Although no details of the signed portion of the certificate can be changed
2167 this can cause problems with some applications: e.g. those using the
2168 certificate fingerprint for blacklists.
2170 1. Reject signatures with non zero unused bits.
2172 If the BIT STRING containing the signature has non zero unused bits reject
2173 the signature. All current signature algorithms require zero unused bits.
2175 2. Check certificate algorithm consistency.
2177 Check the AlgorithmIdentifier inside TBS matches the one in the
2178 certificate signature. NB: this will result in signature failure
2179 errors for some broken certificates.
2181 Thanks to Konrad Kraszewski from Google for reporting this issue.
2183 3. Check DSA/ECDSA signatures use DER.
2185 Re-encode DSA/ECDSA signatures and compare with the original received
2186 signature. Return an error if there is a mismatch.
2188 This will reject various cases including garbage after signature
2189 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2190 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2191 (negative or with leading zeroes).
2193 Further analysis was conducted and fixes were developed by Stephen Henson
2194 of the OpenSSL core team.
2199 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2200 results on some platforms, including x86_64. This bug occurs at random
2201 with a very low probability, and is not known to be exploitable in any
2202 way, though its exact impact is difficult to determine. Thanks to Pieter
2203 Wuille (Blockstream) who reported this issue and also suggested an initial
2204 fix. Further analysis was conducted by the OpenSSL development team and
2205 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2206 the OpenSSL core team.
2210 *) Do not resume sessions on the server if the negotiated protocol
2211 version does not match the session's version. Resuming with a different
2212 version, while not strictly forbidden by the RFC, is of questionable
2213 sanity and breaks all known clients.
2214 [David Benjamin, Emilia Käsper]
2216 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2217 early CCS messages during renegotiation. (Note that because
2218 renegotiation is encrypted, this early CCS was not exploitable.)
2221 *) Tighten client-side session ticket handling during renegotiation:
2222 ensure that the client only accepts a session ticket if the server sends
2223 the extension anew in the ServerHello. Previously, a TLS client would
2224 reuse the old extension state and thus accept a session ticket if one was
2225 announced in the initial ServerHello.
2227 Similarly, ensure that the client requires a session ticket if one
2228 was advertised in the ServerHello. Previously, a TLS client would
2229 ignore a missing NewSessionTicket message.
2232 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2234 *) SRTP Memory Leak.
2236 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2237 sends a carefully crafted handshake message, to cause OpenSSL to fail
2238 to free up to 64k of memory causing a memory leak. This could be
2239 exploited in a Denial Of Service attack. This issue affects OpenSSL
2240 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2241 whether SRTP is used or configured. Implementations of OpenSSL that
2242 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2244 The fix was developed by the OpenSSL team.
2248 *) Session Ticket Memory Leak.
2250 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2251 integrity of that ticket is first verified. In the event of a session
2252 ticket integrity check failing, OpenSSL will fail to free memory
2253 causing a memory leak. By sending a large number of invalid session
2254 tickets an attacker could exploit this issue in a Denial Of Service
2259 *) Build option no-ssl3 is incomplete.
2261 When OpenSSL is configured with "no-ssl3" as a build option, servers
2262 could accept and complete a SSL 3.0 handshake, and clients could be
2263 configured to send them.
2265 [Akamai and the OpenSSL team]
2267 *) Add support for TLS_FALLBACK_SCSV.
2268 Client applications doing fallback retries should call
2269 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2271 [Adam Langley, Bodo Moeller]
2273 *) Add additional DigestInfo checks.
2275 Re-encode DigestInto in DER and check against the original when
2276 verifying RSA signature: this will reject any improperly encoded
2277 DigestInfo structures.
2279 Note: this is a precautionary measure and no attacks are currently known.
2283 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2285 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2286 SRP code can be overrun an internal buffer. Add sanity check that
2287 g, A, B < N to SRP code.
2289 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2290 Group for discovering this issue.
2294 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2295 TLS 1.0 instead of higher protocol versions when the ClientHello message
2296 is badly fragmented. This allows a man-in-the-middle attacker to force a
2297 downgrade to TLS 1.0 even if both the server and the client support a
2298 higher protocol version, by modifying the client's TLS records.
2300 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2301 researching this issue.
2305 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2306 to a denial of service attack. A malicious server can crash the client
2307 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2308 ciphersuite and sending carefully crafted handshake messages.
2310 Thanks to Felix Gröbert (Google) for discovering and researching this
2315 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2316 to leak memory. This can be exploited through a Denial of Service attack.
2317 Thanks to Adam Langley for discovering and researching this issue.
2321 *) An attacker can force openssl to consume large amounts of memory whilst
2322 processing DTLS handshake messages. This can be exploited through a
2323 Denial of Service attack.
2324 Thanks to Adam Langley for discovering and researching this issue.
2328 *) An attacker can force an error condition which causes openssl to crash
2329 whilst processing DTLS packets due to memory being freed twice. This
2330 can be exploited through a Denial of Service attack.
2331 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2336 *) If a multithreaded client connects to a malicious server using a resumed
2337 session and the server sends an ec point format extension it could write
2338 up to 255 bytes to freed memory.
2340 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2345 *) A malicious server can crash an OpenSSL client with a null pointer
2346 dereference (read) by specifying an SRP ciphersuite even though it was not
2347 properly negotiated with the client. This can be exploited through a
2348 Denial of Service attack.
2350 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2351 discovering and researching this issue.
2355 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2356 X509_name_oneline, X509_name_print_ex et al. to leak some information
2357 from the stack. Applications may be affected if they echo pretty printing
2358 output to the attacker.
2360 Thanks to Ivan Fratric (Google) for discovering this issue.
2362 [Emilia Käsper, and Steve Henson]
2364 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2365 for corner cases. (Certain input points at infinity could lead to
2366 bogus results, with non-infinity inputs mapped to infinity too.)
2369 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2371 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2372 handshake can force the use of weak keying material in OpenSSL
2373 SSL/TLS clients and servers.
2375 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2376 researching this issue. (CVE-2014-0224)
2377 [KIKUCHI Masashi, Steve Henson]
2379 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2380 OpenSSL DTLS client the code can be made to recurse eventually crashing
2383 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2385 [Imre Rad, Steve Henson]
2387 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2388 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2389 client or server. This is potentially exploitable to run arbitrary
2390 code on a vulnerable client or server.
2392 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2393 [Jüri Aedla, Steve Henson]
2395 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2396 are subject to a denial of service attack.
2398 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2399 this issue. (CVE-2014-3470)
2400 [Felix Gröbert, Ivan Fratric, Steve Henson]
2402 *) Harmonize version and its documentation. -f flag is used to display
2404 [mancha <mancha1@zoho.com>]
2406 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2407 in i2d_ECPrivateKey.
2408 [mancha <mancha1@zoho.com>]
2410 *) Fix some double frees. These are not thought to be exploitable.
2411 [mancha <mancha1@zoho.com>]
2413 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2415 *) A missing bounds check in the handling of the TLS heartbeat extension
2416 can be used to reveal up to 64k of memory to a connected client or
2419 Thanks for Neel Mehta of Google Security for discovering this bug and to
2420 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2421 preparing the fix (CVE-2014-0160)
2422 [Adam Langley, Bodo Moeller]
2424 *) Fix for the attack described in the paper "Recovering OpenSSL
2425 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2426 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2427 http://eprint.iacr.org/2014/140
2429 Thanks to Yuval Yarom and Naomi Benger for discovering this
2430 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2431 [Yuval Yarom and Naomi Benger]
2433 *) TLS pad extension: draft-agl-tls-padding-03
2435 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2436 TLS client Hello record length value would otherwise be > 255 and
2437 less that 512 pad with a dummy extension containing zeroes so it
2438 is at least 512 bytes long.
2440 [Adam Langley, Steve Henson]
2442 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2444 *) Fix for TLS record tampering bug. A carefully crafted invalid
2445 handshake could crash OpenSSL with a NULL pointer exception.
2446 Thanks to Anton Johansson for reporting this issues.
2449 *) Keep original DTLS digest and encryption contexts in retransmission
2450 structures so we can use the previous session parameters if they need
2451 to be resent. (CVE-2013-6450)
2454 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2455 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2456 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2457 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2458 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2459 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2460 [Rob Stradling, Adam Langley]
2462 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2464 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2465 supporting platforms or when small records were transferred.
2466 [Andy Polyakov, Steve Henson]
2468 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2470 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2472 This addresses the flaw in CBC record processing discovered by
2473 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2474 at: http://www.isg.rhul.ac.uk/tls/
2476 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2477 Security Group at Royal Holloway, University of London
2478 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2479 Emilia Käsper for the initial patch.
2481 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2483 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2484 ciphersuites which can be exploited in a denial of service attack.
2485 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2486 and detecting this bug and to Wolfgang Ettlinger
2487 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2491 *) Return an error when checking OCSP signatures when key is NULL.
2492 This fixes a DoS attack. (CVE-2013-0166)
2495 *) Make openssl verify return errors.
2496 [Chris Palmer <palmer@google.com> and Ben Laurie]
2498 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2499 the right response is stapled. Also change SSL_get_certificate()
2500 so it returns the certificate actually sent.
2501 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2502 [Rob Stradling <rob.stradling@comodo.com>]
2504 *) Fix possible deadlock when decoding public keys.
2507 *) Don't use TLS 1.0 record version number in initial client hello
2511 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2513 *) Sanity check record length before skipping explicit IV in TLS
2514 1.2, 1.1 and DTLS to fix DoS attack.
2516 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2517 fuzzing as a service testing platform.
2521 *) Initialise tkeylen properly when encrypting CMS messages.
2522 Thanks to Solar Designer of Openwall for reporting this issue.
2525 *) In FIPS mode don't try to use composite ciphers as they are not
2529 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2531 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2532 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2533 mean any application compiled against OpenSSL 1.0.0 headers setting
2534 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2535 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2536 0x10000000L Any application which was previously compiled against
2537 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2538 will need to be recompiled as a result. Letting be results in
2539 inability to disable specifically TLS 1.1 and in client context,
2540 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2543 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2544 disable just protocol X, but all protocols above X *if* there are
2545 protocols *below* X still enabled. In more practical terms it means
2546 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2547 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2548 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2552 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2554 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2555 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2556 in CRYPTO_realloc_clean.
2558 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2559 issue and to Adam Langley <agl@chromium.org> for fixing it.
2561 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2563 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2566 *) Workarounds for some broken servers that "hang" if a client hello
2567 record length exceeds 255 bytes.
2569 1. Do not use record version number > TLS 1.0 in initial client
2570 hello: some (but not all) hanging servers will now work.
2571 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2572 the number of ciphers sent in the client hello. This should be
2573 set to an even number, such as 50, for example by passing:
2574 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2575 Most broken servers should now work.
2576 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2577 TLS 1.2 client support entirely.
2580 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2583 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2585 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2586 STRING form instead of a DigestInfo.
2589 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2590 and the RSA_sign/RSA_verify functions. This was made more apparent when
2591 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2592 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2593 the correct format in RSA_verify so both forms transparently work.
2596 *) Some servers which support TLS 1.0 can choke if we initially indicate
2597 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2598 encrypted premaster secret. As a workaround use the maximum permitted
2599 client version in client hello, this should keep such servers happy
2600 and still work with previous versions of OpenSSL.
2603 *) Add support for TLS/DTLS heartbeats.
2604 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2606 *) Add support for SCTP.
2607 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2609 *) Improved PRNG seeding for VOS.
2610 [Paul Green <Paul.Green@stratus.com>]
2612 *) Extensive assembler packs updates, most notably:
2614 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2615 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2616 - x86_64: bit-sliced AES implementation;
2617 - ARM: NEON support, contemporary platforms optimizations;
2618 - s390x: z196 support;
2619 - *: GHASH and GF(2^m) multiplication implementations;
2623 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2624 (removal of unnecessary code)
2625 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2627 *) Add TLS key material exporter from RFC 5705.
2630 *) Add DTLS-SRTP negotiation from RFC 5764.
2633 *) Add Next Protocol Negotiation,
2634 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2635 disabled with a no-npn flag to config or Configure. Code donated
2637 [Adam Langley <agl@google.com> and Ben Laurie]
2639 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2640 NIST-P256, NIST-P521, with constant-time single point multiplication on
2641 typical inputs. Compiler support for the nonstandard type __uint128_t is
2642 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2643 Code made available under Apache License version 2.0.
2645 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2646 line to include this in your build of OpenSSL, and run "make depend" (or
2647 "make update"). This enables the following EC_METHODs:
2649 EC_GFp_nistp224_method()
2650 EC_GFp_nistp256_method()
2651 EC_GFp_nistp521_method()
2653 EC_GROUP_new_by_curve_name() will automatically use these (while
2654 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2656 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2658 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2659 all platforms. Move ssize_t definition from e_os.h to the public
2660 header file e_os2.h as it now appears in public header file cms.h
2663 *) New -sigopt option to the ca, req and x509 utilities. Additional
2664 signature parameters can be passed using this option and in
2668 *) Add RSA PSS signing function. This will generate and set the
2669 appropriate AlgorithmIdentifiers for PSS based on those in the
2670 corresponding EVP_MD_CTX structure. No application support yet.
2673 *) Support for companion algorithm specific ASN1 signing routines.
2674 New function ASN1_item_sign_ctx() signs a pre-initialised
2675 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2676 the appropriate parameters.
2679 *) Add new algorithm specific ASN1 verification initialisation function
2680 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2681 handling will be the same no matter what EVP_PKEY_METHOD is used.
2682 Add a PSS handler to support verification of PSS signatures: checked
2683 against a number of sample certificates.
2686 *) Add signature printing for PSS. Add PSS OIDs.
2687 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2689 *) Add algorithm specific signature printing. An individual ASN1 method
2690 can now print out signatures instead of the standard hex dump.
2692 More complex signatures (e.g. PSS) can print out more meaningful
2693 information. Include DSA version that prints out the signature
2697 *) Password based recipient info support for CMS library: implementing
2701 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2702 neatly separates the code into cipher and PBE sections and is required
2703 for some algorithms that split PBES2 into separate pieces (such as
2704 password based CMS).
2707 *) Session-handling fixes:
2708 - Fix handling of connections that are resuming with a session ID,
2709 but also support Session Tickets.
2710 - Fix a bug that suppressed issuing of a new ticket if the client
2711 presented a ticket with an expired session.
2712 - Try to set the ticket lifetime hint to something reasonable.
2713 - Make tickets shorter by excluding irrelevant information.
2714 - On the client side, don't ignore renewed tickets.
2715 [Adam Langley, Bodo Moeller (Google)]
2717 *) Fix PSK session representation.
2720 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2722 This work was sponsored by Intel.
2725 *) Add GCM support to TLS library. Some custom code is needed to split
2726 the IV between the fixed (from PRF) and explicit (from TLS record)
2727 portions. This adds all GCM ciphersuites supported by RFC5288 and
2728 RFC5289. Generalise some AES* cipherstrings to include GCM and
2729 add a special AESGCM string for GCM only.
2732 *) Expand range of ctrls for AES GCM. Permit setting invocation
2733 field on decrypt and retrieval of invocation field only on encrypt.
2736 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2737 As required by RFC5289 these ciphersuites cannot be used if for
2738 versions of TLS earlier than 1.2.
2741 *) For FIPS capable OpenSSL interpret a NULL default public key method
2742 as unset and return the appropriate default but do *not* set the default.
2743 This means we can return the appropriate method in applications that
2744 switch between FIPS and non-FIPS modes.
2747 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2748 ENGINE is used then we cannot handle that in the FIPS module so we
2749 keep original code iff non-FIPS operations are allowed.
2752 *) Add -attime option to openssl utilities.
2753 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2755 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2758 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2759 FIPS EC methods unconditionally for now.
2762 *) New build option no-ec2m to disable characteristic 2 code.
2765 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2766 all cases can be covered as some introduce binary incompatibilities.
2769 *) Redirect RSA operations to FIPS module including keygen,
2770 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2773 *) Add similar low level API blocking to ciphers.
2776 *) Low level digest APIs are not approved in FIPS mode: any attempt
2777 to use these will cause a fatal error. Applications that *really* want
2778 to use them can use the private_* version instead.
2781 *) Redirect cipher operations to FIPS module for FIPS builds.
2784 *) Redirect digest operations to FIPS module for FIPS builds.
2787 *) Update build system to add "fips" flag which will link in fipscanister.o
2788 for static and shared library builds embedding a signature if needed.
2791 *) Output TLS supported curves in preference order instead of numerical
2792 order. This is currently hardcoded for the highest order curves first.
2793 This should be configurable so applications can judge speed vs strength.
2796 *) Add TLS v1.2 server support for client authentication.
2799 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2803 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2804 FIPS modules versions.
2807 *) Add TLS v1.2 client side support for client authentication. Keep cache
2808 of handshake records longer as we don't know the hash algorithm to use
2809 until after the certificate request message is received.
2812 *) Initial TLS v1.2 client support. Add a default signature algorithms
2813 extension including all the algorithms we support. Parse new signature
2814 format in client key exchange. Relax some ECC signing restrictions for
2815 TLS v1.2 as indicated in RFC5246.
2818 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2819 to new signature format when needed using client digest preference.
2820 All server ciphersuites should now work correctly in TLS v1.2. No client
2821 support yet and no support for client certificates.
2824 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2825 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2826 ciphersuites. At present only RSA key exchange ciphersuites work with
2827 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2828 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2829 and version checking.
2832 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2833 with this defined it will not be affected by any changes to ssl internal
2834 structures. Add several utility functions to allow openssl application
2835 to work with OPENSSL_NO_SSL_INTERN defined.
2839 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2841 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2844 *) Permit abbreviated handshakes when renegotiating using the function
2845 SSL_renegotiate_abbreviated().
2846 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2848 *) Add call to ENGINE_register_all_complete() to
2849 ENGINE_load_builtin_engines(), so some implementations get used
2850 automatically instead of needing explicit application support.
2853 *) Add support for TLS key exporter as described in RFC5705.
2854 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2856 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2857 a few changes are required:
2859 Add SSL_OP_NO_TLSv1_1 flag.
2860 Add TLSv1_1 methods.
2861 Update version checking logic to handle version 1.1.
2862 Add explicit IV handling (ported from DTLS code).
2863 Add command line options to s_client/s_server.
2866 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2868 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2869 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2870 content decryption and always return the same error. Note: this attack
2871 needs on average 2^20 messages so it only affects automated senders. The
2872 old behaviour can be re-enabled in the CMS code by setting the
2873 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2874 an MMA defence is not necessary.
2875 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2876 this issue. (CVE-2012-0884)
2879 *) Fix CVE-2011-4619: make sure we really are receiving a
2880 client hello before rejecting multiple SGC restarts. Thanks to
2881 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2884 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2886 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2887 Thanks to Antonio Martin, Enterprise Secure Access Research and
2888 Development, Cisco Systems, Inc. for discovering this bug and
2889 preparing a fix. (CVE-2012-0050)
2892 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2894 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2895 of the Vaudenay padding oracle attack on CBC mode encryption
2896 which enables an efficient plaintext recovery attack against
2897 the OpenSSL implementation of DTLS. Their attack exploits timing
2898 differences arising during decryption processing. A research
2899 paper describing this attack can be found at:
2900 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2901 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2902 Security Group at Royal Holloway, University of London
2903 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2904 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2905 for preparing the fix. (CVE-2011-4108)
2906 [Robin Seggelmann, Michael Tuexen]
2908 *) Clear bytes used for block padding of SSL 3.0 records.
2910 [Adam Langley (Google)]
2912 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2913 Kadianakis <desnacked@gmail.com> for discovering this issue and
2914 Adam Langley for preparing the fix. (CVE-2011-4619)
2915 [Adam Langley (Google)]
2917 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2918 [Andrey Kulikov <amdeich@gmail.com>]
2920 *) Prevent malformed RFC3779 data triggering an assertion failure.
2921 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2922 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2923 [Rob Austein <sra@hactrn.net>]
2925 *) Improved PRNG seeding for VOS.
2926 [Paul Green <Paul.Green@stratus.com>]
2928 *) Fix ssl_ciph.c set-up race.
2929 [Adam Langley (Google)]
2931 *) Fix spurious failures in ecdsatest.c.
2932 [Emilia Käsper (Google)]
2934 *) Fix the BIO_f_buffer() implementation (which was mixing different
2935 interpretations of the '..._len' fields).
2936 [Adam Langley (Google)]
2938 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2939 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2940 threads won't reuse the same blinding coefficients.
2942 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2943 lock to call BN_BLINDING_invert_ex, and avoids one use of
2944 BN_BLINDING_update for each BN_BLINDING structure (previously,
2945 the last update always remained unused).
2946 [Emilia Käsper (Google)]
2948 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2949 [Bob Buckholz (Google)]
2951 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
2953 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
2954 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
2955 [Kaspar Brand <ossl@velox.ch>]
2957 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2958 for multi-threaded use of ECDH. (CVE-2011-3210)
2959 [Adam Langley (Google)]
2961 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2964 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
2965 signature public key algorithm by using OID xref utilities instead.
2966 Before this you could only use some ECC ciphersuites with SHA1 only.
2969 *) Add protection against ECDSA timing attacks as mentioned in the paper
2970 by Billy Bob Brumley and Nicola Tuveri, see:
2972 http://eprint.iacr.org/2011/232.pdf
2974 [Billy Bob Brumley and Nicola Tuveri]
2976 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
2978 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2979 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2981 *) Fix bug in string printing code: if *any* escaping is enabled we must
2982 escape the escape character (backslash) or the resulting string is
2986 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
2988 *) Disable code workaround for ancient and obsolete Netscape browsers
2989 and servers: an attacker can use it in a ciphersuite downgrade attack.
2990 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2993 *) Fixed J-PAKE implementation error, originally discovered by
2994 Sebastien Martini, further info and confirmation from Stefan
2995 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2998 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3000 *) Fix extension code to avoid race conditions which can result in a buffer
3001 overrun vulnerability: resumed sessions must not be modified as they can
3002 be shared by multiple threads. CVE-2010-3864
3005 *) Fix WIN32 build system to correctly link an ENGINE directory into
3009 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3011 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3013 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3015 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3017 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3018 context. The operation can be customised via the ctrl mechanism in
3019 case ENGINEs want to include additional functionality.
3022 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3025 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3026 output hashes compatible with older versions of OpenSSL.
3027 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3029 *) Fix compression algorithm handling: if resuming a session use the
3030 compression algorithm of the resumed session instead of determining
3031 it from client hello again. Don't allow server to change algorithm.
3034 *) Add load_crls() function to apps tidying load_certs() too. Add option
3035 to verify utility to allow additional CRLs to be included.
3038 *) Update OCSP request code to permit adding custom headers to the request:
3039 some responders need this.
3042 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3044 [Julia Lawall <julia@diku.dk>]
3046 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3047 needlessly dereferenced structures, used obsolete functions and
3048 didn't handle all updated verify codes correctly.
3051 *) Disable MD2 in the default configuration.
3054 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3055 indicate the initial BIO being pushed or popped. This makes it possible
3056 to determine whether the BIO is the one explicitly called or as a result
3057 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3058 it handles reference counts correctly and doesn't zero out the I/O bio
3059 when it is not being explicitly popped. WARNING: applications which
3060 included workarounds for the old buggy behaviour will need to be modified
3061 or they could free up already freed BIOs.
3064 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3065 renaming to all platforms (within the 0.9.8 branch, this was
3066 done conditionally on Netware platforms to avoid a name clash).
3067 [Guenter <lists@gknw.net>]
3069 *) Add ECDHE and PSK support to DTLS.
3070 [Michael Tuexen <tuexen@fh-muenster.de>]
3072 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3076 *) Add "missing" function EVP_MD_flags() (without this the only way to
3077 retrieve a digest flags is by accessing the structure directly. Update
3078 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3079 or cipher is registered as in the "from" argument. Print out all
3080 registered digests in the dgst usage message instead of manually
3081 attempting to work them out.
3084 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3085 this allows the use of compression and extensions. Change default cipher
3086 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3087 by default unless an application cipher string requests it.
3090 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3091 key ids to find matching certificates and keys but some PKCS#12 files
3092 don't follow the (somewhat unwritten) rules and this strategy fails.
3093 Now just gather all certificates together and the first private key
3094 then look for the first certificate that matches the key.
3097 *) Support use of registered digest and cipher names for dgst and cipher
3098 commands instead of having to add each one as a special case. So now
3105 openssl dgst -sha256 foo
3107 and this works for ENGINE based algorithms too.
3111 *) Update Gost ENGINE to support parameter files.
3112 [Victor B. Wagner <vitus@cryptocom.ru>]
3114 *) Support GeneralizedTime in ca utility.
3115 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3117 *) Enhance the hash format used for certificate directory links. The new
3118 form uses the canonical encoding (meaning equivalent names will work
3119 even if they aren't identical) and uses SHA1 instead of MD5. This form
3120 is incompatible with the older format and as a result c_rehash should
3121 be used to rebuild symbolic links.
3124 *) Make PKCS#8 the default write format for private keys, replacing the
3125 traditional format. This form is standardised, more secure and doesn't
3126 include an implicit MD5 dependency.
3129 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3130 committed to OpenSSL should pass this lot as a minimum.
3133 *) Add session ticket override functionality for use by EAP-FAST.
3134 [Jouni Malinen <j@w1.fi>]
3136 *) Modify HMAC functions to return a value. Since these can be implemented
3137 in an ENGINE errors can occur.
3140 *) Type-checked OBJ_bsearch_ex.
3143 *) Type-checked OBJ_bsearch. Also some constification necessitated
3144 by type-checking. Still to come: TXT_DB, bsearch(?),
3145 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3149 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3150 seconds to a tm structure directly, instead of going through OS
3151 specific date routines. This avoids any issues with OS routines such
3152 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3153 and X509_time_adj_ex() to cover the extended range. The existing
3154 X509_time_adj() is still usable and will no longer have any date issues.
3157 *) Delta CRL support. New use deltas option which will attempt to locate
3158 and search any appropriate delta CRLs available.
3160 This work was sponsored by Google.
3163 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3164 code and add additional score elements. Validate alternate CRL paths
3165 as part of the CRL checking and indicate a new error "CRL path validation
3166 error" in this case. Applications wanting additional details can use
3167 the verify callback and check the new "parent" field. If this is not
3168 NULL CRL path validation is taking place. Existing applications won't
3169 see this because it requires extended CRL support which is off by
3172 This work was sponsored by Google.
3175 *) Support for freshest CRL extension.
3177 This work was sponsored by Google.
3180 *) Initial indirect CRL support. Currently only supported in the CRLs
3181 passed directly and not via lookup. Process certificate issuer
3182 CRL entry extension and lookup CRL entries by bother issuer name
3183 and serial number. Check and process CRL issuer entry in IDP extension.
3185 This work was sponsored by Google.
3188 *) Add support for distinct certificate and CRL paths. The CRL issuer
3189 certificate is validated separately in this case. Only enabled if
3190 an extended CRL support flag is set: this flag will enable additional
3191 CRL functionality in future.
3193 This work was sponsored by Google.
3196 *) Add support for policy mappings extension.
3198 This work was sponsored by Google.
3201 *) Fixes to pathlength constraint, self issued certificate handling,
3202 policy processing to align with RFC3280 and PKITS tests.
3204 This work was sponsored by Google.
3207 *) Support for name constraints certificate extension. DN, email, DNS
3208 and URI types are currently supported.
3210 This work was sponsored by Google.
3213 *) To cater for systems that provide a pointer-based thread ID rather
3214 than numeric, deprecate the current numeric thread ID mechanism and
3215 replace it with a structure and associated callback type. This
3216 mechanism allows a numeric "hash" to be extracted from a thread ID in
3217 either case, and on platforms where pointers are larger than 'long',
3218 mixing is done to help ensure the numeric 'hash' is usable even if it
3219 can't be guaranteed unique. The default mechanism is to use "&errno"
3220 as a pointer-based thread ID to distinguish between threads.
3222 Applications that want to provide their own thread IDs should now use
3223 CRYPTO_THREADID_set_callback() to register a callback that will call
3224 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3226 Note that ERR_remove_state() is now deprecated, because it is tied
3227 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3228 to free the current thread's error state should be replaced by
3229 ERR_remove_thread_state(NULL).
3231 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3232 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3233 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3234 application was previously providing a numeric thread callback that
3235 was inappropriate for distinguishing threads, then uniqueness might
3236 have been obtained with &errno that happened immediately in the
3237 intermediate development versions of OpenSSL; this is no longer the
3238 case, the numeric thread callback will now override the automatic use
3240 [Geoff Thorpe, with help from Bodo Moeller]
3242 *) Initial support for different CRL issuing certificates. This covers a
3243 simple case where the self issued certificates in the chain exist and
3244 the real CRL issuer is higher in the existing chain.
3246 This work was sponsored by Google.
3249 *) Removed effectively defunct crypto/store from the build.
3252 *) Revamp of STACK to provide stronger type-checking. Still to come:
3253 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3254 ASN1_STRING, CONF_VALUE.
3257 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3258 RAM on SSL connections. This option can save about 34k per idle SSL.
3261 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3262 STACK, TXT_DB, bsearch, qsort.
3265 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3266 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3267 support for data, signedData, compressedData, digestedData and
3268 encryptedData, envelopedData types included. Scripts to check against
3269 RFC4134 examples draft and interop and consistency checks of many
3270 content types and variants.
3273 *) Add options to enc utility to support use of zlib compression BIO.
3276 *) Extend mk1mf to support importing of options and assembly language
3277 files from Configure script, currently only included in VC-WIN32.
3278 The assembly language rules can now optionally generate the source
3279 files from the associated perl scripts.
3282 *) Implement remaining functionality needed to support GOST ciphersuites.
3283 Interop testing has been performed using CryptoPro implementations.
3284 [Victor B. Wagner <vitus@cryptocom.ru>]
3286 *) s390x assembler pack.
3289 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3293 *) Implement Opaque PRF Input TLS extension as specified in
3294 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3295 official specification yet and no extension type assignment by
3296 IANA exists, this extension (for now) will have to be explicitly
3297 enabled when building OpenSSL by providing the extension number
3298 to use. For example, specify an option
3300 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3302 to the "config" or "Configure" script to enable the extension,
3303 assuming extension number 0x9527 (which is a completely arbitrary
3304 and unofficial assignment based on the MD5 hash of the Internet
3305 Draft). Note that by doing so, you potentially lose
3306 interoperability with other TLS implementations since these might
3307 be using the same extension number for other purposes.
3309 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3310 opaque PRF input value to use in the handshake. This will create
3311 an interal copy of the length-'len' string at 'src', and will
3312 return non-zero for success.
3314 To get more control and flexibility, provide a callback function
3317 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3318 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3322 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3325 Callback function 'cb' will be called in handshakes, and is
3326 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3327 Argument 'arg' is for application purposes (the value as given to
3328 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3329 be provided to the callback function). The callback function
3330 has to return non-zero to report success: usually 1 to use opaque
3331 PRF input just if possible, or 2 to enforce use of the opaque PRF
3332 input. In the latter case, the library will abort the handshake
3333 if opaque PRF input is not successfully negotiated.
3335 Arguments 'peerinput' and 'len' given to the callback function
3336 will always be NULL and 0 in the case of a client. A server will
3337 see the client's opaque PRF input through these variables if
3338 available (NULL and 0 otherwise). Note that if the server
3339 provides an opaque PRF input, the length must be the same as the
3340 length of the client's opaque PRF input.
3342 Note that the callback function will only be called when creating
3343 a new session (session resumption can resume whatever was
3344 previously negotiated), and will not be called in SSL 2.0
3345 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3346 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3347 for applications that need to enforce opaque PRF input.
3351 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3354 [Victor B. Wagner <vitus@cryptocom.ru>]
3356 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3357 RFC4507bis. The encrypted ticket format is an encrypted encoded
3358 SSL_SESSION structure, that way new session features are automatically
3361 If a client application caches session in an SSL_SESSION structure
3362 support is transparent because tickets are now stored in the encoded
3365 The SSL_CTX structure automatically generates keys for ticket
3366 protection in servers so again support should be possible
3367 with no application modification.
3369 If a client or server wishes to disable RFC4507 support then the option
3370 SSL_OP_NO_TICKET can be set.
3372 Add a TLS extension debugging callback to allow the contents of any client
3373 or server extensions to be examined.
3375 This work was sponsored by Google.
3378 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3379 OpenSSL should now compile cleanly on gcc 4.2
3380 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3382 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3383 support including streaming MAC support: this is required for GOST
3384 ciphersuite support.
3385 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3387 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3388 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3389 to output in BER and PEM format.
3392 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3393 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3394 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3395 ENGINE support for HMAC keys which are unextractable. New -mac and
3396 -macopt options to dgst utility.
3399 *) New option -sigopt to dgst utility. Update dgst to use
3400 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3401 alternative signing parameters such as X9.31 or PSS in the dgst
3405 *) Change ssl_cipher_apply_rule(), the internal function that does
3406 the work each time a ciphersuite string requests enabling
3407 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3408 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3409 the order of disabled ciphersuites such that those ciphersuites
3410 that most recently went from enabled to disabled not only stay
3411 in order with respect to each other, but also have higher priority
3412 than other disabled ciphersuites the next time ciphersuites are
3415 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3416 the same ciphersuites as with "HIGH" alone, but in a specific
3417 order where the PSK ciphersuites come first (since they are the
3418 most recently disabled ciphersuites when "HIGH" is parsed).
3420 Also, change ssl_create_cipher_list() (using this new
3421 funcionality) such that between otherwise identical
3422 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3426 *) Change ssl_create_cipher_list() so that it automatically
3427 arranges the ciphersuites in reasonable order before starting
3428 to process the rule string. Thus, the definition for "DEFAULT"
3429 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3430 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3431 This makes it much easier to arrive at a reasonable default order
3432 in applications for which anonymous ciphers are OK (meaning
3433 that you can't actually use DEFAULT).
3434 [Bodo Moeller; suggested by Victor Duchovni]
3436 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3437 processing) into multiple integers instead of setting
3438 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3439 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3440 (These masks as well as the individual bit definitions are hidden
3441 away into the non-exported interface ssl/ssl_locl.h, so this
3442 change to the definition of the SSL_CIPHER structure shouldn't
3443 affect applications.) This give us more bits for each of these
3444 categories, so there is no longer a need to coagulate AES128 and
3445 AES256 into a single algorithm bit, and to coagulate Camellia128
3446 and Camellia256 into a single algorithm bit, which has led to all
3449 Thus, among other things, the kludge introduced in 0.9.7m and
3450 0.9.8e for masking out AES256 independently of AES128 or masking
3451 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3453 With the change, we also introduce new ciphersuite aliases that
3454 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3458 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3459 Use the leftmost N bytes of the signature input if the input is
3460 larger than the prime q (with N being the size in bytes of q).
3463 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3464 it yet and it is largely untested.
3467 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3470 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3471 some compilers (gcc 4.2 and later) reject their use. Safestack is
3472 reimplemented. Update ASN1 to avoid use of legacy functions.
3475 *) Win32/64 targets are linked with Winsock2.
3478 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3479 to external functions. This can be used to increase CRL handling
3480 efficiency especially when CRLs are very large by (for example) storing
3481 the CRL revoked certificates in a database.
3484 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3485 new CRLs added to a directory can be used. New command line option
3486 -verify_return_error to s_client and s_server. This causes real errors
3487 to be returned by the verify callback instead of carrying on no matter
3488 what. This reflects the way a "real world" verify callback would behave.
3491 *) GOST engine, supporting several GOST algorithms and public key formats.
3492 Kindly donated by Cryptocom.
3495 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3496 partitioned by DP are handled but no indirect CRL or reason partitioning
3497 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3498 selected via a scoring technique which handles IDP and AKID in CRLs.
3501 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3502 will ultimately be used for all verify operations: this will remove the
3503 X509_STORE dependency on certificate verification and allow alternative
3504 lookup methods. X509_STORE based implementations of these two callbacks.
3507 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3508 Modify get_crl() to find a valid (unexpired) CRL if possible.
3511 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3512 this would be called X509_CRL_cmp() but that name is already used by
3513 a function that just compares CRL issuer names. Cache several CRL
3514 extensions in X509_CRL structure and cache CRLDP in X509.
3517 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3518 this maps equivalent X509_NAME structures into a consistent structure.
3519 Name comparison can then be performed rapidly using memcmp().
3522 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3526 *) Allow digests to supply their own micalg string for S/MIME type using
3527 the ctrl EVP_MD_CTRL_MICALG.
3530 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3531 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3532 ctrl. It can then customise the structure before and/or after signing
3536 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3537 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3538 to free up any added signature OIDs.
3541 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3542 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3543 digest and cipher tables. New options added to openssl utility:
3544 list-message-digest-algorithms and list-cipher-algorithms.
3547 *) Change the array representation of binary polynomials: the list
3548 of degrees of non-zero coefficients is now terminated with -1.
3549 Previously it was terminated with 0, which was also part of the
3550 value; thus, the array representation was not applicable to
3551 polynomials where t^0 has coefficient zero. This change makes
3552 the array representation useful in a more general context.
3555 *) Various modifications and fixes to SSL/TLS cipher string
3556 handling. For ECC, the code now distinguishes between fixed ECDH
3557 with RSA certificates on the one hand and with ECDSA certificates
3558 on the other hand, since these are separate ciphersuites. The
3559 unused code for Fortezza ciphersuites has been removed.
3561 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3562 (not "ECDHE"). For consistency with the code for DH
3563 certificates, use of ECDH certificates is now considered ECDH
3564 authentication, not RSA or ECDSA authentication (the latter is
3565 merely the CA's signing algorithm and not actively used in the
3568 The temporary ciphersuite alias "ECCdraft" is no longer
3569 available, and ECC ciphersuites are no longer excluded from "ALL"
3570 and "DEFAULT". The following aliases now exist for RFC 4492
3571 ciphersuites, most of these by analogy with the DH case:
3573 kECDHr - ECDH cert, signed with RSA
3574 kECDHe - ECDH cert, signed with ECDSA
3575 kECDH - ECDH cert (signed with either RSA or ECDSA)
3576 kEECDH - ephemeral ECDH
3577 ECDH - ECDH cert or ephemeral ECDH
3583 AECDH - anonymous ECDH
3584 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3588 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3589 Use correct micalg parameters depending on digest(s) in signed message.
3592 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3593 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3596 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3597 an engine to register a method. Add ENGINE lookups for methods and
3598 functional reference processing.
3601 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3602 EVP_{Sign,Verify}* which allow an application to customise the signature
3606 *) New -resign option to smime utility. This adds one or more signers
3607 to an existing PKCS#7 signedData structure. Also -md option to use an
3608 alternative message digest algorithm for signing.
3611 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3612 create PKCS7 structures containing multiple signers. Update smime
3613 application to support multiple signers.
3616 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3620 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3621 Reorganize PBE internals to lookup from a static table using NIDs,
3622 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3623 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3624 PRF which will be automatically used with PBES2.
3627 *) Replace the algorithm specific calls to generate keys in "req" with the
projects / openssl.git / blob