5 Changes between 1.1.0e and 1.1.1 [xx XXX xxxx]
7 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
8 facilitate stack unwinding even from assembly subroutines.
11 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
12 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
15 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
16 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
17 which is the minimum version we support.
20 *) Certificate time validation (X509_cmp_time) enforces stricter
21 compliance with RFC 5280. Fractional seconds and timezone offsets
22 are no longer allowed.
25 *) Add support for SipHash
28 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
29 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
30 prevent issues where no progress is being made and the peer continually
31 sends unrecognised record types, using up resources processing them.
34 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
35 using the algorithm defined in
36 https://www.akkadia.org/drepper/SHA-crypt.txt
39 *) Heartbeat support has been removed; the ABI is changed for now.
40 [Richard Levitte, Rich Salz]
42 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
45 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
47 *) Encrypt-Then-Mac renegotiation crash
49 During a renegotiation handshake if the Encrypt-Then-Mac extension is
50 negotiated where it was not in the original handshake (or vice-versa) then
51 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
52 and servers are affected.
54 This issue was reported to OpenSSL by Joe Orton (Red Hat).
58 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
60 *) Truncated packet could crash via OOB read
62 If one side of an SSL/TLS path is running on a 32-bit host and a specific
63 cipher is being used, then a truncated packet can cause that host to
64 perform an out-of-bounds read, usually resulting in a crash.
66 This issue was reported to OpenSSL by Robert Święcki of Google.
70 *) Bad (EC)DHE parameters cause a client crash
72 If a malicious server supplies bad parameters for a DHE or ECDHE key
73 exchange then this can result in the client attempting to dereference a
74 NULL pointer leading to a client crash. This could be exploited in a Denial
77 This issue was reported to OpenSSL by Guido Vranken.
81 *) BN_mod_exp may produce incorrect results on x86_64
83 There is a carry propagating bug in the x86_64 Montgomery squaring
84 procedure. No EC algorithms are affected. Analysis suggests that attacks
85 against RSA and DSA as a result of this defect would be very difficult to
86 perform and are not believed likely. Attacks against DH are considered just
87 feasible (although very difficult) because most of the work necessary to
88 deduce information about a private key may be performed offline. The amount
89 of resources required for such an attack would be very significant and
90 likely only accessible to a limited number of attackers. An attacker would
91 additionally need online access to an unpatched system using the target
92 private key in a scenario with persistent DH parameters and a private
93 key that is shared between multiple clients. For example this can occur by
94 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
95 similar to CVE-2015-3193 but must be treated as a separate problem.
97 This issue was reported to OpenSSL by the OSS-Fuzz project.
101 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
103 *) ChaCha20/Poly1305 heap-buffer-overflow
105 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
106 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
107 crash. This issue is not considered to be exploitable beyond a DoS.
109 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
113 *) CMS Null dereference
115 Applications parsing invalid CMS structures can crash with a NULL pointer
116 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
117 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
118 structure callback if an attempt is made to free certain invalid encodings.
119 Only CHOICE structures using a callback which do not handle NULL value are
122 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
126 *) Montgomery multiplication may produce incorrect results
128 There is a carry propagating bug in the Broadwell-specific Montgomery
129 multiplication procedure that handles input lengths divisible by, but
130 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
131 and DH private keys are impossible. This is because the subroutine in
132 question is not used in operations with the private key itself and an input
133 of the attacker's direct choice. Otherwise the bug can manifest itself as
134 transient authentication and key negotiation failures or reproducible
135 erroneous outcome of public-key operations with specially crafted input.
136 Among EC algorithms only Brainpool P-512 curves are affected and one
137 presumably can attack ECDH key negotiation. Impact was not analyzed in
138 detail, because pre-requisites for attack are considered unlikely. Namely
139 multiple clients have to choose the curve in question and the server has to
140 share the private key among them, neither of which is default behaviour.
141 Even then only clients that chose the curve will be affected.
143 This issue was publicly reported as transient failures and was not
144 initially recognized as a security issue. Thanks to Richard Morgan for
145 providing reproducible case.
149 *) Removed automatic addition of RPATH in shared libraries and executables,
150 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
153 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
155 *) Fix Use After Free for large message sizes
157 The patch applied to address CVE-2016-6307 resulted in an issue where if a
158 message larger than approx 16k is received then the underlying buffer to
159 store the incoming message is reallocated and moved. Unfortunately a
160 dangling pointer to the old location is left which results in an attempt to
161 write to the previously freed location. This is likely to result in a
162 crash, however it could potentially lead to execution of arbitrary code.
164 This issue only affects OpenSSL 1.1.0a.
166 This issue was reported to OpenSSL by Robert Święcki.
170 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
172 *) OCSP Status Request extension unbounded memory growth
174 A malicious client can send an excessively large OCSP Status Request
175 extension. If that client continually requests renegotiation, sending a
176 large OCSP Status Request extension each time, then there will be unbounded
177 memory growth on the server. This will eventually lead to a Denial Of
178 Service attack through memory exhaustion. Servers with a default
179 configuration are vulnerable even if they do not support OCSP. Builds using
180 the "no-ocsp" build time option are not affected.
182 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
186 *) SSL_peek() hang on empty record
188 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
189 sends an empty record. This could be exploited by a malicious peer in a
190 Denial Of Service attack.
192 This issue was reported to OpenSSL by Alex Gaynor.
196 *) Excessive allocation of memory in tls_get_message_header() and
197 dtls1_preprocess_fragment()
199 A (D)TLS message includes 3 bytes for its length in the header for the
200 message. This would allow for messages up to 16Mb in length. Messages of
201 this length are excessive and OpenSSL includes a check to ensure that a
202 peer is sending reasonably sized messages in order to avoid too much memory
203 being consumed to service a connection. A flaw in the logic of version
204 1.1.0 means that memory for the message is allocated too early, prior to
205 the excessive message length check. Due to way memory is allocated in
206 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
207 to service a connection. This could lead to a Denial of Service through
208 memory exhaustion. However, the excessive message length check still takes
209 place, and this would cause the connection to immediately fail. Assuming
210 that the application calls SSL_free() on the failed conneciton in a timely
211 manner then the 21Mb of allocated memory will then be immediately freed
212 again. Therefore the excessive memory allocation will be transitory in
213 nature. This then means that there is only a security impact if:
215 1) The application does not call SSL_free() in a timely manner in the event
216 that the connection fails
218 2) The application is working in a constrained environment where there is
219 very little free memory
221 3) The attacker initiates multiple connection attempts such that there are
222 multiple connections in a state where memory has been allocated for the
223 connection; SSL_free() has not yet been called; and there is insufficient
224 memory to service the multiple requests.
226 Except in the instance of (1) above any Denial Of Service is likely to be
227 transitory because as soon as the connection fails the memory is
228 subsequently freed again in the SSL_free() call. However there is an
229 increased risk during this period of application crashes due to the lack of
230 memory - which would then mean a more serious Denial of Service.
232 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
233 (CVE-2016-6307 and CVE-2016-6308)
236 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
237 had to be removed. Primary reason is that vendor assembler can't
238 assemble our modules with -KPIC flag. As result it, assembly
239 support, was not even available as option. But its lack means
240 lack of side-channel resistant code, which is incompatible with
241 security by todays standards. Fortunately gcc is readily available
242 prepackaged option, which we firmly point at...
245 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
247 *) Windows command-line tool supports UTF-8 opt-in option for arguments
248 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
249 (to any value) allows Windows user to access PKCS#12 file generated
250 with Windows CryptoAPI and protected with non-ASCII password, as well
251 as files generated under UTF-8 locale on Linux also protected with
255 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
256 have been disabled by default and removed from DEFAULT, just like RC4.
257 See the RC4 item below to re-enable both.
260 *) The method for finding the storage location for the Windows RAND seed file
261 has changed. First we check %RANDFILE%. If that is not set then we check
262 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
263 all else fails we fall back to C:\.
266 *) The EVP_EncryptUpdate() function has had its return type changed from void
267 to int. A return of 0 indicates and error while a return of 1 indicates
271 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
272 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
273 off the constant time implementation for RSA, DSA and DH have been made
274 no-ops and deprecated.
277 *) Windows RAND implementation was simplified to only get entropy by
278 calling CryptGenRandom(). Various other RAND-related tickets
280 [Joseph Wylie Yandle, Rich Salz]
282 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
283 and OPENSSL_LH_, respectively. The old names are available
284 with API compatibility. They new names are now completely documented.
287 *) Unify TYPE_up_ref(obj) methods signature.
288 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
289 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
290 int (instead of void) like all others TYPE_up_ref() methods.
291 So now these methods also check the return value of CRYPTO_atomic_add(),
292 and the validity of object reference counter.
293 [fdasilvayy@gmail.com]
295 *) With Windows Visual Studio builds, the .pdb files are installed
296 alongside the installed libraries and executables. For a static
297 library installation, ossl_static.pdb is the associate compiler
298 generated .pdb file to be used when linking programs.
301 *) Remove openssl.spec. Packaging files belong with the packagers.
304 *) Automatic Darwin/OSX configuration has had a refresh, it will now
305 recognise x86_64 architectures automatically. You can still decide
306 to build for a different bitness with the environment variable
307 KERNEL_BITS (can be 32 or 64), for example:
309 KERNEL_BITS=32 ./config
313 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
314 256 bit AES and HMAC with SHA256.
317 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
320 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
323 *) To enable users to have their own config files and build file templates,
324 Configure looks in the directory indicated by the environment variable
325 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
326 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
327 name and is used as is.
330 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
331 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
332 X509_CERT_FILE_CTX was removed.
335 *) "shared" builds are now the default. To create only static libraries use
336 the "no-shared" Configure option.
339 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
340 All of these option have not worked for some while and are fundamental
344 *) Make various cleanup routines no-ops and mark them as deprecated. Most
345 global cleanup functions are no longer required because they are handled
346 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
347 Explicitly de-initing can cause problems (e.g. where a library that uses
348 OpenSSL de-inits, but an application is still using it). The affected
349 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
350 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
351 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
355 *) --strict-warnings no longer enables runtime debugging options
356 such as REF_DEBUG. Instead, debug options are automatically
357 enabled with '--debug' builds.
358 [Andy Polyakov, Emilia Käsper]
360 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
361 have been moved out of the public header files. New functions for managing
362 these have been added.
365 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
366 objects have been moved out of the public header files. New
367 functions for managing these have been added.
370 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
371 have been moved out of the public header files. New functions for managing
372 these have been added.
375 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
376 moved out of the public header files. New functions for managing these
380 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
383 *) Removed the mk1mf build scripts.
386 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
387 it is always safe to #include a header now.
390 *) Removed the aged BC-32 config and all its supporting scripts
393 *) Removed support for Ultrix, Netware, and OS/2.
396 *) Add support for HKDF.
399 *) Add support for blake2b and blake2s
402 *) Added support for "pipelining". Ciphers that have the
403 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
404 encryptions/decryptions simultaneously. There are currently no built-in
405 ciphers with this property but the expectation is that engines will be able
406 to offer it to significantly improve throughput. Support has been extended
407 into libssl so that multiple records for a single connection can be
408 processed in one go (for >=TLS 1.1).
411 *) Added the AFALG engine. This is an async capable engine which is able to
412 offload work to the Linux kernel. In this initial version it only supports
413 AES128-CBC. The kernel must be version 4.1.0 or greater.
416 *) OpenSSL now uses a new threading API. It is no longer necessary to
417 set locking callbacks to use OpenSSL in a multi-threaded environment. There
418 are two supported threading models: pthreads and windows threads. It is
419 also possible to configure OpenSSL at compile time for "no-threads". The
420 old threading API should no longer be used. The functions have been
421 replaced with "no-op" compatibility macros.
422 [Alessandro Ghedini, Matt Caswell]
424 *) Modify behavior of ALPN to invoke callback after SNI/servername
425 callback, such that updates to the SSL_CTX affect ALPN.
428 *) Add SSL_CIPHER queries for authentication and key-exchange.
431 *) Changes to the DEFAULT cipherlist:
432 - Prefer (EC)DHE handshakes over plain RSA.
433 - Prefer AEAD ciphers over legacy ciphers.
434 - Prefer ECDSA over RSA when both certificates are available.
435 - Prefer TLSv1.2 ciphers/PRF.
436 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
440 *) Change the ECC default curve list to be this, in order: x25519,
441 secp256r1, secp521r1, secp384r1.
444 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
445 disabled by default. They can be re-enabled using the
446 enable-weak-ssl-ciphers option to Configure.
449 *) If the server has ALPN configured, but supports no protocols that the
450 client advertises, send a fatal "no_application_protocol" alert.
451 This behaviour is SHALL in RFC 7301, though it isn't universally
452 implemented by other servers.
455 *) Add X25519 support.
456 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
457 for public and private key encoding using the format documented in
458 draft-ietf-curdle-pkix-02. The coresponding EVP_PKEY method supports
459 key generation and key derivation.
461 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
465 *) Deprecate SRP_VBASE_get_by_user.
466 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
467 In order to fix an unavoidable memory leak (CVE-2016-0798),
468 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
469 seed, even if the seed is configured.
471 Users should use SRP_VBASE_get1_by_user instead. Note that in
472 SRP_VBASE_get1_by_user, caller must free the returned value. Note
473 also that even though configuring the SRP seed attempts to hide
474 invalid usernames by continuing the handshake with fake
475 credentials, this behaviour is not constant time and no strong
476 guarantees are made that the handshake is indistinguishable from
477 that of a valid user.
480 *) Configuration change; it's now possible to build dynamic engines
481 without having to build shared libraries and vice versa. This
482 only applies to the engines in engines/, those in crypto/engine/
483 will always be built into libcrypto (i.e. "static").
485 Building dynamic engines is enabled by default; to disable, use
486 the configuration option "disable-dynamic-engine".
488 The only requirements for building dynamic engines are the
489 presence of the DSO module and building with position independent
490 code, so they will also automatically be disabled if configuring
491 with "disable-dso" or "disable-pic".
493 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
494 are also taken away from openssl/opensslconf.h, as they are
498 *) Configuration change; if there is a known flag to compile
499 position independent code, it will always be applied on the
500 libcrypto and libssl object files, and never on the application
501 object files. This means other libraries that use routines from
502 libcrypto / libssl can be made into shared libraries regardless
503 of how OpenSSL was configured.
505 If this isn't desirable, the configuration options "disable-pic"
506 or "no-pic" can be used to disable the use of PIC. This will
507 also disable building shared libraries and dynamic engines.
510 *) Removed JPAKE code. It was experimental and has no wide use.
513 *) The INSTALL_PREFIX Makefile variable has been renamed to
514 DESTDIR. That makes for less confusion on what this variable
515 is for. Also, the configuration option --install_prefix is
519 *) Heartbeat for TLS has been removed and is disabled by default
520 for DTLS; configure with enable-heartbeats. Code that uses the
521 old #define's might need to be updated.
522 [Emilia Käsper, Rich Salz]
524 *) Rename REF_CHECK to REF_DEBUG.
527 *) New "unified" build system
529 The "unified" build system is aimed to be a common system for all
530 platforms we support. With it comes new support for VMS.
532 This system builds supports building in a different directory tree
533 than the source tree. It produces one Makefile (for unix family
534 or lookalikes), or one descrip.mms (for VMS).
536 The source of information to make the Makefile / descrip.mms is
537 small files called 'build.info', holding the necessary
538 information for each directory with source to compile, and a
539 template in Configurations, like unix-Makefile.tmpl or
542 With this change, the library names were also renamed on Windows
543 and on VMS. They now have names that are closer to the standard
544 on Unix, and include the major version number, and in certain
545 cases, the architecture they are built for. See "Notes on shared
546 libraries" in INSTALL.
548 We rely heavily on the perl module Text::Template.
551 *) Added support for auto-initialisation and de-initialisation of the library.
552 OpenSSL no longer requires explicit init or deinit routines to be called,
553 except in certain circumstances. See the OPENSSL_init_crypto() and
554 OPENSSL_init_ssl() man pages for further information.
557 *) The arguments to the DTLSv1_listen function have changed. Specifically the
558 "peer" argument is now expected to be a BIO_ADDR object.
560 *) Rewrite of BIO networking library. The BIO library lacked consistent
561 support of IPv6, and adding it required some more extensive
562 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
563 which hold all types of addresses and chains of address information.
564 It also introduces a new API, with functions like BIO_socket,
565 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
566 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
567 have been adapted accordingly.
570 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
574 *) CRIME protection: disable compression by default, even if OpenSSL is
575 compiled with zlib enabled. Applications can still enable compression
576 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
577 using the SSL_CONF library to configure compression.
580 *) The signature of the session callback configured with
581 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
582 was explicitly marked as 'const unsigned char*' instead of
586 *) Always DPURIFY. Remove the use of uninitialized memory in the
587 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
590 *) Removed many obsolete configuration items, including
591 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
592 MD2_CHAR, MD2_INT, MD2_LONG
594 IDEA_SHORT, IDEA_LONG
595 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
596 [Rich Salz, with advice from Andy Polyakov]
598 *) Many BN internals have been moved to an internal header file.
599 [Rich Salz with help from Andy Polyakov]
601 *) Configuration and writing out the results from it has changed.
602 Files such as Makefile include/openssl/opensslconf.h and are now
603 produced through general templates, such as Makefile.in and
604 crypto/opensslconf.h.in and some help from the perl module
607 Also, the center of configuration information is no longer
608 Makefile. Instead, Configure produces a perl module in
609 configdata.pm which holds most of the config data (in the hash
610 table %config), the target data that comes from the target
611 configuration in one of the Configurations/*.conf files (in
615 *) To clarify their intended purposes, the Configure options
616 --prefix and --openssldir change their semantics, and become more
617 straightforward and less interdependent.
619 --prefix shall be used exclusively to give the location INSTALLTOP
620 where programs, scripts, libraries, include files and manuals are
621 going to be installed. The default is now /usr/local.
623 --openssldir shall be used exclusively to give the default
624 location OPENSSLDIR where certificates, private keys, CRLs are
625 managed. This is also where the default openssl.cnf gets
627 If the directory given with this option is a relative path, the
628 values of both the --prefix value and the --openssldir value will
629 be combined to become OPENSSLDIR.
630 The default for --openssldir is INSTALLTOP/ssl.
632 Anyone who uses --openssldir to specify where OpenSSL is to be
633 installed MUST change to use --prefix instead.
636 *) The GOST engine was out of date and therefore it has been removed. An up
637 to date GOST engine is now being maintained in an external repository.
638 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
639 support for GOST ciphersuites (these are only activated if a GOST engine
643 *) EGD is no longer supported by default; use enable-egd when
645 [Ben Kaduk and Rich Salz]
647 *) The distribution now has Makefile.in files, which are used to
648 create Makefile's when Configure is run. *Configure must be run
649 before trying to build now.*
652 *) The return value for SSL_CIPHER_description() for error conditions
656 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
658 Obtaining and performing DNSSEC validation of TLSA records is
659 the application's responsibility. The application provides
660 the TLSA records of its choice to OpenSSL, and these are then
661 used to authenticate the peer.
663 The TLSA records need not even come from DNS. They can, for
664 example, be used to implement local end-entity certificate or
665 trust-anchor "pinning", where the "pin" data takes the form
666 of TLSA records, which can augment or replace verification
667 based on the usual WebPKI public certification authorities.
670 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
671 continues to support deprecated interfaces in default builds.
672 However, applications are strongly advised to compile their
673 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
674 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
675 or the 1.1.0 releases.
677 In environments in which all applications have been ported to
678 not use any deprecated interfaces OpenSSL's Configure script
679 should be used with the --api=1.1.0 option to entirely remove
680 support for the deprecated features from the library and
681 unconditionally disable them in the installed headers.
682 Essentially the same effect can be achieved with the "no-deprecated"
683 argument to Configure, except that this will always restrict
684 the build to just the latest API, rather than a fixed API
687 As applications are ported to future revisions of the API,
688 they should update their compile-time OPENSSL_API_COMPAT define
689 accordingly, but in most cases should be able to continue to
690 compile with later releases.
692 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
693 0x10000000L and 0x00908000L, respectively. However those
694 versions did not support the OPENSSL_API_COMPAT feature, and
695 so applications are not typically tested for explicit support
696 of just the undeprecated features of either release.
699 *) Add support for setting the minimum and maximum supported protocol.
700 It can bet set via the SSL_set_min_proto_version() and
701 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
702 MaxProtcol. It's recommended to use the new APIs to disable
703 protocols instead of disabling individual protocols using
704 SSL_set_options() or SSL_CONF's Protocol. This change also
705 removes support for disabling TLS 1.2 in the OpenSSL TLS
706 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
709 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
712 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
713 and integrates ECDSA and ECDH functionality into EC. Implementations can
714 now redirect key generation and no longer need to convert to or from
717 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
718 include the ec.h header file instead.
721 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
722 ciphers who are no longer supported and drops support the ephemeral RSA key
723 exchange. The LOW ciphers currently doesn't have any ciphers in it.
726 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
727 opaque. For HMAC_CTX, the following constructors and destructors
730 HMAC_CTX *HMAC_CTX_new(void);
731 void HMAC_CTX_free(HMAC_CTX *ctx);
733 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
734 destroy such methods has been added. See EVP_MD_meth_new(3) and
735 EVP_CIPHER_meth_new(3) for documentation.
738 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
739 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
740 EVP_MD_CTX_reset() should be called instead to reinitialise
741 an already created structure.
742 2) For consistency with the majority of our object creators and
743 destructors, EVP_MD_CTX_(create|destroy) were renamed to
744 EVP_MD_CTX_(new|free). The old names are retained as macros
745 for deprecated builds.
748 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
749 cryptographic operations to be performed asynchronously as long as an
750 asynchronous capable engine is used. See the ASYNC_start_job() man page for
751 further details. Libssl has also had this capability integrated with the
752 introduction of the new mode SSL_MODE_ASYNC and associated error
753 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
754 pages. This work was developed in partnership with Intel Corp.
757 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
758 always enabled now. If you want to disable the support you should
759 exclude it using the list of supported ciphers. This also means that the
760 "-no_ecdhe" option has been removed from s_server.
763 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
764 SSL_{CTX_}set1_curves() which can set a list.
767 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
768 curve you want to support using SSL_{CTX_}set1_curves().
771 *) State machine rewrite. The state machine code has been significantly
772 refactored in order to remove much duplication of code and solve issues
773 with the old code (see ssl/statem/README for further details). This change
774 does have some associated API changes. Notably the SSL_state() function
775 has been removed and replaced by SSL_get_state which now returns an
776 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
777 altogether. The previous handshake states defined in ssl.h and ssl3.h have
781 *) All instances of the string "ssleay" in the public API were replaced
782 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
783 Some error codes related to internal RSA_eay API's were renamed.
786 *) The demo files in crypto/threads were moved to demo/threads.
789 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
791 [Matt Caswell, Rich Salz]
793 *) New ASN.1 embed macro.
795 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
796 structure is not allocated: it is part of the parent. That is instead of
804 This reduces memory fragmentation and make it impossible to accidentally
805 set a mandatory field to NULL.
807 This currently only works for some fields specifically a SEQUENCE, CHOICE,
808 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
809 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
813 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
816 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
817 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
818 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
819 DES and RC4 ciphersuites.
822 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
823 This changes the decoding behaviour for some invalid messages,
824 though the change is mostly in the more lenient direction, and
825 legacy behaviour is preserved as much as possible.
828 *) Fix no-stdio build.
829 [ David Woodhouse <David.Woodhouse@intel.com> and also
830 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
832 *) New testing framework
833 The testing framework has been largely rewritten and is now using
834 perl and the perl modules Test::Harness and an extended variant of
835 Test::More called OpenSSL::Test to do its work. All test scripts in
836 test/ have been rewritten into test recipes, and all direct calls to
837 executables in test/Makefile have become individual recipes using the
838 simplified testing OpenSSL::Test::Simple.
840 For documentation on our testing modules, do:
842 perldoc test/testlib/OpenSSL/Test/Simple.pm
843 perldoc test/testlib/OpenSSL/Test.pm
847 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
848 are used; the latter aborts on memory leaks (usually checked on exit).
849 Some undocumented "set malloc, etc., hooks" functions were removed
850 and others were changed. All are now documented.
853 *) In DSA_generate_parameters_ex, if the provided seed is too short,
855 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
857 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
858 from RFC4279, RFC4785, RFC5487, RFC5489.
860 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
861 original RSA_PSK patch.
864 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
865 era flag was never set throughout the codebase (only read). Also removed
866 SSL3_FLAGS_POP_BUFFER which was only used if
867 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
870 *) Changed the default name options in the "ca", "crl", "req" and "x509"
871 to be "oneline" instead of "compat".
874 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
875 not aware of clients that still exhibit this bug, and the workaround
876 hasn't been working properly for a while.
879 *) The return type of BIO_number_read() and BIO_number_written() as well as
880 the corresponding num_read and num_write members in the BIO structure has
881 changed from unsigned long to uint64_t. On platforms where an unsigned
882 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
886 *) Given the pervasive nature of TLS extensions it is inadvisable to run
887 OpenSSL without support for them. It also means that maintaining
888 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
889 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
892 *) Removed support for the two export grade static DH ciphersuites
893 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
894 were newly added (along with a number of other static DH ciphersuites) to
895 1.0.2. However the two export ones have *never* worked since they were
896 introduced. It seems strange in any case to be adding new export
897 ciphersuites, and given "logjam" it also does not seem correct to fix them.
900 *) Version negotiation has been rewritten. In particular SSLv23_method(),
901 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
902 and turned into macros which simply call the new preferred function names
903 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
904 should use the new names instead. Also as part of this change the ssl23.h
905 header file has been removed.
908 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
909 code and the associated standard is no longer considered fit-for-purpose.
912 *) RT2547 was closed. When generating a private key, try to make the
913 output file readable only by the owner. This behavior change might
914 be noticeable when interacting with other software.
916 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
920 *) Added HTTP GET support to the ocsp command.
923 *) Changed default digest for the dgst and enc commands from MD5 to
927 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
930 *) Added support for TLS extended master secret from
931 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
932 initial patch which was a great help during development.
935 *) All libssl internal structures have been removed from the public header
936 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
937 now redundant). Users should not attempt to access internal structures
938 directly. Instead they should use the provided API functions.
941 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
942 Access to deprecated functions can be re-enabled by running config with
943 "enable-deprecated". In addition applications wishing to use deprecated
944 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
945 will, by default, disable some transitive includes that previously existed
946 in the header files (e.g. ec.h will no longer, by default, include bn.h)
949 *) Added support for OCB mode. OpenSSL has been granted a patent license
950 compatible with the OpenSSL license for use of OCB. Details are available
951 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
952 for OCB can be removed by calling config with no-ocb.
955 *) SSLv2 support has been removed. It still supports receiving a SSLv2
956 compatible client hello.
959 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
960 done while fixing the error code for the key-too-small case.
961 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
963 *) CA.sh has been removmed; use CA.pl instead.
966 *) Removed old DES API.
969 *) Remove various unsupported platforms:
975 Sinix/ReliantUNIX RM400
980 16-bit platforms such as WIN16
983 *) Clean up OPENSSL_NO_xxx #define's
984 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
985 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
986 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
987 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
988 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
989 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
990 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
991 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
992 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
993 Remove MS_STATIC; it's a relic from platforms <32 bits.
996 *) Cleaned up dead code
997 Remove all but one '#ifdef undef' which is to be looked at.
1000 *) Clean up calling of xxx_free routines.
1001 Just like free(), fix most of the xxx_free routines to accept
1002 NULL. Remove the non-null checks from callers. Save much code.
1005 *) Add secure heap for storage of private keys (when possible).
1006 Add BIO_s_secmem(), CBIGNUM, etc.
1007 Contributed by Akamai Technologies under our Corporate CLA.
1010 *) Experimental support for a new, fast, unbiased prime candidate generator,
1011 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1012 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1014 *) New output format NSS in the sess_id command line tool. This allows
1015 exporting the session id and the master key in NSS keylog format.
1016 [Martin Kaiser <martin@kaiser.cx>]
1018 *) Harmonize version and its documentation. -f flag is used to display
1020 [mancha <mancha1@zoho.com>]
1022 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1023 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1024 [mancha <mancha1@zoho.com>]
1026 *) Fix some double frees. These are not thought to be exploitable.
1027 [mancha <mancha1@zoho.com>]
1029 *) A missing bounds check in the handling of the TLS heartbeat extension
1030 can be used to reveal up to 64k of memory to a connected client or
1033 Thanks for Neel Mehta of Google Security for discovering this bug and to
1034 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1035 preparing the fix (CVE-2014-0160)
1036 [Adam Langley, Bodo Moeller]
1038 *) Fix for the attack described in the paper "Recovering OpenSSL
1039 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1040 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1041 http://eprint.iacr.org/2014/140
1043 Thanks to Yuval Yarom and Naomi Benger for discovering this
1044 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1045 [Yuval Yarom and Naomi Benger]
1047 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1048 this fixes a limitation in previous versions of OpenSSL.
1051 *) Experimental encrypt-then-mac support.
1053 Experimental support for encrypt then mac from
1054 draft-gutmann-tls-encrypt-then-mac-02.txt
1056 To enable it set the appropriate extension number (0x42 for the test
1057 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1059 For non-compliant peers (i.e. just about everything) this should have no
1062 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1066 *) Add EVP support for key wrapping algorithms, to avoid problems with
1067 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1068 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1069 algorithms and include tests cases.
1072 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1076 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1077 MGF1 digest and OAEP label.
1080 *) Make openssl verify return errors.
1081 [Chris Palmer <palmer@google.com> and Ben Laurie]
1083 *) New function ASN1_TIME_diff to calculate the difference between two
1084 ASN1_TIME structures or one structure and the current time.
1087 *) Update fips_test_suite to support multiple command line options. New
1088 test to induce all self test errors in sequence and check expected
1092 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1093 sign or verify all in one operation.
1096 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1097 test programs and fips_test_suite. Includes functionality to parse
1098 the minimal script output of fipsalgest.pl directly.
1101 *) Add authorisation parameter to FIPS_module_mode_set().
1104 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1107 *) Use separate DRBG fields for internal and external flags. New function
1108 FIPS_drbg_health_check() to perform on demand health checking. Add
1109 generation tests to fips_test_suite with reduced health check interval to
1110 demonstrate periodic health checking. Add "nodh" option to
1111 fips_test_suite to skip very slow DH test.
1114 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1118 *) More extensive health check for DRBG checking many more failure modes.
1119 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1120 combination: call this in fips_test_suite.
1123 *) Add support for canonical generation of DSA parameter 'g'. See
1126 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1127 POST to handle HMAC cases.
1130 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1131 to return numerical and string versions of the FIPS module number.
1134 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1135 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1136 outside the validated module in the FIPS capable OpenSSL.
1139 *) Minor change to DRBG entropy callback semantics. In some cases
1140 there is no multiple of the block length between min_len and
1141 max_len. Allow the callback to return more than max_len bytes
1142 of entropy but discard any extra: it is the callback's responsibility
1143 to ensure that the extra data discarded does not impact the
1144 requested amount of entropy.
1147 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1148 information in FIPS186-3, SP800-57 and SP800-131A.
1151 *) CCM support via EVP. Interface is very similar to GCM case except we
1152 must supply all data in one chunk (i.e. no update, final) and the
1153 message length must be supplied if AAD is used. Add algorithm test
1157 *) Initial version of POST overhaul. Add POST callback to allow the status
1158 of POST to be monitored and/or failures induced. Modify fips_test_suite
1159 to use callback. Always run all selftests even if one fails.
1162 *) XTS support including algorithm test driver in the fips_gcmtest program.
1163 Note: this does increase the maximum key length from 32 to 64 bytes but
1164 there should be no binary compatibility issues as existing applications
1165 will never use XTS mode.
1168 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1169 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1170 performs algorithm blocking for unapproved PRNG types. Also do not
1171 set PRNG type in FIPS_mode_set(): leave this to the application.
1172 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1173 the standard OpenSSL PRNG: set additional data to a date time vector.
1176 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1177 This shouldn't present any incompatibility problems because applications
1178 shouldn't be using these directly and any that are will need to rethink
1179 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1182 *) Extensive self tests and health checking required by SP800-90 DRBG.
1183 Remove strength parameter from FIPS_drbg_instantiate and always
1184 instantiate at maximum supported strength.
1187 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1190 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1193 *) New function DH_compute_key_padded() to compute a DH key and pad with
1194 leading zeroes if needed: this complies with SP800-56A et al.
1197 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1198 anything, incomplete, subject to change and largely untested at present.
1201 *) Modify fipscanisteronly build option to only build the necessary object
1202 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1205 *) Add experimental option FIPSSYMS to give all symbols in
1206 fipscanister.o and FIPS or fips prefix. This will avoid
1207 conflicts with future versions of OpenSSL. Add perl script
1208 util/fipsas.pl to preprocess assembly language source files
1209 and rename any affected symbols.
1212 *) Add selftest checks and algorithm block of non-fips algorithms in
1213 FIPS mode. Remove DES2 from selftests.
1216 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1217 return internal method without any ENGINE dependencies. Add new
1218 tiny fips sign and verify functions.
1221 *) New build option no-ec2m to disable characteristic 2 code.
1224 *) New build option "fipscanisteronly". This only builds fipscanister.o
1225 and (currently) associated fips utilities. Uses the file Makefile.fips
1226 instead of Makefile.org as the prototype.
1229 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1230 Update fips_gcmtest to use IV generator.
1233 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1234 setting output buffer to NULL. The *Final function must be
1235 called although it will not retrieve any additional data. The tag
1236 can be set or retrieved with a ctrl. The IV length is by default 12
1237 bytes (96 bits) but can be set to an alternative value. If the IV
1238 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1242 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1243 underlying do_cipher function handles all cipher semantics itself
1244 including padding and finalisation. This is useful if (for example)
1245 an ENGINE cipher handles block padding itself. The behaviour of
1246 do_cipher is subtly changed if this flag is set: the return value
1247 is the number of characters written to the output buffer (zero is
1248 no longer an error code) or a negative error code. Also if the
1249 input buffer is NULL and length 0 finalisation should be performed.
1252 *) If a candidate issuer certificate is already part of the constructed
1253 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1256 *) Improve forward-security support: add functions
1258 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1259 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1261 for use by SSL/TLS servers; the callback function will be called whenever a
1262 new session is created, and gets to decide whether the session may be
1263 cached to make it resumable (return 0) or not (return 1). (As by the
1264 SSL/TLS protocol specifications, the session_id sent by the server will be
1265 empty to indicate that the session is not resumable; also, the server will
1266 not generate RFC 4507 (RFC 5077) session tickets.)
1268 A simple reasonable callback implementation is to return is_forward_secure.
1269 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1270 by the SSL/TLS server library, indicating whether it can provide forward
1272 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1274 *) New -verify_name option in command line utilities to set verification
1278 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1279 Add CMAC pkey methods.
1282 *) Experimental renegotiation in s_server -www mode. If the client
1283 browses /reneg connection is renegotiated. If /renegcert it is
1284 renegotiated requesting a certificate.
1287 *) Add an "external" session cache for debugging purposes to s_server. This
1288 should help trace issues which normally are only apparent in deployed
1289 multi-process servers.
1292 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1293 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1294 BIO_set_cipher() and some obscure PEM functions were changed so they
1295 can now return an error. The RAND changes required a change to the
1296 RAND_METHOD structure.
1299 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1300 a gcc attribute to warn if the result of a function is ignored. This
1301 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1302 whose return value is often ignored.
1305 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1306 These allow SCTs (signed certificate timestamps) to be requested and
1307 validated when establishing a connection.
1308 [Rob Percival <robpercival@google.com>]
1310 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1312 *) Prevent padding oracle in AES-NI CBC MAC check
1314 A MITM attacker can use a padding oracle attack to decrypt traffic
1315 when the connection uses an AES CBC cipher and the server support
1318 This issue was introduced as part of the fix for Lucky 13 padding
1319 attack (CVE-2013-0169). The padding check was rewritten to be in
1320 constant time by making sure that always the same bytes are read and
1321 compared against either the MAC or padding bytes. But it no longer
1322 checked that there was enough data to have both the MAC and padding
1325 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1329 *) Fix EVP_EncodeUpdate overflow
1331 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1332 Base64 encoding of binary data. If an attacker is able to supply very large
1333 amounts of input data then a length check can overflow resulting in a heap
1336 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1337 the PEM_write_bio* family of functions. These are mainly used within the
1338 OpenSSL command line applications, so any application which processes data
1339 from an untrusted source and outputs it as a PEM file should be considered
1340 vulnerable to this issue. User applications that call these APIs directly
1341 with large amounts of untrusted data may also be vulnerable.
1343 This issue was reported by Guido Vranken.
1347 *) Fix EVP_EncryptUpdate overflow
1349 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1350 is able to supply very large amounts of input data after a previous call to
1351 EVP_EncryptUpdate() with a partial block then a length check can overflow
1352 resulting in a heap corruption. Following an analysis of all OpenSSL
1353 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1354 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1355 the first called function after an EVP_EncryptInit(), and therefore that
1356 specific call must be safe. The second form is where the length passed to
1357 EVP_EncryptUpdate() can be seen from the code to be some small value and
1358 therefore there is no possibility of an overflow. Since all instances are
1359 one of these two forms, it is believed that there can be no overflows in
1360 internal code due to this problem. It should be noted that
1361 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1362 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1363 of these calls have also been analysed too and it is believed there are no
1364 instances in internal usage where an overflow could occur.
1366 This issue was reported by Guido Vranken.
1370 *) Prevent ASN.1 BIO excessive memory allocation
1372 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1373 a short invalid encoding can cause allocation of large amounts of memory
1374 potentially consuming excessive resources or exhausting memory.
1376 Any application parsing untrusted data through d2i BIO functions is
1377 affected. The memory based functions such as d2i_X509() are *not* affected.
1378 Since the memory based functions are used by the TLS library, TLS
1379 applications are not affected.
1381 This issue was reported by Brian Carpenter.
1387 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1388 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1389 in arbitrary stack data being returned in the buffer.
1391 This issue was reported by Guido Vranken.
1395 *) Modify behavior of ALPN to invoke callback after SNI/servername
1396 callback, such that updates to the SSL_CTX affect ALPN.
1399 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1403 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1404 methods are enabled and ssl2 is disabled the methods return NULL.
1407 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1409 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1410 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1411 provide any "EXPORT" or "LOW" strength ciphers.
1414 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1415 is by default disabled at build-time. Builds that are not configured with
1416 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1417 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1418 will need to explicitly call either of:
1420 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1422 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1424 as appropriate. Even if either of those is used, or the application
1425 explicitly uses the version-specific SSLv2_method() or its client and
1426 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1427 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1428 ciphers, and SSLv2 56-bit DES are no longer available.
1432 *) Fix a double-free in DSA code
1434 A double free bug was discovered when OpenSSL parses malformed DSA private
1435 keys and could lead to a DoS attack or memory corruption for applications
1436 that receive DSA private keys from untrusted sources. This scenario is
1439 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1444 *) Disable SRP fake user seed to address a server memory leak.
1446 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1448 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1449 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1450 was changed to ignore the "fake user" SRP seed, even if the seed
1453 Users should use SRP_VBASE_get1_by_user instead. Note that in
1454 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1455 also that even though configuring the SRP seed attempts to hide
1456 invalid usernames by continuing the handshake with fake
1457 credentials, this behaviour is not constant time and no strong
1458 guarantees are made that the handshake is indistinguishable from
1459 that of a valid user.
1463 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1465 In the BN_hex2bn function the number of hex digits is calculated using an
1466 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1467 large values of |i| this can result in |bn_expand| not allocating any
1468 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1469 field as NULL leading to a subsequent NULL ptr deref. For very large values
1470 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1471 In this case memory is allocated to the internal BIGNUM data field, but it
1472 is insufficiently sized leading to heap corruption. A similar issue exists
1473 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1474 is ever called by user applications with very large untrusted hex/dec data.
1475 This is anticipated to be a rare occurrence.
1477 All OpenSSL internal usage of these functions use data that is not expected
1478 to be untrusted, e.g. config file data or application command line
1479 arguments. If user developed applications generate config file data based
1480 on untrusted data then it is possible that this could also lead to security
1481 consequences. This is also anticipated to be rare.
1483 This issue was reported to OpenSSL by Guido Vranken.
1487 *) Fix memory issues in BIO_*printf functions
1489 The internal |fmtstr| function used in processing a "%s" format string in
1490 the BIO_*printf functions could overflow while calculating the length of a
1491 string and cause an OOB read when printing very long strings.
1493 Additionally the internal |doapr_outch| function can attempt to write to an
1494 OOB memory location (at an offset from the NULL pointer) in the event of a
1495 memory allocation failure. In 1.0.2 and below this could be caused where
1496 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1497 could be in processing a very long "%s" format string. Memory leaks can
1500 The first issue may mask the second issue dependent on compiler behaviour.
1501 These problems could enable attacks where large amounts of untrusted data
1502 is passed to the BIO_*printf functions. If applications use these functions
1503 in this way then they could be vulnerable. OpenSSL itself uses these
1504 functions when printing out human-readable dumps of ASN.1 data. Therefore
1505 applications that print this data could be vulnerable if the data is from
1506 untrusted sources. OpenSSL command line applications could also be
1507 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1508 as command line arguments.
1510 Libssl is not considered directly vulnerable. Additionally certificates etc
1511 received via remote connections via libssl are also unlikely to be able to
1512 trigger these issues because of message size limits enforced within libssl.
1514 This issue was reported to OpenSSL Guido Vranken.
1518 *) Side channel attack on modular exponentiation
1520 A side-channel attack was found which makes use of cache-bank conflicts on
1521 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1522 of RSA keys. The ability to exploit this issue is limited as it relies on
1523 an attacker who has control of code in a thread running on the same
1524 hyper-threaded core as the victim thread which is performing decryptions.
1526 This issue was reported to OpenSSL by Yuval Yarom, The University of
1527 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1528 Nadia Heninger, University of Pennsylvania with more information at
1529 http://cachebleed.info.
1533 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1534 if no keysize is specified with default_bits. This fixes an
1535 omission in an earlier change that changed all RSA/DSA key generation
1536 apps to use 2048 bits by default.
1539 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1540 *) DH small subgroups
1542 Historically OpenSSL only ever generated DH parameters based on "safe"
1543 primes. More recently (in version 1.0.2) support was provided for
1544 generating X9.42 style parameter files such as those required for RFC 5114
1545 support. The primes used in such files may not be "safe". Where an
1546 application is using DH configured with parameters based on primes that are
1547 not "safe" then an attacker could use this fact to find a peer's private
1548 DH exponent. This attack requires that the attacker complete multiple
1549 handshakes in which the peer uses the same private DH exponent. For example
1550 this could be used to discover a TLS server's private DH exponent if it's
1551 reusing the private DH exponent or it's using a static DH ciphersuite.
1553 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1554 TLS. It is not on by default. If the option is not set then the server
1555 reuses the same private DH exponent for the life of the server process and
1556 would be vulnerable to this attack. It is believed that many popular
1557 applications do set this option and would therefore not be at risk.
1559 The fix for this issue adds an additional check where a "q" parameter is
1560 available (as is the case in X9.42 based parameters). This detects the
1561 only known attack, and is the only possible defense for static DH
1562 ciphersuites. This could have some performance impact.
1564 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1565 default and cannot be disabled. This could have some performance impact.
1567 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1571 *) SSLv2 doesn't block disabled ciphers
1573 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1574 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1575 been disabled, provided that the SSLv2 protocol was not also disabled via
1578 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1579 and Sebastian Schinzel.
1583 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1585 *) BN_mod_exp may produce incorrect results on x86_64
1587 There is a carry propagating bug in the x86_64 Montgomery squaring
1588 procedure. No EC algorithms are affected. Analysis suggests that attacks
1589 against RSA and DSA as a result of this defect would be very difficult to
1590 perform and are not believed likely. Attacks against DH are considered just
1591 feasible (although very difficult) because most of the work necessary to
1592 deduce information about a private key may be performed offline. The amount
1593 of resources required for such an attack would be very significant and
1594 likely only accessible to a limited number of attackers. An attacker would
1595 additionally need online access to an unpatched system using the target
1596 private key in a scenario with persistent DH parameters and a private
1597 key that is shared between multiple clients. For example this can occur by
1598 default in OpenSSL DHE based SSL/TLS ciphersuites.
1600 This issue was reported to OpenSSL by Hanno Böck.
1604 *) Certificate verify crash with missing PSS parameter
1606 The signature verification routines will crash with a NULL pointer
1607 dereference if presented with an ASN.1 signature using the RSA PSS
1608 algorithm and absent mask generation function parameter. Since these
1609 routines are used to verify certificate signature algorithms this can be
1610 used to crash any certificate verification operation and exploited in a
1611 DoS attack. Any application which performs certificate verification is
1612 vulnerable including OpenSSL clients and servers which enable client
1615 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1619 *) X509_ATTRIBUTE memory leak
1621 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1622 memory. This structure is used by the PKCS#7 and CMS routines so any
1623 application which reads PKCS#7 or CMS data from untrusted sources is
1624 affected. SSL/TLS is not affected.
1626 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1631 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1632 This changes the decoding behaviour for some invalid messages,
1633 though the change is mostly in the more lenient direction, and
1634 legacy behaviour is preserved as much as possible.
1637 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1639 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1641 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1643 *) Alternate chains certificate forgery
1645 During certificate verification, OpenSSL will attempt to find an
1646 alternative certificate chain if the first attempt to build such a chain
1647 fails. An error in the implementation of this logic can mean that an
1648 attacker could cause certain checks on untrusted certificates to be
1649 bypassed, such as the CA flag, enabling them to use a valid leaf
1650 certificate to act as a CA and "issue" an invalid certificate.
1652 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1656 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1658 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1659 incompatibility in the handling of HMAC. The previous ABI has now been
1663 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1665 *) Malformed ECParameters causes infinite loop
1667 When processing an ECParameters structure OpenSSL enters an infinite loop
1668 if the curve specified is over a specially malformed binary polynomial
1671 This can be used to perform denial of service against any
1672 system which processes public keys, certificate requests or
1673 certificates. This includes TLS clients and TLS servers with
1674 client authentication enabled.
1676 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1680 *) Exploitable out-of-bounds read in X509_cmp_time
1682 X509_cmp_time does not properly check the length of the ASN1_TIME
1683 string and can read a few bytes out of bounds. In addition,
1684 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1687 An attacker can use this to craft malformed certificates and CRLs of
1688 various sizes and potentially cause a segmentation fault, resulting in
1689 a DoS on applications that verify certificates or CRLs. TLS clients
1690 that verify CRLs are affected. TLS clients and servers with client
1691 authentication enabled may be affected if they use custom verification
1694 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1695 independently by Hanno Böck.
1699 *) PKCS7 crash with missing EnvelopedContent
1701 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1702 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1703 with missing content and trigger a NULL pointer dereference on parsing.
1705 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1706 structures from untrusted sources are affected. OpenSSL clients and
1707 servers are not affected.
1709 This issue was reported to OpenSSL by Michal Zalewski (Google).
1713 *) CMS verify infinite loop with unknown hash function
1715 When verifying a signedData message the CMS code can enter an infinite loop
1716 if presented with an unknown hash function OID. This can be used to perform
1717 denial of service against any system which verifies signedData messages using
1719 This issue was reported to OpenSSL by Johannes Bauer.
1723 *) Race condition handling NewSessionTicket
1725 If a NewSessionTicket is received by a multi-threaded client when attempting to
1726 reuse a previous ticket then a race condition can occur potentially leading to
1727 a double free of the ticket data.
1731 *) Only support 256-bit or stronger elliptic curves with the
1732 'ecdh_auto' setting (server) or by default (client). Of supported
1733 curves, prefer P-256 (both).
1736 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1738 *) ClientHello sigalgs DoS fix
1740 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1741 invalid signature algorithms extension a NULL pointer dereference will
1742 occur. This can be exploited in a DoS attack against the server.
1744 This issue was was reported to OpenSSL by David Ramos of Stanford
1747 [Stephen Henson and Matt Caswell]
1749 *) Multiblock corrupted pointer fix
1751 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1752 feature only applies on 64 bit x86 architecture platforms that support AES
1753 NI instructions. A defect in the implementation of "multiblock" can cause
1754 OpenSSL's internal write buffer to become incorrectly set to NULL when
1755 using non-blocking IO. Typically, when the user application is using a
1756 socket BIO for writing, this will only result in a failed connection.
1757 However if some other BIO is used then it is likely that a segmentation
1758 fault will be triggered, thus enabling a potential DoS attack.
1760 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1764 *) Segmentation fault in DTLSv1_listen fix
1766 The DTLSv1_listen function is intended to be stateless and processes the
1767 initial ClientHello from many peers. It is common for user code to loop
1768 over the call to DTLSv1_listen until a valid ClientHello is received with
1769 an associated cookie. A defect in the implementation of DTLSv1_listen means
1770 that state is preserved in the SSL object from one invocation to the next
1771 that can lead to a segmentation fault. Errors processing the initial
1772 ClientHello can trigger this scenario. An example of such an error could be
1773 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1776 This issue was reported to OpenSSL by Per Allansson.
1780 *) Segmentation fault in ASN1_TYPE_cmp fix
1782 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1783 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1784 certificate signature algorithm consistency this can be used to crash any
1785 certificate verification operation and exploited in a DoS attack. Any
1786 application which performs certificate verification is vulnerable including
1787 OpenSSL clients and servers which enable client authentication.
1791 *) Segmentation fault for invalid PSS parameters fix
1793 The signature verification routines will crash with a NULL pointer
1794 dereference if presented with an ASN.1 signature using the RSA PSS
1795 algorithm and invalid parameters. Since these routines are used to verify
1796 certificate signature algorithms this can be used to crash any
1797 certificate verification operation and exploited in a DoS attack. Any
1798 application which performs certificate verification is vulnerable including
1799 OpenSSL clients and servers which enable client authentication.
1801 This issue was was reported to OpenSSL by Brian Carpenter.
1805 *) ASN.1 structure reuse memory corruption fix
1807 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1808 memory corruption via an invalid write. Such reuse is and has been
1809 strongly discouraged and is believed to be rare.
1811 Applications that parse structures containing CHOICE or ANY DEFINED BY
1812 components may be affected. Certificate parsing (d2i_X509 and related
1813 functions) are however not affected. OpenSSL clients and servers are
1818 *) PKCS7 NULL pointer dereferences fix
1820 The PKCS#7 parsing code does not handle missing outer ContentInfo
1821 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1822 missing content and trigger a NULL pointer dereference on parsing.
1824 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1825 otherwise parse PKCS#7 structures from untrusted sources are
1826 affected. OpenSSL clients and servers are not affected.
1828 This issue was reported to OpenSSL by Michal Zalewski (Google).
1832 *) DoS via reachable assert in SSLv2 servers fix
1834 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1835 servers that both support SSLv2 and enable export cipher suites by sending
1836 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1838 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1839 (OpenSSL development team).
1843 *) Empty CKE with client auth and DHE fix
1845 If client auth is used then a server can seg fault in the event of a DHE
1846 ciphersuite being selected and a zero length ClientKeyExchange message
1847 being sent by the client. This could be exploited in a DoS attack.
1851 *) Handshake with unseeded PRNG fix
1853 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1854 with an unseeded PRNG. The conditions are:
1855 - The client is on a platform where the PRNG has not been seeded
1856 automatically, and the user has not seeded manually
1857 - A protocol specific client method version has been used (i.e. not
1858 SSL_client_methodv23)
1859 - A ciphersuite is used that does not require additional random data from
1860 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1862 If the handshake succeeds then the client random that has been used will
1863 have been generated from a PRNG with insufficient entropy and therefore the
1864 output may be predictable.
1866 For example using the following command with an unseeded openssl will
1867 succeed on an unpatched platform:
1869 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1873 *) Use After Free following d2i_ECPrivatekey error fix
1875 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1876 could cause a use after free condition. This, in turn, could cause a double
1877 free in several private key parsing functions (such as d2i_PrivateKey
1878 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1879 for applications that receive EC private keys from untrusted
1880 sources. This scenario is considered rare.
1882 This issue was discovered by the BoringSSL project and fixed in their
1887 *) X509_to_X509_REQ NULL pointer deref fix
1889 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1890 the certificate key is invalid. This function is rarely used in practice.
1892 This issue was discovered by Brian Carpenter.
1896 *) Removed the export ciphers from the DEFAULT ciphers
1899 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1901 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1902 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1903 So far those who have to target multiple platforms would compromise
1904 and argue that binary targeting say ARMv5 would still execute on
1905 ARMv8. "Universal" build resolves this compromise by providing
1906 near-optimal performance even on newer platforms.
1909 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1910 (other platforms pending).
1911 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1913 *) Add support for the SignedCertificateTimestampList certificate and
1914 OCSP response extensions from RFC6962.
1917 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1918 for corner cases. (Certain input points at infinity could lead to
1919 bogus results, with non-infinity inputs mapped to infinity too.)
1922 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1923 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1924 common cases are optimized and there still is room for further
1925 improvements. Vector Permutation AES for Altivec is also added.
1928 *) Add support for little-endian ppc64 Linux target.
1929 [Marcelo Cerri (IBM)]
1931 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1932 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1933 are optimized and there still is room for further improvements.
1934 Both 32- and 64-bit modes are supported.
1935 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1937 *) Improved ARMv7 NEON support.
1940 *) Support for SPARC Architecture 2011 crypto extensions, first
1941 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1942 SHA256/512, MD5, GHASH and modular exponentiation.
1943 [Andy Polyakov, David Miller]
1945 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1947 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1949 *) Support for new and upcoming Intel processors, including AVX2,
1950 BMI and SHA ISA extensions. This includes additional "stitched"
1951 implementations, AESNI-SHA256 and GCM, and multi-buffer support
1954 This work was sponsored by Intel Corp.
1957 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
1958 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
1959 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
1962 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1963 this fixes a limitation in previous versions of OpenSSL.
1966 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1967 MGF1 digest and OAEP label.
1970 *) Add EVP support for key wrapping algorithms, to avoid problems with
1971 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1972 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1973 algorithms and include tests cases.
1976 *) Add functions to allocate and set the fields of an ECDSA_METHOD
1978 [Douglas E. Engert, Steve Henson]
1980 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
1981 difference in days and seconds between two tm or ASN1_TIME structures.
1984 *) Add -rev test option to s_server to just reverse order of characters
1985 received by client and send back to server. Also prints an abbreviated
1986 summary of the connection parameters.
1989 *) New option -brief for s_client and s_server to print out a brief summary
1990 of connection parameters.
1993 *) Add callbacks for arbitrary TLS extensions.
1994 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
1996 *) New option -crl_download in several openssl utilities to download CRLs
1997 from CRLDP extension in certificates.
2000 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2003 *) New function X509_CRL_diff to generate a delta CRL from the difference
2004 of two full CRLs. Add support to "crl" utility.
2007 *) New functions to set lookup_crls function and to retrieve
2008 X509_STORE from X509_STORE_CTX.
2011 *) Print out deprecated issuer and subject unique ID fields in
2015 *) Extend OCSP I/O functions so they can be used for simple general purpose
2016 HTTP as well as OCSP. New wrapper function which can be used to download
2017 CRLs using the OCSP API.
2020 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2023 *) SSL_CONF* functions. These provide a common framework for application
2024 configuration using configuration files or command lines.
2027 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2028 message callback and prints the results. Needs compile time option
2029 "enable-ssl-trace". New options to s_client and s_server to enable
2033 *) New ctrl and macro to retrieve supported points extensions.
2034 Print out extension in s_server and s_client.
2037 *) New functions to retrieve certificate signature and signature
2041 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2045 *) New Suite B modes for TLS code. These use and enforce the requirements
2046 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2047 only use Suite B curves. The Suite B modes can be set by using the
2048 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2051 *) New chain verification flags for Suite B levels of security. Check
2052 algorithms are acceptable when flags are set in X509_verify_cert.
2055 *) Make tls1_check_chain return a set of flags indicating checks passed
2056 by a certificate chain. Add additional tests to handle client
2057 certificates: checks for matching certificate type and issuer name
2061 *) If an attempt is made to use a signature algorithm not in the peer
2062 preference list abort the handshake. If client has no suitable
2063 signature algorithms in response to a certificate request do not
2064 use the certificate.
2067 *) If server EC tmp key is not in client preference list abort handshake.
2070 *) Add support for certificate stores in CERT structure. This makes it
2071 possible to have different stores per SSL structure or one store in
2072 the parent SSL_CTX. Include distinct stores for certificate chain
2073 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2074 to build and store a certificate chain in CERT structure: returning
2075 an error if the chain cannot be built: this will allow applications
2076 to test if a chain is correctly configured.
2078 Note: if the CERT based stores are not set then the parent SSL_CTX
2079 store is used to retain compatibility with existing behaviour.
2083 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2084 mask based on the current session, check mask when sending client
2085 hello and checking the requested ciphersuite.
2088 *) New ctrls to retrieve and set certificate types in a certificate
2089 request message. Print out received values in s_client. If certificate
2090 types is not set with custom values set sensible values based on
2091 supported signature algorithms.
2094 *) Support for distinct client and server supported signature algorithms.
2097 *) Add certificate callback. If set this is called whenever a certificate
2098 is required by client or server. An application can decide which
2099 certificate chain to present based on arbitrary criteria: for example
2100 supported signature algorithms. Add very simple example to s_server.
2101 This fixes many of the problems and restrictions of the existing client
2102 certificate callback: for example you can now clear an existing
2103 certificate and specify the whole chain.
2106 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2107 the certificate can be used for (if anything). Set valid_flags field
2108 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2109 to have similar checks in it.
2111 Add new "cert_flags" field to CERT structure and include a "strict mode".
2112 This enforces some TLS certificate requirements (such as only permitting
2113 certificate signature algorithms contained in the supported algorithms
2114 extension) which some implementations ignore: this option should be used
2115 with caution as it could cause interoperability issues.
2118 *) Update and tidy signature algorithm extension processing. Work out
2119 shared signature algorithms based on preferences and peer algorithms
2120 and print them out in s_client and s_server. Abort handshake if no
2121 shared signature algorithms.
2124 *) Add new functions to allow customised supported signature algorithms
2125 for SSL and SSL_CTX structures. Add options to s_client and s_server
2129 *) New function SSL_certs_clear() to delete all references to certificates
2130 from an SSL structure. Before this once a certificate had been added
2131 it couldn't be removed.
2134 *) Integrate hostname, email address and IP address checking with certificate
2135 verification. New verify options supporting checking in openssl utility.
2138 *) Fixes and wildcard matching support to hostname and email checking
2139 functions. Add manual page.
2140 [Florian Weimer (Red Hat Product Security Team)]
2142 *) New functions to check a hostname email or IP address against a
2143 certificate. Add options x509 utility to print results of checks against
2147 *) Fix OCSP checking.
2148 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2150 *) Initial experimental support for explicitly trusted non-root CAs.
2151 OpenSSL still tries to build a complete chain to a root but if an
2152 intermediate CA has a trust setting included that is used. The first
2153 setting is used: whether to trust (e.g., -addtrust option to the x509
2157 *) Add -trusted_first option which attempts to find certificates in the
2158 trusted store even if an untrusted chain is also supplied.
2161 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2162 platform support for Linux and Android.
2165 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2168 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2169 When in FIPS mode the approved implementations are used as normal,
2170 when not in FIPS mode the internal unapproved versions are used instead.
2171 This means that the FIPS capable OpenSSL isn't forced to use the
2172 (often lower performance) FIPS implementations outside FIPS mode.
2175 *) Transparently support X9.42 DH parameters when calling
2176 PEM_read_bio_DHparameters. This means existing applications can handle
2177 the new parameter format automatically.
2180 *) Initial experimental support for X9.42 DH parameter format: mainly
2181 to support use of 'q' parameter for RFC5114 parameters.
2184 *) Add DH parameters from RFC5114 including test data to dhtest.
2187 *) Support for automatic EC temporary key parameter selection. If enabled
2188 the most preferred EC parameters are automatically used instead of
2189 hardcoded fixed parameters. Now a server just has to call:
2190 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2191 support ECDH and use the most appropriate parameters.
2194 *) Enhance and tidy EC curve and point format TLS extension code. Use
2195 static structures instead of allocation if default values are used.
2196 New ctrls to set curves we wish to support and to retrieve shared curves.
2197 Print out shared curves in s_server. New options to s_server and s_client
2198 to set list of supported curves.
2201 *) New ctrls to retrieve supported signature algorithms and
2202 supported curve values as an array of NIDs. Extend openssl utility
2203 to print out received values.
2206 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2207 between NIDs and the more common NIST names such as "P-256". Enhance
2208 ecparam utility and ECC method to recognise the NIST names for curves.
2211 *) Enhance SSL/TLS certificate chain handling to support different
2212 chains for each certificate instead of one chain in the parent SSL_CTX.
2215 *) Support for fixed DH ciphersuite client authentication: where both
2216 server and client use DH certificates with common parameters.
2219 *) Support for fixed DH ciphersuites: those requiring DH server
2223 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2225 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2226 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2227 X509_CINF_get_signature were reverted post internal team review.
2229 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2231 *) Build fixes for the Windows and OpenVMS platforms
2232 [Matt Caswell and Richard Levitte]
2234 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2236 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2237 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2238 dereference. This could lead to a Denial Of Service attack. Thanks to
2239 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2243 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2244 dtls1_buffer_record function under certain conditions. In particular this
2245 could occur if an attacker sent repeated DTLS records with the same
2246 sequence number but for the next epoch. The memory leak could be exploited
2247 by an attacker in a Denial of Service attack through memory exhaustion.
2248 Thanks to Chris Mueller for reporting this issue.
2252 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2253 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2254 method would be set to NULL which could later result in a NULL pointer
2255 dereference. Thanks to Frank Schmirler for reporting this issue.
2259 *) Abort handshake if server key exchange message is omitted for ephemeral
2262 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2263 reporting this issue.
2267 *) Remove non-export ephemeral RSA code on client and server. This code
2268 violated the TLS standard by allowing the use of temporary RSA keys in
2269 non-export ciphersuites and could be used by a server to effectively
2270 downgrade the RSA key length used to a value smaller than the server
2271 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2272 INRIA or reporting this issue.
2276 *) Fixed issue where DH client certificates are accepted without verification.
2277 An OpenSSL server will accept a DH certificate for client authentication
2278 without the certificate verify message. This effectively allows a client to
2279 authenticate without the use of a private key. This only affects servers
2280 which trust a client certificate authority which issues certificates
2281 containing DH keys: these are extremely rare and hardly ever encountered.
2282 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2287 *) Ensure that the session ID context of an SSL is updated when its
2288 SSL_CTX is updated via SSL_set_SSL_CTX.
2290 The session ID context is typically set from the parent SSL_CTX,
2291 and can vary with the CTX.
2294 *) Fix various certificate fingerprint issues.
2296 By using non-DER or invalid encodings outside the signed portion of a
2297 certificate the fingerprint can be changed without breaking the signature.
2298 Although no details of the signed portion of the certificate can be changed
2299 this can cause problems with some applications: e.g. those using the
2300 certificate fingerprint for blacklists.
2302 1. Reject signatures with non zero unused bits.
2304 If the BIT STRING containing the signature has non zero unused bits reject
2305 the signature. All current signature algorithms require zero unused bits.
2307 2. Check certificate algorithm consistency.
2309 Check the AlgorithmIdentifier inside TBS matches the one in the
2310 certificate signature. NB: this will result in signature failure
2311 errors for some broken certificates.
2313 Thanks to Konrad Kraszewski from Google for reporting this issue.
2315 3. Check DSA/ECDSA signatures use DER.
2317 Re-encode DSA/ECDSA signatures and compare with the original received
2318 signature. Return an error if there is a mismatch.
2320 This will reject various cases including garbage after signature
2321 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2322 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2323 (negative or with leading zeroes).
2325 Further analysis was conducted and fixes were developed by Stephen Henson
2326 of the OpenSSL core team.
2331 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2332 results on some platforms, including x86_64. This bug occurs at random
2333 with a very low probability, and is not known to be exploitable in any
2334 way, though its exact impact is difficult to determine. Thanks to Pieter
2335 Wuille (Blockstream) who reported this issue and also suggested an initial
2336 fix. Further analysis was conducted by the OpenSSL development team and
2337 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2338 the OpenSSL core team.
2342 *) Do not resume sessions on the server if the negotiated protocol
2343 version does not match the session's version. Resuming with a different
2344 version, while not strictly forbidden by the RFC, is of questionable
2345 sanity and breaks all known clients.
2346 [David Benjamin, Emilia Käsper]
2348 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2349 early CCS messages during renegotiation. (Note that because
2350 renegotiation is encrypted, this early CCS was not exploitable.)
2353 *) Tighten client-side session ticket handling during renegotiation:
2354 ensure that the client only accepts a session ticket if the server sends
2355 the extension anew in the ServerHello. Previously, a TLS client would
2356 reuse the old extension state and thus accept a session ticket if one was
2357 announced in the initial ServerHello.
2359 Similarly, ensure that the client requires a session ticket if one
2360 was advertised in the ServerHello. Previously, a TLS client would
2361 ignore a missing NewSessionTicket message.
2364 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2366 *) SRTP Memory Leak.
2368 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2369 sends a carefully crafted handshake message, to cause OpenSSL to fail
2370 to free up to 64k of memory causing a memory leak. This could be
2371 exploited in a Denial Of Service attack. This issue affects OpenSSL
2372 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2373 whether SRTP is used or configured. Implementations of OpenSSL that
2374 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2376 The fix was developed by the OpenSSL team.
2380 *) Session Ticket Memory Leak.
2382 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2383 integrity of that ticket is first verified. In the event of a session
2384 ticket integrity check failing, OpenSSL will fail to free memory
2385 causing a memory leak. By sending a large number of invalid session
2386 tickets an attacker could exploit this issue in a Denial Of Service
2391 *) Build option no-ssl3 is incomplete.
2393 When OpenSSL is configured with "no-ssl3" as a build option, servers
2394 could accept and complete a SSL 3.0 handshake, and clients could be
2395 configured to send them.
2397 [Akamai and the OpenSSL team]
2399 *) Add support for TLS_FALLBACK_SCSV.
2400 Client applications doing fallback retries should call
2401 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2403 [Adam Langley, Bodo Moeller]
2405 *) Add additional DigestInfo checks.
2407 Re-encode DigestInto in DER and check against the original when
2408 verifying RSA signature: this will reject any improperly encoded
2409 DigestInfo structures.
2411 Note: this is a precautionary measure and no attacks are currently known.
2415 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2417 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2418 SRP code can be overrun an internal buffer. Add sanity check that
2419 g, A, B < N to SRP code.
2421 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2422 Group for discovering this issue.
2426 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2427 TLS 1.0 instead of higher protocol versions when the ClientHello message
2428 is badly fragmented. This allows a man-in-the-middle attacker to force a
2429 downgrade to TLS 1.0 even if both the server and the client support a
2430 higher protocol version, by modifying the client's TLS records.
2432 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2433 researching this issue.
2437 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2438 to a denial of service attack. A malicious server can crash the client
2439 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2440 ciphersuite and sending carefully crafted handshake messages.
2442 Thanks to Felix Gröbert (Google) for discovering and researching this
2447 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2448 to leak memory. This can be exploited through a Denial of Service attack.
2449 Thanks to Adam Langley for discovering and researching this issue.
2453 *) An attacker can force openssl to consume large amounts of memory whilst
2454 processing DTLS handshake messages. This can be exploited through a
2455 Denial of Service attack.
2456 Thanks to Adam Langley for discovering and researching this issue.
2460 *) An attacker can force an error condition which causes openssl to crash
2461 whilst processing DTLS packets due to memory being freed twice. This
2462 can be exploited through a Denial of Service attack.
2463 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2468 *) If a multithreaded client connects to a malicious server using a resumed
2469 session and the server sends an ec point format extension it could write
2470 up to 255 bytes to freed memory.
2472 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2477 *) A malicious server can crash an OpenSSL client with a null pointer
2478 dereference (read) by specifying an SRP ciphersuite even though it was not
2479 properly negotiated with the client. This can be exploited through a
2480 Denial of Service attack.
2482 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2483 discovering and researching this issue.
2487 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2488 X509_name_oneline, X509_name_print_ex et al. to leak some information
2489 from the stack. Applications may be affected if they echo pretty printing
2490 output to the attacker.
2492 Thanks to Ivan Fratric (Google) for discovering this issue.
2494 [Emilia Käsper, and Steve Henson]
2496 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2497 for corner cases. (Certain input points at infinity could lead to
2498 bogus results, with non-infinity inputs mapped to infinity too.)
2501 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2503 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2504 handshake can force the use of weak keying material in OpenSSL
2505 SSL/TLS clients and servers.
2507 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2508 researching this issue. (CVE-2014-0224)
2509 [KIKUCHI Masashi, Steve Henson]
2511 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2512 OpenSSL DTLS client the code can be made to recurse eventually crashing
2515 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2517 [Imre Rad, Steve Henson]
2519 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2520 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2521 client or server. This is potentially exploitable to run arbitrary
2522 code on a vulnerable client or server.
2524 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2525 [Jüri Aedla, Steve Henson]
2527 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2528 are subject to a denial of service attack.
2530 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2531 this issue. (CVE-2014-3470)
2532 [Felix Gröbert, Ivan Fratric, Steve Henson]
2534 *) Harmonize version and its documentation. -f flag is used to display
2536 [mancha <mancha1@zoho.com>]
2538 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2539 in i2d_ECPrivateKey.
2540 [mancha <mancha1@zoho.com>]
2542 *) Fix some double frees. These are not thought to be exploitable.
2543 [mancha <mancha1@zoho.com>]
2545 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2547 *) A missing bounds check in the handling of the TLS heartbeat extension
2548 can be used to reveal up to 64k of memory to a connected client or
2551 Thanks for Neel Mehta of Google Security for discovering this bug and to
2552 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2553 preparing the fix (CVE-2014-0160)
2554 [Adam Langley, Bodo Moeller]
2556 *) Fix for the attack described in the paper "Recovering OpenSSL
2557 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2558 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2559 http://eprint.iacr.org/2014/140
2561 Thanks to Yuval Yarom and Naomi Benger for discovering this
2562 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2563 [Yuval Yarom and Naomi Benger]
2565 *) TLS pad extension: draft-agl-tls-padding-03
2567 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2568 TLS client Hello record length value would otherwise be > 255 and
2569 less that 512 pad with a dummy extension containing zeroes so it
2570 is at least 512 bytes long.
2572 [Adam Langley, Steve Henson]
2574 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2576 *) Fix for TLS record tampering bug. A carefully crafted invalid
2577 handshake could crash OpenSSL with a NULL pointer exception.
2578 Thanks to Anton Johansson for reporting this issues.
2581 *) Keep original DTLS digest and encryption contexts in retransmission
2582 structures so we can use the previous session parameters if they need
2583 to be resent. (CVE-2013-6450)
2586 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2587 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2588 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2589 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2590 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2591 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2592 [Rob Stradling, Adam Langley]
2594 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2596 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2597 supporting platforms or when small records were transferred.
2598 [Andy Polyakov, Steve Henson]
2600 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2602 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2604 This addresses the flaw in CBC record processing discovered by
2605 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2606 at: http://www.isg.rhul.ac.uk/tls/
2608 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2609 Security Group at Royal Holloway, University of London
2610 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2611 Emilia Käsper for the initial patch.
2613 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2615 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2616 ciphersuites which can be exploited in a denial of service attack.
2617 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2618 and detecting this bug and to Wolfgang Ettlinger
2619 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2623 *) Return an error when checking OCSP signatures when key is NULL.
2624 This fixes a DoS attack. (CVE-2013-0166)
2627 *) Make openssl verify return errors.
2628 [Chris Palmer <palmer@google.com> and Ben Laurie]
2630 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2631 the right response is stapled. Also change SSL_get_certificate()
2632 so it returns the certificate actually sent.
2633 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2634 [Rob Stradling <rob.stradling@comodo.com>]
2636 *) Fix possible deadlock when decoding public keys.
2639 *) Don't use TLS 1.0 record version number in initial client hello
2643 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2645 *) Sanity check record length before skipping explicit IV in TLS
2646 1.2, 1.1 and DTLS to fix DoS attack.
2648 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2649 fuzzing as a service testing platform.
2653 *) Initialise tkeylen properly when encrypting CMS messages.
2654 Thanks to Solar Designer of Openwall for reporting this issue.
2657 *) In FIPS mode don't try to use composite ciphers as they are not
2661 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2663 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2664 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2665 mean any application compiled against OpenSSL 1.0.0 headers setting
2666 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2667 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2668 0x10000000L Any application which was previously compiled against
2669 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2670 will need to be recompiled as a result. Letting be results in
2671 inability to disable specifically TLS 1.1 and in client context,
2672 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2675 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2676 disable just protocol X, but all protocols above X *if* there are
2677 protocols *below* X still enabled. In more practical terms it means
2678 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2679 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2680 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2684 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2686 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2687 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2688 in CRYPTO_realloc_clean.
2690 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2691 issue and to Adam Langley <agl@chromium.org> for fixing it.
2693 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2695 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2698 *) Workarounds for some broken servers that "hang" if a client hello
2699 record length exceeds 255 bytes.
2701 1. Do not use record version number > TLS 1.0 in initial client
2702 hello: some (but not all) hanging servers will now work.
2703 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2704 the number of ciphers sent in the client hello. This should be
2705 set to an even number, such as 50, for example by passing:
2706 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2707 Most broken servers should now work.
2708 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2709 TLS 1.2 client support entirely.
2712 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2715 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2717 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2718 STRING form instead of a DigestInfo.
2721 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2722 and the RSA_sign/RSA_verify functions. This was made more apparent when
2723 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2724 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2725 the correct format in RSA_verify so both forms transparently work.
2728 *) Some servers which support TLS 1.0 can choke if we initially indicate
2729 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2730 encrypted premaster secret. As a workaround use the maximum permitted
2731 client version in client hello, this should keep such servers happy
2732 and still work with previous versions of OpenSSL.
2735 *) Add support for TLS/DTLS heartbeats.
2736 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2738 *) Add support for SCTP.
2739 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2741 *) Improved PRNG seeding for VOS.
2742 [Paul Green <Paul.Green@stratus.com>]
2744 *) Extensive assembler packs updates, most notably:
2746 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2747 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2748 - x86_64: bit-sliced AES implementation;
2749 - ARM: NEON support, contemporary platforms optimizations;
2750 - s390x: z196 support;
2751 - *: GHASH and GF(2^m) multiplication implementations;
2755 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2756 (removal of unnecessary code)
2757 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2759 *) Add TLS key material exporter from RFC 5705.
2762 *) Add DTLS-SRTP negotiation from RFC 5764.
2765 *) Add Next Protocol Negotiation,
2766 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2767 disabled with a no-npn flag to config or Configure. Code donated
2769 [Adam Langley <agl@google.com> and Ben Laurie]
2771 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2772 NIST-P256, NIST-P521, with constant-time single point multiplication on
2773 typical inputs. Compiler support for the nonstandard type __uint128_t is
2774 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2775 Code made available under Apache License version 2.0.
2777 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2778 line to include this in your build of OpenSSL, and run "make depend" (or
2779 "make update"). This enables the following EC_METHODs:
2781 EC_GFp_nistp224_method()
2782 EC_GFp_nistp256_method()
2783 EC_GFp_nistp521_method()
2785 EC_GROUP_new_by_curve_name() will automatically use these (while
2786 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2788 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2790 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2791 all platforms. Move ssize_t definition from e_os.h to the public
2792 header file e_os2.h as it now appears in public header file cms.h
2795 *) New -sigopt option to the ca, req and x509 utilities. Additional
2796 signature parameters can be passed using this option and in
2800 *) Add RSA PSS signing function. This will generate and set the
2801 appropriate AlgorithmIdentifiers for PSS based on those in the
2802 corresponding EVP_MD_CTX structure. No application support yet.
2805 *) Support for companion algorithm specific ASN1 signing routines.
2806 New function ASN1_item_sign_ctx() signs a pre-initialised
2807 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2808 the appropriate parameters.
2811 *) Add new algorithm specific ASN1 verification initialisation function
2812 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2813 handling will be the same no matter what EVP_PKEY_METHOD is used.
2814 Add a PSS handler to support verification of PSS signatures: checked
2815 against a number of sample certificates.
2818 *) Add signature printing for PSS. Add PSS OIDs.
2819 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2821 *) Add algorithm specific signature printing. An individual ASN1 method
2822 can now print out signatures instead of the standard hex dump.
2824 More complex signatures (e.g. PSS) can print out more meaningful
2825 information. Include DSA version that prints out the signature
2829 *) Password based recipient info support for CMS library: implementing
2833 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2834 neatly separates the code into cipher and PBE sections and is required
2835 for some algorithms that split PBES2 into separate pieces (such as
2836 password based CMS).
2839 *) Session-handling fixes:
2840 - Fix handling of connections that are resuming with a session ID,
2841 but also support Session Tickets.
2842 - Fix a bug that suppressed issuing of a new ticket if the client
2843 presented a ticket with an expired session.
2844 - Try to set the ticket lifetime hint to something reasonable.
2845 - Make tickets shorter by excluding irrelevant information.
2846 - On the client side, don't ignore renewed tickets.
2847 [Adam Langley, Bodo Moeller (Google)]
2849 *) Fix PSK session representation.
2852 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2854 This work was sponsored by Intel.
2857 *) Add GCM support to TLS library. Some custom code is needed to split
2858 the IV between the fixed (from PRF) and explicit (from TLS record)
2859 portions. This adds all GCM ciphersuites supported by RFC5288 and
2860 RFC5289. Generalise some AES* cipherstrings to include GCM and
2861 add a special AESGCM string for GCM only.
2864 *) Expand range of ctrls for AES GCM. Permit setting invocation
2865 field on decrypt and retrieval of invocation field only on encrypt.
2868 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2869 As required by RFC5289 these ciphersuites cannot be used if for
2870 versions of TLS earlier than 1.2.
2873 *) For FIPS capable OpenSSL interpret a NULL default public key method
2874 as unset and return the appropriate default but do *not* set the default.
2875 This means we can return the appropriate method in applications that
2876 switch between FIPS and non-FIPS modes.
2879 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2880 ENGINE is used then we cannot handle that in the FIPS module so we
2881 keep original code iff non-FIPS operations are allowed.
2884 *) Add -attime option to openssl utilities.
2885 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2887 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2890 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2891 FIPS EC methods unconditionally for now.
2894 *) New build option no-ec2m to disable characteristic 2 code.
2897 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2898 all cases can be covered as some introduce binary incompatibilities.
2901 *) Redirect RSA operations to FIPS module including keygen,
2902 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2905 *) Add similar low level API blocking to ciphers.
2908 *) Low level digest APIs are not approved in FIPS mode: any attempt
2909 to use these will cause a fatal error. Applications that *really* want
2910 to use them can use the private_* version instead.
2913 *) Redirect cipher operations to FIPS module for FIPS builds.
2916 *) Redirect digest operations to FIPS module for FIPS builds.
2919 *) Update build system to add "fips" flag which will link in fipscanister.o
2920 for static and shared library builds embedding a signature if needed.
2923 *) Output TLS supported curves in preference order instead of numerical
2924 order. This is currently hardcoded for the highest order curves first.
2925 This should be configurable so applications can judge speed vs strength.
2928 *) Add TLS v1.2 server support for client authentication.
2931 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2935 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2936 FIPS modules versions.
2939 *) Add TLS v1.2 client side support for client authentication. Keep cache
2940 of handshake records longer as we don't know the hash algorithm to use
2941 until after the certificate request message is received.
2944 *) Initial TLS v1.2 client support. Add a default signature algorithms
2945 extension including all the algorithms we support. Parse new signature
2946 format in client key exchange. Relax some ECC signing restrictions for
2947 TLS v1.2 as indicated in RFC5246.
2950 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2951 to new signature format when needed using client digest preference.
2952 All server ciphersuites should now work correctly in TLS v1.2. No client
2953 support yet and no support for client certificates.
2956 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
2957 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
2958 ciphersuites. At present only RSA key exchange ciphersuites work with
2959 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
2960 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
2961 and version checking.
2964 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
2965 with this defined it will not be affected by any changes to ssl internal
2966 structures. Add several utility functions to allow openssl application
2967 to work with OPENSSL_NO_SSL_INTERN defined.
2971 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
2973 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
2976 *) Permit abbreviated handshakes when renegotiating using the function
2977 SSL_renegotiate_abbreviated().
2978 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2980 *) Add call to ENGINE_register_all_complete() to
2981 ENGINE_load_builtin_engines(), so some implementations get used
2982 automatically instead of needing explicit application support.
2985 *) Add support for TLS key exporter as described in RFC5705.
2986 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
2988 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
2989 a few changes are required:
2991 Add SSL_OP_NO_TLSv1_1 flag.
2992 Add TLSv1_1 methods.
2993 Update version checking logic to handle version 1.1.
2994 Add explicit IV handling (ported from DTLS code).
2995 Add command line options to s_client/s_server.
2998 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3000 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3001 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3002 content decryption and always return the same error. Note: this attack
3003 needs on average 2^20 messages so it only affects automated senders. The
3004 old behaviour can be re-enabled in the CMS code by setting the
3005 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3006 an MMA defence is not necessary.
3007 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3008 this issue. (CVE-2012-0884)
3011 *) Fix CVE-2011-4619: make sure we really are receiving a
3012 client hello before rejecting multiple SGC restarts. Thanks to
3013 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3016 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3018 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3019 Thanks to Antonio Martin, Enterprise Secure Access Research and
3020 Development, Cisco Systems, Inc. for discovering this bug and
3021 preparing a fix. (CVE-2012-0050)
3024 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3026 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3027 of the Vaudenay padding oracle attack on CBC mode encryption
3028 which enables an efficient plaintext recovery attack against
3029 the OpenSSL implementation of DTLS. Their attack exploits timing
3030 differences arising during decryption processing. A research
3031 paper describing this attack can be found at:
3032 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3033 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3034 Security Group at Royal Holloway, University of London
3035 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3036 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3037 for preparing the fix. (CVE-2011-4108)
3038 [Robin Seggelmann, Michael Tuexen]
3040 *) Clear bytes used for block padding of SSL 3.0 records.
3042 [Adam Langley (Google)]
3044 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3045 Kadianakis <desnacked@gmail.com> for discovering this issue and
3046 Adam Langley for preparing the fix. (CVE-2011-4619)
3047 [Adam Langley (Google)]
3049 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3050 [Andrey Kulikov <amdeich@gmail.com>]
3052 *) Prevent malformed RFC3779 data triggering an assertion failure.
3053 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3054 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3055 [Rob Austein <sra@hactrn.net>]
3057 *) Improved PRNG seeding for VOS.
3058 [Paul Green <Paul.Green@stratus.com>]
3060 *) Fix ssl_ciph.c set-up race.
3061 [Adam Langley (Google)]
3063 *) Fix spurious failures in ecdsatest.c.
3064 [Emilia Käsper (Google)]
3066 *) Fix the BIO_f_buffer() implementation (which was mixing different
3067 interpretations of the '..._len' fields).
3068 [Adam Langley (Google)]
3070 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3071 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3072 threads won't reuse the same blinding coefficients.
3074 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3075 lock to call BN_BLINDING_invert_ex, and avoids one use of
3076 BN_BLINDING_update for each BN_BLINDING structure (previously,
3077 the last update always remained unused).
3078 [Emilia Käsper (Google)]
3080 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3081 [Bob Buckholz (Google)]
3083 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3085 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3086 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3087 [Kaspar Brand <ossl@velox.ch>]
3089 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3090 for multi-threaded use of ECDH. (CVE-2011-3210)
3091 [Adam Langley (Google)]
3093 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3096 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3097 signature public key algorithm by using OID xref utilities instead.
3098 Before this you could only use some ECC ciphersuites with SHA1 only.
3101 *) Add protection against ECDSA timing attacks as mentioned in the paper
3102 by Billy Bob Brumley and Nicola Tuveri, see:
3104 http://eprint.iacr.org/2011/232.pdf
3106 [Billy Bob Brumley and Nicola Tuveri]
3108 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3110 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3111 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3113 *) Fix bug in string printing code: if *any* escaping is enabled we must
3114 escape the escape character (backslash) or the resulting string is
3118 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3120 *) Disable code workaround for ancient and obsolete Netscape browsers
3121 and servers: an attacker can use it in a ciphersuite downgrade attack.
3122 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3125 *) Fixed J-PAKE implementation error, originally discovered by
3126 Sebastien Martini, further info and confirmation from Stefan
3127 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3130 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3132 *) Fix extension code to avoid race conditions which can result in a buffer
3133 overrun vulnerability: resumed sessions must not be modified as they can
3134 be shared by multiple threads. CVE-2010-3864
3137 *) Fix WIN32 build system to correctly link an ENGINE directory into
3141 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3143 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3145 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3147 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3149 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3150 context. The operation can be customised via the ctrl mechanism in
3151 case ENGINEs want to include additional functionality.
3154 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3157 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3158 output hashes compatible with older versions of OpenSSL.
3159 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3161 *) Fix compression algorithm handling: if resuming a session use the
3162 compression algorithm of the resumed session instead of determining
3163 it from client hello again. Don't allow server to change algorithm.
3166 *) Add load_crls() function to apps tidying load_certs() too. Add option
3167 to verify utility to allow additional CRLs to be included.
3170 *) Update OCSP request code to permit adding custom headers to the request:
3171 some responders need this.
3174 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3176 [Julia Lawall <julia@diku.dk>]
3178 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3179 needlessly dereferenced structures, used obsolete functions and
3180 didn't handle all updated verify codes correctly.
3183 *) Disable MD2 in the default configuration.
3186 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3187 indicate the initial BIO being pushed or popped. This makes it possible
3188 to determine whether the BIO is the one explicitly called or as a result
3189 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3190 it handles reference counts correctly and doesn't zero out the I/O bio
3191 when it is not being explicitly popped. WARNING: applications which
3192 included workarounds for the old buggy behaviour will need to be modified
3193 or they could free up already freed BIOs.
3196 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3197 renaming to all platforms (within the 0.9.8 branch, this was
3198 done conditionally on Netware platforms to avoid a name clash).
3199 [Guenter <lists@gknw.net>]
3201 *) Add ECDHE and PSK support to DTLS.
3202 [Michael Tuexen <tuexen@fh-muenster.de>]
3204 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3208 *) Add "missing" function EVP_MD_flags() (without this the only way to
3209 retrieve a digest flags is by accessing the structure directly. Update
3210 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3211 or cipher is registered as in the "from" argument. Print out all
3212 registered digests in the dgst usage message instead of manually
3213 attempting to work them out.
3216 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3217 this allows the use of compression and extensions. Change default cipher
3218 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3219 by default unless an application cipher string requests it.
3222 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3223 key ids to find matching certificates and keys but some PKCS#12 files
3224 don't follow the (somewhat unwritten) rules and this strategy fails.
3225 Now just gather all certificates together and the first private key
3226 then look for the first certificate that matches the key.
3229 *) Support use of registered digest and cipher names for dgst and cipher
3230 commands instead of having to add each one as a special case. So now
3237 openssl dgst -sha256 foo
3239 and this works for ENGINE based algorithms too.
3243 *) Update Gost ENGINE to support parameter files.
3244 [Victor B. Wagner <vitus@cryptocom.ru>]
3246 *) Support GeneralizedTime in ca utility.
3247 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3249 *) Enhance the hash format used for certificate directory links. The new
3250 form uses the canonical encoding (meaning equivalent names will work
3251 even if they aren't identical) and uses SHA1 instead of MD5. This form
3252 is incompatible with the older format and as a result c_rehash should
3253 be used to rebuild symbolic links.
3256 *) Make PKCS#8 the default write format for private keys, replacing the
3257 traditional format. This form is standardised, more secure and doesn't
3258 include an implicit MD5 dependency.
3261 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3262 committed to OpenSSL should pass this lot as a minimum.
3265 *) Add session ticket override functionality for use by EAP-FAST.
3266 [Jouni Malinen <j@w1.fi>]
3268 *) Modify HMAC functions to return a value. Since these can be implemented
3269 in an ENGINE errors can occur.
3272 *) Type-checked OBJ_bsearch_ex.
3275 *) Type-checked OBJ_bsearch. Also some constification necessitated
3276 by type-checking. Still to come: TXT_DB, bsearch(?),
3277 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3281 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3282 seconds to a tm structure directly, instead of going through OS
3283 specific date routines. This avoids any issues with OS routines such
3284 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3285 and X509_time_adj_ex() to cover the extended range. The existing
3286 X509_time_adj() is still usable and will no longer have any date issues.
3289 *) Delta CRL support. New use deltas option which will attempt to locate
3290 and search any appropriate delta CRLs available.
3292 This work was sponsored by Google.
3295 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3296 code and add additional score elements. Validate alternate CRL paths
3297 as part of the CRL checking and indicate a new error "CRL path validation
3298 error" in this case. Applications wanting additional details can use
3299 the verify callback and check the new "parent" field. If this is not
3300 NULL CRL path validation is taking place. Existing applications won't
3301 see this because it requires extended CRL support which is off by
3304 This work was sponsored by Google.
3307 *) Support for freshest CRL extension.
3309 This work was sponsored by Google.
3312 *) Initial indirect CRL support. Currently only supported in the CRLs
3313 passed directly and not via lookup. Process certificate issuer
3314 CRL entry extension and lookup CRL entries by bother issuer name
3315 and serial number. Check and process CRL issuer entry in IDP extension.
3317 This work was sponsored by Google.
3320 *) Add support for distinct certificate and CRL paths. The CRL issuer
3321 certificate is validated separately in this case. Only enabled if
3322 an extended CRL support flag is set: this flag will enable additional
3323 CRL functionality in future.
3325 This work was sponsored by Google.
3328 *) Add support for policy mappings extension.
3330 This work was sponsored by Google.
3333 *) Fixes to pathlength constraint, self issued certificate handling,
3334 policy processing to align with RFC3280 and PKITS tests.
3336 This work was sponsored by Google.
3339 *) Support for name constraints certificate extension. DN, email, DNS
3340 and URI types are currently supported.
3342 This work was sponsored by Google.
3345 *) To cater for systems that provide a pointer-based thread ID rather
3346 than numeric, deprecate the current numeric thread ID mechanism and
3347 replace it with a structure and associated callback type. This
3348 mechanism allows a numeric "hash" to be extracted from a thread ID in
3349 either case, and on platforms where pointers are larger than 'long',
3350 mixing is done to help ensure the numeric 'hash' is usable even if it
3351 can't be guaranteed unique. The default mechanism is to use "&errno"
3352 as a pointer-based thread ID to distinguish between threads.
3354 Applications that want to provide their own thread IDs should now use
3355 CRYPTO_THREADID_set_callback() to register a callback that will call
3356 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3358 Note that ERR_remove_state() is now deprecated, because it is tied
3359 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3360 to free the current thread's error state should be replaced by
3361 ERR_remove_thread_state(NULL).
3363 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3364 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3365 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3366 application was previously providing a numeric thread callback that
3367 was inappropriate for distinguishing threads, then uniqueness might
3368 have been obtained with &errno that happened immediately in the
3369 intermediate development versions of OpenSSL; this is no longer the
3370 case, the numeric thread callback will now override the automatic use
3372 [Geoff Thorpe, with help from Bodo Moeller]
3374 *) Initial support for different CRL issuing certificates. This covers a
3375 simple case where the self issued certificates in the chain exist and
3376 the real CRL issuer is higher in the existing chain.
3378 This work was sponsored by Google.
3381 *) Removed effectively defunct crypto/store from the build.
3384 *) Revamp of STACK to provide stronger type-checking. Still to come:
3385 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3386 ASN1_STRING, CONF_VALUE.
3389 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3390 RAM on SSL connections. This option can save about 34k per idle SSL.
3393 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3394 STACK, TXT_DB, bsearch, qsort.
3397 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3398 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3399 support for data, signedData, compressedData, digestedData and
3400 encryptedData, envelopedData types included. Scripts to check against
3401 RFC4134 examples draft and interop and consistency checks of many
3402 content types and variants.
3405 *) Add options to enc utility to support use of zlib compression BIO.
3408 *) Extend mk1mf to support importing of options and assembly language
3409 files from Configure script, currently only included in VC-WIN32.
3410 The assembly language rules can now optionally generate the source
3411 files from the associated perl scripts.
3414 *) Implement remaining functionality needed to support GOST ciphersuites.
3415 Interop testing has been performed using CryptoPro implementations.
3416 [Victor B. Wagner <vitus@cryptocom.ru>]
3418 *) s390x assembler pack.
3421 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3425 *) Implement Opaque PRF Input TLS extension as specified in
3426 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3427 official specification yet and no extension type assignment by
3428 IANA exists, this extension (for now) will have to be explicitly
3429 enabled when building OpenSSL by providing the extension number
3430 to use. For example, specify an option
3432 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3434 to the "config" or "Configure" script to enable the extension,
3435 assuming extension number 0x9527 (which is a completely arbitrary
3436 and unofficial assignment based on the MD5 hash of the Internet
3437 Draft). Note that by doing so, you potentially lose
3438 interoperability with other TLS implementations since these might
3439 be using the same extension number for other purposes.
3441 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3442 opaque PRF input value to use in the handshake. This will create
3443 an interal copy of the length-'len' string at 'src', and will
3444 return non-zero for success.
3446 To get more control and flexibility, provide a callback function
3449 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3450 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3454 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3457 Callback function 'cb' will be called in handshakes, and is
3458 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3459 Argument 'arg' is for application purposes (the value as given to
3460 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3461 be provided to the callback function). The callback function
3462 has to return non-zero to report success: usually 1 to use opaque
3463 PRF input just if possible, or 2 to enforce use of the opaque PRF
3464 input. In the latter case, the library will abort the handshake
3465 if opaque PRF input is not successfully negotiated.
3467 Arguments 'peerinput' and 'len' given to the callback function
3468 will always be NULL and 0 in the case of a client. A server will
3469 see the client's opaque PRF input through these variables if
3470 available (NULL and 0 otherwise). Note that if the server
3471 provides an opaque PRF input, the length must be the same as the
3472 length of the client's opaque PRF input.
3474 Note that the callback function will only be called when creating
3475 a new session (session resumption can resume whatever was
3476 previously negotiated), and will not be called in SSL 2.0
3477 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3478 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3479 for applications that need to enforce opaque PRF input.
3483 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3486 [Victor B. Wagner <vitus@cryptocom.ru>]
3488 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3489 RFC4507bis. The encrypted ticket format is an encrypted encoded
3490 SSL_SESSION structure, that way new session features are automatically
3493 If a client application caches session in an SSL_SESSION structure
3494 support is transparent because tickets are now stored in the encoded
3497 The SSL_CTX structure automatically generates keys for ticket
3498 protection in servers so again support should be possible
3499 with no application modification.
3501 If a client or server wishes to disable RFC4507 support then the option
3502 SSL_OP_NO_TICKET can be set.
3504 Add a TLS extension debugging callback to allow the contents of any client
3505 or server extensions to be examined.
3507 This work was sponsored by Google.
3510 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3511 OpenSSL should now compile cleanly on gcc 4.2
3512 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3514 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3515 support including streaming MAC support: this is required for GOST
3516 ciphersuite support.
3517 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3519 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3520 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3521 to output in BER and PEM format.
3524 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3525 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3526 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3527 ENGINE support for HMAC keys which are unextractable. New -mac and
3528 -macopt options to dgst utility.
3531 *) New option -sigopt to dgst utility. Update dgst to use
3532 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3533 alternative signing parameters such as X9.31 or PSS in the dgst
3537 *) Change ssl_cipher_apply_rule(), the internal function that does
3538 the work each time a ciphersuite string requests enabling
3539 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3540 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3541 the order of disabled ciphersuites such that those ciphersuites
3542 that most recently went from enabled to disabled not only stay
3543 in order with respect to each other, but also have higher priority
3544 than other disabled ciphersuites the next time ciphersuites are
3547 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3548 the same ciphersuites as with "HIGH" alone, but in a specific
3549 order where the PSK ciphersuites come first (since they are the
3550 most recently disabled ciphersuites when "HIGH" is parsed).
3552 Also, change ssl_create_cipher_list() (using this new
3553 funcionality) such that between otherwise identical
3554 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3558 *) Change ssl_create_cipher_list() so that it automatically
3559 arranges the ciphersuites in reasonable order before starting
3560 to process the rule string. Thus, the definition for "DEFAULT"
3561 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3562 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3563 This makes it much easier to arrive at a reasonable default order
3564 in applications for which anonymous ciphers are OK (meaning
3565 that you can't actually use DEFAULT).
3566 [Bodo Moeller; suggested by Victor Duchovni]
3568 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3569 processing) into multiple integers instead of setting
3570 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3571 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3572 (These masks as well as the individual bit definitions are hidden
3573 away into the non-exported interface ssl/ssl_locl.h, so this
3574 change to the definition of the SSL_CIPHER structure shouldn't
3575 affect applications.) This give us more bits for each of these
3576 categories, so there is no longer a need to coagulate AES128 and
3577 AES256 into a single algorithm bit, and to coagulate Camellia128
3578 and Camellia256 into a single algorithm bit, which has led to all
3581 Thus, among other things, the kludge introduced in 0.9.7m and
3582 0.9.8e for masking out AES256 independently of AES128 or masking
3583 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3585 With the change, we also introduce new ciphersuite aliases that
3586 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3590 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3591 Use the leftmost N bytes of the signature input if the input is
3592 larger than the prime q (with N being the size in bytes of q).
3595 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3596 it yet and it is largely untested.
3599 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3602 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3603 some compilers (gcc 4.2 and later) reject their use. Safestack is
3604 reimplemented. Update ASN1 to avoid use of legacy functions.
3607 *) Win32/64 targets are linked with Winsock2.
3610 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3611 to external functions. This can be used to increase CRL handling
3612 efficiency especially when CRLs are very large by (for example) storing
3613 the CRL revoked certificates in a database.
3616 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3617 new CRLs added to a directory can be used. New command line option
3618 -verify_return_error to s_client and s_server. This causes real errors
3619 to be returned by the verify callback instead of carrying on no matter
3620 what. This reflects the way a "real world" verify callback would behave.
3623 *) GOST engine, supporting several GOST algorithms and public key formats.
3624 Kindly donated by Cryptocom.
3627 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3628 partitioned by DP are handled but no indirect CRL or reason partitioning
3629 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
projects / openssl.git / blob