5 Changes between 1.1.0a and 1.1.1 [xx XXX xxxx]
7 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
8 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
9 prevent issues where no progress is being made and the peer continually
10 sends unrecognised record types, using up resources processing them.
13 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
14 using the algorithm defined in
15 https://www.akkadia.org/drepper/SHA-crypt.txt
18 Changes between 1.1.0b and 1.1.0c [xx XXX xxxx]
20 *) ChaCha20/Poly1305 heap-buffer-overflow
22 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
23 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
24 crash. This issue is not considered to be exploitable beyond a DoS.
26 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
30 *) CMS Null dereference
32 Applications parsing invalid CMS structures can crash with a NULL pointer
33 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
34 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
35 structure callback if an attempt is made to free certain invalid encodings.
36 Only CHOICE structures using a callback which do not handle NULL value are
39 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
43 *) Montgomery multiplication may produce incorrect results
45 There is a carry propagating bug in the Broadwell-specific Montgomery
46 multiplication procedure that handles input lengths divisible by, but
47 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
48 and DH private keys are impossible. This is because the subroutine in
49 question is not used in operations with the private key itself and an input
50 of the attacker's direct choice. Otherwise the bug can manifest itself as
51 transient authentication and key negotiation failures or reproducible
52 erroneous outcome of public-key operations with specially crafted input.
53 Among EC algorithms only Brainpool P-512 curves are affected and one
54 presumably can attack ECDH key negotiation. Impact was not analyzed in
55 detail, because pre-requisites for attack are considered unlikely. Namely
56 multiple clients have to choose the curve in question and the server has to
57 share the private key among them, neither of which is default behaviour.
58 Even then only clients that chose the curve will be affected.
60 This issue was publicly reported as transient failures and was not
61 initially recognized as a security issue. Thanks to Richard Morgan for
62 providing reproducible case.
66 *) Removed automatic addition of RPATH in shared libraries and executables,
67 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
70 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
72 *) Fix Use After Free for large message sizes
74 The patch applied to address CVE-2016-6307 resulted in an issue where if a
75 message larger than approx 16k is received then the underlying buffer to
76 store the incoming message is reallocated and moved. Unfortunately a
77 dangling pointer to the old location is left which results in an attempt to
78 write to the previously freed location. This is likely to result in a
79 crash, however it could potentially lead to execution of arbitrary code.
81 This issue only affects OpenSSL 1.1.0a.
83 This issue was reported to OpenSSL by Robert Święcki.
87 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
89 *) OCSP Status Request extension unbounded memory growth
91 A malicious client can send an excessively large OCSP Status Request
92 extension. If that client continually requests renegotiation, sending a
93 large OCSP Status Request extension each time, then there will be unbounded
94 memory growth on the server. This will eventually lead to a Denial Of
95 Service attack through memory exhaustion. Servers with a default
96 configuration are vulnerable even if they do not support OCSP. Builds using
97 the "no-ocsp" build time option are not affected.
99 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
103 *) SSL_peek() hang on empty record
105 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
106 sends an empty record. This could be exploited by a malicious peer in a
107 Denial Of Service attack.
109 This issue was reported to OpenSSL by Alex Gaynor.
113 *) Excessive allocation of memory in tls_get_message_header() and
114 dtls1_preprocess_fragment()
116 A (D)TLS message includes 3 bytes for its length in the header for the
117 message. This would allow for messages up to 16Mb in length. Messages of
118 this length are excessive and OpenSSL includes a check to ensure that a
119 peer is sending reasonably sized messages in order to avoid too much memory
120 being consumed to service a connection. A flaw in the logic of version
121 1.1.0 means that memory for the message is allocated too early, prior to
122 the excessive message length check. Due to way memory is allocated in
123 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
124 to service a connection. This could lead to a Denial of Service through
125 memory exhaustion. However, the excessive message length check still takes
126 place, and this would cause the connection to immediately fail. Assuming
127 that the application calls SSL_free() on the failed conneciton in a timely
128 manner then the 21Mb of allocated memory will then be immediately freed
129 again. Therefore the excessive memory allocation will be transitory in
130 nature. This then means that there is only a security impact if:
132 1) The application does not call SSL_free() in a timely manner in the event
133 that the connection fails
135 2) The application is working in a constrained environment where there is
136 very little free memory
138 3) The attacker initiates multiple connection attempts such that there are
139 multiple connections in a state where memory has been allocated for the
140 connection; SSL_free() has not yet been called; and there is insufficient
141 memory to service the multiple requests.
143 Except in the instance of (1) above any Denial Of Service is likely to be
144 transitory because as soon as the connection fails the memory is
145 subsequently freed again in the SSL_free() call. However there is an
146 increased risk during this period of application crashes due to the lack of
147 memory - which would then mean a more serious Denial of Service.
149 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
150 (CVE-2016-6307 and CVE-2016-6308)
153 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
154 had to be removed. Primary reason is that vendor assembler can't
155 assemble our modules with -KPIC flag. As result it, assembly
156 support, was not even available as option. But its lack means
157 lack of side-channel resistant code, which is incompatible with
158 security by todays standards. Fortunately gcc is readily available
159 prepackaged option, which we firmly point at...
162 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
164 *) Windows command-line tool supports UTF-8 opt-in option for arguments
165 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
166 (to any value) allows Windows user to access PKCS#12 file generated
167 with Windows CryptoAPI and protected with non-ASCII password, as well
168 as files generated under UTF-8 locale on Linux also protected with
172 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
173 have been disabled by default and removed from DEFAULT, just like RC4.
174 See the RC4 item below to re-enable both.
177 *) The method for finding the storage location for the Windows RAND seed file
178 has changed. First we check %RANDFILE%. If that is not set then we check
179 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
180 all else fails we fall back to C:\.
183 *) The EVP_EncryptUpdate() function has had its return type changed from void
184 to int. A return of 0 indicates and error while a return of 1 indicates
188 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
189 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
190 off the constant time implementation for RSA, DSA and DH have been made
191 no-ops and deprecated.
194 *) Windows RAND implementation was simplified to only get entropy by
195 calling CryptGenRandom(). Various other RAND-related tickets
197 [Joseph Wylie Yandle, Rich Salz]
199 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
200 and OPENSSL_LH_, respectively. The old names are available
201 with API compatibility. They new names are now completely documented.
204 *) Unify TYPE_up_ref(obj) methods signature.
205 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
206 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
207 int (instead of void) like all others TYPE_up_ref() methods.
208 So now these methods also check the return value of CRYPTO_atomic_add(),
209 and the validity of object reference counter.
210 [fdasilvayy@gmail.com]
212 *) With Windows Visual Studio builds, the .pdb files are installed
213 alongside the installed libraries and executables. For a static
214 library installation, ossl_static.pdb is the associate compiler
215 generated .pdb file to be used when linking programs.
218 *) Remove openssl.spec. Packaging files belong with the packagers.
221 *) Automatic Darwin/OSX configuration has had a refresh, it will now
222 recognise x86_64 architectures automatically. You can still decide
223 to build for a different bitness with the environment variable
224 KERNEL_BITS (can be 32 or 64), for example:
226 KERNEL_BITS=32 ./config
230 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
231 256 bit AES and HMAC with SHA256.
234 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
237 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
240 *) To enable users to have their own config files and build file templates,
241 Configure looks in the directory indicated by the environment variable
242 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
243 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
244 name and is used as is.
247 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
248 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
249 X509_CERT_FILE_CTX was removed.
252 *) "shared" builds are now the default. To create only static libraries use
253 the "no-shared" Configure option.
256 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
257 All of these option have not worked for some while and are fundamental
261 *) Make various cleanup routines no-ops and mark them as deprecated. Most
262 global cleanup functions are no longer required because they are handled
263 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
264 Explicitly de-initing can cause problems (e.g. where a library that uses
265 OpenSSL de-inits, but an application is still using it). The affected
266 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
267 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
268 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
272 *) --strict-warnings no longer enables runtime debugging options
273 such as REF_DEBUG. Instead, debug options are automatically
274 enabled with '--debug' builds.
275 [Andy Polyakov, Emilia Käsper]
277 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
278 have been moved out of the public header files. New functions for managing
279 these have been added.
282 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
283 objects have been moved out of the public header files. New
284 functions for managing these have been added.
287 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
288 have been moved out of the public header files. New functions for managing
289 these have been added.
292 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
293 moved out of the public header files. New functions for managing these
297 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
300 *) Removed the mk1mf build scripts.
303 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
304 it is always safe to #include a header now.
307 *) Removed the aged BC-32 config and all its supporting scripts
310 *) Removed support for Ultrix, Netware, and OS/2.
313 *) Add support for HKDF.
316 *) Add support for blake2b and blake2s
319 *) Added support for "pipelining". Ciphers that have the
320 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
321 encryptions/decryptions simultaneously. There are currently no built-in
322 ciphers with this property but the expectation is that engines will be able
323 to offer it to significantly improve throughput. Support has been extended
324 into libssl so that multiple records for a single connection can be
325 processed in one go (for >=TLS 1.1).
328 *) Added the AFALG engine. This is an async capable engine which is able to
329 offload work to the Linux kernel. In this initial version it only supports
330 AES128-CBC. The kernel must be version 4.1.0 or greater.
333 *) OpenSSL now uses a new threading API. It is no longer necessary to
334 set locking callbacks to use OpenSSL in a multi-threaded environment. There
335 are two supported threading models: pthreads and windows threads. It is
336 also possible to configure OpenSSL at compile time for "no-threads". The
337 old threading API should no longer be used. The functions have been
338 replaced with "no-op" compatibility macros.
339 [Alessandro Ghedini, Matt Caswell]
341 *) Modify behavior of ALPN to invoke callback after SNI/servername
342 callback, such that updates to the SSL_CTX affect ALPN.
345 *) Add SSL_CIPHER queries for authentication and key-exchange.
348 *) Changes to the DEFAULT cipherlist:
349 - Prefer (EC)DHE handshakes over plain RSA.
350 - Prefer AEAD ciphers over legacy ciphers.
351 - Prefer ECDSA over RSA when both certificates are available.
352 - Prefer TLSv1.2 ciphers/PRF.
353 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
357 *) Change the ECC default curve list to be this, in order: x25519,
358 secp256r1, secp521r1, secp384r1.
361 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
362 disabled by default. They can be re-enabled using the
363 enable-weak-ssl-ciphers option to Configure.
366 *) If the server has ALPN configured, but supports no protocols that the
367 client advertises, send a fatal "no_application_protocol" alert.
368 This behaviour is SHALL in RFC 7301, though it isn't universally
369 implemented by other servers.
372 *) Add X25519 support.
373 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
374 for public and private key encoding using the format documented in
375 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
376 key generation and key derivation.
378 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
382 *) Deprecate SRP_VBASE_get_by_user.
383 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
384 In order to fix an unavoidable memory leak (CVE-2016-0798),
385 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
386 seed, even if the seed is configured.
388 Users should use SRP_VBASE_get1_by_user instead. Note that in
389 SRP_VBASE_get1_by_user, caller must free the returned value. Note
390 also that even though configuring the SRP seed attempts to hide
391 invalid usernames by continuing the handshake with fake
392 credentials, this behaviour is not constant time and no strong
393 guarantees are made that the handshake is indistinguishable from
394 that of a valid user.
397 *) Configuration change; it's now possible to build dynamic engines
398 without having to build shared libraries and vice versa. This
399 only applies to the engines in engines/, those in crypto/engine/
400 will always be built into libcrypto (i.e. "static").
402 Building dynamic engines is enabled by default; to disable, use
403 the configuration option "disable-dynamic-engine".
405 The only requirements for building dynamic engines are the
406 presence of the DSO module and building with position independent
407 code, so they will also automatically be disabled if configuring
408 with "disable-dso" or "disable-pic".
410 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
411 are also taken away from openssl/opensslconf.h, as they are
415 *) Configuration change; if there is a known flag to compile
416 position independent code, it will always be applied on the
417 libcrypto and libssl object files, and never on the application
418 object files. This means other libraries that use routines from
419 libcrypto / libssl can be made into shared libraries regardless
420 of how OpenSSL was configured.
422 If this isn't desirable, the configuration options "disable-pic"
423 or "no-pic" can be used to disable the use of PIC. This will
424 also disable building shared libraries and dynamic engines.
427 *) Removed JPAKE code. It was experimental and has no wide use.
430 *) The INSTALL_PREFIX Makefile variable has been renamed to
431 DESTDIR. That makes for less confusion on what this variable
432 is for. Also, the configuration option --install_prefix is
436 *) Heartbeat for TLS has been removed and is disabled by default
437 for DTLS; configure with enable-heartbeats. Code that uses the
438 old #define's might need to be updated.
439 [Emilia Käsper, Rich Salz]
441 *) Rename REF_CHECK to REF_DEBUG.
444 *) New "unified" build system
446 The "unified" build system is aimed to be a common system for all
447 platforms we support. With it comes new support for VMS.
449 This system builds supports building in a different directory tree
450 than the source tree. It produces one Makefile (for unix family
451 or lookalikes), or one descrip.mms (for VMS).
453 The source of information to make the Makefile / descrip.mms is
454 small files called 'build.info', holding the necessary
455 information for each directory with source to compile, and a
456 template in Configurations, like unix-Makefile.tmpl or
459 With this change, the library names were also renamed on Windows
460 and on VMS. They now have names that are closer to the standard
461 on Unix, and include the major version number, and in certain
462 cases, the architecture they are built for. See "Notes on shared
463 libraries" in INSTALL.
465 We rely heavily on the perl module Text::Template.
468 *) Added support for auto-initialisation and de-initialisation of the library.
469 OpenSSL no longer requires explicit init or deinit routines to be called,
470 except in certain circumstances. See the OPENSSL_init_crypto() and
471 OPENSSL_init_ssl() man pages for further information.
474 *) The arguments to the DTLSv1_listen function have changed. Specifically the
475 "peer" argument is now expected to be a BIO_ADDR object.
477 *) Rewrite of BIO networking library. The BIO library lacked consistent
478 support of IPv6, and adding it required some more extensive
479 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
480 which hold all types of addresses and chains of address information.
481 It also introduces a new API, with functions like BIO_socket,
482 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
483 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
484 have been adapted accordingly.
487 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
491 *) CRIME protection: disable compression by default, even if OpenSSL is
492 compiled with zlib enabled. Applications can still enable compression
493 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
494 using the SSL_CONF library to configure compression.
497 *) The signature of the session callback configured with
498 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
499 was explicitly marked as 'const unsigned char*' instead of
503 *) Always DPURIFY. Remove the use of uninitialized memory in the
504 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
507 *) Removed many obsolete configuration items, including
508 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
509 MD2_CHAR, MD2_INT, MD2_LONG
511 IDEA_SHORT, IDEA_LONG
512 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
513 [Rich Salz, with advice from Andy Polyakov]
515 *) Many BN internals have been moved to an internal header file.
516 [Rich Salz with help from Andy Polyakov]
518 *) Configuration and writing out the results from it has changed.
519 Files such as Makefile include/openssl/opensslconf.h and are now
520 produced through general templates, such as Makefile.in and
521 crypto/opensslconf.h.in and some help from the perl module
524 Also, the center of configuration information is no longer
525 Makefile. Instead, Configure produces a perl module in
526 configdata.pm which holds most of the config data (in the hash
527 table %config), the target data that comes from the target
528 configuration in one of the Configurations/*.conf files (in
532 *) To clarify their intended purposes, the Configure options
533 --prefix and --openssldir change their semantics, and become more
534 straightforward and less interdependent.
536 --prefix shall be used exclusively to give the location INSTALLTOP
537 where programs, scripts, libraries, include files and manuals are
538 going to be installed. The default is now /usr/local.
540 --openssldir shall be used exclusively to give the default
541 location OPENSSLDIR where certificates, private keys, CRLs are
542 managed. This is also where the default openssl.cnf gets
544 If the directory given with this option is a relative path, the
545 values of both the --prefix value and the --openssldir value will
546 be combined to become OPENSSLDIR.
547 The default for --openssldir is INSTALLTOP/ssl.
549 Anyone who uses --openssldir to specify where OpenSSL is to be
550 installed MUST change to use --prefix instead.
553 *) The GOST engine was out of date and therefore it has been removed. An up
554 to date GOST engine is now being maintained in an external repository.
555 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
556 support for GOST ciphersuites (these are only activated if a GOST engine
560 *) EGD is no longer supported by default; use enable-egd when
562 [Ben Kaduk and Rich Salz]
564 *) The distribution now has Makefile.in files, which are used to
565 create Makefile's when Configure is run. *Configure must be run
566 before trying to build now.*
569 *) The return value for SSL_CIPHER_description() for error conditions
573 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
575 Obtaining and performing DNSSEC validation of TLSA records is
576 the application's responsibility. The application provides
577 the TLSA records of its choice to OpenSSL, and these are then
578 used to authenticate the peer.
580 The TLSA records need not even come from DNS. They can, for
581 example, be used to implement local end-entity certificate or
582 trust-anchor "pinning", where the "pin" data takes the form
583 of TLSA records, which can augment or replace verification
584 based on the usual WebPKI public certification authorities.
587 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
588 continues to support deprecated interfaces in default builds.
589 However, applications are strongly advised to compile their
590 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
591 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
592 or the 1.1.0 releases.
594 In environments in which all applications have been ported to
595 not use any deprecated interfaces OpenSSL's Configure script
596 should be used with the --api=1.1.0 option to entirely remove
597 support for the deprecated features from the library and
598 unconditionally disable them in the installed headers.
599 Essentially the same effect can be achieved with the "no-deprecated"
600 argument to Configure, except that this will always restrict
601 the build to just the latest API, rather than a fixed API
604 As applications are ported to future revisions of the API,
605 they should update their compile-time OPENSSL_API_COMPAT define
606 accordingly, but in most cases should be able to continue to
607 compile with later releases.
609 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
610 0x10000000L and 0x00908000L, respectively. However those
611 versions did not support the OPENSSL_API_COMPAT feature, and
612 so applications are not typically tested for explicit support
613 of just the undeprecated features of either release.
616 *) Add support for setting the minimum and maximum supported protocol.
617 It can bet set via the SSL_set_min_proto_version() and
618 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
619 MaxProtcol. It's recommended to use the new APIs to disable
620 protocols instead of disabling individual protocols using
621 SSL_set_options() or SSL_CONF's Protocol. This change also
622 removes support for disabling TLS 1.2 in the OpenSSL TLS
623 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
626 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
629 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
630 and integrates ECDSA and ECDH functionality into EC. Implementations can
631 now redirect key generation and no longer need to convert to or from
634 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
635 include the ec.h header file instead.
638 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
639 ciphers who are no longer supported and drops support the ephemeral RSA key
640 exchange. The LOW ciphers currently doesn't have any ciphers in it.
643 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
644 opaque. For HMAC_CTX, the following constructors and destructors
647 HMAC_CTX *HMAC_CTX_new(void);
648 void HMAC_CTX_free(HMAC_CTX *ctx);
650 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
651 destroy such methods has been added. See EVP_MD_meth_new(3) and
652 EVP_CIPHER_meth_new(3) for documentation.
655 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
656 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
657 EVP_MD_CTX_reset() should be called instead to reinitialise
658 an already created structure.
659 2) For consistency with the majority of our object creators and
660 destructors, EVP_MD_CTX_(create|destroy) were renamed to
661 EVP_MD_CTX_(new|free). The old names are retained as macros
662 for deprecated builds.
665 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
666 cryptographic operations to be performed asynchronously as long as an
667 asynchronous capable engine is used. See the ASYNC_start_job() man page for
668 further details. Libssl has also had this capability integrated with the
669 introduction of the new mode SSL_MODE_ASYNC and associated error
670 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
671 pages. This work was developed in partnership with Intel Corp.
674 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
675 always enabled now. If you want to disable the support you should
676 exclude it using the list of supported ciphers. This also means that the
677 "-no_ecdhe" option has been removed from s_server.
680 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
681 SSL_{CTX_}set1_curves() which can set a list.
684 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
685 curve you want to support using SSL_{CTX_}set1_curves().
688 *) State machine rewrite. The state machine code has been significantly
689 refactored in order to remove much duplication of code and solve issues
690 with the old code (see ssl/statem/README for further details). This change
691 does have some associated API changes. Notably the SSL_state() function
692 has been removed and replaced by SSL_get_state which now returns an
693 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
694 altogether. The previous handshake states defined in ssl.h and ssl3.h have
698 *) All instances of the string "ssleay" in the public API were replaced
699 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
700 Some error codes related to internal RSA_eay API's were renamed.
703 *) The demo files in crypto/threads were moved to demo/threads.
706 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
708 [Matt Caswell, Rich Salz]
710 *) New ASN.1 embed macro.
712 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
713 structure is not allocated: it is part of the parent. That is instead of
721 This reduces memory fragmentation and make it impossible to accidentally
722 set a mandatory field to NULL.
724 This currently only works for some fields specifically a SEQUENCE, CHOICE,
725 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
726 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
730 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
733 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
734 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
735 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
736 DES and RC4 ciphersuites.
739 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
740 This changes the decoding behaviour for some invalid messages,
741 though the change is mostly in the more lenient direction, and
742 legacy behaviour is preserved as much as possible.
745 *) Fix no-stdio build.
746 [ David Woodhouse <David.Woodhouse@intel.com> and also
747 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
749 *) New testing framework
750 The testing framework has been largely rewritten and is now using
751 perl and the perl modules Test::Harness and an extended variant of
752 Test::More called OpenSSL::Test to do its work. All test scripts in
753 test/ have been rewritten into test recipes, and all direct calls to
754 executables in test/Makefile have become individual recipes using the
755 simplified testing OpenSSL::Test::Simple.
757 For documentation on our testing modules, do:
759 perldoc test/testlib/OpenSSL/Test/Simple.pm
760 perldoc test/testlib/OpenSSL/Test.pm
764 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
765 are used; the latter aborts on memory leaks (usually checked on exit).
766 Some undocumented "set malloc, etc., hooks" functions were removed
767 and others were changed. All are now documented.
770 *) In DSA_generate_parameters_ex, if the provided seed is too short,
772 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
774 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
775 from RFC4279, RFC4785, RFC5487, RFC5489.
777 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
778 original RSA_PSK patch.
781 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
782 era flag was never set throughout the codebase (only read). Also removed
783 SSL3_FLAGS_POP_BUFFER which was only used if
784 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
787 *) Changed the default name options in the "ca", "crl", "req" and "x509"
788 to be "oneline" instead of "compat".
791 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
792 not aware of clients that still exhibit this bug, and the workaround
793 hasn't been working properly for a while.
796 *) The return type of BIO_number_read() and BIO_number_written() as well as
797 the corresponding num_read and num_write members in the BIO structure has
798 changed from unsigned long to uint64_t. On platforms where an unsigned
799 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
803 *) Given the pervasive nature of TLS extensions it is inadvisable to run
804 OpenSSL without support for them. It also means that maintaining
805 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
806 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
809 *) Removed support for the two export grade static DH ciphersuites
810 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
811 were newly added (along with a number of other static DH ciphersuites) to
812 1.0.2. However the two export ones have *never* worked since they were
813 introduced. It seems strange in any case to be adding new export
814 ciphersuites, and given "logjam" it also does not seem correct to fix them.
817 *) Version negotiation has been rewritten. In particular SSLv23_method(),
818 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
819 and turned into macros which simply call the new preferred function names
820 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
821 should use the new names instead. Also as part of this change the ssl23.h
822 header file has been removed.
825 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
826 code and the associated standard is no longer considered fit-for-purpose.
829 *) RT2547 was closed. When generating a private key, try to make the
830 output file readable only by the owner. This behavior change might
831 be noticeable when interacting with other software.
833 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
837 *) Added HTTP GET support to the ocsp command.
840 *) Changed default digest for the dgst and enc commands from MD5 to
844 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
847 *) Added support for TLS extended master secret from
848 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
849 initial patch which was a great help during development.
852 *) All libssl internal structures have been removed from the public header
853 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
854 now redundant). Users should not attempt to access internal structures
855 directly. Instead they should use the provided API functions.
858 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
859 Access to deprecated functions can be re-enabled by running config with
860 "enable-deprecated". In addition applications wishing to use deprecated
861 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
862 will, by default, disable some transitive includes that previously existed
863 in the header files (e.g. ec.h will no longer, by default, include bn.h)
866 *) Added support for OCB mode. OpenSSL has been granted a patent license
867 compatible with the OpenSSL license for use of OCB. Details are available
868 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
869 for OCB can be removed by calling config with no-ocb.
872 *) SSLv2 support has been removed. It still supports receiving a SSLv2
873 compatible client hello.
876 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
877 done while fixing the error code for the key-too-small case.
878 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
880 *) CA.sh has been removmed; use CA.pl instead.
883 *) Removed old DES API.
886 *) Remove various unsupported platforms:
892 Sinix/ReliantUNIX RM400
897 16-bit platforms such as WIN16
900 *) Clean up OPENSSL_NO_xxx #define's
901 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
902 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
903 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
904 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
905 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
906 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
907 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
908 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
909 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
910 Remove MS_STATIC; it's a relic from platforms <32 bits.
913 *) Cleaned up dead code
914 Remove all but one '#ifdef undef' which is to be looked at.
917 *) Clean up calling of xxx_free routines.
918 Just like free(), fix most of the xxx_free routines to accept
919 NULL. Remove the non-null checks from callers. Save much code.
922 *) Add secure heap for storage of private keys (when possible).
923 Add BIO_s_secmem(), CBIGNUM, etc.
924 Contributed by Akamai Technologies under our Corporate CLA.
927 *) Experimental support for a new, fast, unbiased prime candidate generator,
928 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
929 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
931 *) New output format NSS in the sess_id command line tool. This allows
932 exporting the session id and the master key in NSS keylog format.
933 [Martin Kaiser <martin@kaiser.cx>]
935 *) Harmonize version and its documentation. -f flag is used to display
937 [mancha <mancha1@zoho.com>]
939 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
940 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
941 [mancha <mancha1@zoho.com>]
943 *) Fix some double frees. These are not thought to be exploitable.
944 [mancha <mancha1@zoho.com>]
946 *) A missing bounds check in the handling of the TLS heartbeat extension
947 can be used to reveal up to 64k of memory to a connected client or
950 Thanks for Neel Mehta of Google Security for discovering this bug and to
951 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
952 preparing the fix (CVE-2014-0160)
953 [Adam Langley, Bodo Moeller]
955 *) Fix for the attack described in the paper "Recovering OpenSSL
956 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
957 by Yuval Yarom and Naomi Benger. Details can be obtained from:
958 http://eprint.iacr.org/2014/140
960 Thanks to Yuval Yarom and Naomi Benger for discovering this
961 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
962 [Yuval Yarom and Naomi Benger]
964 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
965 this fixes a limitation in previous versions of OpenSSL.
968 *) Experimental encrypt-then-mac support.
970 Experimental support for encrypt then mac from
971 draft-gutmann-tls-encrypt-then-mac-02.txt
973 To enable it set the appropriate extension number (0x42 for the test
974 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
976 For non-compliant peers (i.e. just about everything) this should have no
979 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
983 *) Add EVP support for key wrapping algorithms, to avoid problems with
984 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
985 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
986 algorithms and include tests cases.
989 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
993 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
994 MGF1 digest and OAEP label.
997 *) Make openssl verify return errors.
998 [Chris Palmer <palmer@google.com> and Ben Laurie]
1000 *) New function ASN1_TIME_diff to calculate the difference between two
1001 ASN1_TIME structures or one structure and the current time.
1004 *) Update fips_test_suite to support multiple command line options. New
1005 test to induce all self test errors in sequence and check expected
1009 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1010 sign or verify all in one operation.
1013 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1014 test programs and fips_test_suite. Includes functionality to parse
1015 the minimal script output of fipsalgest.pl directly.
1018 *) Add authorisation parameter to FIPS_module_mode_set().
1021 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1024 *) Use separate DRBG fields for internal and external flags. New function
1025 FIPS_drbg_health_check() to perform on demand health checking. Add
1026 generation tests to fips_test_suite with reduced health check interval to
1027 demonstrate periodic health checking. Add "nodh" option to
1028 fips_test_suite to skip very slow DH test.
1031 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1035 *) More extensive health check for DRBG checking many more failure modes.
1036 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1037 combination: call this in fips_test_suite.
1040 *) Add support for canonical generation of DSA parameter 'g'. See
1043 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1044 POST to handle HMAC cases.
1047 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1048 to return numerical and string versions of the FIPS module number.
1051 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1052 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1053 outside the validated module in the FIPS capable OpenSSL.
1056 *) Minor change to DRBG entropy callback semantics. In some cases
1057 there is no multiple of the block length between min_len and
1058 max_len. Allow the callback to return more than max_len bytes
1059 of entropy but discard any extra: it is the callback's responsibility
1060 to ensure that the extra data discarded does not impact the
1061 requested amount of entropy.
1064 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1065 information in FIPS186-3, SP800-57 and SP800-131A.
1068 *) CCM support via EVP. Interface is very similar to GCM case except we
1069 must supply all data in one chunk (i.e. no update, final) and the
1070 message length must be supplied if AAD is used. Add algorithm test
1074 *) Initial version of POST overhaul. Add POST callback to allow the status
1075 of POST to be monitored and/or failures induced. Modify fips_test_suite
1076 to use callback. Always run all selftests even if one fails.
1079 *) XTS support including algorithm test driver in the fips_gcmtest program.
1080 Note: this does increase the maximum key length from 32 to 64 bytes but
1081 there should be no binary compatibility issues as existing applications
1082 will never use XTS mode.
1085 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1086 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1087 performs algorithm blocking for unapproved PRNG types. Also do not
1088 set PRNG type in FIPS_mode_set(): leave this to the application.
1089 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1090 the standard OpenSSL PRNG: set additional data to a date time vector.
1093 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1094 This shouldn't present any incompatibility problems because applications
1095 shouldn't be using these directly and any that are will need to rethink
1096 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1099 *) Extensive self tests and health checking required by SP800-90 DRBG.
1100 Remove strength parameter from FIPS_drbg_instantiate and always
1101 instantiate at maximum supported strength.
1104 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1107 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1110 *) New function DH_compute_key_padded() to compute a DH key and pad with
1111 leading zeroes if needed: this complies with SP800-56A et al.
1114 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1115 anything, incomplete, subject to change and largely untested at present.
1118 *) Modify fipscanisteronly build option to only build the necessary object
1119 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1122 *) Add experimental option FIPSSYMS to give all symbols in
1123 fipscanister.o and FIPS or fips prefix. This will avoid
1124 conflicts with future versions of OpenSSL. Add perl script
1125 util/fipsas.pl to preprocess assembly language source files
1126 and rename any affected symbols.
1129 *) Add selftest checks and algorithm block of non-fips algorithms in
1130 FIPS mode. Remove DES2 from selftests.
1133 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1134 return internal method without any ENGINE dependencies. Add new
1135 tiny fips sign and verify functions.
1138 *) New build option no-ec2m to disable characteristic 2 code.
1141 *) New build option "fipscanisteronly". This only builds fipscanister.o
1142 and (currently) associated fips utilities. Uses the file Makefile.fips
1143 instead of Makefile.org as the prototype.
1146 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1147 Update fips_gcmtest to use IV generator.
1150 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1151 setting output buffer to NULL. The *Final function must be
1152 called although it will not retrieve any additional data. The tag
1153 can be set or retrieved with a ctrl. The IV length is by default 12
1154 bytes (96 bits) but can be set to an alternative value. If the IV
1155 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1159 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1160 underlying do_cipher function handles all cipher semantics itself
1161 including padding and finalisation. This is useful if (for example)
1162 an ENGINE cipher handles block padding itself. The behaviour of
1163 do_cipher is subtly changed if this flag is set: the return value
1164 is the number of characters written to the output buffer (zero is
1165 no longer an error code) or a negative error code. Also if the
1166 input buffer is NULL and length 0 finalisation should be performed.
1169 *) If a candidate issuer certificate is already part of the constructed
1170 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1173 *) Improve forward-security support: add functions
1175 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1176 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1178 for use by SSL/TLS servers; the callback function will be called whenever a
1179 new session is created, and gets to decide whether the session may be
1180 cached to make it resumable (return 0) or not (return 1). (As by the
1181 SSL/TLS protocol specifications, the session_id sent by the server will be
1182 empty to indicate that the session is not resumable; also, the server will
1183 not generate RFC 4507 (RFC 5077) session tickets.)
1185 A simple reasonable callback implementation is to return is_forward_secure.
1186 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1187 by the SSL/TLS server library, indicating whether it can provide forward
1189 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1191 *) New -verify_name option in command line utilities to set verification
1195 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1196 Add CMAC pkey methods.
1199 *) Experimental renegotiation in s_server -www mode. If the client
1200 browses /reneg connection is renegotiated. If /renegcert it is
1201 renegotiated requesting a certificate.
1204 *) Add an "external" session cache for debugging purposes to s_server. This
1205 should help trace issues which normally are only apparent in deployed
1206 multi-process servers.
1209 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1210 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1211 BIO_set_cipher() and some obscure PEM functions were changed so they
1212 can now return an error. The RAND changes required a change to the
1213 RAND_METHOD structure.
1216 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1217 a gcc attribute to warn if the result of a function is ignored. This
1218 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1219 whose return value is often ignored.
1222 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1223 These allow SCTs (signed certificate timestamps) to be requested and
1224 validated when establishing a connection.
1225 [Rob Percival <robpercival@google.com>]
1227 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1229 *) Prevent padding oracle in AES-NI CBC MAC check
1231 A MITM attacker can use a padding oracle attack to decrypt traffic
1232 when the connection uses an AES CBC cipher and the server support
1235 This issue was introduced as part of the fix for Lucky 13 padding
1236 attack (CVE-2013-0169). The padding check was rewritten to be in
1237 constant time by making sure that always the same bytes are read and
1238 compared against either the MAC or padding bytes. But it no longer
1239 checked that there was enough data to have both the MAC and padding
1242 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1246 *) Fix EVP_EncodeUpdate overflow
1248 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1249 Base64 encoding of binary data. If an attacker is able to supply very large
1250 amounts of input data then a length check can overflow resulting in a heap
1253 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1254 the PEM_write_bio* family of functions. These are mainly used within the
1255 OpenSSL command line applications, so any application which processes data
1256 from an untrusted source and outputs it as a PEM file should be considered
1257 vulnerable to this issue. User applications that call these APIs directly
1258 with large amounts of untrusted data may also be vulnerable.
1260 This issue was reported by Guido Vranken.
1264 *) Fix EVP_EncryptUpdate overflow
1266 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1267 is able to supply very large amounts of input data after a previous call to
1268 EVP_EncryptUpdate() with a partial block then a length check can overflow
1269 resulting in a heap corruption. Following an analysis of all OpenSSL
1270 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1271 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1272 the first called function after an EVP_EncryptInit(), and therefore that
1273 specific call must be safe. The second form is where the length passed to
1274 EVP_EncryptUpdate() can be seen from the code to be some small value and
1275 therefore there is no possibility of an overflow. Since all instances are
1276 one of these two forms, it is believed that there can be no overflows in
1277 internal code due to this problem. It should be noted that
1278 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1279 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1280 of these calls have also been analysed too and it is believed there are no
1281 instances in internal usage where an overflow could occur.
1283 This issue was reported by Guido Vranken.
1287 *) Prevent ASN.1 BIO excessive memory allocation
1289 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1290 a short invalid encoding can cause allocation of large amounts of memory
1291 potentially consuming excessive resources or exhausting memory.
1293 Any application parsing untrusted data through d2i BIO functions is
1294 affected. The memory based functions such as d2i_X509() are *not* affected.
1295 Since the memory based functions are used by the TLS library, TLS
1296 applications are not affected.
1298 This issue was reported by Brian Carpenter.
1304 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1305 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1306 in arbitrary stack data being returned in the buffer.
1308 This issue was reported by Guido Vranken.
1312 *) Modify behavior of ALPN to invoke callback after SNI/servername
1313 callback, such that updates to the SSL_CTX affect ALPN.
1316 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1320 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1321 methods are enabled and ssl2 is disabled the methods return NULL.
1324 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1326 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1327 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1328 provide any "EXPORT" or "LOW" strength ciphers.
1331 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1332 is by default disabled at build-time. Builds that are not configured with
1333 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1334 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1335 will need to explicitly call either of:
1337 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1339 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1341 as appropriate. Even if either of those is used, or the application
1342 explicitly uses the version-specific SSLv2_method() or its client and
1343 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1344 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1345 ciphers, and SSLv2 56-bit DES are no longer available.
1349 *) Fix a double-free in DSA code
1351 A double free bug was discovered when OpenSSL parses malformed DSA private
1352 keys and could lead to a DoS attack or memory corruption for applications
1353 that receive DSA private keys from untrusted sources. This scenario is
1356 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1361 *) Disable SRP fake user seed to address a server memory leak.
1363 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1365 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1366 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1367 was changed to ignore the "fake user" SRP seed, even if the seed
1370 Users should use SRP_VBASE_get1_by_user instead. Note that in
1371 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1372 also that even though configuring the SRP seed attempts to hide
1373 invalid usernames by continuing the handshake with fake
1374 credentials, this behaviour is not constant time and no strong
1375 guarantees are made that the handshake is indistinguishable from
1376 that of a valid user.
1380 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1382 In the BN_hex2bn function the number of hex digits is calculated using an
1383 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1384 large values of |i| this can result in |bn_expand| not allocating any
1385 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1386 field as NULL leading to a subsequent NULL ptr deref. For very large values
1387 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1388 In this case memory is allocated to the internal BIGNUM data field, but it
1389 is insufficiently sized leading to heap corruption. A similar issue exists
1390 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1391 is ever called by user applications with very large untrusted hex/dec data.
1392 This is anticipated to be a rare occurrence.
1394 All OpenSSL internal usage of these functions use data that is not expected
1395 to be untrusted, e.g. config file data or application command line
1396 arguments. If user developed applications generate config file data based
1397 on untrusted data then it is possible that this could also lead to security
1398 consequences. This is also anticipated to be rare.
1400 This issue was reported to OpenSSL by Guido Vranken.
1404 *) Fix memory issues in BIO_*printf functions
1406 The internal |fmtstr| function used in processing a "%s" format string in
1407 the BIO_*printf functions could overflow while calculating the length of a
1408 string and cause an OOB read when printing very long strings.
1410 Additionally the internal |doapr_outch| function can attempt to write to an
1411 OOB memory location (at an offset from the NULL pointer) in the event of a
1412 memory allocation failure. In 1.0.2 and below this could be caused where
1413 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1414 could be in processing a very long "%s" format string. Memory leaks can
1417 The first issue may mask the second issue dependent on compiler behaviour.
1418 These problems could enable attacks where large amounts of untrusted data
1419 is passed to the BIO_*printf functions. If applications use these functions
1420 in this way then they could be vulnerable. OpenSSL itself uses these
1421 functions when printing out human-readable dumps of ASN.1 data. Therefore
1422 applications that print this data could be vulnerable if the data is from
1423 untrusted sources. OpenSSL command line applications could also be
1424 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1425 as command line arguments.
1427 Libssl is not considered directly vulnerable. Additionally certificates etc
1428 received via remote connections via libssl are also unlikely to be able to
1429 trigger these issues because of message size limits enforced within libssl.
1431 This issue was reported to OpenSSL Guido Vranken.
1435 *) Side channel attack on modular exponentiation
1437 A side-channel attack was found which makes use of cache-bank conflicts on
1438 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1439 of RSA keys. The ability to exploit this issue is limited as it relies on
1440 an attacker who has control of code in a thread running on the same
1441 hyper-threaded core as the victim thread which is performing decryptions.
1443 This issue was reported to OpenSSL by Yuval Yarom, The University of
1444 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1445 Nadia Heninger, University of Pennsylvania with more information at
1446 http://cachebleed.info.
1450 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1451 if no keysize is specified with default_bits. This fixes an
1452 omission in an earlier change that changed all RSA/DSA key generation
1453 apps to use 2048 bits by default.
1456 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1457 *) DH small subgroups
1459 Historically OpenSSL only ever generated DH parameters based on "safe"
1460 primes. More recently (in version 1.0.2) support was provided for
1461 generating X9.42 style parameter files such as those required for RFC 5114
1462 support. The primes used in such files may not be "safe". Where an
1463 application is using DH configured with parameters based on primes that are
1464 not "safe" then an attacker could use this fact to find a peer's private
1465 DH exponent. This attack requires that the attacker complete multiple
1466 handshakes in which the peer uses the same private DH exponent. For example
1467 this could be used to discover a TLS server's private DH exponent if it's
1468 reusing the private DH exponent or it's using a static DH ciphersuite.
1470 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1471 TLS. It is not on by default. If the option is not set then the server
1472 reuses the same private DH exponent for the life of the server process and
1473 would be vulnerable to this attack. It is believed that many popular
1474 applications do set this option and would therefore not be at risk.
1476 The fix for this issue adds an additional check where a "q" parameter is
1477 available (as is the case in X9.42 based parameters). This detects the
1478 only known attack, and is the only possible defense for static DH
1479 ciphersuites. This could have some performance impact.
1481 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1482 default and cannot be disabled. This could have some performance impact.
1484 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1488 *) SSLv2 doesn't block disabled ciphers
1490 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1491 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1492 been disabled, provided that the SSLv2 protocol was not also disabled via
1495 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1496 and Sebastian Schinzel.
1500 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1502 *) BN_mod_exp may produce incorrect results on x86_64
1504 There is a carry propagating bug in the x86_64 Montgomery squaring
1505 procedure. No EC algorithms are affected. Analysis suggests that attacks
1506 against RSA and DSA as a result of this defect would be very difficult to
1507 perform and are not believed likely. Attacks against DH are considered just
1508 feasible (although very difficult) because most of the work necessary to
1509 deduce information about a private key may be performed offline. The amount
1510 of resources required for such an attack would be very significant and
1511 likely only accessible to a limited number of attackers. An attacker would
1512 additionally need online access to an unpatched system using the target
1513 private key in a scenario with persistent DH parameters and a private
1514 key that is shared between multiple clients. For example this can occur by
1515 default in OpenSSL DHE based SSL/TLS ciphersuites.
1517 This issue was reported to OpenSSL by Hanno Böck.
1521 *) Certificate verify crash with missing PSS parameter
1523 The signature verification routines will crash with a NULL pointer
1524 dereference if presented with an ASN.1 signature using the RSA PSS
1525 algorithm and absent mask generation function parameter. Since these
1526 routines are used to verify certificate signature algorithms this can be
1527 used to crash any certificate verification operation and exploited in a
1528 DoS attack. Any application which performs certificate verification is
1529 vulnerable including OpenSSL clients and servers which enable client
1532 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1536 *) X509_ATTRIBUTE memory leak
1538 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1539 memory. This structure is used by the PKCS#7 and CMS routines so any
1540 application which reads PKCS#7 or CMS data from untrusted sources is
1541 affected. SSL/TLS is not affected.
1543 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1548 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1549 This changes the decoding behaviour for some invalid messages,
1550 though the change is mostly in the more lenient direction, and
1551 legacy behaviour is preserved as much as possible.
1554 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1556 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1558 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1560 *) Alternate chains certificate forgery
1562 During certificate verification, OpenSSL will attempt to find an
1563 alternative certificate chain if the first attempt to build such a chain
1564 fails. An error in the implementation of this logic can mean that an
1565 attacker could cause certain checks on untrusted certificates to be
1566 bypassed, such as the CA flag, enabling them to use a valid leaf
1567 certificate to act as a CA and "issue" an invalid certificate.
1569 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1573 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1575 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1576 incompatibility in the handling of HMAC. The previous ABI has now been
1580 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1582 *) Malformed ECParameters causes infinite loop
1584 When processing an ECParameters structure OpenSSL enters an infinite loop
1585 if the curve specified is over a specially malformed binary polynomial
1588 This can be used to perform denial of service against any
1589 system which processes public keys, certificate requests or
1590 certificates. This includes TLS clients and TLS servers with
1591 client authentication enabled.
1593 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1597 *) Exploitable out-of-bounds read in X509_cmp_time
1599 X509_cmp_time does not properly check the length of the ASN1_TIME
1600 string and can read a few bytes out of bounds. In addition,
1601 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1604 An attacker can use this to craft malformed certificates and CRLs of
1605 various sizes and potentially cause a segmentation fault, resulting in
1606 a DoS on applications that verify certificates or CRLs. TLS clients
1607 that verify CRLs are affected. TLS clients and servers with client
1608 authentication enabled may be affected if they use custom verification
1611 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1612 independently by Hanno Böck.
1616 *) PKCS7 crash with missing EnvelopedContent
1618 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1619 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1620 with missing content and trigger a NULL pointer dereference on parsing.
1622 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1623 structures from untrusted sources are affected. OpenSSL clients and
1624 servers are not affected.
1626 This issue was reported to OpenSSL by Michal Zalewski (Google).
1630 *) CMS verify infinite loop with unknown hash function
1632 When verifying a signedData message the CMS code can enter an infinite loop
1633 if presented with an unknown hash function OID. This can be used to perform
1634 denial of service against any system which verifies signedData messages using
1636 This issue was reported to OpenSSL by Johannes Bauer.
1640 *) Race condition handling NewSessionTicket
1642 If a NewSessionTicket is received by a multi-threaded client when attempting to
1643 reuse a previous ticket then a race condition can occur potentially leading to
1644 a double free of the ticket data.
1648 *) Only support 256-bit or stronger elliptic curves with the
1649 'ecdh_auto' setting (server) or by default (client). Of supported
1650 curves, prefer P-256 (both).
1653 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1655 *) ClientHello sigalgs DoS fix
1657 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1658 invalid signature algorithms extension a NULL pointer dereference will
1659 occur. This can be exploited in a DoS attack against the server.
1661 This issue was was reported to OpenSSL by David Ramos of Stanford
1664 [Stephen Henson and Matt Caswell]
1666 *) Multiblock corrupted pointer fix
1668 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1669 feature only applies on 64 bit x86 architecture platforms that support AES
1670 NI instructions. A defect in the implementation of "multiblock" can cause
1671 OpenSSL's internal write buffer to become incorrectly set to NULL when
1672 using non-blocking IO. Typically, when the user application is using a
1673 socket BIO for writing, this will only result in a failed connection.
1674 However if some other BIO is used then it is likely that a segmentation
1675 fault will be triggered, thus enabling a potential DoS attack.
1677 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1681 *) Segmentation fault in DTLSv1_listen fix
1683 The DTLSv1_listen function is intended to be stateless and processes the
1684 initial ClientHello from many peers. It is common for user code to loop
1685 over the call to DTLSv1_listen until a valid ClientHello is received with
1686 an associated cookie. A defect in the implementation of DTLSv1_listen means
1687 that state is preserved in the SSL object from one invocation to the next
1688 that can lead to a segmentation fault. Errors processing the initial
1689 ClientHello can trigger this scenario. An example of such an error could be
1690 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1693 This issue was reported to OpenSSL by Per Allansson.
1697 *) Segmentation fault in ASN1_TYPE_cmp fix
1699 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1700 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1701 certificate signature algorithm consistency this can be used to crash any
1702 certificate verification operation and exploited in a DoS attack. Any
1703 application which performs certificate verification is vulnerable including
1704 OpenSSL clients and servers which enable client authentication.
1708 *) Segmentation fault for invalid PSS parameters fix
1710 The signature verification routines will crash with a NULL pointer
1711 dereference if presented with an ASN.1 signature using the RSA PSS
1712 algorithm and invalid parameters. Since these routines are used to verify
1713 certificate signature algorithms this can be used to crash any
1714 certificate verification operation and exploited in a DoS attack. Any
1715 application which performs certificate verification is vulnerable including
1716 OpenSSL clients and servers which enable client authentication.
1718 This issue was was reported to OpenSSL by Brian Carpenter.
1722 *) ASN.1 structure reuse memory corruption fix
1724 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1725 memory corruption via an invalid write. Such reuse is and has been
1726 strongly discouraged and is believed to be rare.
1728 Applications that parse structures containing CHOICE or ANY DEFINED BY
1729 components may be affected. Certificate parsing (d2i_X509 and related
1730 functions) are however not affected. OpenSSL clients and servers are
1735 *) PKCS7 NULL pointer dereferences fix
1737 The PKCS#7 parsing code does not handle missing outer ContentInfo
1738 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1739 missing content and trigger a NULL pointer dereference on parsing.
1741 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1742 otherwise parse PKCS#7 structures from untrusted sources are
1743 affected. OpenSSL clients and servers are not affected.
1745 This issue was reported to OpenSSL by Michal Zalewski (Google).
1749 *) DoS via reachable assert in SSLv2 servers fix
1751 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1752 servers that both support SSLv2 and enable export cipher suites by sending
1753 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1755 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1756 (OpenSSL development team).
1760 *) Empty CKE with client auth and DHE fix
1762 If client auth is used then a server can seg fault in the event of a DHE
1763 ciphersuite being selected and a zero length ClientKeyExchange message
1764 being sent by the client. This could be exploited in a DoS attack.
1768 *) Handshake with unseeded PRNG fix
1770 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1771 with an unseeded PRNG. The conditions are:
1772 - The client is on a platform where the PRNG has not been seeded
1773 automatically, and the user has not seeded manually
1774 - A protocol specific client method version has been used (i.e. not
1775 SSL_client_methodv23)
1776 - A ciphersuite is used that does not require additional random data from
1777 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1779 If the handshake succeeds then the client random that has been used will
1780 have been generated from a PRNG with insufficient entropy and therefore the
1781 output may be predictable.
1783 For example using the following command with an unseeded openssl will
1784 succeed on an unpatched platform:
1786 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1790 *) Use After Free following d2i_ECPrivatekey error fix
1792 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1793 could cause a use after free condition. This, in turn, could cause a double
1794 free in several private key parsing functions (such as d2i_PrivateKey
1795 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1796 for applications that receive EC private keys from untrusted
1797 sources. This scenario is considered rare.
1799 This issue was discovered by the BoringSSL project and fixed in their
1804 *) X509_to_X509_REQ NULL pointer deref fix
1806 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1807 the certificate key is invalid. This function is rarely used in practice.
1809 This issue was discovered by Brian Carpenter.
1813 *) Removed the export ciphers from the DEFAULT ciphers
1816 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1818 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1819 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1820 So far those who have to target multiple platforms would compromise
1821 and argue that binary targeting say ARMv5 would still execute on
1822 ARMv8. "Universal" build resolves this compromise by providing
1823 near-optimal performance even on newer platforms.
1826 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1827 (other platforms pending).
1828 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1830 *) Add support for the SignedCertificateTimestampList certificate and
1831 OCSP response extensions from RFC6962.
1834 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1835 for corner cases. (Certain input points at infinity could lead to
1836 bogus results, with non-infinity inputs mapped to infinity too.)
1839 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1840 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1841 common cases are optimized and there still is room for further
1842 improvements. Vector Permutation AES for Altivec is also added.
1845 *) Add support for little-endian ppc64 Linux target.
1846 [Marcelo Cerri (IBM)]
1848 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1849 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1850 are optimized and there still is room for further improvements.
1851 Both 32- and 64-bit modes are supported.
1852 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1854 *) Improved ARMv7 NEON support.
1857 *) Support for SPARC Architecture 2011 crypto extensions, first
1858 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1859 SHA256/512, MD5, GHASH and modular exponentiation.
1860 [Andy Polyakov, David Miller]
1862 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1864 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1866 *) Support for new and upcoming Intel processors, including AVX2,
1867 BMI and SHA ISA extensions. This includes additional "stitched"
1868 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1871 This work was sponsored by Intel Corp.
1874 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1875 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1876 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1879 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1880 this fixes a limitation in previous versions of OpenSSL.
1883 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1884 MGF1 digest and OAEP label.
1887 *) Add EVP support for key wrapping algorithms, to avoid problems with
1888 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1889 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1890 algorithms and include tests cases.
1893 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1895 [Douglas E. Engert, Steve Henson]
1897 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1898 difference in days and seconds between two tm or ASN1_TIME structures.
1901 *) Add -rev test option to s_server to just reverse order of characters
1902 received by client and send back to server. Also prints an abbreviated
1903 summary of the connection parameters.
1906 *) New option -brief for s_client and s_server to print out a brief summary
1907 of connection parameters.
1910 *) Add callbacks for arbitrary TLS extensions.
1911 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1913 *) New option -crl_download in several openssl utilities to download CRLs
1914 from CRLDP extension in certificates.
1917 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
1920 *) New function X509_CRL_diff to generate a delta CRL from the difference
1921 of two full CRLs. Add support to "crl" utility.
1924 *) New functions to set lookup_crls function and to retrieve
1925 X509_STORE from X509_STORE_CTX.
1928 *) Print out deprecated issuer and subject unique ID fields in
1932 *) Extend OCSP I/O functions so they can be used for simple general purpose
1933 HTTP as well as OCSP. New wrapper function which can be used to download
1934 CRLs using the OCSP API.
1937 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
1940 *) SSL_CONF* functions. These provide a common framework for application
1941 configuration using configuration files or command lines.
1944 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
1945 message callback and prints the results. Needs compile time option
1946 "enable-ssl-trace". New options to s_client and s_server to enable
1950 *) New ctrl and macro to retrieve supported points extensions.
1951 Print out extension in s_server and s_client.
1954 *) New functions to retrieve certificate signature and signature
1958 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
1962 *) New Suite B modes for TLS code. These use and enforce the requirements
1963 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
1964 only use Suite B curves. The Suite B modes can be set by using the
1965 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
1968 *) New chain verification flags for Suite B levels of security. Check
1969 algorithms are acceptable when flags are set in X509_verify_cert.
1972 *) Make tls1_check_chain return a set of flags indicating checks passed
1973 by a certificate chain. Add additional tests to handle client
1974 certificates: checks for matching certificate type and issuer name
1978 *) If an attempt is made to use a signature algorithm not in the peer
1979 preference list abort the handshake. If client has no suitable
1980 signature algorithms in response to a certificate request do not
1981 use the certificate.
1984 *) If server EC tmp key is not in client preference list abort handshake.
1987 *) Add support for certificate stores in CERT structure. This makes it
1988 possible to have different stores per SSL structure or one store in
1989 the parent SSL_CTX. Include distinct stores for certificate chain
1990 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
1991 to build and store a certificate chain in CERT structure: returning
1992 an error if the chain cannot be built: this will allow applications
1993 to test if a chain is correctly configured.
1995 Note: if the CERT based stores are not set then the parent SSL_CTX
1996 store is used to retain compatibility with existing behaviour.
2000 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2001 mask based on the current session, check mask when sending client
2002 hello and checking the requested ciphersuite.
2005 *) New ctrls to retrieve and set certificate types in a certificate
2006 request message. Print out received values in s_client. If certificate
2007 types is not set with custom values set sensible values based on
2008 supported signature algorithms.
2011 *) Support for distinct client and server supported signature algorithms.
2014 *) Add certificate callback. If set this is called whenever a certificate
2015 is required by client or server. An application can decide which
2016 certificate chain to present based on arbitrary criteria: for example
2017 supported signature algorithms. Add very simple example to s_server.
2018 This fixes many of the problems and restrictions of the existing client
2019 certificate callback: for example you can now clear an existing
2020 certificate and specify the whole chain.
2023 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2024 the certificate can be used for (if anything). Set valid_flags field
2025 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2026 to have similar checks in it.
2028 Add new "cert_flags" field to CERT structure and include a "strict mode".
2029 This enforces some TLS certificate requirements (such as only permitting
2030 certificate signature algorithms contained in the supported algorithms
2031 extension) which some implementations ignore: this option should be used
2032 with caution as it could cause interoperability issues.
2035 *) Update and tidy signature algorithm extension processing. Work out
2036 shared signature algorithms based on preferences and peer algorithms
2037 and print them out in s_client and s_server. Abort handshake if no
2038 shared signature algorithms.
2041 *) Add new functions to allow customised supported signature algorithms
2042 for SSL and SSL_CTX structures. Add options to s_client and s_server
2046 *) New function SSL_certs_clear() to delete all references to certificates
2047 from an SSL structure. Before this once a certificate had been added
2048 it couldn't be removed.
2051 *) Integrate hostname, email address and IP address checking with certificate
2052 verification. New verify options supporting checking in openssl utility.
2055 *) Fixes and wildcard matching support to hostname and email checking
2056 functions. Add manual page.
2057 [Florian Weimer (Red Hat Product Security Team)]
2059 *) New functions to check a hostname email or IP address against a
2060 certificate. Add options x509 utility to print results of checks against
2064 *) Fix OCSP checking.
2065 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2067 *) Initial experimental support for explicitly trusted non-root CAs.
2068 OpenSSL still tries to build a complete chain to a root but if an
2069 intermediate CA has a trust setting included that is used. The first
2070 setting is used: whether to trust (e.g., -addtrust option to the x509
2074 *) Add -trusted_first option which attempts to find certificates in the
2075 trusted store even if an untrusted chain is also supplied.
2078 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2079 platform support for Linux and Android.
2082 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2085 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2086 When in FIPS mode the approved implementations are used as normal,
2087 when not in FIPS mode the internal unapproved versions are used instead.
2088 This means that the FIPS capable OpenSSL isn't forced to use the
2089 (often lower performance) FIPS implementations outside FIPS mode.
2092 *) Transparently support X9.42 DH parameters when calling
2093 PEM_read_bio_DHparameters. This means existing applications can handle
2094 the new parameter format automatically.
2097 *) Initial experimental support for X9.42 DH parameter format: mainly
2098 to support use of 'q' parameter for RFC5114 parameters.
2101 *) Add DH parameters from RFC5114 including test data to dhtest.
2104 *) Support for automatic EC temporary key parameter selection. If enabled
2105 the most preferred EC parameters are automatically used instead of
2106 hardcoded fixed parameters. Now a server just has to call:
2107 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2108 support ECDH and use the most appropriate parameters.
2111 *) Enhance and tidy EC curve and point format TLS extension code. Use
2112 static structures instead of allocation if default values are used.
2113 New ctrls to set curves we wish to support and to retrieve shared curves.
2114 Print out shared curves in s_server. New options to s_server and s_client
2115 to set list of supported curves.
2118 *) New ctrls to retrieve supported signature algorithms and
2119 supported curve values as an array of NIDs. Extend openssl utility
2120 to print out received values.
2123 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2124 between NIDs and the more common NIST names such as "P-256". Enhance
2125 ecparam utility and ECC method to recognise the NIST names for curves.
2128 *) Enhance SSL/TLS certificate chain handling to support different
2129 chains for each certificate instead of one chain in the parent SSL_CTX.
2132 *) Support for fixed DH ciphersuite client authentication: where both
2133 server and client use DH certificates with common parameters.
2136 *) Support for fixed DH ciphersuites: those requiring DH server
2140 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2142 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2143 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2144 X509_CINF_get_signature were reverted post internal team review.
2146 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2148 *) Build fixes for the Windows and OpenVMS platforms
2149 [Matt Caswell and Richard Levitte]
2151 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2153 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2154 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2155 dereference. This could lead to a Denial Of Service attack. Thanks to
2156 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2160 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2161 dtls1_buffer_record function under certain conditions. In particular this
2162 could occur if an attacker sent repeated DTLS records with the same
2163 sequence number but for the next epoch. The memory leak could be exploited
2164 by an attacker in a Denial of Service attack through memory exhaustion.
2165 Thanks to Chris Mueller for reporting this issue.
2169 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2170 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2171 method would be set to NULL which could later result in a NULL pointer
2172 dereference. Thanks to Frank Schmirler for reporting this issue.
2176 *) Abort handshake if server key exchange message is omitted for ephemeral
2179 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2180 reporting this issue.
2184 *) Remove non-export ephemeral RSA code on client and server. This code
2185 violated the TLS standard by allowing the use of temporary RSA keys in
2186 non-export ciphersuites and could be used by a server to effectively
2187 downgrade the RSA key length used to a value smaller than the server
2188 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2189 INRIA or reporting this issue.
2193 *) Fixed issue where DH client certificates are accepted without verification.
2194 An OpenSSL server will accept a DH certificate for client authentication
2195 without the certificate verify message. This effectively allows a client to
2196 authenticate without the use of a private key. This only affects servers
2197 which trust a client certificate authority which issues certificates
2198 containing DH keys: these are extremely rare and hardly ever encountered.
2199 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2204 *) Ensure that the session ID context of an SSL is updated when its
2205 SSL_CTX is updated via SSL_set_SSL_CTX.
2207 The session ID context is typically set from the parent SSL_CTX,
2208 and can vary with the CTX.
2211 *) Fix various certificate fingerprint issues.
2213 By using non-DER or invalid encodings outside the signed portion of a
2214 certificate the fingerprint can be changed without breaking the signature.
2215 Although no details of the signed portion of the certificate can be changed
2216 this can cause problems with some applications: e.g. those using the
2217 certificate fingerprint for blacklists.
2219 1. Reject signatures with non zero unused bits.
2221 If the BIT STRING containing the signature has non zero unused bits reject
2222 the signature. All current signature algorithms require zero unused bits.
2224 2. Check certificate algorithm consistency.
2226 Check the AlgorithmIdentifier inside TBS matches the one in the
2227 certificate signature. NB: this will result in signature failure
2228 errors for some broken certificates.
2230 Thanks to Konrad Kraszewski from Google for reporting this issue.
2232 3. Check DSA/ECDSA signatures use DER.
2234 Re-encode DSA/ECDSA signatures and compare with the original received
2235 signature. Return an error if there is a mismatch.
2237 This will reject various cases including garbage after signature
2238 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2239 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2240 (negative or with leading zeroes).
2242 Further analysis was conducted and fixes were developed by Stephen Henson
2243 of the OpenSSL core team.
2248 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2249 results on some platforms, including x86_64. This bug occurs at random
2250 with a very low probability, and is not known to be exploitable in any
2251 way, though its exact impact is difficult to determine. Thanks to Pieter
2252 Wuille (Blockstream) who reported this issue and also suggested an initial
2253 fix. Further analysis was conducted by the OpenSSL development team and
2254 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2255 the OpenSSL core team.
2259 *) Do not resume sessions on the server if the negotiated protocol
2260 version does not match the session's version. Resuming with a different
2261 version, while not strictly forbidden by the RFC, is of questionable
2262 sanity and breaks all known clients.
2263 [David Benjamin, Emilia Käsper]
2265 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2266 early CCS messages during renegotiation. (Note that because
2267 renegotiation is encrypted, this early CCS was not exploitable.)
2270 *) Tighten client-side session ticket handling during renegotiation:
2271 ensure that the client only accepts a session ticket if the server sends
2272 the extension anew in the ServerHello. Previously, a TLS client would
2273 reuse the old extension state and thus accept a session ticket if one was
2274 announced in the initial ServerHello.
2276 Similarly, ensure that the client requires a session ticket if one
2277 was advertised in the ServerHello. Previously, a TLS client would
2278 ignore a missing NewSessionTicket message.
2281 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2283 *) SRTP Memory Leak.
2285 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2286 sends a carefully crafted handshake message, to cause OpenSSL to fail
2287 to free up to 64k of memory causing a memory leak. This could be
2288 exploited in a Denial Of Service attack. This issue affects OpenSSL
2289 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2290 whether SRTP is used or configured. Implementations of OpenSSL that
2291 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2293 The fix was developed by the OpenSSL team.
2297 *) Session Ticket Memory Leak.
2299 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2300 integrity of that ticket is first verified. In the event of a session
2301 ticket integrity check failing, OpenSSL will fail to free memory
2302 causing a memory leak. By sending a large number of invalid session
2303 tickets an attacker could exploit this issue in a Denial Of Service
2308 *) Build option no-ssl3 is incomplete.
2310 When OpenSSL is configured with "no-ssl3" as a build option, servers
2311 could accept and complete a SSL 3.0 handshake, and clients could be
2312 configured to send them.
2314 [Akamai and the OpenSSL team]
2316 *) Add support for TLS_FALLBACK_SCSV.
2317 Client applications doing fallback retries should call
2318 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2320 [Adam Langley, Bodo Moeller]
2322 *) Add additional DigestInfo checks.
2324 Re-encode DigestInto in DER and check against the original when
2325 verifying RSA signature: this will reject any improperly encoded
2326 DigestInfo structures.
2328 Note: this is a precautionary measure and no attacks are currently known.
2332 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2334 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2335 SRP code can be overrun an internal buffer. Add sanity check that
2336 g, A, B < N to SRP code.
2338 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2339 Group for discovering this issue.
2343 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2344 TLS 1.0 instead of higher protocol versions when the ClientHello message
2345 is badly fragmented. This allows a man-in-the-middle attacker to force a
2346 downgrade to TLS 1.0 even if both the server and the client support a
2347 higher protocol version, by modifying the client's TLS records.
2349 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2350 researching this issue.
2354 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2355 to a denial of service attack. A malicious server can crash the client
2356 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2357 ciphersuite and sending carefully crafted handshake messages.
2359 Thanks to Felix Gröbert (Google) for discovering and researching this
2364 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2365 to leak memory. This can be exploited through a Denial of Service attack.
2366 Thanks to Adam Langley for discovering and researching this issue.
2370 *) An attacker can force openssl to consume large amounts of memory whilst
2371 processing DTLS handshake messages. This can be exploited through a
2372 Denial of Service attack.
2373 Thanks to Adam Langley for discovering and researching this issue.
2377 *) An attacker can force an error condition which causes openssl to crash
2378 whilst processing DTLS packets due to memory being freed twice. This
2379 can be exploited through a Denial of Service attack.
2380 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2385 *) If a multithreaded client connects to a malicious server using a resumed
2386 session and the server sends an ec point format extension it could write
2387 up to 255 bytes to freed memory.
2389 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2394 *) A malicious server can crash an OpenSSL client with a null pointer
2395 dereference (read) by specifying an SRP ciphersuite even though it was not
2396 properly negotiated with the client. This can be exploited through a
2397 Denial of Service attack.
2399 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2400 discovering and researching this issue.
2404 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2405 X509_name_oneline, X509_name_print_ex et al. to leak some information
2406 from the stack. Applications may be affected if they echo pretty printing
2407 output to the attacker.
2409 Thanks to Ivan Fratric (Google) for discovering this issue.
2411 [Emilia Käsper, and Steve Henson]
2413 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2414 for corner cases. (Certain input points at infinity could lead to
2415 bogus results, with non-infinity inputs mapped to infinity too.)
2418 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2420 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2421 handshake can force the use of weak keying material in OpenSSL
2422 SSL/TLS clients and servers.
2424 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2425 researching this issue. (CVE-2014-0224)
2426 [KIKUCHI Masashi, Steve Henson]
2428 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2429 OpenSSL DTLS client the code can be made to recurse eventually crashing
2432 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2434 [Imre Rad, Steve Henson]
2436 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2437 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2438 client or server. This is potentially exploitable to run arbitrary
2439 code on a vulnerable client or server.
2441 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2442 [Jüri Aedla, Steve Henson]
2444 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2445 are subject to a denial of service attack.
2447 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2448 this issue. (CVE-2014-3470)
2449 [Felix Gröbert, Ivan Fratric, Steve Henson]
2451 *) Harmonize version and its documentation. -f flag is used to display
2453 [mancha <mancha1@zoho.com>]
2455 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2456 in i2d_ECPrivateKey.
2457 [mancha <mancha1@zoho.com>]
2459 *) Fix some double frees. These are not thought to be exploitable.
2460 [mancha <mancha1@zoho.com>]
2462 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2464 *) A missing bounds check in the handling of the TLS heartbeat extension
2465 can be used to reveal up to 64k of memory to a connected client or
2468 Thanks for Neel Mehta of Google Security for discovering this bug and to
2469 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2470 preparing the fix (CVE-2014-0160)
2471 [Adam Langley, Bodo Moeller]
2473 *) Fix for the attack described in the paper "Recovering OpenSSL
2474 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2475 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2476 http://eprint.iacr.org/2014/140
2478 Thanks to Yuval Yarom and Naomi Benger for discovering this
2479 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2480 [Yuval Yarom and Naomi Benger]
2482 *) TLS pad extension: draft-agl-tls-padding-03
2484 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2485 TLS client Hello record length value would otherwise be > 255 and
2486 less that 512 pad with a dummy extension containing zeroes so it
2487 is at least 512 bytes long.
2489 [Adam Langley, Steve Henson]
2491 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2493 *) Fix for TLS record tampering bug. A carefully crafted invalid
2494 handshake could crash OpenSSL with a NULL pointer exception.
2495 Thanks to Anton Johansson for reporting this issues.
2498 *) Keep original DTLS digest and encryption contexts in retransmission
2499 structures so we can use the previous session parameters if they need
2500 to be resent. (CVE-2013-6450)
2503 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2504 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2505 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2506 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2507 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2508 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2509 [Rob Stradling, Adam Langley]
2511 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2513 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2514 supporting platforms or when small records were transferred.
2515 [Andy Polyakov, Steve Henson]
2517 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2519 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2521 This addresses the flaw in CBC record processing discovered by
2522 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2523 at: http://www.isg.rhul.ac.uk/tls/
2525 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2526 Security Group at Royal Holloway, University of London
2527 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2528 Emilia Käsper for the initial patch.
2530 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2532 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2533 ciphersuites which can be exploited in a denial of service attack.
2534 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2535 and detecting this bug and to Wolfgang Ettlinger
2536 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2540 *) Return an error when checking OCSP signatures when key is NULL.
2541 This fixes a DoS attack. (CVE-2013-0166)
2544 *) Make openssl verify return errors.
2545 [Chris Palmer <palmer@google.com> and Ben Laurie]
2547 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2548 the right response is stapled. Also change SSL_get_certificate()
2549 so it returns the certificate actually sent.
2550 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2551 [Rob Stradling <rob.stradling@comodo.com>]
2553 *) Fix possible deadlock when decoding public keys.
2556 *) Don't use TLS 1.0 record version number in initial client hello
2560 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2562 *) Sanity check record length before skipping explicit IV in TLS
2563 1.2, 1.1 and DTLS to fix DoS attack.
2565 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2566 fuzzing as a service testing platform.
2570 *) Initialise tkeylen properly when encrypting CMS messages.
2571 Thanks to Solar Designer of Openwall for reporting this issue.
2574 *) In FIPS mode don't try to use composite ciphers as they are not
2578 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2580 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2581 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2582 mean any application compiled against OpenSSL 1.0.0 headers setting
2583 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2584 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2585 0x10000000L Any application which was previously compiled against
2586 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2587 will need to be recompiled as a result. Letting be results in
2588 inability to disable specifically TLS 1.1 and in client context,
2589 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2592 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2593 disable just protocol X, but all protocols above X *if* there are
2594 protocols *below* X still enabled. In more practical terms it means
2595 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2596 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2597 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2601 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2603 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2604 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2605 in CRYPTO_realloc_clean.
2607 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2608 issue and to Adam Langley <agl@chromium.org> for fixing it.
2610 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2612 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2615 *) Workarounds for some broken servers that "hang" if a client hello
2616 record length exceeds 255 bytes.
2618 1. Do not use record version number > TLS 1.0 in initial client
2619 hello: some (but not all) hanging servers will now work.
2620 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2621 the number of ciphers sent in the client hello. This should be
2622 set to an even number, such as 50, for example by passing:
2623 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2624 Most broken servers should now work.
2625 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2626 TLS 1.2 client support entirely.
2629 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2632 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2634 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2635 STRING form instead of a DigestInfo.
2638 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2639 and the RSA_sign/RSA_verify functions. This was made more apparent when
2640 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2641 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2642 the correct format in RSA_verify so both forms transparently work.
2645 *) Some servers which support TLS 1.0 can choke if we initially indicate
2646 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2647 encrypted premaster secret. As a workaround use the maximum permitted
2648 client version in client hello, this should keep such servers happy
2649 and still work with previous versions of OpenSSL.
2652 *) Add support for TLS/DTLS heartbeats.
2653 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2655 *) Add support for SCTP.
2656 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2658 *) Improved PRNG seeding for VOS.
2659 [Paul Green <Paul.Green@stratus.com>]
2661 *) Extensive assembler packs updates, most notably:
2663 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2664 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2665 - x86_64: bit-sliced AES implementation;
2666 - ARM: NEON support, contemporary platforms optimizations;
2667 - s390x: z196 support;
2668 - *: GHASH and GF(2^m) multiplication implementations;
2672 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2673 (removal of unnecessary code)
2674 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2676 *) Add TLS key material exporter from RFC 5705.
2679 *) Add DTLS-SRTP negotiation from RFC 5764.
2682 *) Add Next Protocol Negotiation,
2683 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2684 disabled with a no-npn flag to config or Configure. Code donated
2686 [Adam Langley <agl@google.com> and Ben Laurie]
2688 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2689 NIST-P256, NIST-P521, with constant-time single point multiplication on
2690 typical inputs. Compiler support for the nonstandard type __uint128_t is
2691 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2692 Code made available under Apache License version 2.0.
2694 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2695 line to include this in your build of OpenSSL, and run "make depend" (or
2696 "make update"). This enables the following EC_METHODs:
2698 EC_GFp_nistp224_method()
2699 EC_GFp_nistp256_method()
2700 EC_GFp_nistp521_method()
2702 EC_GROUP_new_by_curve_name() will automatically use these (while
2703 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2705 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2707 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2708 all platforms. Move ssize_t definition from e_os.h to the public
2709 header file e_os2.h as it now appears in public header file cms.h
2712 *) New -sigopt option to the ca, req and x509 utilities. Additional
2713 signature parameters can be passed using this option and in
2717 *) Add RSA PSS signing function. This will generate and set the
2718 appropriate AlgorithmIdentifiers for PSS based on those in the
2719 corresponding EVP_MD_CTX structure. No application support yet.
2722 *) Support for companion algorithm specific ASN1 signing routines.
2723 New function ASN1_item_sign_ctx() signs a pre-initialised
2724 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2725 the appropriate parameters.
2728 *) Add new algorithm specific ASN1 verification initialisation function
2729 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2730 handling will be the same no matter what EVP_PKEY_METHOD is used.
2731 Add a PSS handler to support verification of PSS signatures: checked
2732 against a number of sample certificates.
2735 *) Add signature printing for PSS. Add PSS OIDs.
2736 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2738 *) Add algorithm specific signature printing. An individual ASN1 method
2739 can now print out signatures instead of the standard hex dump.
2741 More complex signatures (e.g. PSS) can print out more meaningful
2742 information. Include DSA version that prints out the signature
2746 *) Password based recipient info support for CMS library: implementing
2750 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2751 neatly separates the code into cipher and PBE sections and is required
2752 for some algorithms that split PBES2 into separate pieces (such as
2753 password based CMS).
2756 *) Session-handling fixes:
2757 - Fix handling of connections that are resuming with a session ID,
2758 but also support Session Tickets.
2759 - Fix a bug that suppressed issuing of a new ticket if the client
2760 presented a ticket with an expired session.
2761 - Try to set the ticket lifetime hint to something reasonable.
2762 - Make tickets shorter by excluding irrelevant information.
2763 - On the client side, don't ignore renewed tickets.
2764 [Adam Langley, Bodo Moeller (Google)]
2766 *) Fix PSK session representation.
2769 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2771 This work was sponsored by Intel.
2774 *) Add GCM support to TLS library. Some custom code is needed to split
2775 the IV between the fixed (from PRF) and explicit (from TLS record)
2776 portions. This adds all GCM ciphersuites supported by RFC5288 and
2777 RFC5289. Generalise some AES* cipherstrings to include GCM and
2778 add a special AESGCM string for GCM only.
2781 *) Expand range of ctrls for AES GCM. Permit setting invocation
2782 field on decrypt and retrieval of invocation field only on encrypt.
2785 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2786 As required by RFC5289 these ciphersuites cannot be used if for
2787 versions of TLS earlier than 1.2.
2790 *) For FIPS capable OpenSSL interpret a NULL default public key method
2791 as unset and return the appropriate default but do *not* set the default.
2792 This means we can return the appropriate method in applications that
2793 switch between FIPS and non-FIPS modes.
2796 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2797 ENGINE is used then we cannot handle that in the FIPS module so we
2798 keep original code iff non-FIPS operations are allowed.
2801 *) Add -attime option to openssl utilities.
2802 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2804 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2807 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2808 FIPS EC methods unconditionally for now.
2811 *) New build option no-ec2m to disable characteristic 2 code.
2814 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2815 all cases can be covered as some introduce binary incompatibilities.
2818 *) Redirect RSA operations to FIPS module including keygen,
2819 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2822 *) Add similar low level API blocking to ciphers.
2825 *) Low level digest APIs are not approved in FIPS mode: any attempt
2826 to use these will cause a fatal error. Applications that *really* want
2827 to use them can use the private_* version instead.
2830 *) Redirect cipher operations to FIPS module for FIPS builds.
2833 *) Redirect digest operations to FIPS module for FIPS builds.
2836 *) Update build system to add "fips" flag which will link in fipscanister.o
2837 for static and shared library builds embedding a signature if needed.
2840 *) Output TLS supported curves in preference order instead of numerical
2841 order. This is currently hardcoded for the highest order curves first.
2842 This should be configurable so applications can judge speed vs strength.
2845 *) Add TLS v1.2 server support for client authentication.
2848 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2852 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2853 FIPS modules versions.
2856 *) Add TLS v1.2 client side support for client authentication. Keep cache
2857 of handshake records longer as we don't know the hash algorithm to use
2858 until after the certificate request message is received.
2861 *) Initial TLS v1.2 client support. Add a default signature algorithms
2862 extension including all the algorithms we support. Parse new signature
2863 format in client key exchange. Relax some ECC signing restrictions for
2864 TLS v1.2 as indicated in RFC5246.
2867 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2868 to new signature format when needed using client digest preference.
2869 All server ciphersuites should now work correctly in TLS v1.2. No client
2870 support yet and no support for client certificates.
2873 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2874 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2875 ciphersuites. At present only RSA key exchange ciphersuites work with
2876 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2877 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2878 and version checking.
2881 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2882 with this defined it will not be affected by any changes to ssl internal
2883 structures. Add several utility functions to allow openssl application
2884 to work with OPENSSL_NO_SSL_INTERN defined.
2888 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2890 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2893 *) Permit abbreviated handshakes when renegotiating using the function
2894 SSL_renegotiate_abbreviated().
2895 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2897 *) Add call to ENGINE_register_all_complete() to
2898 ENGINE_load_builtin_engines(), so some implementations get used
2899 automatically instead of needing explicit application support.
2902 *) Add support for TLS key exporter as described in RFC5705.
2903 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2905 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2906 a few changes are required:
2908 Add SSL_OP_NO_TLSv1_1 flag.
2909 Add TLSv1_1 methods.
2910 Update version checking logic to handle version 1.1.
2911 Add explicit IV handling (ported from DTLS code).
2912 Add command line options to s_client/s_server.
2915 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
2917 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2918 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2919 content decryption and always return the same error. Note: this attack
2920 needs on average 2^20 messages so it only affects automated senders. The
2921 old behaviour can be re-enabled in the CMS code by setting the
2922 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2923 an MMA defence is not necessary.
2924 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2925 this issue. (CVE-2012-0884)
2928 *) Fix CVE-2011-4619: make sure we really are receiving a
2929 client hello before rejecting multiple SGC restarts. Thanks to
2930 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2933 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
2935 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2936 Thanks to Antonio Martin, Enterprise Secure Access Research and
2937 Development, Cisco Systems, Inc. for discovering this bug and
2938 preparing a fix. (CVE-2012-0050)
2941 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
2943 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2944 of the Vaudenay padding oracle attack on CBC mode encryption
2945 which enables an efficient plaintext recovery attack against
2946 the OpenSSL implementation of DTLS. Their attack exploits timing
2947 differences arising during decryption processing. A research
2948 paper describing this attack can be found at:
2949 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2950 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2951 Security Group at Royal Holloway, University of London
2952 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2953 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2954 for preparing the fix. (CVE-2011-4108)
2955 [Robin Seggelmann, Michael Tuexen]
2957 *) Clear bytes used for block padding of SSL 3.0 records.
2959 [Adam Langley (Google)]
2961 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2962 Kadianakis <desnacked@gmail.com> for discovering this issue and
2963 Adam Langley for preparing the fix. (CVE-2011-4619)
2964 [Adam Langley (Google)]
2966 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
2967 [Andrey Kulikov <amdeich@gmail.com>]
2969 *) Prevent malformed RFC3779 data triggering an assertion failure.
2970 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2971 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2972 [Rob Austein <sra@hactrn.net>]
2974 *) Improved PRNG seeding for VOS.
2975 [Paul Green <Paul.Green@stratus.com>]
2977 *) Fix ssl_ciph.c set-up race.
2978 [Adam Langley (Google)]
2980 *) Fix spurious failures in ecdsatest.c.
2981 [Emilia Käsper (Google)]
2983 *) Fix the BIO_f_buffer() implementation (which was mixing different
2984 interpretations of the '..._len' fields).
2985 [Adam Langley (Google)]
2987 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2988 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2989 threads won't reuse the same blinding coefficients.
2991 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2992 lock to call BN_BLINDING_invert_ex, and avoids one use of
2993 BN_BLINDING_update for each BN_BLINDING structure (previously,
2994 the last update always remained unused).
2995 [Emilia Käsper (Google)]
2997 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
2998 [Bob Buckholz (Google)]
3000 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3002 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3003 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3004 [Kaspar Brand <ossl@velox.ch>]
3006 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3007 for multi-threaded use of ECDH. (CVE-2011-3210)
3008 [Adam Langley (Google)]
3010 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3013 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3014 signature public key algorithm by using OID xref utilities instead.
3015 Before this you could only use some ECC ciphersuites with SHA1 only.
3018 *) Add protection against ECDSA timing attacks as mentioned in the paper
3019 by Billy Bob Brumley and Nicola Tuveri, see:
3021 http://eprint.iacr.org/2011/232.pdf
3023 [Billy Bob Brumley and Nicola Tuveri]
3025 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3027 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3028 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3030 *) Fix bug in string printing code: if *any* escaping is enabled we must
3031 escape the escape character (backslash) or the resulting string is
3035 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3037 *) Disable code workaround for ancient and obsolete Netscape browsers
3038 and servers: an attacker can use it in a ciphersuite downgrade attack.
3039 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3042 *) Fixed J-PAKE implementation error, originally discovered by
3043 Sebastien Martini, further info and confirmation from Stefan
3044 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3047 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3049 *) Fix extension code to avoid race conditions which can result in a buffer
3050 overrun vulnerability: resumed sessions must not be modified as they can
3051 be shared by multiple threads. CVE-2010-3864
3054 *) Fix WIN32 build system to correctly link an ENGINE directory into
3058 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3060 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3062 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3064 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3066 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3067 context. The operation can be customised via the ctrl mechanism in
3068 case ENGINEs want to include additional functionality.
3071 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3074 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3075 output hashes compatible with older versions of OpenSSL.
3076 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3078 *) Fix compression algorithm handling: if resuming a session use the
3079 compression algorithm of the resumed session instead of determining
3080 it from client hello again. Don't allow server to change algorithm.
3083 *) Add load_crls() function to apps tidying load_certs() too. Add option
3084 to verify utility to allow additional CRLs to be included.
3087 *) Update OCSP request code to permit adding custom headers to the request:
3088 some responders need this.
3091 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3093 [Julia Lawall <julia@diku.dk>]
3095 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3096 needlessly dereferenced structures, used obsolete functions and
3097 didn't handle all updated verify codes correctly.
3100 *) Disable MD2 in the default configuration.
3103 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3104 indicate the initial BIO being pushed or popped. This makes it possible
3105 to determine whether the BIO is the one explicitly called or as a result
3106 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3107 it handles reference counts correctly and doesn't zero out the I/O bio
3108 when it is not being explicitly popped. WARNING: applications which
3109 included workarounds for the old buggy behaviour will need to be modified
3110 or they could free up already freed BIOs.
3113 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3114 renaming to all platforms (within the 0.9.8 branch, this was
3115 done conditionally on Netware platforms to avoid a name clash).
3116 [Guenter <lists@gknw.net>]
3118 *) Add ECDHE and PSK support to DTLS.
3119 [Michael Tuexen <tuexen@fh-muenster.de>]
3121 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3125 *) Add "missing" function EVP_MD_flags() (without this the only way to
3126 retrieve a digest flags is by accessing the structure directly. Update
3127 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3128 or cipher is registered as in the "from" argument. Print out all
3129 registered digests in the dgst usage message instead of manually
3130 attempting to work them out.
3133 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3134 this allows the use of compression and extensions. Change default cipher
3135 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3136 by default unless an application cipher string requests it.
3139 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3140 key ids to find matching certificates and keys but some PKCS#12 files
3141 don't follow the (somewhat unwritten) rules and this strategy fails.
3142 Now just gather all certificates together and the first private key
3143 then look for the first certificate that matches the key.
3146 *) Support use of registered digest and cipher names for dgst and cipher
3147 commands instead of having to add each one as a special case. So now
3154 openssl dgst -sha256 foo
3156 and this works for ENGINE based algorithms too.
3160 *) Update Gost ENGINE to support parameter files.
3161 [Victor B. Wagner <vitus@cryptocom.ru>]
3163 *) Support GeneralizedTime in ca utility.
3164 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3166 *) Enhance the hash format used for certificate directory links. The new
3167 form uses the canonical encoding (meaning equivalent names will work
3168 even if they aren't identical) and uses SHA1 instead of MD5. This form
3169 is incompatible with the older format and as a result c_rehash should
3170 be used to rebuild symbolic links.
3173 *) Make PKCS#8 the default write format for private keys, replacing the
3174 traditional format. This form is standardised, more secure and doesn't
3175 include an implicit MD5 dependency.
3178 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3179 committed to OpenSSL should pass this lot as a minimum.
3182 *) Add session ticket override functionality for use by EAP-FAST.
3183 [Jouni Malinen <j@w1.fi>]
3185 *) Modify HMAC functions to return a value. Since these can be implemented
3186 in an ENGINE errors can occur.
3189 *) Type-checked OBJ_bsearch_ex.
3192 *) Type-checked OBJ_bsearch. Also some constification necessitated
3193 by type-checking. Still to come: TXT_DB, bsearch(?),
3194 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3198 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3199 seconds to a tm structure directly, instead of going through OS
3200 specific date routines. This avoids any issues with OS routines such
3201 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3202 and X509_time_adj_ex() to cover the extended range. The existing
3203 X509_time_adj() is still usable and will no longer have any date issues.
3206 *) Delta CRL support. New use deltas option which will attempt to locate
3207 and search any appropriate delta CRLs available.
3209 This work was sponsored by Google.
3212 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3213 code and add additional score elements. Validate alternate CRL paths
3214 as part of the CRL checking and indicate a new error "CRL path validation
3215 error" in this case. Applications wanting additional details can use
3216 the verify callback and check the new "parent" field. If this is not
3217 NULL CRL path validation is taking place. Existing applications won't
3218 see this because it requires extended CRL support which is off by
3221 This work was sponsored by Google.
3224 *) Support for freshest CRL extension.
3226 This work was sponsored by Google.
3229 *) Initial indirect CRL support. Currently only supported in the CRLs
3230 passed directly and not via lookup. Process certificate issuer
3231 CRL entry extension and lookup CRL entries by bother issuer name
3232 and serial number. Check and process CRL issuer entry in IDP extension.
3234 This work was sponsored by Google.
3237 *) Add support for distinct certificate and CRL paths. The CRL issuer
3238 certificate is validated separately in this case. Only enabled if
3239 an extended CRL support flag is set: this flag will enable additional
3240 CRL functionality in future.
3242 This work was sponsored by Google.
3245 *) Add support for policy mappings extension.
3247 This work was sponsored by Google.
3250 *) Fixes to pathlength constraint, self issued certificate handling,
3251 policy processing to align with RFC3280 and PKITS tests.
3253 This work was sponsored by Google.
3256 *) Support for name constraints certificate extension. DN, email, DNS
3257 and URI types are currently supported.
3259 This work was sponsored by Google.
3262 *) To cater for systems that provide a pointer-based thread ID rather
3263 than numeric, deprecate the current numeric thread ID mechanism and
3264 replace it with a structure and associated callback type. This
3265 mechanism allows a numeric "hash" to be extracted from a thread ID in
3266 either case, and on platforms where pointers are larger than 'long',
3267 mixing is done to help ensure the numeric 'hash' is usable even if it
3268 can't be guaranteed unique. The default mechanism is to use "&errno"
3269 as a pointer-based thread ID to distinguish between threads.
3271 Applications that want to provide their own thread IDs should now use
3272 CRYPTO_THREADID_set_callback() to register a callback that will call
3273 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3275 Note that ERR_remove_state() is now deprecated, because it is tied
3276 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3277 to free the current thread's error state should be replaced by
3278 ERR_remove_thread_state(NULL).
3280 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3281 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3282 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3283 application was previously providing a numeric thread callback that
3284 was inappropriate for distinguishing threads, then uniqueness might
3285 have been obtained with &errno that happened immediately in the
3286 intermediate development versions of OpenSSL; this is no longer the
3287 case, the numeric thread callback will now override the automatic use
3289 [Geoff Thorpe, with help from Bodo Moeller]
3291 *) Initial support for different CRL issuing certificates. This covers a
3292 simple case where the self issued certificates in the chain exist and
3293 the real CRL issuer is higher in the existing chain.
3295 This work was sponsored by Google.
3298 *) Removed effectively defunct crypto/store from the build.
3301 *) Revamp of STACK to provide stronger type-checking. Still to come:
3302 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3303 ASN1_STRING, CONF_VALUE.
3306 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3307 RAM on SSL connections. This option can save about 34k per idle SSL.
3310 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3311 STACK, TXT_DB, bsearch, qsort.
3314 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3315 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3316 support for data, signedData, compressedData, digestedData and
3317 encryptedData, envelopedData types included. Scripts to check against
3318 RFC4134 examples draft and interop and consistency checks of many
3319 content types and variants.
3322 *) Add options to enc utility to support use of zlib compression BIO.
3325 *) Extend mk1mf to support importing of options and assembly language
3326 files from Configure script, currently only included in VC-WIN32.
3327 The assembly language rules can now optionally generate the source
3328 files from the associated perl scripts.
3331 *) Implement remaining functionality needed to support GOST ciphersuites.
3332 Interop testing has been performed using CryptoPro implementations.
3333 [Victor B. Wagner <vitus@cryptocom.ru>]
3335 *) s390x assembler pack.
3338 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3342 *) Implement Opaque PRF Input TLS extension as specified in
3343 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3344 official specification yet and no extension type assignment by
3345 IANA exists, this extension (for now) will have to be explicitly
3346 enabled when building OpenSSL by providing the extension number
3347 to use. For example, specify an option
3349 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3351 to the "config" or "Configure" script to enable the extension,
3352 assuming extension number 0x9527 (which is a completely arbitrary
3353 and unofficial assignment based on the MD5 hash of the Internet
3354 Draft). Note that by doing so, you potentially lose
3355 interoperability with other TLS implementations since these might
3356 be using the same extension number for other purposes.
3358 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3359 opaque PRF input value to use in the handshake. This will create
3360 an interal copy of the length-'len' string at 'src', and will
3361 return non-zero for success.
3363 To get more control and flexibility, provide a callback function
3366 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3367 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3371 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3374 Callback function 'cb' will be called in handshakes, and is
3375 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3376 Argument 'arg' is for application purposes (the value as given to
3377 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3378 be provided to the callback function). The callback function
3379 has to return non-zero to report success: usually 1 to use opaque
3380 PRF input just if possible, or 2 to enforce use of the opaque PRF
3381 input. In the latter case, the library will abort the handshake
3382 if opaque PRF input is not successfully negotiated.
3384 Arguments 'peerinput' and 'len' given to the callback function
3385 will always be NULL and 0 in the case of a client. A server will
3386 see the client's opaque PRF input through these variables if
3387 available (NULL and 0 otherwise). Note that if the server
3388 provides an opaque PRF input, the length must be the same as the
3389 length of the client's opaque PRF input.
3391 Note that the callback function will only be called when creating
3392 a new session (session resumption can resume whatever was
3393 previously negotiated), and will not be called in SSL 2.0
3394 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3395 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3396 for applications that need to enforce opaque PRF input.
3400 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3403 [Victor B. Wagner <vitus@cryptocom.ru>]
3405 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3406 RFC4507bis. The encrypted ticket format is an encrypted encoded
3407 SSL_SESSION structure, that way new session features are automatically
3410 If a client application caches session in an SSL_SESSION structure
3411 support is transparent because tickets are now stored in the encoded
3414 The SSL_CTX structure automatically generates keys for ticket
3415 protection in servers so again support should be possible
3416 with no application modification.
3418 If a client or server wishes to disable RFC4507 support then the option
3419 SSL_OP_NO_TICKET can be set.
3421 Add a TLS extension debugging callback to allow the contents of any client
3422 or server extensions to be examined.
3424 This work was sponsored by Google.
3427 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3428 OpenSSL should now compile cleanly on gcc 4.2
3429 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3431 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3432 support including streaming MAC support: this is required for GOST
3433 ciphersuite support.
3434 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3436 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3437 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3438 to output in BER and PEM format.
3441 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3442 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3443 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3444 ENGINE support for HMAC keys which are unextractable. New -mac and
3445 -macopt options to dgst utility.
3448 *) New option -sigopt to dgst utility. Update dgst to use
3449 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3450 alternative signing parameters such as X9.31 or PSS in the dgst
3454 *) Change ssl_cipher_apply_rule(), the internal function that does
3455 the work each time a ciphersuite string requests enabling
3456 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3457 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3458 the order of disabled ciphersuites such that those ciphersuites
3459 that most recently went from enabled to disabled not only stay
3460 in order with respect to each other, but also have higher priority
3461 than other disabled ciphersuites the next time ciphersuites are
3464 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3465 the same ciphersuites as with "HIGH" alone, but in a specific
3466 order where the PSK ciphersuites come first (since they are the
3467 most recently disabled ciphersuites when "HIGH" is parsed).
3469 Also, change ssl_create_cipher_list() (using this new
3470 funcionality) such that between otherwise identical
3471 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3475 *) Change ssl_create_cipher_list() so that it automatically
3476 arranges the ciphersuites in reasonable order before starting
3477 to process the rule string. Thus, the definition for "DEFAULT"
3478 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3479 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3480 This makes it much easier to arrive at a reasonable default order
3481 in applications for which anonymous ciphers are OK (meaning
3482 that you can't actually use DEFAULT).
3483 [Bodo Moeller; suggested by Victor Duchovni]
3485 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3486 processing) into multiple integers instead of setting
3487 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3488 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3489 (These masks as well as the individual bit definitions are hidden
3490 away into the non-exported interface ssl/ssl_locl.h, so this
3491 change to the definition of the SSL_CIPHER structure shouldn't
3492 affect applications.) This give us more bits for each of these
3493 categories, so there is no longer a need to coagulate AES128 and
3494 AES256 into a single algorithm bit, and to coagulate Camellia128
3495 and Camellia256 into a single algorithm bit, which has led to all
3498 Thus, among other things, the kludge introduced in 0.9.7m and
3499 0.9.8e for masking out AES256 independently of AES128 or masking
3500 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3502 With the change, we also introduce new ciphersuite aliases that
3503 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3507 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3508 Use the leftmost N bytes of the signature input if the input is
3509 larger than the prime q (with N being the size in bytes of q).
3512 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3513 it yet and it is largely untested.
3516 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3519 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3520 some compilers (gcc 4.2 and later) reject their use. Safestack is
3521 reimplemented. Update ASN1 to avoid use of legacy functions.
3524 *) Win32/64 targets are linked with Winsock2.
3527 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3528 to external functions. This can be used to increase CRL handling
3529 efficiency especially when CRLs are very large by (for example) storing
3530 the CRL revoked certificates in a database.
3533 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3534 new CRLs added to a directory can be used. New command line option
3535 -verify_return_error to s_client and s_server. This causes real errors
3536 to be returned by the verify callback instead of carrying on no matter
3537 what. This reflects the way a "real world" verify callback would behave.
3540 *) GOST engine, supporting several GOST algorithms and public key formats.
3541 Kindly donated by Cryptocom.
3544 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3545 partitioned by DP are handled but no indirect CRL or reason partitioning
3546 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3547 selected via a scoring technique which handles IDP and AKID in CRLs.
3550 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3551 will ultimately be used for all verify operations: this will remove the
3552 X509_STORE dependency on certificate verification and allow alternative
3553 lookup methods. X509_STORE based implementations of these two callbacks.
3556 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3557 Modify get_crl() to find a valid (unexpired) CRL if possible.
3560 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3561 this would be called X509_CRL_cmp() but that name is already used by
3562 a function that just compares CRL issuer names. Cache several CRL
3563 extensions in X509_CRL structure and cache CRLDP in X509.
3566 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3567 this maps equivalent X509_NAME structures into a consistent structure.
3568 Name comparison can then be performed rapidly using memcmp().
3571 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3575 *) Allow digests to supply their own micalg string for S/MIME type using
3576 the ctrl EVP_MD_CTRL_MICALG.
3579 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3580 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3581 ctrl. It can then customise the structure before and/or after signing
3585 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3586 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3587 to free up any added signature OIDs.
3590 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3591 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3592 digest and cipher tables. New options added to openssl utility:
3593 list-message-digest-algorithms and list-cipher-algorithms.
3596 *) Change the array representation of binary polynomials: the list
3597 of degrees of non-zero coefficients is now terminated with -1.
3598 Previously it was terminated with 0, which was also part of the
3599 value; thus, the array representation was not applicable to
3600 polynomials where t^0 has coefficient zero. This change makes
3601 the array representation useful in a more general context.
3604 *) Various modifications and fixes to SSL/TLS cipher string
3605 handling. For ECC, the code now distinguishes between fixed ECDH
3606 with RSA certificates on the one hand and with ECDSA certificates
3607 on the other hand, since these are separate ciphersuites. The
3608 unused code for Fortezza ciphersuites has been removed.
3610 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3611 (not "ECDHE"). For consistency with the code for DH
3612 certificates, use of ECDH certificates is now considered ECDH
3613 authentication, not RSA or ECDSA authentication (the latter is
3614 merely the CA's signing algorithm and not actively used in the
3617 The temporary ciphersuite alias "ECCdraft" is no longer
3618 available, and ECC ciphersuites are no longer excluded from "ALL"
3619 and "DEFAULT". The following aliases now exist for RFC 4492
3620 ciphersuites, most of these by analogy with the DH case:
3622 kECDHr - ECDH cert, signed with RSA
3623 kECDHe - ECDH cert, signed with ECDSA
3624 kECDH - ECDH cert (signed with either RSA or ECDSA)
3625 kEECDH - ephemeral ECDH
3626 ECDH - ECDH cert or ephemeral ECDH
3632 AECDH - anonymous ECDH
3633 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3637 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3638 Use correct micalg parameters depending on digest(s) in signed message.
3641 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3642 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3645 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3646 an engine to register a method. Add ENGINE lookups for methods and
3647 functional reference processing.
3650 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3651 EVP_{Sign,Verify}* which allow an application to customise the signature
3655 *) New -resign option to smime utility. This adds one or more signers
3656 to an existing PKCS#7 signedData structure. Also -md option to use an
3657 alternative message digest algorithm for signing.
3660 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3661 create PKCS7 structures containing multiple signers. Update smime
3662 application to support multiple signers.
3665 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3669 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3670 Reorganize PBE internals to lookup from a static table using NIDs,
3671 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3672 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3673 PRF which will be automatically used with PBES2.
3676 *) Replace the algorithm specific calls to generate keys in "req" with the
3680 *) Update PKCS#7 enveloped data routines to use new API. This is now
3681 supported by any public key method supporting the encrypt operation. A
3682 ctrl is added to allow the public key algorithm to examine or modify
3683 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3687 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3688 a default digest type to use. In most cases this will be SHA1 but some
3689 algorithms (such as GOST) need to specify an alternative digest. The
3690 return value indicates how strong the preference is 1 means optional and
3691 2 is mandatory (that is it is the only supported type). Modify
3692 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3693 use the default md. Update openssl utilities to use the default digest
3694 type for signing if it is not explicitly indicated.
3697 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3698 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3699 signing method from the key type. This effectively removes the link
3700 between digests and public key types.
3703 *) Add an OID cross reference table and utility functions. Its purpose is to
3704 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3705 rsaEncryption. This will allow some of the algorithm specific hackery
3706 needed to use the correct OID to be removed.
3709 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3710 structures for PKCS7_sign(). They are now set up by the relevant public
3714 *) Add provisional EC pkey method with support for ECDSA and ECDH.
projects / openssl.git / blob