5 Changes between 1.1.0e and 1.1.1 [xx XXX xxxx]
7 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
8 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
9 alerts across multiple records (some of which could be empty). In practice
10 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
11 prohibts this altogether and other libraries (BoringSSL, NSS) do not
12 support this at all. Supporting it adds significant complexity to the
13 record layer, and its removal is unlikely to cause inter-operability
17 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
18 with Z. These are meant to replace LONG and ZLONG and to be size safe.
19 The use of LONG and ZLONG is discouraged and scheduled for deprecation
23 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
24 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
25 [Richard Levitte, Andy Polyakov]
27 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
31 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
32 platform rather than 'mingw'.
35 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
36 success if they are asked to add an object which already exists
37 in the store. This change cascades to other functions which load
38 certificates and CRLs.
41 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
42 facilitate stack unwinding even from assembly subroutines.
45 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
46 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
49 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
50 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
51 which is the minimum version we support.
54 *) Certificate time validation (X509_cmp_time) enforces stricter
55 compliance with RFC 5280. Fractional seconds and timezone offsets
56 are no longer allowed.
59 *) Add support for ARIA
62 *) s_client will now send the Server Name Indication (SNI) extension by
63 default unless the new "-noservername" option is used. The server name is
64 based on the host provided to the "-connect" option unless overridden by
68 *) Add support for SipHash
71 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
72 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
73 prevent issues where no progress is being made and the peer continually
74 sends unrecognised record types, using up resources processing them.
77 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
78 using the algorithm defined in
79 https://www.akkadia.org/drepper/SHA-crypt.txt
82 *) Heartbeat support has been removed; the ABI is changed for now.
83 [Richard Levitte, Rich Salz]
85 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
88 *) The RSA "null" method, which was partially supported to avoid patent
89 issues, has been replaced to always returns NULL.
92 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
94 *) Encrypt-Then-Mac renegotiation crash
96 During a renegotiation handshake if the Encrypt-Then-Mac extension is
97 negotiated where it was not in the original handshake (or vice-versa) then
98 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
99 and servers are affected.
101 This issue was reported to OpenSSL by Joe Orton (Red Hat).
105 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
107 *) Truncated packet could crash via OOB read
109 If one side of an SSL/TLS path is running on a 32-bit host and a specific
110 cipher is being used, then a truncated packet can cause that host to
111 perform an out-of-bounds read, usually resulting in a crash.
113 This issue was reported to OpenSSL by Robert Święcki of Google.
117 *) Bad (EC)DHE parameters cause a client crash
119 If a malicious server supplies bad parameters for a DHE or ECDHE key
120 exchange then this can result in the client attempting to dereference a
121 NULL pointer leading to a client crash. This could be exploited in a Denial
124 This issue was reported to OpenSSL by Guido Vranken.
128 *) BN_mod_exp may produce incorrect results on x86_64
130 There is a carry propagating bug in the x86_64 Montgomery squaring
131 procedure. No EC algorithms are affected. Analysis suggests that attacks
132 against RSA and DSA as a result of this defect would be very difficult to
133 perform and are not believed likely. Attacks against DH are considered just
134 feasible (although very difficult) because most of the work necessary to
135 deduce information about a private key may be performed offline. The amount
136 of resources required for such an attack would be very significant and
137 likely only accessible to a limited number of attackers. An attacker would
138 additionally need online access to an unpatched system using the target
139 private key in a scenario with persistent DH parameters and a private
140 key that is shared between multiple clients. For example this can occur by
141 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
142 similar to CVE-2015-3193 but must be treated as a separate problem.
144 This issue was reported to OpenSSL by the OSS-Fuzz project.
148 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
150 *) ChaCha20/Poly1305 heap-buffer-overflow
152 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
153 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
154 crash. This issue is not considered to be exploitable beyond a DoS.
156 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
160 *) CMS Null dereference
162 Applications parsing invalid CMS structures can crash with a NULL pointer
163 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
164 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
165 structure callback if an attempt is made to free certain invalid encodings.
166 Only CHOICE structures using a callback which do not handle NULL value are
169 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
173 *) Montgomery multiplication may produce incorrect results
175 There is a carry propagating bug in the Broadwell-specific Montgomery
176 multiplication procedure that handles input lengths divisible by, but
177 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
178 and DH private keys are impossible. This is because the subroutine in
179 question is not used in operations with the private key itself and an input
180 of the attacker's direct choice. Otherwise the bug can manifest itself as
181 transient authentication and key negotiation failures or reproducible
182 erroneous outcome of public-key operations with specially crafted input.
183 Among EC algorithms only Brainpool P-512 curves are affected and one
184 presumably can attack ECDH key negotiation. Impact was not analyzed in
185 detail, because pre-requisites for attack are considered unlikely. Namely
186 multiple clients have to choose the curve in question and the server has to
187 share the private key among them, neither of which is default behaviour.
188 Even then only clients that chose the curve will be affected.
190 This issue was publicly reported as transient failures and was not
191 initially recognized as a security issue. Thanks to Richard Morgan for
192 providing reproducible case.
196 *) Removed automatic addition of RPATH in shared libraries and executables,
197 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
200 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
202 *) Fix Use After Free for large message sizes
204 The patch applied to address CVE-2016-6307 resulted in an issue where if a
205 message larger than approx 16k is received then the underlying buffer to
206 store the incoming message is reallocated and moved. Unfortunately a
207 dangling pointer to the old location is left which results in an attempt to
208 write to the previously freed location. This is likely to result in a
209 crash, however it could potentially lead to execution of arbitrary code.
211 This issue only affects OpenSSL 1.1.0a.
213 This issue was reported to OpenSSL by Robert Święcki.
217 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
219 *) OCSP Status Request extension unbounded memory growth
221 A malicious client can send an excessively large OCSP Status Request
222 extension. If that client continually requests renegotiation, sending a
223 large OCSP Status Request extension each time, then there will be unbounded
224 memory growth on the server. This will eventually lead to a Denial Of
225 Service attack through memory exhaustion. Servers with a default
226 configuration are vulnerable even if they do not support OCSP. Builds using
227 the "no-ocsp" build time option are not affected.
229 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
233 *) SSL_peek() hang on empty record
235 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
236 sends an empty record. This could be exploited by a malicious peer in a
237 Denial Of Service attack.
239 This issue was reported to OpenSSL by Alex Gaynor.
243 *) Excessive allocation of memory in tls_get_message_header() and
244 dtls1_preprocess_fragment()
246 A (D)TLS message includes 3 bytes for its length in the header for the
247 message. This would allow for messages up to 16Mb in length. Messages of
248 this length are excessive and OpenSSL includes a check to ensure that a
249 peer is sending reasonably sized messages in order to avoid too much memory
250 being consumed to service a connection. A flaw in the logic of version
251 1.1.0 means that memory for the message is allocated too early, prior to
252 the excessive message length check. Due to way memory is allocated in
253 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
254 to service a connection. This could lead to a Denial of Service through
255 memory exhaustion. However, the excessive message length check still takes
256 place, and this would cause the connection to immediately fail. Assuming
257 that the application calls SSL_free() on the failed connection in a timely
258 manner then the 21Mb of allocated memory will then be immediately freed
259 again. Therefore the excessive memory allocation will be transitory in
260 nature. This then means that there is only a security impact if:
262 1) The application does not call SSL_free() in a timely manner in the event
263 that the connection fails
265 2) The application is working in a constrained environment where there is
266 very little free memory
268 3) The attacker initiates multiple connection attempts such that there are
269 multiple connections in a state where memory has been allocated for the
270 connection; SSL_free() has not yet been called; and there is insufficient
271 memory to service the multiple requests.
273 Except in the instance of (1) above any Denial Of Service is likely to be
274 transitory because as soon as the connection fails the memory is
275 subsequently freed again in the SSL_free() call. However there is an
276 increased risk during this period of application crashes due to the lack of
277 memory - which would then mean a more serious Denial of Service.
279 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
280 (CVE-2016-6307 and CVE-2016-6308)
283 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
284 had to be removed. Primary reason is that vendor assembler can't
285 assemble our modules with -KPIC flag. As result it, assembly
286 support, was not even available as option. But its lack means
287 lack of side-channel resistant code, which is incompatible with
288 security by todays standards. Fortunately gcc is readily available
289 prepackaged option, which we firmly point at...
292 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
294 *) Windows command-line tool supports UTF-8 opt-in option for arguments
295 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
296 (to any value) allows Windows user to access PKCS#12 file generated
297 with Windows CryptoAPI and protected with non-ASCII password, as well
298 as files generated under UTF-8 locale on Linux also protected with
302 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
303 have been disabled by default and removed from DEFAULT, just like RC4.
304 See the RC4 item below to re-enable both.
307 *) The method for finding the storage location for the Windows RAND seed file
308 has changed. First we check %RANDFILE%. If that is not set then we check
309 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
310 all else fails we fall back to C:\.
313 *) The EVP_EncryptUpdate() function has had its return type changed from void
314 to int. A return of 0 indicates and error while a return of 1 indicates
318 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
319 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
320 off the constant time implementation for RSA, DSA and DH have been made
321 no-ops and deprecated.
324 *) Windows RAND implementation was simplified to only get entropy by
325 calling CryptGenRandom(). Various other RAND-related tickets
327 [Joseph Wylie Yandle, Rich Salz]
329 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
330 and OPENSSL_LH_, respectively. The old names are available
331 with API compatibility. They new names are now completely documented.
334 *) Unify TYPE_up_ref(obj) methods signature.
335 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
336 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
337 int (instead of void) like all others TYPE_up_ref() methods.
338 So now these methods also check the return value of CRYPTO_atomic_add(),
339 and the validity of object reference counter.
340 [fdasilvayy@gmail.com]
342 *) With Windows Visual Studio builds, the .pdb files are installed
343 alongside the installed libraries and executables. For a static
344 library installation, ossl_static.pdb is the associate compiler
345 generated .pdb file to be used when linking programs.
348 *) Remove openssl.spec. Packaging files belong with the packagers.
351 *) Automatic Darwin/OSX configuration has had a refresh, it will now
352 recognise x86_64 architectures automatically. You can still decide
353 to build for a different bitness with the environment variable
354 KERNEL_BITS (can be 32 or 64), for example:
356 KERNEL_BITS=32 ./config
360 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
361 256 bit AES and HMAC with SHA256.
364 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
367 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
370 *) To enable users to have their own config files and build file templates,
371 Configure looks in the directory indicated by the environment variable
372 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
373 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
374 name and is used as is.
377 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
378 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
379 X509_CERT_FILE_CTX was removed.
382 *) "shared" builds are now the default. To create only static libraries use
383 the "no-shared" Configure option.
386 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
387 All of these option have not worked for some while and are fundamental
391 *) Make various cleanup routines no-ops and mark them as deprecated. Most
392 global cleanup functions are no longer required because they are handled
393 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
394 Explicitly de-initing can cause problems (e.g. where a library that uses
395 OpenSSL de-inits, but an application is still using it). The affected
396 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
397 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
398 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
402 *) --strict-warnings no longer enables runtime debugging options
403 such as REF_DEBUG. Instead, debug options are automatically
404 enabled with '--debug' builds.
405 [Andy Polyakov, Emilia Käsper]
407 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
408 have been moved out of the public header files. New functions for managing
409 these have been added.
412 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
413 objects have been moved out of the public header files. New
414 functions for managing these have been added.
417 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
418 have been moved out of the public header files. New functions for managing
419 these have been added.
422 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
423 moved out of the public header files. New functions for managing these
427 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
430 *) Removed the mk1mf build scripts.
433 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
434 it is always safe to #include a header now.
437 *) Removed the aged BC-32 config and all its supporting scripts
440 *) Removed support for Ultrix, Netware, and OS/2.
443 *) Add support for HKDF.
446 *) Add support for blake2b and blake2s
449 *) Added support for "pipelining". Ciphers that have the
450 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
451 encryptions/decryptions simultaneously. There are currently no built-in
452 ciphers with this property but the expectation is that engines will be able
453 to offer it to significantly improve throughput. Support has been extended
454 into libssl so that multiple records for a single connection can be
455 processed in one go (for >=TLS 1.1).
458 *) Added the AFALG engine. This is an async capable engine which is able to
459 offload work to the Linux kernel. In this initial version it only supports
460 AES128-CBC. The kernel must be version 4.1.0 or greater.
463 *) OpenSSL now uses a new threading API. It is no longer necessary to
464 set locking callbacks to use OpenSSL in a multi-threaded environment. There
465 are two supported threading models: pthreads and windows threads. It is
466 also possible to configure OpenSSL at compile time for "no-threads". The
467 old threading API should no longer be used. The functions have been
468 replaced with "no-op" compatibility macros.
469 [Alessandro Ghedini, Matt Caswell]
471 *) Modify behavior of ALPN to invoke callback after SNI/servername
472 callback, such that updates to the SSL_CTX affect ALPN.
475 *) Add SSL_CIPHER queries for authentication and key-exchange.
478 *) Changes to the DEFAULT cipherlist:
479 - Prefer (EC)DHE handshakes over plain RSA.
480 - Prefer AEAD ciphers over legacy ciphers.
481 - Prefer ECDSA over RSA when both certificates are available.
482 - Prefer TLSv1.2 ciphers/PRF.
483 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
487 *) Change the ECC default curve list to be this, in order: x25519,
488 secp256r1, secp521r1, secp384r1.
491 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
492 disabled by default. They can be re-enabled using the
493 enable-weak-ssl-ciphers option to Configure.
496 *) If the server has ALPN configured, but supports no protocols that the
497 client advertises, send a fatal "no_application_protocol" alert.
498 This behaviour is SHALL in RFC 7301, though it isn't universally
499 implemented by other servers.
502 *) Add X25519 support.
503 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
504 for public and private key encoding using the format documented in
505 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
506 key generation and key derivation.
508 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
512 *) Deprecate SRP_VBASE_get_by_user.
513 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
514 In order to fix an unavoidable memory leak (CVE-2016-0798),
515 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
516 seed, even if the seed is configured.
518 Users should use SRP_VBASE_get1_by_user instead. Note that in
519 SRP_VBASE_get1_by_user, caller must free the returned value. Note
520 also that even though configuring the SRP seed attempts to hide
521 invalid usernames by continuing the handshake with fake
522 credentials, this behaviour is not constant time and no strong
523 guarantees are made that the handshake is indistinguishable from
524 that of a valid user.
527 *) Configuration change; it's now possible to build dynamic engines
528 without having to build shared libraries and vice versa. This
529 only applies to the engines in engines/, those in crypto/engine/
530 will always be built into libcrypto (i.e. "static").
532 Building dynamic engines is enabled by default; to disable, use
533 the configuration option "disable-dynamic-engine".
535 The only requirements for building dynamic engines are the
536 presence of the DSO module and building with position independent
537 code, so they will also automatically be disabled if configuring
538 with "disable-dso" or "disable-pic".
540 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
541 are also taken away from openssl/opensslconf.h, as they are
545 *) Configuration change; if there is a known flag to compile
546 position independent code, it will always be applied on the
547 libcrypto and libssl object files, and never on the application
548 object files. This means other libraries that use routines from
549 libcrypto / libssl can be made into shared libraries regardless
550 of how OpenSSL was configured.
552 If this isn't desirable, the configuration options "disable-pic"
553 or "no-pic" can be used to disable the use of PIC. This will
554 also disable building shared libraries and dynamic engines.
557 *) Removed JPAKE code. It was experimental and has no wide use.
560 *) The INSTALL_PREFIX Makefile variable has been renamed to
561 DESTDIR. That makes for less confusion on what this variable
562 is for. Also, the configuration option --install_prefix is
566 *) Heartbeat for TLS has been removed and is disabled by default
567 for DTLS; configure with enable-heartbeats. Code that uses the
568 old #define's might need to be updated.
569 [Emilia Käsper, Rich Salz]
571 *) Rename REF_CHECK to REF_DEBUG.
574 *) New "unified" build system
576 The "unified" build system is aimed to be a common system for all
577 platforms we support. With it comes new support for VMS.
579 This system builds supports building in a different directory tree
580 than the source tree. It produces one Makefile (for unix family
581 or lookalikes), or one descrip.mms (for VMS).
583 The source of information to make the Makefile / descrip.mms is
584 small files called 'build.info', holding the necessary
585 information for each directory with source to compile, and a
586 template in Configurations, like unix-Makefile.tmpl or
589 With this change, the library names were also renamed on Windows
590 and on VMS. They now have names that are closer to the standard
591 on Unix, and include the major version number, and in certain
592 cases, the architecture they are built for. See "Notes on shared
593 libraries" in INSTALL.
595 We rely heavily on the perl module Text::Template.
598 *) Added support for auto-initialisation and de-initialisation of the library.
599 OpenSSL no longer requires explicit init or deinit routines to be called,
600 except in certain circumstances. See the OPENSSL_init_crypto() and
601 OPENSSL_init_ssl() man pages for further information.
604 *) The arguments to the DTLSv1_listen function have changed. Specifically the
605 "peer" argument is now expected to be a BIO_ADDR object.
607 *) Rewrite of BIO networking library. The BIO library lacked consistent
608 support of IPv6, and adding it required some more extensive
609 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
610 which hold all types of addresses and chains of address information.
611 It also introduces a new API, with functions like BIO_socket,
612 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
613 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
614 have been adapted accordingly.
617 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
621 *) CRIME protection: disable compression by default, even if OpenSSL is
622 compiled with zlib enabled. Applications can still enable compression
623 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
624 using the SSL_CONF library to configure compression.
627 *) The signature of the session callback configured with
628 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
629 was explicitly marked as 'const unsigned char*' instead of
633 *) Always DPURIFY. Remove the use of uninitialized memory in the
634 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
637 *) Removed many obsolete configuration items, including
638 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
639 MD2_CHAR, MD2_INT, MD2_LONG
641 IDEA_SHORT, IDEA_LONG
642 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
643 [Rich Salz, with advice from Andy Polyakov]
645 *) Many BN internals have been moved to an internal header file.
646 [Rich Salz with help from Andy Polyakov]
648 *) Configuration and writing out the results from it has changed.
649 Files such as Makefile include/openssl/opensslconf.h and are now
650 produced through general templates, such as Makefile.in and
651 crypto/opensslconf.h.in and some help from the perl module
654 Also, the center of configuration information is no longer
655 Makefile. Instead, Configure produces a perl module in
656 configdata.pm which holds most of the config data (in the hash
657 table %config), the target data that comes from the target
658 configuration in one of the Configurations/*.conf files (in
662 *) To clarify their intended purposes, the Configure options
663 --prefix and --openssldir change their semantics, and become more
664 straightforward and less interdependent.
666 --prefix shall be used exclusively to give the location INSTALLTOP
667 where programs, scripts, libraries, include files and manuals are
668 going to be installed. The default is now /usr/local.
670 --openssldir shall be used exclusively to give the default
671 location OPENSSLDIR where certificates, private keys, CRLs are
672 managed. This is also where the default openssl.cnf gets
674 If the directory given with this option is a relative path, the
675 values of both the --prefix value and the --openssldir value will
676 be combined to become OPENSSLDIR.
677 The default for --openssldir is INSTALLTOP/ssl.
679 Anyone who uses --openssldir to specify where OpenSSL is to be
680 installed MUST change to use --prefix instead.
683 *) The GOST engine was out of date and therefore it has been removed. An up
684 to date GOST engine is now being maintained in an external repository.
685 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
686 support for GOST ciphersuites (these are only activated if a GOST engine
690 *) EGD is no longer supported by default; use enable-egd when
692 [Ben Kaduk and Rich Salz]
694 *) The distribution now has Makefile.in files, which are used to
695 create Makefile's when Configure is run. *Configure must be run
696 before trying to build now.*
699 *) The return value for SSL_CIPHER_description() for error conditions
703 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
705 Obtaining and performing DNSSEC validation of TLSA records is
706 the application's responsibility. The application provides
707 the TLSA records of its choice to OpenSSL, and these are then
708 used to authenticate the peer.
710 The TLSA records need not even come from DNS. They can, for
711 example, be used to implement local end-entity certificate or
712 trust-anchor "pinning", where the "pin" data takes the form
713 of TLSA records, which can augment or replace verification
714 based on the usual WebPKI public certification authorities.
717 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
718 continues to support deprecated interfaces in default builds.
719 However, applications are strongly advised to compile their
720 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
721 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
722 or the 1.1.0 releases.
724 In environments in which all applications have been ported to
725 not use any deprecated interfaces OpenSSL's Configure script
726 should be used with the --api=1.1.0 option to entirely remove
727 support for the deprecated features from the library and
728 unconditionally disable them in the installed headers.
729 Essentially the same effect can be achieved with the "no-deprecated"
730 argument to Configure, except that this will always restrict
731 the build to just the latest API, rather than a fixed API
734 As applications are ported to future revisions of the API,
735 they should update their compile-time OPENSSL_API_COMPAT define
736 accordingly, but in most cases should be able to continue to
737 compile with later releases.
739 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
740 0x10000000L and 0x00908000L, respectively. However those
741 versions did not support the OPENSSL_API_COMPAT feature, and
742 so applications are not typically tested for explicit support
743 of just the undeprecated features of either release.
746 *) Add support for setting the minimum and maximum supported protocol.
747 It can bet set via the SSL_set_min_proto_version() and
748 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
749 MaxProtcol. It's recommended to use the new APIs to disable
750 protocols instead of disabling individual protocols using
751 SSL_set_options() or SSL_CONF's Protocol. This change also
752 removes support for disabling TLS 1.2 in the OpenSSL TLS
753 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
756 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
759 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
760 and integrates ECDSA and ECDH functionality into EC. Implementations can
761 now redirect key generation and no longer need to convert to or from
764 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
765 include the ec.h header file instead.
768 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
769 ciphers who are no longer supported and drops support the ephemeral RSA key
770 exchange. The LOW ciphers currently doesn't have any ciphers in it.
773 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
774 opaque. For HMAC_CTX, the following constructors and destructors
777 HMAC_CTX *HMAC_CTX_new(void);
778 void HMAC_CTX_free(HMAC_CTX *ctx);
780 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
781 destroy such methods has been added. See EVP_MD_meth_new(3) and
782 EVP_CIPHER_meth_new(3) for documentation.
785 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
786 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
787 EVP_MD_CTX_reset() should be called instead to reinitialise
788 an already created structure.
789 2) For consistency with the majority of our object creators and
790 destructors, EVP_MD_CTX_(create|destroy) were renamed to
791 EVP_MD_CTX_(new|free). The old names are retained as macros
792 for deprecated builds.
795 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
796 cryptographic operations to be performed asynchronously as long as an
797 asynchronous capable engine is used. See the ASYNC_start_job() man page for
798 further details. Libssl has also had this capability integrated with the
799 introduction of the new mode SSL_MODE_ASYNC and associated error
800 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
801 pages. This work was developed in partnership with Intel Corp.
804 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
805 always enabled now. If you want to disable the support you should
806 exclude it using the list of supported ciphers. This also means that the
807 "-no_ecdhe" option has been removed from s_server.
810 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
811 SSL_{CTX_}set1_curves() which can set a list.
814 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
815 curve you want to support using SSL_{CTX_}set1_curves().
818 *) State machine rewrite. The state machine code has been significantly
819 refactored in order to remove much duplication of code and solve issues
820 with the old code (see ssl/statem/README for further details). This change
821 does have some associated API changes. Notably the SSL_state() function
822 has been removed and replaced by SSL_get_state which now returns an
823 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
824 altogether. The previous handshake states defined in ssl.h and ssl3.h have
828 *) All instances of the string "ssleay" in the public API were replaced
829 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
830 Some error codes related to internal RSA_eay API's were renamed.
833 *) The demo files in crypto/threads were moved to demo/threads.
836 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
838 [Matt Caswell, Rich Salz]
840 *) New ASN.1 embed macro.
842 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
843 structure is not allocated: it is part of the parent. That is instead of
851 This reduces memory fragmentation and make it impossible to accidentally
852 set a mandatory field to NULL.
854 This currently only works for some fields specifically a SEQUENCE, CHOICE,
855 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
856 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
860 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
863 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
864 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
865 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
866 DES and RC4 ciphersuites.
869 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
870 This changes the decoding behaviour for some invalid messages,
871 though the change is mostly in the more lenient direction, and
872 legacy behaviour is preserved as much as possible.
875 *) Fix no-stdio build.
876 [ David Woodhouse <David.Woodhouse@intel.com> and also
877 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
879 *) New testing framework
880 The testing framework has been largely rewritten and is now using
881 perl and the perl modules Test::Harness and an extended variant of
882 Test::More called OpenSSL::Test to do its work. All test scripts in
883 test/ have been rewritten into test recipes, and all direct calls to
884 executables in test/Makefile have become individual recipes using the
885 simplified testing OpenSSL::Test::Simple.
887 For documentation on our testing modules, do:
889 perldoc test/testlib/OpenSSL/Test/Simple.pm
890 perldoc test/testlib/OpenSSL/Test.pm
894 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
895 are used; the latter aborts on memory leaks (usually checked on exit).
896 Some undocumented "set malloc, etc., hooks" functions were removed
897 and others were changed. All are now documented.
900 *) In DSA_generate_parameters_ex, if the provided seed is too short,
902 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
904 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
905 from RFC4279, RFC4785, RFC5487, RFC5489.
907 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
908 original RSA_PSK patch.
911 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
912 era flag was never set throughout the codebase (only read). Also removed
913 SSL3_FLAGS_POP_BUFFER which was only used if
914 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
917 *) Changed the default name options in the "ca", "crl", "req" and "x509"
918 to be "oneline" instead of "compat".
921 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
922 not aware of clients that still exhibit this bug, and the workaround
923 hasn't been working properly for a while.
926 *) The return type of BIO_number_read() and BIO_number_written() as well as
927 the corresponding num_read and num_write members in the BIO structure has
928 changed from unsigned long to uint64_t. On platforms where an unsigned
929 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
933 *) Given the pervasive nature of TLS extensions it is inadvisable to run
934 OpenSSL without support for them. It also means that maintaining
935 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
936 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
939 *) Removed support for the two export grade static DH ciphersuites
940 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
941 were newly added (along with a number of other static DH ciphersuites) to
942 1.0.2. However the two export ones have *never* worked since they were
943 introduced. It seems strange in any case to be adding new export
944 ciphersuites, and given "logjam" it also does not seem correct to fix them.
947 *) Version negotiation has been rewritten. In particular SSLv23_method(),
948 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
949 and turned into macros which simply call the new preferred function names
950 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
951 should use the new names instead. Also as part of this change the ssl23.h
952 header file has been removed.
955 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
956 code and the associated standard is no longer considered fit-for-purpose.
959 *) RT2547 was closed. When generating a private key, try to make the
960 output file readable only by the owner. This behavior change might
961 be noticeable when interacting with other software.
963 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
967 *) Added HTTP GET support to the ocsp command.
970 *) Changed default digest for the dgst and enc commands from MD5 to
974 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
977 *) Added support for TLS extended master secret from
978 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
979 initial patch which was a great help during development.
982 *) All libssl internal structures have been removed from the public header
983 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
984 now redundant). Users should not attempt to access internal structures
985 directly. Instead they should use the provided API functions.
988 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
989 Access to deprecated functions can be re-enabled by running config with
990 "enable-deprecated". In addition applications wishing to use deprecated
991 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
992 will, by default, disable some transitive includes that previously existed
993 in the header files (e.g. ec.h will no longer, by default, include bn.h)
996 *) Added support for OCB mode. OpenSSL has been granted a patent license
997 compatible with the OpenSSL license for use of OCB. Details are available
998 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
999 for OCB can be removed by calling config with no-ocb.
1002 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1003 compatible client hello.
1006 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1007 done while fixing the error code for the key-too-small case.
1008 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1010 *) CA.sh has been removed; use CA.pl instead.
1013 *) Removed old DES API.
1016 *) Remove various unsupported platforms:
1022 Sinix/ReliantUNIX RM400
1027 16-bit platforms such as WIN16
1030 *) Clean up OPENSSL_NO_xxx #define's
1031 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1032 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1033 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1034 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1035 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1036 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1037 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1038 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1039 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1040 Remove MS_STATIC; it's a relic from platforms <32 bits.
1043 *) Cleaned up dead code
1044 Remove all but one '#ifdef undef' which is to be looked at.
1047 *) Clean up calling of xxx_free routines.
1048 Just like free(), fix most of the xxx_free routines to accept
1049 NULL. Remove the non-null checks from callers. Save much code.
1052 *) Add secure heap for storage of private keys (when possible).
1053 Add BIO_s_secmem(), CBIGNUM, etc.
1054 Contributed by Akamai Technologies under our Corporate CLA.
1057 *) Experimental support for a new, fast, unbiased prime candidate generator,
1058 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1059 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1061 *) New output format NSS in the sess_id command line tool. This allows
1062 exporting the session id and the master key in NSS keylog format.
1063 [Martin Kaiser <martin@kaiser.cx>]
1065 *) Harmonize version and its documentation. -f flag is used to display
1067 [mancha <mancha1@zoho.com>]
1069 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1070 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1071 [mancha <mancha1@zoho.com>]
1073 *) Fix some double frees. These are not thought to be exploitable.
1074 [mancha <mancha1@zoho.com>]
1076 *) A missing bounds check in the handling of the TLS heartbeat extension
1077 can be used to reveal up to 64k of memory to a connected client or
1080 Thanks for Neel Mehta of Google Security for discovering this bug and to
1081 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1082 preparing the fix (CVE-2014-0160)
1083 [Adam Langley, Bodo Moeller]
1085 *) Fix for the attack described in the paper "Recovering OpenSSL
1086 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1087 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1088 http://eprint.iacr.org/2014/140
1090 Thanks to Yuval Yarom and Naomi Benger for discovering this
1091 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1092 [Yuval Yarom and Naomi Benger]
1094 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1095 this fixes a limitation in previous versions of OpenSSL.
1098 *) Experimental encrypt-then-mac support.
1100 Experimental support for encrypt then mac from
1101 draft-gutmann-tls-encrypt-then-mac-02.txt
1103 To enable it set the appropriate extension number (0x42 for the test
1104 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1106 For non-compliant peers (i.e. just about everything) this should have no
1109 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1113 *) Add EVP support for key wrapping algorithms, to avoid problems with
1114 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1115 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1116 algorithms and include tests cases.
1119 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1123 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1124 MGF1 digest and OAEP label.
1127 *) Make openssl verify return errors.
1128 [Chris Palmer <palmer@google.com> and Ben Laurie]
1130 *) New function ASN1_TIME_diff to calculate the difference between two
1131 ASN1_TIME structures or one structure and the current time.
1134 *) Update fips_test_suite to support multiple command line options. New
1135 test to induce all self test errors in sequence and check expected
1139 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1140 sign or verify all in one operation.
1143 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1144 test programs and fips_test_suite. Includes functionality to parse
1145 the minimal script output of fipsalgest.pl directly.
1148 *) Add authorisation parameter to FIPS_module_mode_set().
1151 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1154 *) Use separate DRBG fields for internal and external flags. New function
1155 FIPS_drbg_health_check() to perform on demand health checking. Add
1156 generation tests to fips_test_suite with reduced health check interval to
1157 demonstrate periodic health checking. Add "nodh" option to
1158 fips_test_suite to skip very slow DH test.
1161 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1165 *) More extensive health check for DRBG checking many more failure modes.
1166 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1167 combination: call this in fips_test_suite.
1170 *) Add support for canonical generation of DSA parameter 'g'. See
1173 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1174 POST to handle HMAC cases.
1177 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1178 to return numerical and string versions of the FIPS module number.
1181 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1182 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1183 outside the validated module in the FIPS capable OpenSSL.
1186 *) Minor change to DRBG entropy callback semantics. In some cases
1187 there is no multiple of the block length between min_len and
1188 max_len. Allow the callback to return more than max_len bytes
1189 of entropy but discard any extra: it is the callback's responsibility
1190 to ensure that the extra data discarded does not impact the
1191 requested amount of entropy.
1194 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1195 information in FIPS186-3, SP800-57 and SP800-131A.
1198 *) CCM support via EVP. Interface is very similar to GCM case except we
1199 must supply all data in one chunk (i.e. no update, final) and the
1200 message length must be supplied if AAD is used. Add algorithm test
1204 *) Initial version of POST overhaul. Add POST callback to allow the status
1205 of POST to be monitored and/or failures induced. Modify fips_test_suite
1206 to use callback. Always run all selftests even if one fails.
1209 *) XTS support including algorithm test driver in the fips_gcmtest program.
1210 Note: this does increase the maximum key length from 32 to 64 bytes but
1211 there should be no binary compatibility issues as existing applications
1212 will never use XTS mode.
1215 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1216 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1217 performs algorithm blocking for unapproved PRNG types. Also do not
1218 set PRNG type in FIPS_mode_set(): leave this to the application.
1219 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1220 the standard OpenSSL PRNG: set additional data to a date time vector.
1223 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1224 This shouldn't present any incompatibility problems because applications
1225 shouldn't be using these directly and any that are will need to rethink
1226 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1229 *) Extensive self tests and health checking required by SP800-90 DRBG.
1230 Remove strength parameter from FIPS_drbg_instantiate and always
1231 instantiate at maximum supported strength.
1234 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1237 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1240 *) New function DH_compute_key_padded() to compute a DH key and pad with
1241 leading zeroes if needed: this complies with SP800-56A et al.
1244 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1245 anything, incomplete, subject to change and largely untested at present.
1248 *) Modify fipscanisteronly build option to only build the necessary object
1249 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1252 *) Add experimental option FIPSSYMS to give all symbols in
1253 fipscanister.o and FIPS or fips prefix. This will avoid
1254 conflicts with future versions of OpenSSL. Add perl script
1255 util/fipsas.pl to preprocess assembly language source files
1256 and rename any affected symbols.
1259 *) Add selftest checks and algorithm block of non-fips algorithms in
1260 FIPS mode. Remove DES2 from selftests.
1263 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1264 return internal method without any ENGINE dependencies. Add new
1265 tiny fips sign and verify functions.
1268 *) New build option no-ec2m to disable characteristic 2 code.
1271 *) New build option "fipscanisteronly". This only builds fipscanister.o
1272 and (currently) associated fips utilities. Uses the file Makefile.fips
1273 instead of Makefile.org as the prototype.
1276 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1277 Update fips_gcmtest to use IV generator.
1280 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1281 setting output buffer to NULL. The *Final function must be
1282 called although it will not retrieve any additional data. The tag
1283 can be set or retrieved with a ctrl. The IV length is by default 12
1284 bytes (96 bits) but can be set to an alternative value. If the IV
1285 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1289 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1290 underlying do_cipher function handles all cipher semantics itself
1291 including padding and finalisation. This is useful if (for example)
1292 an ENGINE cipher handles block padding itself. The behaviour of
1293 do_cipher is subtly changed if this flag is set: the return value
1294 is the number of characters written to the output buffer (zero is
1295 no longer an error code) or a negative error code. Also if the
1296 input buffer is NULL and length 0 finalisation should be performed.
1299 *) If a candidate issuer certificate is already part of the constructed
1300 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1303 *) Improve forward-security support: add functions
1305 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1306 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1308 for use by SSL/TLS servers; the callback function will be called whenever a
1309 new session is created, and gets to decide whether the session may be
1310 cached to make it resumable (return 0) or not (return 1). (As by the
1311 SSL/TLS protocol specifications, the session_id sent by the server will be
1312 empty to indicate that the session is not resumable; also, the server will
1313 not generate RFC 4507 (RFC 5077) session tickets.)
1315 A simple reasonable callback implementation is to return is_forward_secure.
1316 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1317 by the SSL/TLS server library, indicating whether it can provide forward
1319 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1321 *) New -verify_name option in command line utilities to set verification
1325 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1326 Add CMAC pkey methods.
1329 *) Experimental renegotiation in s_server -www mode. If the client
1330 browses /reneg connection is renegotiated. If /renegcert it is
1331 renegotiated requesting a certificate.
1334 *) Add an "external" session cache for debugging purposes to s_server. This
1335 should help trace issues which normally are only apparent in deployed
1336 multi-process servers.
1339 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1340 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1341 BIO_set_cipher() and some obscure PEM functions were changed so they
1342 can now return an error. The RAND changes required a change to the
1343 RAND_METHOD structure.
1346 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1347 a gcc attribute to warn if the result of a function is ignored. This
1348 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1349 whose return value is often ignored.
1352 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1353 These allow SCTs (signed certificate timestamps) to be requested and
1354 validated when establishing a connection.
1355 [Rob Percival <robpercival@google.com>]
1357 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1359 *) Prevent padding oracle in AES-NI CBC MAC check
1361 A MITM attacker can use a padding oracle attack to decrypt traffic
1362 when the connection uses an AES CBC cipher and the server support
1365 This issue was introduced as part of the fix for Lucky 13 padding
1366 attack (CVE-2013-0169). The padding check was rewritten to be in
1367 constant time by making sure that always the same bytes are read and
1368 compared against either the MAC or padding bytes. But it no longer
1369 checked that there was enough data to have both the MAC and padding
1372 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1376 *) Fix EVP_EncodeUpdate overflow
1378 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1379 Base64 encoding of binary data. If an attacker is able to supply very large
1380 amounts of input data then a length check can overflow resulting in a heap
1383 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1384 the PEM_write_bio* family of functions. These are mainly used within the
1385 OpenSSL command line applications, so any application which processes data
1386 from an untrusted source and outputs it as a PEM file should be considered
1387 vulnerable to this issue. User applications that call these APIs directly
1388 with large amounts of untrusted data may also be vulnerable.
1390 This issue was reported by Guido Vranken.
1394 *) Fix EVP_EncryptUpdate overflow
1396 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1397 is able to supply very large amounts of input data after a previous call to
1398 EVP_EncryptUpdate() with a partial block then a length check can overflow
1399 resulting in a heap corruption. Following an analysis of all OpenSSL
1400 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1401 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1402 the first called function after an EVP_EncryptInit(), and therefore that
1403 specific call must be safe. The second form is where the length passed to
1404 EVP_EncryptUpdate() can be seen from the code to be some small value and
1405 therefore there is no possibility of an overflow. Since all instances are
1406 one of these two forms, it is believed that there can be no overflows in
1407 internal code due to this problem. It should be noted that
1408 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1409 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1410 of these calls have also been analysed too and it is believed there are no
1411 instances in internal usage where an overflow could occur.
1413 This issue was reported by Guido Vranken.
1417 *) Prevent ASN.1 BIO excessive memory allocation
1419 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1420 a short invalid encoding can cause allocation of large amounts of memory
1421 potentially consuming excessive resources or exhausting memory.
1423 Any application parsing untrusted data through d2i BIO functions is
1424 affected. The memory based functions such as d2i_X509() are *not* affected.
1425 Since the memory based functions are used by the TLS library, TLS
1426 applications are not affected.
1428 This issue was reported by Brian Carpenter.
1434 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1435 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1436 in arbitrary stack data being returned in the buffer.
1438 This issue was reported by Guido Vranken.
1442 *) Modify behavior of ALPN to invoke callback after SNI/servername
1443 callback, such that updates to the SSL_CTX affect ALPN.
1446 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1450 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1451 methods are enabled and ssl2 is disabled the methods return NULL.
1454 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1456 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1457 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1458 provide any "EXPORT" or "LOW" strength ciphers.
1461 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1462 is by default disabled at build-time. Builds that are not configured with
1463 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1464 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1465 will need to explicitly call either of:
1467 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1469 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1471 as appropriate. Even if either of those is used, or the application
1472 explicitly uses the version-specific SSLv2_method() or its client and
1473 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1474 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1475 ciphers, and SSLv2 56-bit DES are no longer available.
1479 *) Fix a double-free in DSA code
1481 A double free bug was discovered when OpenSSL parses malformed DSA private
1482 keys and could lead to a DoS attack or memory corruption for applications
1483 that receive DSA private keys from untrusted sources. This scenario is
1486 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1491 *) Disable SRP fake user seed to address a server memory leak.
1493 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1495 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1496 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1497 was changed to ignore the "fake user" SRP seed, even if the seed
1500 Users should use SRP_VBASE_get1_by_user instead. Note that in
1501 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1502 also that even though configuring the SRP seed attempts to hide
1503 invalid usernames by continuing the handshake with fake
1504 credentials, this behaviour is not constant time and no strong
1505 guarantees are made that the handshake is indistinguishable from
1506 that of a valid user.
1510 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1512 In the BN_hex2bn function the number of hex digits is calculated using an
1513 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1514 large values of |i| this can result in |bn_expand| not allocating any
1515 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1516 field as NULL leading to a subsequent NULL ptr deref. For very large values
1517 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1518 In this case memory is allocated to the internal BIGNUM data field, but it
1519 is insufficiently sized leading to heap corruption. A similar issue exists
1520 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1521 is ever called by user applications with very large untrusted hex/dec data.
1522 This is anticipated to be a rare occurrence.
1524 All OpenSSL internal usage of these functions use data that is not expected
1525 to be untrusted, e.g. config file data or application command line
1526 arguments. If user developed applications generate config file data based
1527 on untrusted data then it is possible that this could also lead to security
1528 consequences. This is also anticipated to be rare.
1530 This issue was reported to OpenSSL by Guido Vranken.
1534 *) Fix memory issues in BIO_*printf functions
1536 The internal |fmtstr| function used in processing a "%s" format string in
1537 the BIO_*printf functions could overflow while calculating the length of a
1538 string and cause an OOB read when printing very long strings.
1540 Additionally the internal |doapr_outch| function can attempt to write to an
1541 OOB memory location (at an offset from the NULL pointer) in the event of a
1542 memory allocation failure. In 1.0.2 and below this could be caused where
1543 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1544 could be in processing a very long "%s" format string. Memory leaks can
1547 The first issue may mask the second issue dependent on compiler behaviour.
1548 These problems could enable attacks where large amounts of untrusted data
1549 is passed to the BIO_*printf functions. If applications use these functions
1550 in this way then they could be vulnerable. OpenSSL itself uses these
1551 functions when printing out human-readable dumps of ASN.1 data. Therefore
1552 applications that print this data could be vulnerable if the data is from
1553 untrusted sources. OpenSSL command line applications could also be
1554 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1555 as command line arguments.
1557 Libssl is not considered directly vulnerable. Additionally certificates etc
1558 received via remote connections via libssl are also unlikely to be able to
1559 trigger these issues because of message size limits enforced within libssl.
1561 This issue was reported to OpenSSL Guido Vranken.
1565 *) Side channel attack on modular exponentiation
1567 A side-channel attack was found which makes use of cache-bank conflicts on
1568 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1569 of RSA keys. The ability to exploit this issue is limited as it relies on
1570 an attacker who has control of code in a thread running on the same
1571 hyper-threaded core as the victim thread which is performing decryptions.
1573 This issue was reported to OpenSSL by Yuval Yarom, The University of
1574 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1575 Nadia Heninger, University of Pennsylvania with more information at
1576 http://cachebleed.info.
1580 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1581 if no keysize is specified with default_bits. This fixes an
1582 omission in an earlier change that changed all RSA/DSA key generation
1583 apps to use 2048 bits by default.
1586 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1587 *) DH small subgroups
1589 Historically OpenSSL only ever generated DH parameters based on "safe"
1590 primes. More recently (in version 1.0.2) support was provided for
1591 generating X9.42 style parameter files such as those required for RFC 5114
1592 support. The primes used in such files may not be "safe". Where an
1593 application is using DH configured with parameters based on primes that are
1594 not "safe" then an attacker could use this fact to find a peer's private
1595 DH exponent. This attack requires that the attacker complete multiple
1596 handshakes in which the peer uses the same private DH exponent. For example
1597 this could be used to discover a TLS server's private DH exponent if it's
1598 reusing the private DH exponent or it's using a static DH ciphersuite.
1600 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1601 TLS. It is not on by default. If the option is not set then the server
1602 reuses the same private DH exponent for the life of the server process and
1603 would be vulnerable to this attack. It is believed that many popular
1604 applications do set this option and would therefore not be at risk.
1606 The fix for this issue adds an additional check where a "q" parameter is
1607 available (as is the case in X9.42 based parameters). This detects the
1608 only known attack, and is the only possible defense for static DH
1609 ciphersuites. This could have some performance impact.
1611 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1612 default and cannot be disabled. This could have some performance impact.
1614 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1618 *) SSLv2 doesn't block disabled ciphers
1620 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1621 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1622 been disabled, provided that the SSLv2 protocol was not also disabled via
1625 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1626 and Sebastian Schinzel.
1630 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1632 *) BN_mod_exp may produce incorrect results on x86_64
1634 There is a carry propagating bug in the x86_64 Montgomery squaring
1635 procedure. No EC algorithms are affected. Analysis suggests that attacks
1636 against RSA and DSA as a result of this defect would be very difficult to
1637 perform and are not believed likely. Attacks against DH are considered just
1638 feasible (although very difficult) because most of the work necessary to
1639 deduce information about a private key may be performed offline. The amount
1640 of resources required for such an attack would be very significant and
1641 likely only accessible to a limited number of attackers. An attacker would
1642 additionally need online access to an unpatched system using the target
1643 private key in a scenario with persistent DH parameters and a private
1644 key that is shared between multiple clients. For example this can occur by
1645 default in OpenSSL DHE based SSL/TLS ciphersuites.
1647 This issue was reported to OpenSSL by Hanno Böck.
1651 *) Certificate verify crash with missing PSS parameter
1653 The signature verification routines will crash with a NULL pointer
1654 dereference if presented with an ASN.1 signature using the RSA PSS
1655 algorithm and absent mask generation function parameter. Since these
1656 routines are used to verify certificate signature algorithms this can be
1657 used to crash any certificate verification operation and exploited in a
1658 DoS attack. Any application which performs certificate verification is
1659 vulnerable including OpenSSL clients and servers which enable client
1662 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1666 *) X509_ATTRIBUTE memory leak
1668 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1669 memory. This structure is used by the PKCS#7 and CMS routines so any
1670 application which reads PKCS#7 or CMS data from untrusted sources is
1671 affected. SSL/TLS is not affected.
1673 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1678 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1679 This changes the decoding behaviour for some invalid messages,
1680 though the change is mostly in the more lenient direction, and
1681 legacy behaviour is preserved as much as possible.
1684 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1686 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1688 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1690 *) Alternate chains certificate forgery
1692 During certificate verification, OpenSSL will attempt to find an
1693 alternative certificate chain if the first attempt to build such a chain
1694 fails. An error in the implementation of this logic can mean that an
1695 attacker could cause certain checks on untrusted certificates to be
1696 bypassed, such as the CA flag, enabling them to use a valid leaf
1697 certificate to act as a CA and "issue" an invalid certificate.
1699 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1703 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1705 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1706 incompatibility in the handling of HMAC. The previous ABI has now been
1710 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1712 *) Malformed ECParameters causes infinite loop
1714 When processing an ECParameters structure OpenSSL enters an infinite loop
1715 if the curve specified is over a specially malformed binary polynomial
1718 This can be used to perform denial of service against any
1719 system which processes public keys, certificate requests or
1720 certificates. This includes TLS clients and TLS servers with
1721 client authentication enabled.
1723 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1727 *) Exploitable out-of-bounds read in X509_cmp_time
1729 X509_cmp_time does not properly check the length of the ASN1_TIME
1730 string and can read a few bytes out of bounds. In addition,
1731 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1734 An attacker can use this to craft malformed certificates and CRLs of
1735 various sizes and potentially cause a segmentation fault, resulting in
1736 a DoS on applications that verify certificates or CRLs. TLS clients
1737 that verify CRLs are affected. TLS clients and servers with client
1738 authentication enabled may be affected if they use custom verification
1741 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1742 independently by Hanno Böck.
1746 *) PKCS7 crash with missing EnvelopedContent
1748 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1749 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1750 with missing content and trigger a NULL pointer dereference on parsing.
1752 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1753 structures from untrusted sources are affected. OpenSSL clients and
1754 servers are not affected.
1756 This issue was reported to OpenSSL by Michal Zalewski (Google).
1760 *) CMS verify infinite loop with unknown hash function
1762 When verifying a signedData message the CMS code can enter an infinite loop
1763 if presented with an unknown hash function OID. This can be used to perform
1764 denial of service against any system which verifies signedData messages using
1766 This issue was reported to OpenSSL by Johannes Bauer.
1770 *) Race condition handling NewSessionTicket
1772 If a NewSessionTicket is received by a multi-threaded client when attempting to
1773 reuse a previous ticket then a race condition can occur potentially leading to
1774 a double free of the ticket data.
1778 *) Only support 256-bit or stronger elliptic curves with the
1779 'ecdh_auto' setting (server) or by default (client). Of supported
1780 curves, prefer P-256 (both).
1783 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1785 *) ClientHello sigalgs DoS fix
1787 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1788 invalid signature algorithms extension a NULL pointer dereference will
1789 occur. This can be exploited in a DoS attack against the server.
1791 This issue was was reported to OpenSSL by David Ramos of Stanford
1794 [Stephen Henson and Matt Caswell]
1796 *) Multiblock corrupted pointer fix
1798 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1799 feature only applies on 64 bit x86 architecture platforms that support AES
1800 NI instructions. A defect in the implementation of "multiblock" can cause
1801 OpenSSL's internal write buffer to become incorrectly set to NULL when
1802 using non-blocking IO. Typically, when the user application is using a
1803 socket BIO for writing, this will only result in a failed connection.
1804 However if some other BIO is used then it is likely that a segmentation
1805 fault will be triggered, thus enabling a potential DoS attack.
1807 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1811 *) Segmentation fault in DTLSv1_listen fix
1813 The DTLSv1_listen function is intended to be stateless and processes the
1814 initial ClientHello from many peers. It is common for user code to loop
1815 over the call to DTLSv1_listen until a valid ClientHello is received with
1816 an associated cookie. A defect in the implementation of DTLSv1_listen means
1817 that state is preserved in the SSL object from one invocation to the next
1818 that can lead to a segmentation fault. Errors processing the initial
1819 ClientHello can trigger this scenario. An example of such an error could be
1820 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1823 This issue was reported to OpenSSL by Per Allansson.
1827 *) Segmentation fault in ASN1_TYPE_cmp fix
1829 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
1830 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
1831 certificate signature algorithm consistency this can be used to crash any
1832 certificate verification operation and exploited in a DoS attack. Any
1833 application which performs certificate verification is vulnerable including
1834 OpenSSL clients and servers which enable client authentication.
1838 *) Segmentation fault for invalid PSS parameters fix
1840 The signature verification routines will crash with a NULL pointer
1841 dereference if presented with an ASN.1 signature using the RSA PSS
1842 algorithm and invalid parameters. Since these routines are used to verify
1843 certificate signature algorithms this can be used to crash any
1844 certificate verification operation and exploited in a DoS attack. Any
1845 application which performs certificate verification is vulnerable including
1846 OpenSSL clients and servers which enable client authentication.
1848 This issue was was reported to OpenSSL by Brian Carpenter.
1852 *) ASN.1 structure reuse memory corruption fix
1854 Reusing a structure in ASN.1 parsing may allow an attacker to cause
1855 memory corruption via an invalid write. Such reuse is and has been
1856 strongly discouraged and is believed to be rare.
1858 Applications that parse structures containing CHOICE or ANY DEFINED BY
1859 components may be affected. Certificate parsing (d2i_X509 and related
1860 functions) are however not affected. OpenSSL clients and servers are
1865 *) PKCS7 NULL pointer dereferences fix
1867 The PKCS#7 parsing code does not handle missing outer ContentInfo
1868 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
1869 missing content and trigger a NULL pointer dereference on parsing.
1871 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
1872 otherwise parse PKCS#7 structures from untrusted sources are
1873 affected. OpenSSL clients and servers are not affected.
1875 This issue was reported to OpenSSL by Michal Zalewski (Google).
1879 *) DoS via reachable assert in SSLv2 servers fix
1881 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
1882 servers that both support SSLv2 and enable export cipher suites by sending
1883 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
1885 This issue was discovered by Sean Burford (Google) and Emilia Käsper
1886 (OpenSSL development team).
1890 *) Empty CKE with client auth and DHE fix
1892 If client auth is used then a server can seg fault in the event of a DHE
1893 ciphersuite being selected and a zero length ClientKeyExchange message
1894 being sent by the client. This could be exploited in a DoS attack.
1898 *) Handshake with unseeded PRNG fix
1900 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
1901 with an unseeded PRNG. The conditions are:
1902 - The client is on a platform where the PRNG has not been seeded
1903 automatically, and the user has not seeded manually
1904 - A protocol specific client method version has been used (i.e. not
1905 SSL_client_methodv23)
1906 - A ciphersuite is used that does not require additional random data from
1907 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
1909 If the handshake succeeds then the client random that has been used will
1910 have been generated from a PRNG with insufficient entropy and therefore the
1911 output may be predictable.
1913 For example using the following command with an unseeded openssl will
1914 succeed on an unpatched platform:
1916 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
1920 *) Use After Free following d2i_ECPrivatekey error fix
1922 A malformed EC private key file consumed via the d2i_ECPrivateKey function
1923 could cause a use after free condition. This, in turn, could cause a double
1924 free in several private key parsing functions (such as d2i_PrivateKey
1925 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
1926 for applications that receive EC private keys from untrusted
1927 sources. This scenario is considered rare.
1929 This issue was discovered by the BoringSSL project and fixed in their
1934 *) X509_to_X509_REQ NULL pointer deref fix
1936 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
1937 the certificate key is invalid. This function is rarely used in practice.
1939 This issue was discovered by Brian Carpenter.
1943 *) Removed the export ciphers from the DEFAULT ciphers
1946 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
1948 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
1949 ARMv5 through ARMv8, as opposite to "locking" it to single one.
1950 So far those who have to target multiple platforms would compromise
1951 and argue that binary targeting say ARMv5 would still execute on
1952 ARMv8. "Universal" build resolves this compromise by providing
1953 near-optimal performance even on newer platforms.
1956 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
1957 (other platforms pending).
1958 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
1960 *) Add support for the SignedCertificateTimestampList certificate and
1961 OCSP response extensions from RFC6962.
1964 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
1965 for corner cases. (Certain input points at infinity could lead to
1966 bogus results, with non-infinity inputs mapped to infinity too.)
1969 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
1970 This covers AES, SHA256/512 and GHASH. "Initial" means that most
1971 common cases are optimized and there still is room for further
1972 improvements. Vector Permutation AES for Altivec is also added.
1975 *) Add support for little-endian ppc64 Linux target.
1976 [Marcelo Cerri (IBM)]
1978 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
1979 SHA1, SHA256 and GHASH. "Initial" means that most common cases
1980 are optimized and there still is room for further improvements.
1981 Both 32- and 64-bit modes are supported.
1982 [Andy Polyakov, Ard Biesheuvel (Linaro)]
1984 *) Improved ARMv7 NEON support.
1987 *) Support for SPARC Architecture 2011 crypto extensions, first
1988 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
1989 SHA256/512, MD5, GHASH and modular exponentiation.
1990 [Andy Polyakov, David Miller]
1992 *) Accelerated modular exponentiation for Intel processors, a.k.a.
1994 [Shay Gueron & Vlad Krasnov (Intel Corp)]
1996 *) Support for new and upcoming Intel processors, including AVX2,
1997 BMI and SHA ISA extensions. This includes additional "stitched"
1998 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2001 This work was sponsored by Intel Corp.
2004 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2005 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2006 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2009 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2010 this fixes a limitation in previous versions of OpenSSL.
2013 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2014 MGF1 digest and OAEP label.
2017 *) Add EVP support for key wrapping algorithms, to avoid problems with
2018 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2019 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2020 algorithms and include tests cases.
2023 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2025 [Douglas E. Engert, Steve Henson]
2027 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2028 difference in days and seconds between two tm or ASN1_TIME structures.
2031 *) Add -rev test option to s_server to just reverse order of characters
2032 received by client and send back to server. Also prints an abbreviated
2033 summary of the connection parameters.
2036 *) New option -brief for s_client and s_server to print out a brief summary
2037 of connection parameters.
2040 *) Add callbacks for arbitrary TLS extensions.
2041 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2043 *) New option -crl_download in several openssl utilities to download CRLs
2044 from CRLDP extension in certificates.
2047 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2050 *) New function X509_CRL_diff to generate a delta CRL from the difference
2051 of two full CRLs. Add support to "crl" utility.
2054 *) New functions to set lookup_crls function and to retrieve
2055 X509_STORE from X509_STORE_CTX.
2058 *) Print out deprecated issuer and subject unique ID fields in
2062 *) Extend OCSP I/O functions so they can be used for simple general purpose
2063 HTTP as well as OCSP. New wrapper function which can be used to download
2064 CRLs using the OCSP API.
2067 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2070 *) SSL_CONF* functions. These provide a common framework for application
2071 configuration using configuration files or command lines.
2074 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2075 message callback and prints the results. Needs compile time option
2076 "enable-ssl-trace". New options to s_client and s_server to enable
2080 *) New ctrl and macro to retrieve supported points extensions.
2081 Print out extension in s_server and s_client.
2084 *) New functions to retrieve certificate signature and signature
2088 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2092 *) New Suite B modes for TLS code. These use and enforce the requirements
2093 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2094 only use Suite B curves. The Suite B modes can be set by using the
2095 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2098 *) New chain verification flags for Suite B levels of security. Check
2099 algorithms are acceptable when flags are set in X509_verify_cert.
2102 *) Make tls1_check_chain return a set of flags indicating checks passed
2103 by a certificate chain. Add additional tests to handle client
2104 certificates: checks for matching certificate type and issuer name
2108 *) If an attempt is made to use a signature algorithm not in the peer
2109 preference list abort the handshake. If client has no suitable
2110 signature algorithms in response to a certificate request do not
2111 use the certificate.
2114 *) If server EC tmp key is not in client preference list abort handshake.
2117 *) Add support for certificate stores in CERT structure. This makes it
2118 possible to have different stores per SSL structure or one store in
2119 the parent SSL_CTX. Include distinct stores for certificate chain
2120 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2121 to build and store a certificate chain in CERT structure: returning
2122 an error if the chain cannot be built: this will allow applications
2123 to test if a chain is correctly configured.
2125 Note: if the CERT based stores are not set then the parent SSL_CTX
2126 store is used to retain compatibility with existing behaviour.
2130 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2131 mask based on the current session, check mask when sending client
2132 hello and checking the requested ciphersuite.
2135 *) New ctrls to retrieve and set certificate types in a certificate
2136 request message. Print out received values in s_client. If certificate
2137 types is not set with custom values set sensible values based on
2138 supported signature algorithms.
2141 *) Support for distinct client and server supported signature algorithms.
2144 *) Add certificate callback. If set this is called whenever a certificate
2145 is required by client or server. An application can decide which
2146 certificate chain to present based on arbitrary criteria: for example
2147 supported signature algorithms. Add very simple example to s_server.
2148 This fixes many of the problems and restrictions of the existing client
2149 certificate callback: for example you can now clear an existing
2150 certificate and specify the whole chain.
2153 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2154 the certificate can be used for (if anything). Set valid_flags field
2155 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2156 to have similar checks in it.
2158 Add new "cert_flags" field to CERT structure and include a "strict mode".
2159 This enforces some TLS certificate requirements (such as only permitting
2160 certificate signature algorithms contained in the supported algorithms
2161 extension) which some implementations ignore: this option should be used
2162 with caution as it could cause interoperability issues.
2165 *) Update and tidy signature algorithm extension processing. Work out
2166 shared signature algorithms based on preferences and peer algorithms
2167 and print them out in s_client and s_server. Abort handshake if no
2168 shared signature algorithms.
2171 *) Add new functions to allow customised supported signature algorithms
2172 for SSL and SSL_CTX structures. Add options to s_client and s_server
2176 *) New function SSL_certs_clear() to delete all references to certificates
2177 from an SSL structure. Before this once a certificate had been added
2178 it couldn't be removed.
2181 *) Integrate hostname, email address and IP address checking with certificate
2182 verification. New verify options supporting checking in openssl utility.
2185 *) Fixes and wildcard matching support to hostname and email checking
2186 functions. Add manual page.
2187 [Florian Weimer (Red Hat Product Security Team)]
2189 *) New functions to check a hostname email or IP address against a
2190 certificate. Add options x509 utility to print results of checks against
2194 *) Fix OCSP checking.
2195 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2197 *) Initial experimental support for explicitly trusted non-root CAs.
2198 OpenSSL still tries to build a complete chain to a root but if an
2199 intermediate CA has a trust setting included that is used. The first
2200 setting is used: whether to trust (e.g., -addtrust option to the x509
2204 *) Add -trusted_first option which attempts to find certificates in the
2205 trusted store even if an untrusted chain is also supplied.
2208 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2209 platform support for Linux and Android.
2212 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2215 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2216 When in FIPS mode the approved implementations are used as normal,
2217 when not in FIPS mode the internal unapproved versions are used instead.
2218 This means that the FIPS capable OpenSSL isn't forced to use the
2219 (often lower performance) FIPS implementations outside FIPS mode.
2222 *) Transparently support X9.42 DH parameters when calling
2223 PEM_read_bio_DHparameters. This means existing applications can handle
2224 the new parameter format automatically.
2227 *) Initial experimental support for X9.42 DH parameter format: mainly
2228 to support use of 'q' parameter for RFC5114 parameters.
2231 *) Add DH parameters from RFC5114 including test data to dhtest.
2234 *) Support for automatic EC temporary key parameter selection. If enabled
2235 the most preferred EC parameters are automatically used instead of
2236 hardcoded fixed parameters. Now a server just has to call:
2237 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2238 support ECDH and use the most appropriate parameters.
2241 *) Enhance and tidy EC curve and point format TLS extension code. Use
2242 static structures instead of allocation if default values are used.
2243 New ctrls to set curves we wish to support and to retrieve shared curves.
2244 Print out shared curves in s_server. New options to s_server and s_client
2245 to set list of supported curves.
2248 *) New ctrls to retrieve supported signature algorithms and
2249 supported curve values as an array of NIDs. Extend openssl utility
2250 to print out received values.
2253 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2254 between NIDs and the more common NIST names such as "P-256". Enhance
2255 ecparam utility and ECC method to recognise the NIST names for curves.
2258 *) Enhance SSL/TLS certificate chain handling to support different
2259 chains for each certificate instead of one chain in the parent SSL_CTX.
2262 *) Support for fixed DH ciphersuite client authentication: where both
2263 server and client use DH certificates with common parameters.
2266 *) Support for fixed DH ciphersuites: those requiring DH server
2270 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2272 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2273 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2274 X509_CINF_get_signature were reverted post internal team review.
2276 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2278 *) Build fixes for the Windows and OpenVMS platforms
2279 [Matt Caswell and Richard Levitte]
2281 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2283 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2284 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2285 dereference. This could lead to a Denial Of Service attack. Thanks to
2286 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2290 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2291 dtls1_buffer_record function under certain conditions. In particular this
2292 could occur if an attacker sent repeated DTLS records with the same
2293 sequence number but for the next epoch. The memory leak could be exploited
2294 by an attacker in a Denial of Service attack through memory exhaustion.
2295 Thanks to Chris Mueller for reporting this issue.
2299 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2300 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2301 method would be set to NULL which could later result in a NULL pointer
2302 dereference. Thanks to Frank Schmirler for reporting this issue.
2306 *) Abort handshake if server key exchange message is omitted for ephemeral
2309 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2310 reporting this issue.
2314 *) Remove non-export ephemeral RSA code on client and server. This code
2315 violated the TLS standard by allowing the use of temporary RSA keys in
2316 non-export ciphersuites and could be used by a server to effectively
2317 downgrade the RSA key length used to a value smaller than the server
2318 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2319 INRIA or reporting this issue.
2323 *) Fixed issue where DH client certificates are accepted without verification.
2324 An OpenSSL server will accept a DH certificate for client authentication
2325 without the certificate verify message. This effectively allows a client to
2326 authenticate without the use of a private key. This only affects servers
2327 which trust a client certificate authority which issues certificates
2328 containing DH keys: these are extremely rare and hardly ever encountered.
2329 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2334 *) Ensure that the session ID context of an SSL is updated when its
2335 SSL_CTX is updated via SSL_set_SSL_CTX.
2337 The session ID context is typically set from the parent SSL_CTX,
2338 and can vary with the CTX.
2341 *) Fix various certificate fingerprint issues.
2343 By using non-DER or invalid encodings outside the signed portion of a
2344 certificate the fingerprint can be changed without breaking the signature.
2345 Although no details of the signed portion of the certificate can be changed
2346 this can cause problems with some applications: e.g. those using the
2347 certificate fingerprint for blacklists.
2349 1. Reject signatures with non zero unused bits.
2351 If the BIT STRING containing the signature has non zero unused bits reject
2352 the signature. All current signature algorithms require zero unused bits.
2354 2. Check certificate algorithm consistency.
2356 Check the AlgorithmIdentifier inside TBS matches the one in the
2357 certificate signature. NB: this will result in signature failure
2358 errors for some broken certificates.
2360 Thanks to Konrad Kraszewski from Google for reporting this issue.
2362 3. Check DSA/ECDSA signatures use DER.
2364 Re-encode DSA/ECDSA signatures and compare with the original received
2365 signature. Return an error if there is a mismatch.
2367 This will reject various cases including garbage after signature
2368 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2369 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2370 (negative or with leading zeroes).
2372 Further analysis was conducted and fixes were developed by Stephen Henson
2373 of the OpenSSL core team.
2378 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2379 results on some platforms, including x86_64. This bug occurs at random
2380 with a very low probability, and is not known to be exploitable in any
2381 way, though its exact impact is difficult to determine. Thanks to Pieter
2382 Wuille (Blockstream) who reported this issue and also suggested an initial
2383 fix. Further analysis was conducted by the OpenSSL development team and
2384 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2385 the OpenSSL core team.
2389 *) Do not resume sessions on the server if the negotiated protocol
2390 version does not match the session's version. Resuming with a different
2391 version, while not strictly forbidden by the RFC, is of questionable
2392 sanity and breaks all known clients.
2393 [David Benjamin, Emilia Käsper]
2395 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2396 early CCS messages during renegotiation. (Note that because
2397 renegotiation is encrypted, this early CCS was not exploitable.)
2400 *) Tighten client-side session ticket handling during renegotiation:
2401 ensure that the client only accepts a session ticket if the server sends
2402 the extension anew in the ServerHello. Previously, a TLS client would
2403 reuse the old extension state and thus accept a session ticket if one was
2404 announced in the initial ServerHello.
2406 Similarly, ensure that the client requires a session ticket if one
2407 was advertised in the ServerHello. Previously, a TLS client would
2408 ignore a missing NewSessionTicket message.
2411 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2413 *) SRTP Memory Leak.
2415 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2416 sends a carefully crafted handshake message, to cause OpenSSL to fail
2417 to free up to 64k of memory causing a memory leak. This could be
2418 exploited in a Denial Of Service attack. This issue affects OpenSSL
2419 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2420 whether SRTP is used or configured. Implementations of OpenSSL that
2421 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2423 The fix was developed by the OpenSSL team.
2427 *) Session Ticket Memory Leak.
2429 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2430 integrity of that ticket is first verified. In the event of a session
2431 ticket integrity check failing, OpenSSL will fail to free memory
2432 causing a memory leak. By sending a large number of invalid session
2433 tickets an attacker could exploit this issue in a Denial Of Service
2438 *) Build option no-ssl3 is incomplete.
2440 When OpenSSL is configured with "no-ssl3" as a build option, servers
2441 could accept and complete a SSL 3.0 handshake, and clients could be
2442 configured to send them.
2444 [Akamai and the OpenSSL team]
2446 *) Add support for TLS_FALLBACK_SCSV.
2447 Client applications doing fallback retries should call
2448 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2450 [Adam Langley, Bodo Moeller]
2452 *) Add additional DigestInfo checks.
2454 Re-encode DigestInto in DER and check against the original when
2455 verifying RSA signature: this will reject any improperly encoded
2456 DigestInfo structures.
2458 Note: this is a precautionary measure and no attacks are currently known.
2462 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2464 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2465 SRP code can be overrun an internal buffer. Add sanity check that
2466 g, A, B < N to SRP code.
2468 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2469 Group for discovering this issue.
2473 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2474 TLS 1.0 instead of higher protocol versions when the ClientHello message
2475 is badly fragmented. This allows a man-in-the-middle attacker to force a
2476 downgrade to TLS 1.0 even if both the server and the client support a
2477 higher protocol version, by modifying the client's TLS records.
2479 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2480 researching this issue.
2484 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2485 to a denial of service attack. A malicious server can crash the client
2486 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2487 ciphersuite and sending carefully crafted handshake messages.
2489 Thanks to Felix Gröbert (Google) for discovering and researching this
2494 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2495 to leak memory. This can be exploited through a Denial of Service attack.
2496 Thanks to Adam Langley for discovering and researching this issue.
2500 *) An attacker can force openssl to consume large amounts of memory whilst
2501 processing DTLS handshake messages. This can be exploited through a
2502 Denial of Service attack.
2503 Thanks to Adam Langley for discovering and researching this issue.
2507 *) An attacker can force an error condition which causes openssl to crash
2508 whilst processing DTLS packets due to memory being freed twice. This
2509 can be exploited through a Denial of Service attack.
2510 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2515 *) If a multithreaded client connects to a malicious server using a resumed
2516 session and the server sends an ec point format extension it could write
2517 up to 255 bytes to freed memory.
2519 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2524 *) A malicious server can crash an OpenSSL client with a null pointer
2525 dereference (read) by specifying an SRP ciphersuite even though it was not
2526 properly negotiated with the client. This can be exploited through a
2527 Denial of Service attack.
2529 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2530 discovering and researching this issue.
2534 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2535 X509_name_oneline, X509_name_print_ex et al. to leak some information
2536 from the stack. Applications may be affected if they echo pretty printing
2537 output to the attacker.
2539 Thanks to Ivan Fratric (Google) for discovering this issue.
2541 [Emilia Käsper, and Steve Henson]
2543 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2544 for corner cases. (Certain input points at infinity could lead to
2545 bogus results, with non-infinity inputs mapped to infinity too.)
2548 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2550 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2551 handshake can force the use of weak keying material in OpenSSL
2552 SSL/TLS clients and servers.
2554 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2555 researching this issue. (CVE-2014-0224)
2556 [KIKUCHI Masashi, Steve Henson]
2558 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2559 OpenSSL DTLS client the code can be made to recurse eventually crashing
2562 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2564 [Imre Rad, Steve Henson]
2566 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2567 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2568 client or server. This is potentially exploitable to run arbitrary
2569 code on a vulnerable client or server.
2571 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2572 [Jüri Aedla, Steve Henson]
2574 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2575 are subject to a denial of service attack.
2577 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2578 this issue. (CVE-2014-3470)
2579 [Felix Gröbert, Ivan Fratric, Steve Henson]
2581 *) Harmonize version and its documentation. -f flag is used to display
2583 [mancha <mancha1@zoho.com>]
2585 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2586 in i2d_ECPrivateKey.
2587 [mancha <mancha1@zoho.com>]
2589 *) Fix some double frees. These are not thought to be exploitable.
2590 [mancha <mancha1@zoho.com>]
2592 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2594 *) A missing bounds check in the handling of the TLS heartbeat extension
2595 can be used to reveal up to 64k of memory to a connected client or
2598 Thanks for Neel Mehta of Google Security for discovering this bug and to
2599 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2600 preparing the fix (CVE-2014-0160)
2601 [Adam Langley, Bodo Moeller]
2603 *) Fix for the attack described in the paper "Recovering OpenSSL
2604 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2605 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2606 http://eprint.iacr.org/2014/140
2608 Thanks to Yuval Yarom and Naomi Benger for discovering this
2609 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2610 [Yuval Yarom and Naomi Benger]
2612 *) TLS pad extension: draft-agl-tls-padding-03
2614 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2615 TLS client Hello record length value would otherwise be > 255 and
2616 less that 512 pad with a dummy extension containing zeroes so it
2617 is at least 512 bytes long.
2619 [Adam Langley, Steve Henson]
2621 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2623 *) Fix for TLS record tampering bug. A carefully crafted invalid
2624 handshake could crash OpenSSL with a NULL pointer exception.
2625 Thanks to Anton Johansson for reporting this issues.
2628 *) Keep original DTLS digest and encryption contexts in retransmission
2629 structures so we can use the previous session parameters if they need
2630 to be resent. (CVE-2013-6450)
2633 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2634 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2635 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2636 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2637 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2638 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2639 [Rob Stradling, Adam Langley]
2641 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2643 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2644 supporting platforms or when small records were transferred.
2645 [Andy Polyakov, Steve Henson]
2647 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2649 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2651 This addresses the flaw in CBC record processing discovered by
2652 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2653 at: http://www.isg.rhul.ac.uk/tls/
2655 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2656 Security Group at Royal Holloway, University of London
2657 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2658 Emilia Käsper for the initial patch.
2660 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2662 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2663 ciphersuites which can be exploited in a denial of service attack.
2664 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2665 and detecting this bug and to Wolfgang Ettlinger
2666 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2670 *) Return an error when checking OCSP signatures when key is NULL.
2671 This fixes a DoS attack. (CVE-2013-0166)
2674 *) Make openssl verify return errors.
2675 [Chris Palmer <palmer@google.com> and Ben Laurie]
2677 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2678 the right response is stapled. Also change SSL_get_certificate()
2679 so it returns the certificate actually sent.
2680 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2681 [Rob Stradling <rob.stradling@comodo.com>]
2683 *) Fix possible deadlock when decoding public keys.
2686 *) Don't use TLS 1.0 record version number in initial client hello
2690 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2692 *) Sanity check record length before skipping explicit IV in TLS
2693 1.2, 1.1 and DTLS to fix DoS attack.
2695 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2696 fuzzing as a service testing platform.
2700 *) Initialise tkeylen properly when encrypting CMS messages.
2701 Thanks to Solar Designer of Openwall for reporting this issue.
2704 *) In FIPS mode don't try to use composite ciphers as they are not
2708 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2710 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2711 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2712 mean any application compiled against OpenSSL 1.0.0 headers setting
2713 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2714 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2715 0x10000000L Any application which was previously compiled against
2716 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2717 will need to be recompiled as a result. Letting be results in
2718 inability to disable specifically TLS 1.1 and in client context,
2719 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2722 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2723 disable just protocol X, but all protocols above X *if* there are
2724 protocols *below* X still enabled. In more practical terms it means
2725 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2726 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2727 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2731 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2733 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2734 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2735 in CRYPTO_realloc_clean.
2737 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2738 issue and to Adam Langley <agl@chromium.org> for fixing it.
2740 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2742 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2745 *) Workarounds for some broken servers that "hang" if a client hello
2746 record length exceeds 255 bytes.
2748 1. Do not use record version number > TLS 1.0 in initial client
2749 hello: some (but not all) hanging servers will now work.
2750 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2751 the number of ciphers sent in the client hello. This should be
2752 set to an even number, such as 50, for example by passing:
2753 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2754 Most broken servers should now work.
2755 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2756 TLS 1.2 client support entirely.
2759 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2762 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2764 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2765 STRING form instead of a DigestInfo.
2768 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2769 and the RSA_sign/RSA_verify functions. This was made more apparent when
2770 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2771 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2772 the correct format in RSA_verify so both forms transparently work.
2775 *) Some servers which support TLS 1.0 can choke if we initially indicate
2776 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2777 encrypted premaster secret. As a workaround use the maximum permitted
2778 client version in client hello, this should keep such servers happy
2779 and still work with previous versions of OpenSSL.
2782 *) Add support for TLS/DTLS heartbeats.
2783 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2785 *) Add support for SCTP.
2786 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2788 *) Improved PRNG seeding for VOS.
2789 [Paul Green <Paul.Green@stratus.com>]
2791 *) Extensive assembler packs updates, most notably:
2793 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2794 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2795 - x86_64: bit-sliced AES implementation;
2796 - ARM: NEON support, contemporary platforms optimizations;
2797 - s390x: z196 support;
2798 - *: GHASH and GF(2^m) multiplication implementations;
2802 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2803 (removal of unnecessary code)
2804 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2806 *) Add TLS key material exporter from RFC 5705.
2809 *) Add DTLS-SRTP negotiation from RFC 5764.
2812 *) Add Next Protocol Negotiation,
2813 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2814 disabled with a no-npn flag to config or Configure. Code donated
2816 [Adam Langley <agl@google.com> and Ben Laurie]
2818 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2819 NIST-P256, NIST-P521, with constant-time single point multiplication on
2820 typical inputs. Compiler support for the nonstandard type __uint128_t is
2821 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2822 Code made available under Apache License version 2.0.
2824 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2825 line to include this in your build of OpenSSL, and run "make depend" (or
2826 "make update"). This enables the following EC_METHODs:
2828 EC_GFp_nistp224_method()
2829 EC_GFp_nistp256_method()
2830 EC_GFp_nistp521_method()
2832 EC_GROUP_new_by_curve_name() will automatically use these (while
2833 EC_GROUP_new_curve_GFp() currently prefers the more flexible
2835 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
2837 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
2838 all platforms. Move ssize_t definition from e_os.h to the public
2839 header file e_os2.h as it now appears in public header file cms.h
2842 *) New -sigopt option to the ca, req and x509 utilities. Additional
2843 signature parameters can be passed using this option and in
2847 *) Add RSA PSS signing function. This will generate and set the
2848 appropriate AlgorithmIdentifiers for PSS based on those in the
2849 corresponding EVP_MD_CTX structure. No application support yet.
2852 *) Support for companion algorithm specific ASN1 signing routines.
2853 New function ASN1_item_sign_ctx() signs a pre-initialised
2854 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
2855 the appropriate parameters.
2858 *) Add new algorithm specific ASN1 verification initialisation function
2859 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
2860 handling will be the same no matter what EVP_PKEY_METHOD is used.
2861 Add a PSS handler to support verification of PSS signatures: checked
2862 against a number of sample certificates.
2865 *) Add signature printing for PSS. Add PSS OIDs.
2866 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
2868 *) Add algorithm specific signature printing. An individual ASN1 method
2869 can now print out signatures instead of the standard hex dump.
2871 More complex signatures (e.g. PSS) can print out more meaningful
2872 information. Include DSA version that prints out the signature
2876 *) Password based recipient info support for CMS library: implementing
2880 *) Split password based encryption into PBES2 and PBKDF2 functions. This
2881 neatly separates the code into cipher and PBE sections and is required
2882 for some algorithms that split PBES2 into separate pieces (such as
2883 password based CMS).
2886 *) Session-handling fixes:
2887 - Fix handling of connections that are resuming with a session ID,
2888 but also support Session Tickets.
2889 - Fix a bug that suppressed issuing of a new ticket if the client
2890 presented a ticket with an expired session.
2891 - Try to set the ticket lifetime hint to something reasonable.
2892 - Make tickets shorter by excluding irrelevant information.
2893 - On the client side, don't ignore renewed tickets.
2894 [Adam Langley, Bodo Moeller (Google)]
2896 *) Fix PSK session representation.
2899 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
2901 This work was sponsored by Intel.
2904 *) Add GCM support to TLS library. Some custom code is needed to split
2905 the IV between the fixed (from PRF) and explicit (from TLS record)
2906 portions. This adds all GCM ciphersuites supported by RFC5288 and
2907 RFC5289. Generalise some AES* cipherstrings to include GCM and
2908 add a special AESGCM string for GCM only.
2911 *) Expand range of ctrls for AES GCM. Permit setting invocation
2912 field on decrypt and retrieval of invocation field only on encrypt.
2915 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
2916 As required by RFC5289 these ciphersuites cannot be used if for
2917 versions of TLS earlier than 1.2.
2920 *) For FIPS capable OpenSSL interpret a NULL default public key method
2921 as unset and return the appropriate default but do *not* set the default.
2922 This means we can return the appropriate method in applications that
2923 switch between FIPS and non-FIPS modes.
2926 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
2927 ENGINE is used then we cannot handle that in the FIPS module so we
2928 keep original code iff non-FIPS operations are allowed.
2931 *) Add -attime option to openssl utilities.
2932 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
2934 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
2937 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
2938 FIPS EC methods unconditionally for now.
2941 *) New build option no-ec2m to disable characteristic 2 code.
2944 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
2945 all cases can be covered as some introduce binary incompatibilities.
2948 *) Redirect RSA operations to FIPS module including keygen,
2949 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
2952 *) Add similar low level API blocking to ciphers.
2955 *) Low level digest APIs are not approved in FIPS mode: any attempt
2956 to use these will cause a fatal error. Applications that *really* want
2957 to use them can use the private_* version instead.
2960 *) Redirect cipher operations to FIPS module for FIPS builds.
2963 *) Redirect digest operations to FIPS module for FIPS builds.
2966 *) Update build system to add "fips" flag which will link in fipscanister.o
2967 for static and shared library builds embedding a signature if needed.
2970 *) Output TLS supported curves in preference order instead of numerical
2971 order. This is currently hardcoded for the highest order curves first.
2972 This should be configurable so applications can judge speed vs strength.
2975 *) Add TLS v1.2 server support for client authentication.
2978 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
2982 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
2983 FIPS modules versions.
2986 *) Add TLS v1.2 client side support for client authentication. Keep cache
2987 of handshake records longer as we don't know the hash algorithm to use
2988 until after the certificate request message is received.
2991 *) Initial TLS v1.2 client support. Add a default signature algorithms
2992 extension including all the algorithms we support. Parse new signature
2993 format in client key exchange. Relax some ECC signing restrictions for
2994 TLS v1.2 as indicated in RFC5246.
2997 *) Add server support for TLS v1.2 signature algorithms extension. Switch
2998 to new signature format when needed using client digest preference.
2999 All server ciphersuites should now work correctly in TLS v1.2. No client
3000 support yet and no support for client certificates.
3003 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3004 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3005 ciphersuites. At present only RSA key exchange ciphersuites work with
3006 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3007 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3008 and version checking.
3011 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3012 with this defined it will not be affected by any changes to ssl internal
3013 structures. Add several utility functions to allow openssl application
3014 to work with OPENSSL_NO_SSL_INTERN defined.
3018 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3020 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3023 *) Permit abbreviated handshakes when renegotiating using the function
3024 SSL_renegotiate_abbreviated().
3025 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3027 *) Add call to ENGINE_register_all_complete() to
3028 ENGINE_load_builtin_engines(), so some implementations get used
3029 automatically instead of needing explicit application support.
3032 *) Add support for TLS key exporter as described in RFC5705.
3033 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3035 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3036 a few changes are required:
3038 Add SSL_OP_NO_TLSv1_1 flag.
3039 Add TLSv1_1 methods.
3040 Update version checking logic to handle version 1.1.
3041 Add explicit IV handling (ported from DTLS code).
3042 Add command line options to s_client/s_server.
3045 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3047 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3048 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3049 content decryption and always return the same error. Note: this attack
3050 needs on average 2^20 messages so it only affects automated senders. The
3051 old behaviour can be re-enabled in the CMS code by setting the
3052 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3053 an MMA defence is not necessary.
3054 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3055 this issue. (CVE-2012-0884)
3058 *) Fix CVE-2011-4619: make sure we really are receiving a
3059 client hello before rejecting multiple SGC restarts. Thanks to
3060 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3063 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3065 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3066 Thanks to Antonio Martin, Enterprise Secure Access Research and
3067 Development, Cisco Systems, Inc. for discovering this bug and
3068 preparing a fix. (CVE-2012-0050)
3071 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3073 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3074 of the Vaudenay padding oracle attack on CBC mode encryption
3075 which enables an efficient plaintext recovery attack against
3076 the OpenSSL implementation of DTLS. Their attack exploits timing
3077 differences arising during decryption processing. A research
3078 paper describing this attack can be found at:
3079 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3080 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3081 Security Group at Royal Holloway, University of London
3082 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3083 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3084 for preparing the fix. (CVE-2011-4108)
3085 [Robin Seggelmann, Michael Tuexen]
3087 *) Clear bytes used for block padding of SSL 3.0 records.
3089 [Adam Langley (Google)]
3091 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3092 Kadianakis <desnacked@gmail.com> for discovering this issue and
3093 Adam Langley for preparing the fix. (CVE-2011-4619)
3094 [Adam Langley (Google)]
3096 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3097 [Andrey Kulikov <amdeich@gmail.com>]
3099 *) Prevent malformed RFC3779 data triggering an assertion failure.
3100 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3101 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3102 [Rob Austein <sra@hactrn.net>]
3104 *) Improved PRNG seeding for VOS.
3105 [Paul Green <Paul.Green@stratus.com>]
3107 *) Fix ssl_ciph.c set-up race.
3108 [Adam Langley (Google)]
3110 *) Fix spurious failures in ecdsatest.c.
3111 [Emilia Käsper (Google)]
3113 *) Fix the BIO_f_buffer() implementation (which was mixing different
3114 interpretations of the '..._len' fields).
3115 [Adam Langley (Google)]
3117 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3118 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3119 threads won't reuse the same blinding coefficients.
3121 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3122 lock to call BN_BLINDING_invert_ex, and avoids one use of
3123 BN_BLINDING_update for each BN_BLINDING structure (previously,
3124 the last update always remained unused).
3125 [Emilia Käsper (Google)]
3127 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3128 [Bob Buckholz (Google)]
3130 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3132 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3133 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3134 [Kaspar Brand <ossl@velox.ch>]
3136 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3137 for multi-threaded use of ECDH. (CVE-2011-3210)
3138 [Adam Langley (Google)]
3140 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3143 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3144 signature public key algorithm by using OID xref utilities instead.
3145 Before this you could only use some ECC ciphersuites with SHA1 only.
3148 *) Add protection against ECDSA timing attacks as mentioned in the paper
3149 by Billy Bob Brumley and Nicola Tuveri, see:
3151 http://eprint.iacr.org/2011/232.pdf
3153 [Billy Bob Brumley and Nicola Tuveri]
3155 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3157 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3158 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3160 *) Fix bug in string printing code: if *any* escaping is enabled we must
3161 escape the escape character (backslash) or the resulting string is
3165 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3167 *) Disable code workaround for ancient and obsolete Netscape browsers
3168 and servers: an attacker can use it in a ciphersuite downgrade attack.
3169 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3172 *) Fixed J-PAKE implementation error, originally discovered by
3173 Sebastien Martini, further info and confirmation from Stefan
3174 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3177 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3179 *) Fix extension code to avoid race conditions which can result in a buffer
3180 overrun vulnerability: resumed sessions must not be modified as they can
3181 be shared by multiple threads. CVE-2010-3864
3184 *) Fix WIN32 build system to correctly link an ENGINE directory into
3188 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3190 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3192 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3194 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3196 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3197 context. The operation can be customised via the ctrl mechanism in
3198 case ENGINEs want to include additional functionality.
3201 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3204 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3205 output hashes compatible with older versions of OpenSSL.
3206 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3208 *) Fix compression algorithm handling: if resuming a session use the
3209 compression algorithm of the resumed session instead of determining
3210 it from client hello again. Don't allow server to change algorithm.
3213 *) Add load_crls() function to apps tidying load_certs() too. Add option
3214 to verify utility to allow additional CRLs to be included.
3217 *) Update OCSP request code to permit adding custom headers to the request:
3218 some responders need this.
3221 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3223 [Julia Lawall <julia@diku.dk>]
3225 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3226 needlessly dereferenced structures, used obsolete functions and
3227 didn't handle all updated verify codes correctly.
3230 *) Disable MD2 in the default configuration.
3233 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3234 indicate the initial BIO being pushed or popped. This makes it possible
3235 to determine whether the BIO is the one explicitly called or as a result
3236 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3237 it handles reference counts correctly and doesn't zero out the I/O bio
3238 when it is not being explicitly popped. WARNING: applications which
3239 included workarounds for the old buggy behaviour will need to be modified
3240 or they could free up already freed BIOs.
3243 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3244 renaming to all platforms (within the 0.9.8 branch, this was
3245 done conditionally on Netware platforms to avoid a name clash).
3246 [Guenter <lists@gknw.net>]
3248 *) Add ECDHE and PSK support to DTLS.
3249 [Michael Tuexen <tuexen@fh-muenster.de>]
3251 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3255 *) Add "missing" function EVP_MD_flags() (without this the only way to
3256 retrieve a digest flags is by accessing the structure directly. Update
3257 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3258 or cipher is registered as in the "from" argument. Print out all
3259 registered digests in the dgst usage message instead of manually
3260 attempting to work them out.
3263 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3264 this allows the use of compression and extensions. Change default cipher
3265 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3266 by default unless an application cipher string requests it.
3269 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3270 key ids to find matching certificates and keys but some PKCS#12 files
3271 don't follow the (somewhat unwritten) rules and this strategy fails.
3272 Now just gather all certificates together and the first private key
3273 then look for the first certificate that matches the key.
3276 *) Support use of registered digest and cipher names for dgst and cipher
3277 commands instead of having to add each one as a special case. So now
3284 openssl dgst -sha256 foo
3286 and this works for ENGINE based algorithms too.
3290 *) Update Gost ENGINE to support parameter files.
3291 [Victor B. Wagner <vitus@cryptocom.ru>]
3293 *) Support GeneralizedTime in ca utility.
3294 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3296 *) Enhance the hash format used for certificate directory links. The new
3297 form uses the canonical encoding (meaning equivalent names will work
3298 even if they aren't identical) and uses SHA1 instead of MD5. This form
3299 is incompatible with the older format and as a result c_rehash should
3300 be used to rebuild symbolic links.
3303 *) Make PKCS#8 the default write format for private keys, replacing the
3304 traditional format. This form is standardised, more secure and doesn't
3305 include an implicit MD5 dependency.
3308 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3309 committed to OpenSSL should pass this lot as a minimum.
3312 *) Add session ticket override functionality for use by EAP-FAST.
3313 [Jouni Malinen <j@w1.fi>]
3315 *) Modify HMAC functions to return a value. Since these can be implemented
3316 in an ENGINE errors can occur.
3319 *) Type-checked OBJ_bsearch_ex.
3322 *) Type-checked OBJ_bsearch. Also some constification necessitated
3323 by type-checking. Still to come: TXT_DB, bsearch(?),
3324 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3328 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3329 seconds to a tm structure directly, instead of going through OS
3330 specific date routines. This avoids any issues with OS routines such
3331 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3332 and X509_time_adj_ex() to cover the extended range. The existing
3333 X509_time_adj() is still usable and will no longer have any date issues.
3336 *) Delta CRL support. New use deltas option which will attempt to locate
3337 and search any appropriate delta CRLs available.
3339 This work was sponsored by Google.
3342 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3343 code and add additional score elements. Validate alternate CRL paths
3344 as part of the CRL checking and indicate a new error "CRL path validation
3345 error" in this case. Applications wanting additional details can use
3346 the verify callback and check the new "parent" field. If this is not
3347 NULL CRL path validation is taking place. Existing applications won't
3348 see this because it requires extended CRL support which is off by
3351 This work was sponsored by Google.
3354 *) Support for freshest CRL extension.
3356 This work was sponsored by Google.
3359 *) Initial indirect CRL support. Currently only supported in the CRLs
3360 passed directly and not via lookup. Process certificate issuer
3361 CRL entry extension and lookup CRL entries by bother issuer name
3362 and serial number. Check and process CRL issuer entry in IDP extension.
3364 This work was sponsored by Google.
3367 *) Add support for distinct certificate and CRL paths. The CRL issuer
3368 certificate is validated separately in this case. Only enabled if
3369 an extended CRL support flag is set: this flag will enable additional
3370 CRL functionality in future.
3372 This work was sponsored by Google.
3375 *) Add support for policy mappings extension.
3377 This work was sponsored by Google.
3380 *) Fixes to pathlength constraint, self issued certificate handling,
3381 policy processing to align with RFC3280 and PKITS tests.
3383 This work was sponsored by Google.
3386 *) Support for name constraints certificate extension. DN, email, DNS
3387 and URI types are currently supported.
3389 This work was sponsored by Google.
3392 *) To cater for systems that provide a pointer-based thread ID rather
3393 than numeric, deprecate the current numeric thread ID mechanism and
3394 replace it with a structure and associated callback type. This
3395 mechanism allows a numeric "hash" to be extracted from a thread ID in
3396 either case, and on platforms where pointers are larger than 'long',
3397 mixing is done to help ensure the numeric 'hash' is usable even if it
3398 can't be guaranteed unique. The default mechanism is to use "&errno"
3399 as a pointer-based thread ID to distinguish between threads.
3401 Applications that want to provide their own thread IDs should now use
3402 CRYPTO_THREADID_set_callback() to register a callback that will call
3403 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3405 Note that ERR_remove_state() is now deprecated, because it is tied
3406 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3407 to free the current thread's error state should be replaced by
3408 ERR_remove_thread_state(NULL).
3410 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3411 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3412 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3413 application was previously providing a numeric thread callback that
3414 was inappropriate for distinguishing threads, then uniqueness might
3415 have been obtained with &errno that happened immediately in the
3416 intermediate development versions of OpenSSL; this is no longer the
3417 case, the numeric thread callback will now override the automatic use
3419 [Geoff Thorpe, with help from Bodo Moeller]
3421 *) Initial support for different CRL issuing certificates. This covers a
3422 simple case where the self issued certificates in the chain exist and
3423 the real CRL issuer is higher in the existing chain.
3425 This work was sponsored by Google.
3428 *) Removed effectively defunct crypto/store from the build.
3431 *) Revamp of STACK to provide stronger type-checking. Still to come:
3432 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3433 ASN1_STRING, CONF_VALUE.
3436 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3437 RAM on SSL connections. This option can save about 34k per idle SSL.
3440 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3441 STACK, TXT_DB, bsearch, qsort.
3444 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3445 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3446 support for data, signedData, compressedData, digestedData and
3447 encryptedData, envelopedData types included. Scripts to check against
3448 RFC4134 examples draft and interop and consistency checks of many
3449 content types and variants.
3452 *) Add options to enc utility to support use of zlib compression BIO.
3455 *) Extend mk1mf to support importing of options and assembly language
3456 files from Configure script, currently only included in VC-WIN32.
3457 The assembly language rules can now optionally generate the source
3458 files from the associated perl scripts.
3461 *) Implement remaining functionality needed to support GOST ciphersuites.
3462 Interop testing has been performed using CryptoPro implementations.
3463 [Victor B. Wagner <vitus@cryptocom.ru>]
3465 *) s390x assembler pack.
3468 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3472 *) Implement Opaque PRF Input TLS extension as specified in
3473 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3474 official specification yet and no extension type assignment by
3475 IANA exists, this extension (for now) will have to be explicitly
3476 enabled when building OpenSSL by providing the extension number
3477 to use. For example, specify an option
3479 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3481 to the "config" or "Configure" script to enable the extension,
3482 assuming extension number 0x9527 (which is a completely arbitrary
3483 and unofficial assignment based on the MD5 hash of the Internet
3484 Draft). Note that by doing so, you potentially lose
3485 interoperability with other TLS implementations since these might
3486 be using the same extension number for other purposes.
3488 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3489 opaque PRF input value to use in the handshake. This will create
3490 an interal copy of the length-'len' string at 'src', and will
3491 return non-zero for success.
3493 To get more control and flexibility, provide a callback function
3496 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3497 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3501 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3504 Callback function 'cb' will be called in handshakes, and is
3505 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3506 Argument 'arg' is for application purposes (the value as given to
3507 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3508 be provided to the callback function). The callback function
3509 has to return non-zero to report success: usually 1 to use opaque
3510 PRF input just if possible, or 2 to enforce use of the opaque PRF
3511 input. In the latter case, the library will abort the handshake
3512 if opaque PRF input is not successfully negotiated.
3514 Arguments 'peerinput' and 'len' given to the callback function
3515 will always be NULL and 0 in the case of a client. A server will
3516 see the client's opaque PRF input through these variables if
3517 available (NULL and 0 otherwise). Note that if the server
3518 provides an opaque PRF input, the length must be the same as the
3519 length of the client's opaque PRF input.
3521 Note that the callback function will only be called when creating
3522 a new session (session resumption can resume whatever was
3523 previously negotiated), and will not be called in SSL 2.0
3524 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3525 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3526 for applications that need to enforce opaque PRF input.
3530 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3533 [Victor B. Wagner <vitus@cryptocom.ru>]
3535 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3536 RFC4507bis. The encrypted ticket format is an encrypted encoded
3537 SSL_SESSION structure, that way new session features are automatically
3540 If a client application caches session in an SSL_SESSION structure
3541 support is transparent because tickets are now stored in the encoded
3544 The SSL_CTX structure automatically generates keys for ticket
3545 protection in servers so again support should be possible
3546 with no application modification.
3548 If a client or server wishes to disable RFC4507 support then the option
3549 SSL_OP_NO_TICKET can be set.
3551 Add a TLS extension debugging callback to allow the contents of any client
3552 or server extensions to be examined.
3554 This work was sponsored by Google.
3557 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3558 OpenSSL should now compile cleanly on gcc 4.2
3559 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3561 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3562 support including streaming MAC support: this is required for GOST
3563 ciphersuite support.
3564 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3566 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3567 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3568 to output in BER and PEM format.
3571 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3572 allows HMAC to be handled via the EVP_DigestSign*() interface. The
3573 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
3574 ENGINE support for HMAC keys which are unextractable. New -mac and
3575 -macopt options to dgst utility.
3578 *) New option -sigopt to dgst utility. Update dgst to use
3579 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
3580 alternative signing parameters such as X9.31 or PSS in the dgst
3584 *) Change ssl_cipher_apply_rule(), the internal function that does
3585 the work each time a ciphersuite string requests enabling
3586 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
3587 removing ("!foo+bar") a class of ciphersuites: Now it maintains
3588 the order of disabled ciphersuites such that those ciphersuites
3589 that most recently went from enabled to disabled not only stay
3590 in order with respect to each other, but also have higher priority
3591 than other disabled ciphersuites the next time ciphersuites are
3594 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
3595 the same ciphersuites as with "HIGH" alone, but in a specific
3596 order where the PSK ciphersuites come first (since they are the
3597 most recently disabled ciphersuites when "HIGH" is parsed).
3599 Also, change ssl_create_cipher_list() (using this new
3600 funcionality) such that between otherwise identical
3601 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
3605 *) Change ssl_create_cipher_list() so that it automatically
3606 arranges the ciphersuites in reasonable order before starting
3607 to process the rule string. Thus, the definition for "DEFAULT"
3608 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
3609 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
3610 This makes it much easier to arrive at a reasonable default order
3611 in applications for which anonymous ciphers are OK (meaning
3612 that you can't actually use DEFAULT).
3613 [Bodo Moeller; suggested by Victor Duchovni]
3615 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
3616 processing) into multiple integers instead of setting
3617 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
3618 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
3619 (These masks as well as the individual bit definitions are hidden
3620 away into the non-exported interface ssl/ssl_locl.h, so this
3621 change to the definition of the SSL_CIPHER structure shouldn't
3622 affect applications.) This give us more bits for each of these
3623 categories, so there is no longer a need to coagulate AES128 and
3624 AES256 into a single algorithm bit, and to coagulate Camellia128
3625 and Camellia256 into a single algorithm bit, which has led to all
3628 Thus, among other things, the kludge introduced in 0.9.7m and
3629 0.9.8e for masking out AES256 independently of AES128 or masking
3630 out Camellia256 independently of AES256 is not needed here in 0.9.9.
3632 With the change, we also introduce new ciphersuite aliases that
3633 so far were missing: "AES128", "AES256", "CAMELLIA128", and
3637 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
3638 Use the leftmost N bytes of the signature input if the input is
3639 larger than the prime q (with N being the size in bytes of q).
3642 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
3643 it yet and it is largely untested.
3646 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
3649 *) Initial incomplete changes to avoid need for function casts in OpenSSL
3650 some compilers (gcc 4.2 and later) reject their use. Safestack is
3651 reimplemented. Update ASN1 to avoid use of legacy functions.
3654 *) Win32/64 targets are linked with Winsock2.
3657 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
3658 to external functions. This can be used to increase CRL handling
3659 efficiency especially when CRLs are very large by (for example) storing
3660 the CRL revoked certificates in a database.
3663 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
3664 new CRLs added to a directory can be used. New command line option
3665 -verify_return_error to s_client and s_server. This causes real errors
3666 to be returned by the verify callback instead of carrying on no matter
3667 what. This reflects the way a "real world" verify callback would behave.
3670 *) GOST engine, supporting several GOST algorithms and public key formats.
3671 Kindly donated by Cryptocom.
3674 *) Partial support for Issuing Distribution Point CRL extension. CRLs
3675 partitioned by DP are handled but no indirect CRL or reason partitioning
3676 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
3677 selected via a scoring technique which handles IDP and AKID in CRLs.
3680 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
3681 will ultimately be used for all verify operations: this will remove the
3682 X509_STORE dependency on certificate verification and allow alternative
3683 lookup methods. X509_STORE based implementations of these two callbacks.
3686 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
3687 Modify get_crl() to find a valid (unexpired) CRL if possible.
3690 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
3691 this would be called X509_CRL_cmp() but that name is already used by
3692 a function that just compares CRL issuer names. Cache several CRL
3693 extensions in X509_CRL structure and cache CRLDP in X509.
3696 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
3697 this maps equivalent X509_NAME structures into a consistent structure.
3698 Name comparison can then be performed rapidly using memcmp().
3701 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
3705 *) Allow digests to supply their own micalg string for S/MIME type using
3706 the ctrl EVP_MD_CTRL_MICALG.
3709 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
3710 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
3711 ctrl. It can then customise the structure before and/or after signing
3715 *) New function OBJ_add_sigid() to allow application defined signature OIDs
3716 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
3717 to free up any added signature OIDs.
3720 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
3721 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
3722 digest and cipher tables. New options added to openssl utility:
3723 list-message-digest-algorithms and list-cipher-algorithms.
3726 *) Change the array representation of binary polynomials: the list
3727 of degrees of non-zero coefficients is now terminated with -1.
3728 Previously it was terminated with 0, which was also part of the
3729 value; thus, the array representation was not applicable to
3730 polynomials where t^0 has coefficient zero. This change makes
3731 the array representation useful in a more general context.
3734 *) Various modifications and fixes to SSL/TLS cipher string
3735 handling. For ECC, the code now distinguishes between fixed ECDH
3736 with RSA certificates on the one hand and with ECDSA certificates
3737 on the other hand, since these are separate ciphersuites. The
3738 unused code for Fortezza ciphersuites has been removed.
3740 For consistency with EDH, ephemeral ECDH is now called "EECDH"
3741 (not "ECDHE"). For consistency with the code for DH
3742 certificates, use of ECDH certificates is now considered ECDH
3743 authentication, not RSA or ECDSA authentication (the latter is
3744 merely the CA's signing algorithm and not actively used in the
3747 The temporary ciphersuite alias "ECCdraft" is no longer
3748 available, and ECC ciphersuites are no longer excluded from "ALL"
3749 and "DEFAULT". The following aliases now exist for RFC 4492
3750 ciphersuites, most of these by analogy with the DH case:
3752 kECDHr - ECDH cert, signed with RSA
3753 kECDHe - ECDH cert, signed with ECDSA
3754 kECDH - ECDH cert (signed with either RSA or ECDSA)
3755 kEECDH - ephemeral ECDH
3756 ECDH - ECDH cert or ephemeral ECDH
3762 AECDH - anonymous ECDH
3763 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
3767 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
3768 Use correct micalg parameters depending on digest(s) in signed message.
3771 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
3772 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
3775 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
3776 an engine to register a method. Add ENGINE lookups for methods and
3777 functional reference processing.
3780 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
3781 EVP_{Sign,Verify}* which allow an application to customise the signature
3785 *) New -resign option to smime utility. This adds one or more signers
3786 to an existing PKCS#7 signedData structure. Also -md option to use an
3787 alternative message digest algorithm for signing.
3790 *) Tidy up PKCS#7 routines and add new functions to make it easier to
3791 create PKCS7 structures containing multiple signers. Update smime
3792 application to support multiple signers.
3795 *) New -macalg option to pkcs12 utility to allow setting of an alternative
3799 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
3800 Reorganize PBE internals to lookup from a static table using NIDs,
3801 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
3802 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
3803 PRF which will be automatically used with PBES2.
3806 *) Replace the algorithm specific calls to generate keys in "req" with the
3810 *) Update PKCS#7 enveloped data routines to use new API. This is now
3811 supported by any public key method supporting the encrypt operation. A
3812 ctrl is added to allow the public key algorithm to examine or modify
3813 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
3817 *) Add a ctrl to asn1 method to allow a public key algorithm to express
3818 a default digest type to use. In most cases this will be SHA1 but some
3819 algorithms (such as GOST) need to specify an alternative digest. The
3820 return value indicates how strong the preference is 1 means optional and
3821 2 is mandatory (that is it is the only supported type). Modify
3822 ASN1_item_sign() to accept a NULL digest argument to indicate it should
3823 use the default md. Update openssl utilities to use the default digest
3824 type for signing if it is not explicitly indicated.
3827 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
3828 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
3829 signing method from the key type. This effectively removes the link
3830 between digests and public key types.
3833 *) Add an OID cross reference table and utility functions. Its purpose is to
3834 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
3835 rsaEncryption. This will allow some of the algorithm specific hackery
3836 needed to use the correct OID to be removed.
3839 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
3840 structures for PKCS7_sign(). They are now set up by the relevant public
3844 *) Add provisional EC pkey method with support for ECDSA and ECDH.
3847 *) Add support for key derivation (agreement) in the API, DH method and
3851 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
3852 public and private key formats. As a side effect these add additional
3853 command line functionality not previously available: DSA signatures can be
3854 generated and verified using pkeyutl and DH key support and generation in
3859 [Oliver Tappe <zooey@hirschkaefer.de>]
3861 *) New make target "install_html_docs" installs HTML renditions of the
3863 [Oliver Tappe <zooey@hirschkaefer.de>]
3865 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
3866 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
3867 support key and parameter generation and add initial key generation
3868 functionality for RSA.
3871 *) Add functions for main EVP_PKEY_method operations. The undocumented
3872 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
3873 EVP_PKEY_{encrypt,decrypt}_old.
3876 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
3877 key API, doesn't do much yet.
3880 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
3881 public key algorithms. New option to openssl utility:
3882 "list-public-key-algorithms" to print out info.
3885 *) Implement the Supported Elliptic Curves Extension for
3886 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3889 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
3890 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
3893 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
3894 utilities such as rsa, dsa, dsaparam etc except they process any key
3898 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
3899 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
3900 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
3904 *) Initial support for pluggable public key ASN1.
3905 De-spaghettify the public key ASN1 handling. Move public and private
3906 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
3907 algorithm specific handling to a single module within the relevant
3908 algorithm directory. Add functions to allow (near) opaque processing
3909 of public and private key structures.
3912 *) Implement the Supported Point Formats Extension for
3913 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
3916 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
3917 for the psk identity [hint] and the psk callback functions to the
3918 SSL_SESSION, SSL and SSL_CTX structure.
3921 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
3925 SSL_CTX_use_psk_identity_hint
3926 SSL_get_psk_identity_hint
3927 SSL_get_psk_identity
3928 SSL_use_psk_identity_hint
3930 [Mika Kousa and Pasi Eronen of Nokia Corporation]
3932 *) Add RFC 3161 compliant time stamp request creation, response generation
3933 and response verification functionality.
3934 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
3936 *) Add initial support for TLS extensions, specifically for the server_name
3937 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3938 have new members for a host name. The SSL data structure has an
3939 additional member SSL_CTX *initial_ctx so that new sessions can be
3940 stored in that context to allow for session resumption, even after the
3941 SSL has been switched to a new SSL_CTX in reaction to a client's
3942 server_name extension.
3944 New functions (subject to change):
3946 SSL_get_servername()
3947 SSL_get_servername_type()
3950 New CTRL codes and macros (subject to change):
3952 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3953 - SSL_CTX_set_tlsext_servername_callback()
3954 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3955 - SSL_CTX_set_tlsext_servername_arg()
3956 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3958 openssl s_client has a new '-servername ...' option.
3960 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3961 '-key2 ...', '-servername_fatal' (subject to change). This allows
3962 testing the HostName extension for a specific single host name ('-cert'
3963 and '-key' remain fallbacks for handshakes without HostName
3964 negotiation). If the unrecognized_name alert has to be sent, this by
3965 default is a warning; it becomes fatal with the '-servername_fatal'
3968 [Peter Sylvester, Remy Allais, Christophe Renou]
3970 *) Whirlpool hash implementation is added.
3973 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
3974 bn(64,32). Because of instruction set limitations it doesn't have
3975 any negative impact on performance. This was done mostly in order
3976 to make it possible to share assembler modules, such as bn_mul_mont
3977 implementations, between 32- and 64-bit builds without hassle.
3980 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
3981 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
3985 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
3986 dedicated Montgomery multiplication procedure, is introduced.
3987 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
3988 "64-bit" performance on certain 32-bit targets.
3991 *) New option SSL_OP_NO_COMP to disable use of compression selectively
3992 in SSL structures. New SSL ctrl to set maximum send fragment size.
3993 Save memory by seeting the I/O buffer sizes dynamically instead of
3994 using the maximum available value.
3997 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
3998 in addition to the text details.
4001 *) Very, very preliminary EXPERIMENTAL support for printing of general
4002 ASN1 structures. This currently produces rather ugly output and doesn't
4003 handle several customised structures at all.
4006 *) Integrated support for PVK file format and some related formats such
4007 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4008 these in the 'rsa' and 'dsa' utilities.
4011 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4014 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4015 place for the (very old) "NETSCAPE" format certificates which are now
4016 handled using new ASN1 code equivalents.
4019 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4020 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4021 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4024 *) Modify CRL distribution points extension code to print out previously
4025 unsupported fields. Enhance extension setting code to allow setting of
4029 *) Add print and set support for Issuing Distribution Point CRL extension.
4032 *) Change 'Configure' script to enable Camellia by default.
4035 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4037 *) When rejecting SSL/TLS records due to an incorrect version number, never
4038 update s->server with a new major version number. As of
4039 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4040 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4041 the previous behavior could result in a read attempt at NULL when
4042 receiving specific incorrect SSL/TLS records once record payload
4043 protection is active. (CVE-2010-0740)
4044 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4046 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4047 could be crashed if the relevant tables were not present (e.g. chrooted).
4048 [Tomas Hoger <thoger@redhat.com>]
4050 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4052 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
4053 [Martin Olsson, Neel Mehta]
4055 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4056 accommodate for stack sorting, always a write lock!).
4059 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4060 excessive delays in the RAND_poll(): over a minute. As a workaround
4061 include a time check in the inner Heap32Next loop too.
4064 *) The code that handled flushing of data in SSL/TLS originally used the
4065 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4066 the problem outlined in PR#1949. The fix suggested there however can
4067 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4068 of Apache). So instead simplify the code to flush unconditionally.
4069 This should be fine since flushing with no data to flush is a no op.
4072 *) Handle TLS versions 2.0 and later properly and correctly use the
4073 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4074 off ancient servers have a habit of sticking around for a while...
4077 *) Modify compression code so it frees up structures without using the
4078 ex_data callbacks. This works around a problem where some applications
4079 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4080 restarting) then use compression (e.g. SSL with compression) later.
4081 This results in significant per-connection memory leaks and
4082 has caused some security issues including CVE-2008-1678 and
4086 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4087 change when encrypting or decrypting.
4090 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4091 connect and renegotiate with servers which do not support RI.
4092 Until RI is more widely deployed this option is enabled by default.
4095 *) Add "missing" ssl ctrls to clear options and mode.
4098 *) If client attempts to renegotiate and doesn't support RI respond with
4099 a no_renegotiation alert as required by RFC5746. Some renegotiating
4100 TLS clients will continue a connection gracefully when they receive
4101 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4102 waiting for a server hello which it will never receive. Now we treat a
4103 received no_renegotiation alert as a fatal error. This is because
4104 applications requesting a renegotiation might well expect it to succeed
4105 and would have no code in place to handle the server denying it so the
4106 only safe thing to do is to terminate the connection.
4109 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4110 peer supports secure renegotiation and 0 otherwise. Print out peer
4111 renegotiation support in s_client/s_server.
4114 *) Replace the highly broken and deprecated SPKAC certification method with
4115 the updated NID creation version. This should correctly handle UTF8.
4118 *) Implement RFC5746. Re-enable renegotiation but require the extension
4119 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4120 turns out to be a bad idea. It has been replaced by
4121 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4122 SSL_CTX_set_options(). This is really not recommended unless you
4123 know what you are doing.
4124 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4126 *) Fixes to stateless session resumption handling. Use initial_ctx when
4127 issuing and attempting to decrypt tickets in case it has changed during
4128 servername handling. Use a non-zero length session ID when attempting
4129 stateless session resumption: this makes it possible to determine if
4130 a resumption has occurred immediately after receiving server hello
4131 (several places in OpenSSL subtly assume this) instead of later in
4135 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4136 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4137 fixes for a few places where the return code is not checked
4139 [Julia Lawall <julia@diku.dk>]
4141 *) Add --strict-warnings option to Configure script to include devteam
4142 warnings in other configurations.
4145 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4146 makes it possible to install openssl libraries in locations which
4147 have names other than "lib", for example "/usr/lib64" which some
4149 [Steve Henson, based on patch from Jeremy Utley]
4151 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4152 X690 8.9.12 and can produce some misleading textual output of OIDs.
4153 [Steve Henson, reported by Dan Kaminsky]
4155 *) Delete MD2 from algorithm tables. This follows the recommendation in
4156 several standards that it is not used in new applications due to
4157 several cryptographic weaknesses. For binary compatibility reasons
4158 the MD2 API is still compiled in by default.
4161 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4165 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4166 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4168 [Guenter <lists@gknw.net>]
4170 *) Fix the server certificate chain building code to use X509_verify_cert(),
4171 it used to have an ad-hoc builder which was unable to cope with anything
4172 other than a simple chain.
4173 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4175 *) Don't check self signed certificate signatures in X509_verify_cert()
4176 by default (a flag can override this): it just wastes time without
4177 adding any security. As a useful side effect self signed root CAs
4178 with non-FIPS digests are now usable in FIPS mode.
4181 *) In dtls1_process_out_of_seq_message() the check if the current message
4182 is already buffered was missing. For every new message was memory
4183 allocated, allowing an attacker to perform an denial of service attack
4184 with sending out of seq handshake messages until there is no memory
4185 left. Additionally every future messege was buffered, even if the
4186 sequence number made no sense and would be part of another handshake.
4187 So only messages with sequence numbers less than 10 in advance will be
4188 buffered. (CVE-2009-1378)
4189 [Robin Seggelmann, discovered by Daniel Mentz]
4191 *) Records are buffered if they arrive with a future epoch to be
4192 processed after finishing the corresponding handshake. There is
4193 currently no limitation to this buffer allowing an attacker to perform
4194 a DOS attack with sending records with future epochs until there is no
4195 memory left. This patch adds the pqueue_size() function to determine
4196 the size of a buffer and limits the record buffer to 100 entries.
4198 [Robin Seggelmann, discovered by Daniel Mentz]
4200 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4201 parent structure is freed. (CVE-2009-1379)
4204 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4205 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4208 [Ilya O. <vrghost@gmail.com>]
4210 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4212 *) Disable renegotiation completely - this fixes a severe security
4213 problem (CVE-2009-3555) at the cost of breaking all
4214 renegotiation. Renegotiation can be re-enabled by setting
4215 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4216 run-time. This is really not recommended unless you know what
4220 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4222 *) Don't set val to NULL when freeing up structures, it is freed up by
4223 underlying code. If sizeof(void *) > sizeof(long) this can result in
4224 zeroing past the valid field. (CVE-2009-0789)
4225 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4227 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4228 checked correctly. This would allow some invalid signed attributes to
4229 appear to verify correctly. (CVE-2009-0591)
4230 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4232 *) Reject UniversalString and BMPString types with invalid lengths. This
4233 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4234 a legal length. (CVE-2009-0590)
4237 *) Set S/MIME signing as the default purpose rather than setting it
4238 unconditionally. This allows applications to override it at the store
4242 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4243 to handle some structures.
4246 *) Improve efficiency of mem_gets: don't search whole buffer each time
4248 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4250 *) New -hex option for openssl rand.
4253 *) Print out UTF8String and NumericString when parsing ASN1.
4256 *) Support NumericString type for name components.
4259 *) Allow CC in the environment to override the automatically chosen
4260 compiler. Note that nothing is done to ensure flags work with the