5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.0f and 1.1.1 [xx XXX xxxx]
12 *) Added processing of "make variables" to Configure.
15 *) Added SHA512/224 and SHA512/256 algorithm support.
18 *) The last traces of Netware support, first removed in 1.1.0, have
22 *) Get rid of Makefile.shared, and in the process, make the processing
23 of certain files (rc.obj, or the .def/.map/.opt files produced from
24 the ordinal files) more visible and hopefully easier to trace and
25 debug (or make silent).
28 *) Make it possible to have environment variable assignments as
29 arguments to config / Configure.
32 *) Add multi-prime RSA (RFC 8017) support.
35 *) Add SM3 implemented according to GB/T 32905-2016
36 [ Jack Lloyd <jack.lloyd@ribose.com>,
37 Ronald Tse <ronald.tse@ribose.com>,
38 Erick Borsboom <erick.borsboom@ribose.com> ]
40 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
41 as documented in RFC6066.
42 Based on a patch from Tomasz Moń
43 [Filipe Raimundo da Silva]
45 *) Add SM4 implemented according to GB/T 32907-2016.
46 [ Jack Lloyd <jack.lloyd@ribose.com>,
47 Ronald Tse <ronald.tse@ribose.com>,
48 Erick Borsboom <erick.borsboom@ribose.com> ]
50 *) Reimplement -newreq-nodes and ERR_error_string_n; the
51 original author does not agree with the license change.
54 *) Add ARIA AEAD TLS support.
57 *) Some macro definitions to support VS6 have been removed. Visual
58 Studio 6 has not worked since 1.1.0
61 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
62 without clearing the errors.
65 *) Add "atfork" functions. If building on a system that without
66 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
67 requirements. The RAND facility now uses/requires this.
73 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
74 not possible to disable entirely. However, it's still possible to
75 disable the console reading UI method, UI_OpenSSL() (use UI_null()
78 To disable, configure with 'no-ui-console'. 'no-ui' is still
79 possible to use as an alias. Check at compile time with the
80 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
81 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
84 *) Add a STORE module, which implements a uniform and URI based reader of
85 stores that can contain keys, certificates, CRLs and numerous other
86 objects. The main API is loosely based on a few stdio functions,
87 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
88 OSSL_STORE_error and OSSL_STORE_close.
89 The implementation uses backends called "loaders" to implement arbitrary
90 URI schemes. There is one built in "loader" for the 'file' scheme.
93 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
94 then adjusted to work on FreeBSD 8.4 as well.
95 Enable by configuring with 'enable-devcryptoeng'. This is done by default
96 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
99 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
100 util/mkerr.pl, which is adapted to allow those prefixes, leading to
101 error code calls like this:
103 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
105 With this change, we claim the namespaces OSSL and OPENSSL in a manner
106 that can be encoded in C. For the foreseeable future, this will only
108 [Richard Levitte and Tim Hudson]
110 *) Removed BSD cryptodev engine.
113 *) Add a build target 'build_all_generated', to build all generated files
114 and only that. This can be used to prepare everything that requires
115 things like perl for a system that lacks perl and then move everything
116 to that system and do the rest of the build there.
119 *) In the UI interface, make it possible to duplicate the user data. This
120 can be used by engines that need to retain the data for a longer time
121 than just the call where this user data is passed.
124 *) Ignore the '-named_curve auto' value for compatibility of applications
126 [Tomas Mraz <tmraz@fedoraproject.org>]
128 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
129 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
130 alerts across multiple records (some of which could be empty). In practice
131 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
132 prohibts this altogether and other libraries (BoringSSL, NSS) do not
133 support this at all. Supporting it adds significant complexity to the
134 record layer, and its removal is unlikely to cause inter-operability
138 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
139 with Z. These are meant to replace LONG and ZLONG and to be size safe.
140 The use of LONG and ZLONG is discouraged and scheduled for deprecation
144 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
145 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
146 [Richard Levitte, Andy Polyakov]
148 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
152 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
153 platform rather than 'mingw'.
156 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
157 success if they are asked to add an object which already exists
158 in the store. This change cascades to other functions which load
159 certificates and CRLs.
162 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
163 facilitate stack unwinding even from assembly subroutines.
166 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
167 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
170 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
171 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
172 which is the minimum version we support.
175 *) Certificate time validation (X509_cmp_time) enforces stricter
176 compliance with RFC 5280. Fractional seconds and timezone offsets
177 are no longer allowed.
180 *) Add support for ARIA
183 *) s_client will now send the Server Name Indication (SNI) extension by
184 default unless the new "-noservername" option is used. The server name is
185 based on the host provided to the "-connect" option unless overridden by
189 *) Add support for SipHash
192 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
193 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
194 prevent issues where no progress is being made and the peer continually
195 sends unrecognised record types, using up resources processing them.
198 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
199 using the algorithm defined in
200 https://www.akkadia.org/drepper/SHA-crypt.txt
203 *) Heartbeat support has been removed; the ABI is changed for now.
204 [Richard Levitte, Rich Salz]
206 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
209 *) The RSA "null" method, which was partially supported to avoid patent
210 issues, has been replaced to always returns NULL.
213 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
215 *) Removed the OS390-Unix config target. It relied on a script that doesn't
219 *) rsaz_1024_mul_avx2 overflow bug on x86_64
221 There is an overflow bug in the AVX2 Montgomery multiplication procedure
222 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
223 Analysis suggests that attacks against RSA and DSA as a result of this
224 defect would be very difficult to perform and are not believed likely.
225 Attacks against DH1024 are considered just feasible, because most of the
226 work necessary to deduce information about a private key may be performed
227 offline. The amount of resources required for such an attack would be
228 significant. However, for an attack on TLS to be meaningful, the server
229 would have to share the DH1024 private key among multiple clients, which is
230 no longer an option since CVE-2016-0701.
232 This only affects processors that support the AVX2 but not ADX extensions
233 like Intel Haswell (4th generation).
235 This issue was reported to OpenSSL by David Benjamin (Google). The issue
236 was originally found via the OSS-Fuzz project.
240 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
242 *) bn_sqrx8x_internal carry bug on x86_64
244 There is a carry propagating bug in the x86_64 Montgomery squaring
245 procedure. No EC algorithms are affected. Analysis suggests that attacks
246 against RSA and DSA as a result of this defect would be very difficult to
247 perform and are not believed likely. Attacks against DH are considered just
248 feasible (although very difficult) because most of the work necessary to
249 deduce information about a private key may be performed offline. The amount
250 of resources required for such an attack would be very significant and
251 likely only accessible to a limited number of attackers. An attacker would
252 additionally need online access to an unpatched system using the target
253 private key in a scenario with persistent DH parameters and a private
254 key that is shared between multiple clients.
256 This only affects processors that support the BMI1, BMI2 and ADX extensions
257 like Intel Broadwell (5th generation) and later or AMD Ryzen.
259 This issue was reported to OpenSSL by the OSS-Fuzz project.
263 *) Malformed X.509 IPAddressFamily could cause OOB read
265 If an X.509 certificate has a malformed IPAddressFamily extension,
266 OpenSSL could do a one-byte buffer overread. The most likely result
267 would be an erroneous display of the certificate in text format.
269 This issue was reported to OpenSSL by the OSS-Fuzz project.
273 Changes between 1.1.0e and 1.1.0f [25 May 2017]
275 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
276 platform rather than 'mingw'.
279 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
280 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
281 which is the minimum version we support.
284 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
286 *) Encrypt-Then-Mac renegotiation crash
288 During a renegotiation handshake if the Encrypt-Then-Mac extension is
289 negotiated where it was not in the original handshake (or vice-versa) then
290 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
291 and servers are affected.
293 This issue was reported to OpenSSL by Joe Orton (Red Hat).
297 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
299 *) Truncated packet could crash via OOB read
301 If one side of an SSL/TLS path is running on a 32-bit host and a specific
302 cipher is being used, then a truncated packet can cause that host to
303 perform an out-of-bounds read, usually resulting in a crash.
305 This issue was reported to OpenSSL by Robert Święcki of Google.
309 *) Bad (EC)DHE parameters cause a client crash
311 If a malicious server supplies bad parameters for a DHE or ECDHE key
312 exchange then this can result in the client attempting to dereference a
313 NULL pointer leading to a client crash. This could be exploited in a Denial
316 This issue was reported to OpenSSL by Guido Vranken.
320 *) BN_mod_exp may produce incorrect results on x86_64
322 There is a carry propagating bug in the x86_64 Montgomery squaring
323 procedure. No EC algorithms are affected. Analysis suggests that attacks
324 against RSA and DSA as a result of this defect would be very difficult to
325 perform and are not believed likely. Attacks against DH are considered just
326 feasible (although very difficult) because most of the work necessary to
327 deduce information about a private key may be performed offline. The amount
328 of resources required for such an attack would be very significant and
329 likely only accessible to a limited number of attackers. An attacker would
330 additionally need online access to an unpatched system using the target
331 private key in a scenario with persistent DH parameters and a private
332 key that is shared between multiple clients. For example this can occur by
333 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
334 similar to CVE-2015-3193 but must be treated as a separate problem.
336 This issue was reported to OpenSSL by the OSS-Fuzz project.
340 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
342 *) ChaCha20/Poly1305 heap-buffer-overflow
344 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
345 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
346 crash. This issue is not considered to be exploitable beyond a DoS.
348 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
352 *) CMS Null dereference
354 Applications parsing invalid CMS structures can crash with a NULL pointer
355 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
356 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
357 structure callback if an attempt is made to free certain invalid encodings.
358 Only CHOICE structures using a callback which do not handle NULL value are
361 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
365 *) Montgomery multiplication may produce incorrect results
367 There is a carry propagating bug in the Broadwell-specific Montgomery
368 multiplication procedure that handles input lengths divisible by, but
369 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
370 and DH private keys are impossible. This is because the subroutine in
371 question is not used in operations with the private key itself and an input
372 of the attacker's direct choice. Otherwise the bug can manifest itself as
373 transient authentication and key negotiation failures or reproducible
374 erroneous outcome of public-key operations with specially crafted input.
375 Among EC algorithms only Brainpool P-512 curves are affected and one
376 presumably can attack ECDH key negotiation. Impact was not analyzed in
377 detail, because pre-requisites for attack are considered unlikely. Namely
378 multiple clients have to choose the curve in question and the server has to
379 share the private key among them, neither of which is default behaviour.
380 Even then only clients that chose the curve will be affected.
382 This issue was publicly reported as transient failures and was not
383 initially recognized as a security issue. Thanks to Richard Morgan for
384 providing reproducible case.
388 *) Removed automatic addition of RPATH in shared libraries and executables,
389 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
392 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
394 *) Fix Use After Free for large message sizes
396 The patch applied to address CVE-2016-6307 resulted in an issue where if a
397 message larger than approx 16k is received then the underlying buffer to
398 store the incoming message is reallocated and moved. Unfortunately a
399 dangling pointer to the old location is left which results in an attempt to
400 write to the previously freed location. This is likely to result in a
401 crash, however it could potentially lead to execution of arbitrary code.
403 This issue only affects OpenSSL 1.1.0a.
405 This issue was reported to OpenSSL by Robert Święcki.
409 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
411 *) OCSP Status Request extension unbounded memory growth
413 A malicious client can send an excessively large OCSP Status Request
414 extension. If that client continually requests renegotiation, sending a
415 large OCSP Status Request extension each time, then there will be unbounded
416 memory growth on the server. This will eventually lead to a Denial Of
417 Service attack through memory exhaustion. Servers with a default
418 configuration are vulnerable even if they do not support OCSP. Builds using
419 the "no-ocsp" build time option are not affected.
421 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
425 *) SSL_peek() hang on empty record
427 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
428 sends an empty record. This could be exploited by a malicious peer in a
429 Denial Of Service attack.
431 This issue was reported to OpenSSL by Alex Gaynor.
435 *) Excessive allocation of memory in tls_get_message_header() and
436 dtls1_preprocess_fragment()
438 A (D)TLS message includes 3 bytes for its length in the header for the
439 message. This would allow for messages up to 16Mb in length. Messages of
440 this length are excessive and OpenSSL includes a check to ensure that a
441 peer is sending reasonably sized messages in order to avoid too much memory
442 being consumed to service a connection. A flaw in the logic of version
443 1.1.0 means that memory for the message is allocated too early, prior to
444 the excessive message length check. Due to way memory is allocated in
445 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
446 to service a connection. This could lead to a Denial of Service through
447 memory exhaustion. However, the excessive message length check still takes
448 place, and this would cause the connection to immediately fail. Assuming
449 that the application calls SSL_free() on the failed connection in a timely
450 manner then the 21Mb of allocated memory will then be immediately freed
451 again. Therefore the excessive memory allocation will be transitory in
452 nature. This then means that there is only a security impact if:
454 1) The application does not call SSL_free() in a timely manner in the event
455 that the connection fails
457 2) The application is working in a constrained environment where there is
458 very little free memory
460 3) The attacker initiates multiple connection attempts such that there are
461 multiple connections in a state where memory has been allocated for the
462 connection; SSL_free() has not yet been called; and there is insufficient
463 memory to service the multiple requests.
465 Except in the instance of (1) above any Denial Of Service is likely to be
466 transitory because as soon as the connection fails the memory is
467 subsequently freed again in the SSL_free() call. However there is an
468 increased risk during this period of application crashes due to the lack of
469 memory - which would then mean a more serious Denial of Service.
471 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
472 (CVE-2016-6307 and CVE-2016-6308)
475 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
476 had to be removed. Primary reason is that vendor assembler can't
477 assemble our modules with -KPIC flag. As result it, assembly
478 support, was not even available as option. But its lack means
479 lack of side-channel resistant code, which is incompatible with
480 security by todays standards. Fortunately gcc is readily available
481 prepackaged option, which we firmly point at...
484 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
486 *) Windows command-line tool supports UTF-8 opt-in option for arguments
487 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
488 (to any value) allows Windows user to access PKCS#12 file generated
489 with Windows CryptoAPI and protected with non-ASCII password, as well
490 as files generated under UTF-8 locale on Linux also protected with
494 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
495 have been disabled by default and removed from DEFAULT, just like RC4.
496 See the RC4 item below to re-enable both.
499 *) The method for finding the storage location for the Windows RAND seed file
500 has changed. First we check %RANDFILE%. If that is not set then we check
501 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
502 all else fails we fall back to C:\.
505 *) The EVP_EncryptUpdate() function has had its return type changed from void
506 to int. A return of 0 indicates and error while a return of 1 indicates
510 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
511 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
512 off the constant time implementation for RSA, DSA and DH have been made
513 no-ops and deprecated.
516 *) Windows RAND implementation was simplified to only get entropy by
517 calling CryptGenRandom(). Various other RAND-related tickets
519 [Joseph Wylie Yandle, Rich Salz]
521 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
522 and OPENSSL_LH_, respectively. The old names are available
523 with API compatibility. They new names are now completely documented.
526 *) Unify TYPE_up_ref(obj) methods signature.
527 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
528 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
529 int (instead of void) like all others TYPE_up_ref() methods.
530 So now these methods also check the return value of CRYPTO_atomic_add(),
531 and the validity of object reference counter.
532 [fdasilvayy@gmail.com]
534 *) With Windows Visual Studio builds, the .pdb files are installed
535 alongside the installed libraries and executables. For a static
536 library installation, ossl_static.pdb is the associate compiler
537 generated .pdb file to be used when linking programs.
540 *) Remove openssl.spec. Packaging files belong with the packagers.
543 *) Automatic Darwin/OSX configuration has had a refresh, it will now
544 recognise x86_64 architectures automatically. You can still decide
545 to build for a different bitness with the environment variable
546 KERNEL_BITS (can be 32 or 64), for example:
548 KERNEL_BITS=32 ./config
552 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
553 256 bit AES and HMAC with SHA256.
556 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
559 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
562 *) To enable users to have their own config files and build file templates,
563 Configure looks in the directory indicated by the environment variable
564 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
565 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
566 name and is used as is.
569 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
570 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
571 X509_CERT_FILE_CTX was removed.
574 *) "shared" builds are now the default. To create only static libraries use
575 the "no-shared" Configure option.
578 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
579 All of these option have not worked for some while and are fundamental
583 *) Make various cleanup routines no-ops and mark them as deprecated. Most
584 global cleanup functions are no longer required because they are handled
585 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
586 Explicitly de-initing can cause problems (e.g. where a library that uses
587 OpenSSL de-inits, but an application is still using it). The affected
588 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
589 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
590 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
594 *) --strict-warnings no longer enables runtime debugging options
595 such as REF_DEBUG. Instead, debug options are automatically
596 enabled with '--debug' builds.
597 [Andy Polyakov, Emilia Käsper]
599 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
600 have been moved out of the public header files. New functions for managing
601 these have been added.
604 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
605 objects have been moved out of the public header files. New
606 functions for managing these have been added.
609 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
610 have been moved out of the public header files. New functions for managing
611 these have been added.
614 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
615 moved out of the public header files. New functions for managing these
619 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
622 *) Removed the mk1mf build scripts.
625 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
626 it is always safe to #include a header now.
629 *) Removed the aged BC-32 config and all its supporting scripts
632 *) Removed support for Ultrix, Netware, and OS/2.
635 *) Add support for HKDF.
638 *) Add support for blake2b and blake2s
641 *) Added support for "pipelining". Ciphers that have the
642 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
643 encryptions/decryptions simultaneously. There are currently no built-in
644 ciphers with this property but the expectation is that engines will be able
645 to offer it to significantly improve throughput. Support has been extended
646 into libssl so that multiple records for a single connection can be
647 processed in one go (for >=TLS 1.1).
650 *) Added the AFALG engine. This is an async capable engine which is able to
651 offload work to the Linux kernel. In this initial version it only supports
652 AES128-CBC. The kernel must be version 4.1.0 or greater.
655 *) OpenSSL now uses a new threading API. It is no longer necessary to
656 set locking callbacks to use OpenSSL in a multi-threaded environment. There
657 are two supported threading models: pthreads and windows threads. It is
658 also possible to configure OpenSSL at compile time for "no-threads". The
659 old threading API should no longer be used. The functions have been
660 replaced with "no-op" compatibility macros.
661 [Alessandro Ghedini, Matt Caswell]
663 *) Modify behavior of ALPN to invoke callback after SNI/servername
664 callback, such that updates to the SSL_CTX affect ALPN.
667 *) Add SSL_CIPHER queries for authentication and key-exchange.
670 *) Changes to the DEFAULT cipherlist:
671 - Prefer (EC)DHE handshakes over plain RSA.
672 - Prefer AEAD ciphers over legacy ciphers.
673 - Prefer ECDSA over RSA when both certificates are available.
674 - Prefer TLSv1.2 ciphers/PRF.
675 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
679 *) Change the ECC default curve list to be this, in order: x25519,
680 secp256r1, secp521r1, secp384r1.
683 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
684 disabled by default. They can be re-enabled using the
685 enable-weak-ssl-ciphers option to Configure.
688 *) If the server has ALPN configured, but supports no protocols that the
689 client advertises, send a fatal "no_application_protocol" alert.
690 This behaviour is SHALL in RFC 7301, though it isn't universally
691 implemented by other servers.
694 *) Add X25519 support.
695 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
696 for public and private key encoding using the format documented in
697 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
698 key generation and key derivation.
700 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
704 *) Deprecate SRP_VBASE_get_by_user.
705 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
706 In order to fix an unavoidable memory leak (CVE-2016-0798),
707 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
708 seed, even if the seed is configured.
710 Users should use SRP_VBASE_get1_by_user instead. Note that in
711 SRP_VBASE_get1_by_user, caller must free the returned value. Note
712 also that even though configuring the SRP seed attempts to hide
713 invalid usernames by continuing the handshake with fake
714 credentials, this behaviour is not constant time and no strong
715 guarantees are made that the handshake is indistinguishable from
716 that of a valid user.
719 *) Configuration change; it's now possible to build dynamic engines
720 without having to build shared libraries and vice versa. This
721 only applies to the engines in engines/, those in crypto/engine/
722 will always be built into libcrypto (i.e. "static").
724 Building dynamic engines is enabled by default; to disable, use
725 the configuration option "disable-dynamic-engine".
727 The only requirements for building dynamic engines are the
728 presence of the DSO module and building with position independent
729 code, so they will also automatically be disabled if configuring
730 with "disable-dso" or "disable-pic".
732 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
733 are also taken away from openssl/opensslconf.h, as they are
737 *) Configuration change; if there is a known flag to compile
738 position independent code, it will always be applied on the
739 libcrypto and libssl object files, and never on the application
740 object files. This means other libraries that use routines from
741 libcrypto / libssl can be made into shared libraries regardless
742 of how OpenSSL was configured.
744 If this isn't desirable, the configuration options "disable-pic"
745 or "no-pic" can be used to disable the use of PIC. This will
746 also disable building shared libraries and dynamic engines.
749 *) Removed JPAKE code. It was experimental and has no wide use.
752 *) The INSTALL_PREFIX Makefile variable has been renamed to
753 DESTDIR. That makes for less confusion on what this variable
754 is for. Also, the configuration option --install_prefix is
758 *) Heartbeat for TLS has been removed and is disabled by default
759 for DTLS; configure with enable-heartbeats. Code that uses the
760 old #define's might need to be updated.
761 [Emilia Käsper, Rich Salz]
763 *) Rename REF_CHECK to REF_DEBUG.
766 *) New "unified" build system
768 The "unified" build system is aimed to be a common system for all
769 platforms we support. With it comes new support for VMS.
771 This system builds supports building in a different directory tree
772 than the source tree. It produces one Makefile (for unix family
773 or lookalikes), or one descrip.mms (for VMS).
775 The source of information to make the Makefile / descrip.mms is
776 small files called 'build.info', holding the necessary
777 information for each directory with source to compile, and a
778 template in Configurations, like unix-Makefile.tmpl or
781 With this change, the library names were also renamed on Windows
782 and on VMS. They now have names that are closer to the standard
783 on Unix, and include the major version number, and in certain
784 cases, the architecture they are built for. See "Notes on shared
785 libraries" in INSTALL.
787 We rely heavily on the perl module Text::Template.
790 *) Added support for auto-initialisation and de-initialisation of the library.
791 OpenSSL no longer requires explicit init or deinit routines to be called,
792 except in certain circumstances. See the OPENSSL_init_crypto() and
793 OPENSSL_init_ssl() man pages for further information.
796 *) The arguments to the DTLSv1_listen function have changed. Specifically the
797 "peer" argument is now expected to be a BIO_ADDR object.
799 *) Rewrite of BIO networking library. The BIO library lacked consistent
800 support of IPv6, and adding it required some more extensive
801 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
802 which hold all types of addresses and chains of address information.
803 It also introduces a new API, with functions like BIO_socket,
804 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
805 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
806 have been adapted accordingly.
809 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
813 *) CRIME protection: disable compression by default, even if OpenSSL is
814 compiled with zlib enabled. Applications can still enable compression
815 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
816 using the SSL_CONF library to configure compression.
819 *) The signature of the session callback configured with
820 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
821 was explicitly marked as 'const unsigned char*' instead of
825 *) Always DPURIFY. Remove the use of uninitialized memory in the
826 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
829 *) Removed many obsolete configuration items, including
830 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
831 MD2_CHAR, MD2_INT, MD2_LONG
833 IDEA_SHORT, IDEA_LONG
834 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
835 [Rich Salz, with advice from Andy Polyakov]
837 *) Many BN internals have been moved to an internal header file.
838 [Rich Salz with help from Andy Polyakov]
840 *) Configuration and writing out the results from it has changed.
841 Files such as Makefile include/openssl/opensslconf.h and are now
842 produced through general templates, such as Makefile.in and
843 crypto/opensslconf.h.in and some help from the perl module
846 Also, the center of configuration information is no longer
847 Makefile. Instead, Configure produces a perl module in
848 configdata.pm which holds most of the config data (in the hash
849 table %config), the target data that comes from the target
850 configuration in one of the Configurations/*.conf files (in
854 *) To clarify their intended purposes, the Configure options
855 --prefix and --openssldir change their semantics, and become more
856 straightforward and less interdependent.
858 --prefix shall be used exclusively to give the location INSTALLTOP
859 where programs, scripts, libraries, include files and manuals are
860 going to be installed. The default is now /usr/local.
862 --openssldir shall be used exclusively to give the default
863 location OPENSSLDIR where certificates, private keys, CRLs are
864 managed. This is also where the default openssl.cnf gets
866 If the directory given with this option is a relative path, the
867 values of both the --prefix value and the --openssldir value will
868 be combined to become OPENSSLDIR.
869 The default for --openssldir is INSTALLTOP/ssl.
871 Anyone who uses --openssldir to specify where OpenSSL is to be
872 installed MUST change to use --prefix instead.
875 *) The GOST engine was out of date and therefore it has been removed. An up
876 to date GOST engine is now being maintained in an external repository.
877 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
878 support for GOST ciphersuites (these are only activated if a GOST engine
882 *) EGD is no longer supported by default; use enable-egd when
884 [Ben Kaduk and Rich Salz]
886 *) The distribution now has Makefile.in files, which are used to
887 create Makefile's when Configure is run. *Configure must be run
888 before trying to build now.*
891 *) The return value for SSL_CIPHER_description() for error conditions
895 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
897 Obtaining and performing DNSSEC validation of TLSA records is
898 the application's responsibility. The application provides
899 the TLSA records of its choice to OpenSSL, and these are then
900 used to authenticate the peer.
902 The TLSA records need not even come from DNS. They can, for
903 example, be used to implement local end-entity certificate or
904 trust-anchor "pinning", where the "pin" data takes the form
905 of TLSA records, which can augment or replace verification
906 based on the usual WebPKI public certification authorities.
909 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
910 continues to support deprecated interfaces in default builds.
911 However, applications are strongly advised to compile their
912 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
913 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
914 or the 1.1.0 releases.
916 In environments in which all applications have been ported to
917 not use any deprecated interfaces OpenSSL's Configure script
918 should be used with the --api=1.1.0 option to entirely remove
919 support for the deprecated features from the library and
920 unconditionally disable them in the installed headers.
921 Essentially the same effect can be achieved with the "no-deprecated"
922 argument to Configure, except that this will always restrict
923 the build to just the latest API, rather than a fixed API
926 As applications are ported to future revisions of the API,
927 they should update their compile-time OPENSSL_API_COMPAT define
928 accordingly, but in most cases should be able to continue to
929 compile with later releases.
931 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
932 0x10000000L and 0x00908000L, respectively. However those
933 versions did not support the OPENSSL_API_COMPAT feature, and
934 so applications are not typically tested for explicit support
935 of just the undeprecated features of either release.
938 *) Add support for setting the minimum and maximum supported protocol.
939 It can bet set via the SSL_set_min_proto_version() and
940 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
941 MaxProtocol. It's recommended to use the new APIs to disable
942 protocols instead of disabling individual protocols using
943 SSL_set_options() or SSL_CONF's Protocol. This change also
944 removes support for disabling TLS 1.2 in the OpenSSL TLS
945 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
948 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
951 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
952 and integrates ECDSA and ECDH functionality into EC. Implementations can
953 now redirect key generation and no longer need to convert to or from
956 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
957 include the ec.h header file instead.
960 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
961 ciphers who are no longer supported and drops support the ephemeral RSA key
962 exchange. The LOW ciphers currently doesn't have any ciphers in it.
965 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
966 opaque. For HMAC_CTX, the following constructors and destructors
969 HMAC_CTX *HMAC_CTX_new(void);
970 void HMAC_CTX_free(HMAC_CTX *ctx);
972 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
973 destroy such methods has been added. See EVP_MD_meth_new(3) and
974 EVP_CIPHER_meth_new(3) for documentation.
977 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
978 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
979 EVP_MD_CTX_reset() should be called instead to reinitialise
980 an already created structure.
981 2) For consistency with the majority of our object creators and
982 destructors, EVP_MD_CTX_(create|destroy) were renamed to
983 EVP_MD_CTX_(new|free). The old names are retained as macros
984 for deprecated builds.
987 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
988 cryptographic operations to be performed asynchronously as long as an
989 asynchronous capable engine is used. See the ASYNC_start_job() man page for
990 further details. Libssl has also had this capability integrated with the
991 introduction of the new mode SSL_MODE_ASYNC and associated error
992 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
993 pages. This work was developed in partnership with Intel Corp.
996 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
997 always enabled now. If you want to disable the support you should
998 exclude it using the list of supported ciphers. This also means that the
999 "-no_ecdhe" option has been removed from s_server.
1002 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1003 SSL_{CTX_}set1_curves() which can set a list.
1006 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1007 curve you want to support using SSL_{CTX_}set1_curves().
1010 *) State machine rewrite. The state machine code has been significantly
1011 refactored in order to remove much duplication of code and solve issues
1012 with the old code (see ssl/statem/README for further details). This change
1013 does have some associated API changes. Notably the SSL_state() function
1014 has been removed and replaced by SSL_get_state which now returns an
1015 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1016 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1020 *) All instances of the string "ssleay" in the public API were replaced
1021 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1022 Some error codes related to internal RSA_eay API's were renamed.
1025 *) The demo files in crypto/threads were moved to demo/threads.
1028 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1030 [Matt Caswell, Rich Salz]
1032 *) New ASN.1 embed macro.
1034 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1035 structure is not allocated: it is part of the parent. That is instead of
1043 This reduces memory fragmentation and make it impossible to accidentally
1044 set a mandatory field to NULL.
1046 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1047 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1048 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1052 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1055 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1056 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1057 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1058 DES and RC4 ciphersuites.
1061 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1062 This changes the decoding behaviour for some invalid messages,
1063 though the change is mostly in the more lenient direction, and
1064 legacy behaviour is preserved as much as possible.
1067 *) Fix no-stdio build.
1068 [ David Woodhouse <David.Woodhouse@intel.com> and also
1069 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1071 *) New testing framework
1072 The testing framework has been largely rewritten and is now using
1073 perl and the perl modules Test::Harness and an extended variant of
1074 Test::More called OpenSSL::Test to do its work. All test scripts in
1075 test/ have been rewritten into test recipes, and all direct calls to
1076 executables in test/Makefile have become individual recipes using the
1077 simplified testing OpenSSL::Test::Simple.
1079 For documentation on our testing modules, do:
1081 perldoc test/testlib/OpenSSL/Test/Simple.pm
1082 perldoc test/testlib/OpenSSL/Test.pm
1086 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1087 are used; the latter aborts on memory leaks (usually checked on exit).
1088 Some undocumented "set malloc, etc., hooks" functions were removed
1089 and others were changed. All are now documented.
1092 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1094 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1096 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1097 from RFC4279, RFC4785, RFC5487, RFC5489.
1099 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1100 original RSA_PSK patch.
1103 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1104 era flag was never set throughout the codebase (only read). Also removed
1105 SSL3_FLAGS_POP_BUFFER which was only used if
1106 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1109 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1110 to be "oneline" instead of "compat".
1113 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1114 not aware of clients that still exhibit this bug, and the workaround
1115 hasn't been working properly for a while.
1118 *) The return type of BIO_number_read() and BIO_number_written() as well as
1119 the corresponding num_read and num_write members in the BIO structure has
1120 changed from unsigned long to uint64_t. On platforms where an unsigned
1121 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1125 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1126 OpenSSL without support for them. It also means that maintaining
1127 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1128 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1131 *) Removed support for the two export grade static DH ciphersuites
1132 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1133 were newly added (along with a number of other static DH ciphersuites) to
1134 1.0.2. However the two export ones have *never* worked since they were
1135 introduced. It seems strange in any case to be adding new export
1136 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1139 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1140 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1141 and turned into macros which simply call the new preferred function names
1142 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1143 should use the new names instead. Also as part of this change the ssl23.h
1144 header file has been removed.
1147 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1148 code and the associated standard is no longer considered fit-for-purpose.
1151 *) RT2547 was closed. When generating a private key, try to make the
1152 output file readable only by the owner. This behavior change might
1153 be noticeable when interacting with other software.
1155 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1159 *) Added HTTP GET support to the ocsp command.
1162 *) Changed default digest for the dgst and enc commands from MD5 to
1166 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1169 *) Added support for TLS extended master secret from
1170 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1171 initial patch which was a great help during development.
1174 *) All libssl internal structures have been removed from the public header
1175 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1176 now redundant). Users should not attempt to access internal structures
1177 directly. Instead they should use the provided API functions.
1180 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1181 Access to deprecated functions can be re-enabled by running config with
1182 "enable-deprecated". In addition applications wishing to use deprecated
1183 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1184 will, by default, disable some transitive includes that previously existed
1185 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1188 *) Added support for OCB mode. OpenSSL has been granted a patent license
1189 compatible with the OpenSSL license for use of OCB. Details are available
1190 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1191 for OCB can be removed by calling config with no-ocb.
1194 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1195 compatible client hello.
1198 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1199 done while fixing the error code for the key-too-small case.
1200 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1202 *) CA.sh has been removed; use CA.pl instead.
1205 *) Removed old DES API.
1208 *) Remove various unsupported platforms:
1214 Sinix/ReliantUNIX RM400
1219 16-bit platforms such as WIN16
1222 *) Clean up OPENSSL_NO_xxx #define's
1223 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1224 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1225 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1226 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1227 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1228 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1229 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1230 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1231 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1232 Remove MS_STATIC; it's a relic from platforms <32 bits.
1235 *) Cleaned up dead code
1236 Remove all but one '#ifdef undef' which is to be looked at.
1239 *) Clean up calling of xxx_free routines.
1240 Just like free(), fix most of the xxx_free routines to accept
1241 NULL. Remove the non-null checks from callers. Save much code.
1244 *) Add secure heap for storage of private keys (when possible).
1245 Add BIO_s_secmem(), CBIGNUM, etc.
1246 Contributed by Akamai Technologies under our Corporate CLA.
1249 *) Experimental support for a new, fast, unbiased prime candidate generator,
1250 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1251 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1253 *) New output format NSS in the sess_id command line tool. This allows
1254 exporting the session id and the master key in NSS keylog format.
1255 [Martin Kaiser <martin@kaiser.cx>]
1257 *) Harmonize version and its documentation. -f flag is used to display
1259 [mancha <mancha1@zoho.com>]
1261 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1262 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1263 [mancha <mancha1@zoho.com>]
1265 *) Fix some double frees. These are not thought to be exploitable.
1266 [mancha <mancha1@zoho.com>]
1268 *) A missing bounds check in the handling of the TLS heartbeat extension
1269 can be used to reveal up to 64k of memory to a connected client or
1272 Thanks for Neel Mehta of Google Security for discovering this bug and to
1273 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1274 preparing the fix (CVE-2014-0160)
1275 [Adam Langley, Bodo Moeller]
1277 *) Fix for the attack described in the paper "Recovering OpenSSL
1278 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1279 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1280 http://eprint.iacr.org/2014/140
1282 Thanks to Yuval Yarom and Naomi Benger for discovering this
1283 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1284 [Yuval Yarom and Naomi Benger]
1286 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1287 this fixes a limitation in previous versions of OpenSSL.
1290 *) Experimental encrypt-then-mac support.
1292 Experimental support for encrypt then mac from
1293 draft-gutmann-tls-encrypt-then-mac-02.txt
1295 To enable it set the appropriate extension number (0x42 for the test
1296 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1298 For non-compliant peers (i.e. just about everything) this should have no
1301 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1305 *) Add EVP support for key wrapping algorithms, to avoid problems with
1306 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1307 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1308 algorithms and include tests cases.
1311 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1315 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1316 MGF1 digest and OAEP label.
1319 *) Make openssl verify return errors.
1320 [Chris Palmer <palmer@google.com> and Ben Laurie]
1322 *) New function ASN1_TIME_diff to calculate the difference between two
1323 ASN1_TIME structures or one structure and the current time.
1326 *) Update fips_test_suite to support multiple command line options. New
1327 test to induce all self test errors in sequence and check expected
1331 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1332 sign or verify all in one operation.
1335 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1336 test programs and fips_test_suite. Includes functionality to parse
1337 the minimal script output of fipsalgest.pl directly.
1340 *) Add authorisation parameter to FIPS_module_mode_set().
1343 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1346 *) Use separate DRBG fields for internal and external flags. New function
1347 FIPS_drbg_health_check() to perform on demand health checking. Add
1348 generation tests to fips_test_suite with reduced health check interval to
1349 demonstrate periodic health checking. Add "nodh" option to
1350 fips_test_suite to skip very slow DH test.
1353 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1357 *) More extensive health check for DRBG checking many more failure modes.
1358 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1359 combination: call this in fips_test_suite.
1362 *) Add support for canonical generation of DSA parameter 'g'. See
1365 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1366 POST to handle HMAC cases.
1369 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1370 to return numerical and string versions of the FIPS module number.
1373 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1374 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1375 outside the validated module in the FIPS capable OpenSSL.
1378 *) Minor change to DRBG entropy callback semantics. In some cases
1379 there is no multiple of the block length between min_len and
1380 max_len. Allow the callback to return more than max_len bytes
1381 of entropy but discard any extra: it is the callback's responsibility
1382 to ensure that the extra data discarded does not impact the
1383 requested amount of entropy.
1386 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1387 information in FIPS186-3, SP800-57 and SP800-131A.
1390 *) CCM support via EVP. Interface is very similar to GCM case except we
1391 must supply all data in one chunk (i.e. no update, final) and the
1392 message length must be supplied if AAD is used. Add algorithm test
1396 *) Initial version of POST overhaul. Add POST callback to allow the status
1397 of POST to be monitored and/or failures induced. Modify fips_test_suite
1398 to use callback. Always run all selftests even if one fails.
1401 *) XTS support including algorithm test driver in the fips_gcmtest program.
1402 Note: this does increase the maximum key length from 32 to 64 bytes but
1403 there should be no binary compatibility issues as existing applications
1404 will never use XTS mode.
1407 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1408 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1409 performs algorithm blocking for unapproved PRNG types. Also do not
1410 set PRNG type in FIPS_mode_set(): leave this to the application.
1411 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1412 the standard OpenSSL PRNG: set additional data to a date time vector.
1415 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1416 This shouldn't present any incompatibility problems because applications
1417 shouldn't be using these directly and any that are will need to rethink
1418 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1421 *) Extensive self tests and health checking required by SP800-90 DRBG.
1422 Remove strength parameter from FIPS_drbg_instantiate and always
1423 instantiate at maximum supported strength.
1426 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1429 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1432 *) New function DH_compute_key_padded() to compute a DH key and pad with
1433 leading zeroes if needed: this complies with SP800-56A et al.
1436 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1437 anything, incomplete, subject to change and largely untested at present.
1440 *) Modify fipscanisteronly build option to only build the necessary object
1441 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1444 *) Add experimental option FIPSSYMS to give all symbols in
1445 fipscanister.o and FIPS or fips prefix. This will avoid
1446 conflicts with future versions of OpenSSL. Add perl script
1447 util/fipsas.pl to preprocess assembly language source files
1448 and rename any affected symbols.
1451 *) Add selftest checks and algorithm block of non-fips algorithms in
1452 FIPS mode. Remove DES2 from selftests.
1455 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1456 return internal method without any ENGINE dependencies. Add new
1457 tiny fips sign and verify functions.
1460 *) New build option no-ec2m to disable characteristic 2 code.
1463 *) New build option "fipscanisteronly". This only builds fipscanister.o
1464 and (currently) associated fips utilities. Uses the file Makefile.fips
1465 instead of Makefile.org as the prototype.
1468 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1469 Update fips_gcmtest to use IV generator.
1472 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1473 setting output buffer to NULL. The *Final function must be
1474 called although it will not retrieve any additional data. The tag
1475 can be set or retrieved with a ctrl. The IV length is by default 12
1476 bytes (96 bits) but can be set to an alternative value. If the IV
1477 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1481 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1482 underlying do_cipher function handles all cipher semantics itself
1483 including padding and finalisation. This is useful if (for example)
1484 an ENGINE cipher handles block padding itself. The behaviour of
1485 do_cipher is subtly changed if this flag is set: the return value
1486 is the number of characters written to the output buffer (zero is
1487 no longer an error code) or a negative error code. Also if the
1488 input buffer is NULL and length 0 finalisation should be performed.
1491 *) If a candidate issuer certificate is already part of the constructed
1492 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1495 *) Improve forward-security support: add functions
1497 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1498 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1500 for use by SSL/TLS servers; the callback function will be called whenever a
1501 new session is created, and gets to decide whether the session may be
1502 cached to make it resumable (return 0) or not (return 1). (As by the
1503 SSL/TLS protocol specifications, the session_id sent by the server will be
1504 empty to indicate that the session is not resumable; also, the server will
1505 not generate RFC 4507 (RFC 5077) session tickets.)
1507 A simple reasonable callback implementation is to return is_forward_secure.
1508 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1509 by the SSL/TLS server library, indicating whether it can provide forward
1511 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1513 *) New -verify_name option in command line utilities to set verification
1517 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1518 Add CMAC pkey methods.
1521 *) Experimental renegotiation in s_server -www mode. If the client
1522 browses /reneg connection is renegotiated. If /renegcert it is
1523 renegotiated requesting a certificate.
1526 *) Add an "external" session cache for debugging purposes to s_server. This
1527 should help trace issues which normally are only apparent in deployed
1528 multi-process servers.
1531 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1532 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1533 BIO_set_cipher() and some obscure PEM functions were changed so they
1534 can now return an error. The RAND changes required a change to the
1535 RAND_METHOD structure.
1538 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1539 a gcc attribute to warn if the result of a function is ignored. This
1540 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1541 whose return value is often ignored.
1544 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1545 These allow SCTs (signed certificate timestamps) to be requested and
1546 validated when establishing a connection.
1547 [Rob Percival <robpercival@google.com>]
1549 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1551 *) Prevent padding oracle in AES-NI CBC MAC check
1553 A MITM attacker can use a padding oracle attack to decrypt traffic
1554 when the connection uses an AES CBC cipher and the server support
1557 This issue was introduced as part of the fix for Lucky 13 padding
1558 attack (CVE-2013-0169). The padding check was rewritten to be in
1559 constant time by making sure that always the same bytes are read and
1560 compared against either the MAC or padding bytes. But it no longer
1561 checked that there was enough data to have both the MAC and padding
1564 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1568 *) Fix EVP_EncodeUpdate overflow
1570 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1571 Base64 encoding of binary data. If an attacker is able to supply very large
1572 amounts of input data then a length check can overflow resulting in a heap
1575 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1576 the PEM_write_bio* family of functions. These are mainly used within the
1577 OpenSSL command line applications, so any application which processes data
1578 from an untrusted source and outputs it as a PEM file should be considered
1579 vulnerable to this issue. User applications that call these APIs directly
1580 with large amounts of untrusted data may also be vulnerable.
1582 This issue was reported by Guido Vranken.
1586 *) Fix EVP_EncryptUpdate overflow
1588 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1589 is able to supply very large amounts of input data after a previous call to
1590 EVP_EncryptUpdate() with a partial block then a length check can overflow
1591 resulting in a heap corruption. Following an analysis of all OpenSSL
1592 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1593 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1594 the first called function after an EVP_EncryptInit(), and therefore that
1595 specific call must be safe. The second form is where the length passed to
1596 EVP_EncryptUpdate() can be seen from the code to be some small value and
1597 therefore there is no possibility of an overflow. Since all instances are
1598 one of these two forms, it is believed that there can be no overflows in
1599 internal code due to this problem. It should be noted that
1600 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1601 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1602 of these calls have also been analysed too and it is believed there are no
1603 instances in internal usage where an overflow could occur.
1605 This issue was reported by Guido Vranken.
1609 *) Prevent ASN.1 BIO excessive memory allocation
1611 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1612 a short invalid encoding can cause allocation of large amounts of memory
1613 potentially consuming excessive resources or exhausting memory.
1615 Any application parsing untrusted data through d2i BIO functions is
1616 affected. The memory based functions such as d2i_X509() are *not* affected.
1617 Since the memory based functions are used by the TLS library, TLS
1618 applications are not affected.
1620 This issue was reported by Brian Carpenter.
1626 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1627 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1628 in arbitrary stack data being returned in the buffer.
1630 This issue was reported by Guido Vranken.
1634 *) Modify behavior of ALPN to invoke callback after SNI/servername
1635 callback, such that updates to the SSL_CTX affect ALPN.
1638 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1642 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1643 methods are enabled and ssl2 is disabled the methods return NULL.
1646 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1648 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1649 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1650 provide any "EXPORT" or "LOW" strength ciphers.
1653 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1654 is by default disabled at build-time. Builds that are not configured with
1655 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1656 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1657 will need to explicitly call either of:
1659 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1661 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1663 as appropriate. Even if either of those is used, or the application
1664 explicitly uses the version-specific SSLv2_method() or its client and
1665 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1666 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1667 ciphers, and SSLv2 56-bit DES are no longer available.
1671 *) Fix a double-free in DSA code
1673 A double free bug was discovered when OpenSSL parses malformed DSA private
1674 keys and could lead to a DoS attack or memory corruption for applications
1675 that receive DSA private keys from untrusted sources. This scenario is
1678 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1683 *) Disable SRP fake user seed to address a server memory leak.
1685 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1687 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1688 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1689 was changed to ignore the "fake user" SRP seed, even if the seed
1692 Users should use SRP_VBASE_get1_by_user instead. Note that in
1693 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1694 also that even though configuring the SRP seed attempts to hide
1695 invalid usernames by continuing the handshake with fake
1696 credentials, this behaviour is not constant time and no strong
1697 guarantees are made that the handshake is indistinguishable from
1698 that of a valid user.
1702 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1704 In the BN_hex2bn function the number of hex digits is calculated using an
1705 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1706 large values of |i| this can result in |bn_expand| not allocating any
1707 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1708 field as NULL leading to a subsequent NULL ptr deref. For very large values
1709 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1710 In this case memory is allocated to the internal BIGNUM data field, but it
1711 is insufficiently sized leading to heap corruption. A similar issue exists
1712 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1713 is ever called by user applications with very large untrusted hex/dec data.
1714 This is anticipated to be a rare occurrence.
1716 All OpenSSL internal usage of these functions use data that is not expected
1717 to be untrusted, e.g. config file data or application command line
1718 arguments. If user developed applications generate config file data based
1719 on untrusted data then it is possible that this could also lead to security
1720 consequences. This is also anticipated to be rare.
1722 This issue was reported to OpenSSL by Guido Vranken.
1726 *) Fix memory issues in BIO_*printf functions
1728 The internal |fmtstr| function used in processing a "%s" format string in
1729 the BIO_*printf functions could overflow while calculating the length of a
1730 string and cause an OOB read when printing very long strings.
1732 Additionally the internal |doapr_outch| function can attempt to write to an
1733 OOB memory location (at an offset from the NULL pointer) in the event of a
1734 memory allocation failure. In 1.0.2 and below this could be caused where
1735 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1736 could be in processing a very long "%s" format string. Memory leaks can
1739 The first issue may mask the second issue dependent on compiler behaviour.
1740 These problems could enable attacks where large amounts of untrusted data
1741 is passed to the BIO_*printf functions. If applications use these functions
1742 in this way then they could be vulnerable. OpenSSL itself uses these
1743 functions when printing out human-readable dumps of ASN.1 data. Therefore
1744 applications that print this data could be vulnerable if the data is from
1745 untrusted sources. OpenSSL command line applications could also be
1746 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1747 as command line arguments.
1749 Libssl is not considered directly vulnerable. Additionally certificates etc
1750 received via remote connections via libssl are also unlikely to be able to
1751 trigger these issues because of message size limits enforced within libssl.
1753 This issue was reported to OpenSSL Guido Vranken.
1757 *) Side channel attack on modular exponentiation
1759 A side-channel attack was found which makes use of cache-bank conflicts on
1760 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1761 of RSA keys. The ability to exploit this issue is limited as it relies on
1762 an attacker who has control of code in a thread running on the same
1763 hyper-threaded core as the victim thread which is performing decryptions.
1765 This issue was reported to OpenSSL by Yuval Yarom, The University of
1766 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1767 Nadia Heninger, University of Pennsylvania with more information at
1768 http://cachebleed.info.
1772 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1773 if no keysize is specified with default_bits. This fixes an
1774 omission in an earlier change that changed all RSA/DSA key generation
1775 apps to use 2048 bits by default.
1778 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1779 *) DH small subgroups
1781 Historically OpenSSL only ever generated DH parameters based on "safe"
1782 primes. More recently (in version 1.0.2) support was provided for
1783 generating X9.42 style parameter files such as those required for RFC 5114
1784 support. The primes used in such files may not be "safe". Where an
1785 application is using DH configured with parameters based on primes that are
1786 not "safe" then an attacker could use this fact to find a peer's private
1787 DH exponent. This attack requires that the attacker complete multiple
1788 handshakes in which the peer uses the same private DH exponent. For example
1789 this could be used to discover a TLS server's private DH exponent if it's
1790 reusing the private DH exponent or it's using a static DH ciphersuite.
1792 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1793 TLS. It is not on by default. If the option is not set then the server
1794 reuses the same private DH exponent for the life of the server process and
1795 would be vulnerable to this attack. It is believed that many popular
1796 applications do set this option and would therefore not be at risk.
1798 The fix for this issue adds an additional check where a "q" parameter is
1799 available (as is the case in X9.42 based parameters). This detects the
1800 only known attack, and is the only possible defense for static DH
1801 ciphersuites. This could have some performance impact.
1803 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1804 default and cannot be disabled. This could have some performance impact.
1806 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1810 *) SSLv2 doesn't block disabled ciphers
1812 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1813 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1814 been disabled, provided that the SSLv2 protocol was not also disabled via
1817 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1818 and Sebastian Schinzel.
1822 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1824 *) BN_mod_exp may produce incorrect results on x86_64
1826 There is a carry propagating bug in the x86_64 Montgomery squaring
1827 procedure. No EC algorithms are affected. Analysis suggests that attacks
1828 against RSA and DSA as a result of this defect would be very difficult to
1829 perform and are not believed likely. Attacks against DH are considered just
1830 feasible (although very difficult) because most of the work necessary to
1831 deduce information about a private key may be performed offline. The amount
1832 of resources required for such an attack would be very significant and
1833 likely only accessible to a limited number of attackers. An attacker would
1834 additionally need online access to an unpatched system using the target
1835 private key in a scenario with persistent DH parameters and a private
1836 key that is shared between multiple clients. For example this can occur by
1837 default in OpenSSL DHE based SSL/TLS ciphersuites.
1839 This issue was reported to OpenSSL by Hanno Böck.
1843 *) Certificate verify crash with missing PSS parameter
1845 The signature verification routines will crash with a NULL pointer
1846 dereference if presented with an ASN.1 signature using the RSA PSS
1847 algorithm and absent mask generation function parameter. Since these
1848 routines are used to verify certificate signature algorithms this can be
1849 used to crash any certificate verification operation and exploited in a
1850 DoS attack. Any application which performs certificate verification is
1851 vulnerable including OpenSSL clients and servers which enable client
1854 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1858 *) X509_ATTRIBUTE memory leak
1860 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1861 memory. This structure is used by the PKCS#7 and CMS routines so any
1862 application which reads PKCS#7 or CMS data from untrusted sources is
1863 affected. SSL/TLS is not affected.
1865 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1870 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1871 This changes the decoding behaviour for some invalid messages,
1872 though the change is mostly in the more lenient direction, and
1873 legacy behaviour is preserved as much as possible.
1876 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1878 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1880 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1882 *) Alternate chains certificate forgery
1884 During certificate verification, OpenSSL will attempt to find an
1885 alternative certificate chain if the first attempt to build such a chain
1886 fails. An error in the implementation of this logic can mean that an
1887 attacker could cause certain checks on untrusted certificates to be
1888 bypassed, such as the CA flag, enabling them to use a valid leaf
1889 certificate to act as a CA and "issue" an invalid certificate.
1891 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1895 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1897 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1898 incompatibility in the handling of HMAC. The previous ABI has now been
1902 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1904 *) Malformed ECParameters causes infinite loop
1906 When processing an ECParameters structure OpenSSL enters an infinite loop
1907 if the curve specified is over a specially malformed binary polynomial
1910 This can be used to perform denial of service against any
1911 system which processes public keys, certificate requests or
1912 certificates. This includes TLS clients and TLS servers with
1913 client authentication enabled.
1915 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1919 *) Exploitable out-of-bounds read in X509_cmp_time
1921 X509_cmp_time does not properly check the length of the ASN1_TIME
1922 string and can read a few bytes out of bounds. In addition,
1923 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1926 An attacker can use this to craft malformed certificates and CRLs of
1927 various sizes and potentially cause a segmentation fault, resulting in
1928 a DoS on applications that verify certificates or CRLs. TLS clients
1929 that verify CRLs are affected. TLS clients and servers with client
1930 authentication enabled may be affected if they use custom verification
1933 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1934 independently by Hanno Böck.
1938 *) PKCS7 crash with missing EnvelopedContent
1940 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1941 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1942 with missing content and trigger a NULL pointer dereference on parsing.
1944 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1945 structures from untrusted sources are affected. OpenSSL clients and
1946 servers are not affected.
1948 This issue was reported to OpenSSL by Michal Zalewski (Google).
1952 *) CMS verify infinite loop with unknown hash function
1954 When verifying a signedData message the CMS code can enter an infinite loop
1955 if presented with an unknown hash function OID. This can be used to perform
1956 denial of service against any system which verifies signedData messages using
1958 This issue was reported to OpenSSL by Johannes Bauer.
1962 *) Race condition handling NewSessionTicket
1964 If a NewSessionTicket is received by a multi-threaded client when attempting to
1965 reuse a previous ticket then a race condition can occur potentially leading to
1966 a double free of the ticket data.
1970 *) Only support 256-bit or stronger elliptic curves with the
1971 'ecdh_auto' setting (server) or by default (client). Of supported
1972 curves, prefer P-256 (both).
1975 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1977 *) ClientHello sigalgs DoS fix
1979 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1980 invalid signature algorithms extension a NULL pointer dereference will
1981 occur. This can be exploited in a DoS attack against the server.
1983 This issue was was reported to OpenSSL by David Ramos of Stanford
1986 [Stephen Henson and Matt Caswell]
1988 *) Multiblock corrupted pointer fix
1990 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1991 feature only applies on 64 bit x86 architecture platforms that support AES
1992 NI instructions. A defect in the implementation of "multiblock" can cause
1993 OpenSSL's internal write buffer to become incorrectly set to NULL when
1994 using non-blocking IO. Typically, when the user application is using a
1995 socket BIO for writing, this will only result in a failed connection.
1996 However if some other BIO is used then it is likely that a segmentation
1997 fault will be triggered, thus enabling a potential DoS attack.
1999 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2003 *) Segmentation fault in DTLSv1_listen fix
2005 The DTLSv1_listen function is intended to be stateless and processes the
2006 initial ClientHello from many peers. It is common for user code to loop
2007 over the call to DTLSv1_listen until a valid ClientHello is received with
2008 an associated cookie. A defect in the implementation of DTLSv1_listen means
2009 that state is preserved in the SSL object from one invocation to the next
2010 that can lead to a segmentation fault. Errors processing the initial
2011 ClientHello can trigger this scenario. An example of such an error could be
2012 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2015 This issue was reported to OpenSSL by Per Allansson.
2019 *) Segmentation fault in ASN1_TYPE_cmp fix
2021 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2022 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2023 certificate signature algorithm consistency this can be used to crash any
2024 certificate verification operation and exploited in a DoS attack. Any
2025 application which performs certificate verification is vulnerable including
2026 OpenSSL clients and servers which enable client authentication.
2030 *) Segmentation fault for invalid PSS parameters fix
2032 The signature verification routines will crash with a NULL pointer
2033 dereference if presented with an ASN.1 signature using the RSA PSS
2034 algorithm and invalid parameters. Since these routines are used to verify
2035 certificate signature algorithms this can be used to crash any
2036 certificate verification operation and exploited in a DoS attack. Any
2037 application which performs certificate verification is vulnerable including
2038 OpenSSL clients and servers which enable client authentication.
2040 This issue was was reported to OpenSSL by Brian Carpenter.
2044 *) ASN.1 structure reuse memory corruption fix
2046 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2047 memory corruption via an invalid write. Such reuse is and has been
2048 strongly discouraged and is believed to be rare.
2050 Applications that parse structures containing CHOICE or ANY DEFINED BY
2051 components may be affected. Certificate parsing (d2i_X509 and related
2052 functions) are however not affected. OpenSSL clients and servers are
2057 *) PKCS7 NULL pointer dereferences fix
2059 The PKCS#7 parsing code does not handle missing outer ContentInfo
2060 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2061 missing content and trigger a NULL pointer dereference on parsing.
2063 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2064 otherwise parse PKCS#7 structures from untrusted sources are
2065 affected. OpenSSL clients and servers are not affected.
2067 This issue was reported to OpenSSL by Michal Zalewski (Google).
2071 *) DoS via reachable assert in SSLv2 servers fix
2073 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2074 servers that both support SSLv2 and enable export cipher suites by sending
2075 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2077 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2078 (OpenSSL development team).
2082 *) Empty CKE with client auth and DHE fix
2084 If client auth is used then a server can seg fault in the event of a DHE
2085 ciphersuite being selected and a zero length ClientKeyExchange message
2086 being sent by the client. This could be exploited in a DoS attack.
2090 *) Handshake with unseeded PRNG fix
2092 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2093 with an unseeded PRNG. The conditions are:
2094 - The client is on a platform where the PRNG has not been seeded
2095 automatically, and the user has not seeded manually
2096 - A protocol specific client method version has been used (i.e. not
2097 SSL_client_methodv23)
2098 - A ciphersuite is used that does not require additional random data from
2099 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2101 If the handshake succeeds then the client random that has been used will
2102 have been generated from a PRNG with insufficient entropy and therefore the
2103 output may be predictable.
2105 For example using the following command with an unseeded openssl will
2106 succeed on an unpatched platform:
2108 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2112 *) Use After Free following d2i_ECPrivatekey error fix
2114 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2115 could cause a use after free condition. This, in turn, could cause a double
2116 free in several private key parsing functions (such as d2i_PrivateKey
2117 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2118 for applications that receive EC private keys from untrusted
2119 sources. This scenario is considered rare.
2121 This issue was discovered by the BoringSSL project and fixed in their
2126 *) X509_to_X509_REQ NULL pointer deref fix
2128 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2129 the certificate key is invalid. This function is rarely used in practice.
2131 This issue was discovered by Brian Carpenter.
2135 *) Removed the export ciphers from the DEFAULT ciphers
2138 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2140 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2141 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2142 So far those who have to target multiple platforms would compromise
2143 and argue that binary targeting say ARMv5 would still execute on
2144 ARMv8. "Universal" build resolves this compromise by providing
2145 near-optimal performance even on newer platforms.
2148 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2149 (other platforms pending).
2150 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2152 *) Add support for the SignedCertificateTimestampList certificate and
2153 OCSP response extensions from RFC6962.
2156 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2157 for corner cases. (Certain input points at infinity could lead to
2158 bogus results, with non-infinity inputs mapped to infinity too.)
2161 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2162 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2163 common cases are optimized and there still is room for further
2164 improvements. Vector Permutation AES for Altivec is also added.
2167 *) Add support for little-endian ppc64 Linux target.
2168 [Marcelo Cerri (IBM)]
2170 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2171 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2172 are optimized and there still is room for further improvements.
2173 Both 32- and 64-bit modes are supported.
2174 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2176 *) Improved ARMv7 NEON support.
2179 *) Support for SPARC Architecture 2011 crypto extensions, first
2180 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2181 SHA256/512, MD5, GHASH and modular exponentiation.
2182 [Andy Polyakov, David Miller]
2184 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2186 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2188 *) Support for new and upcoming Intel processors, including AVX2,
2189 BMI and SHA ISA extensions. This includes additional "stitched"
2190 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2193 This work was sponsored by Intel Corp.
2196 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2197 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2198 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2201 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2202 this fixes a limitation in previous versions of OpenSSL.
2205 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2206 MGF1 digest and OAEP label.
2209 *) Add EVP support for key wrapping algorithms, to avoid problems with
2210 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2211 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2212 algorithms and include tests cases.
2215 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2217 [Douglas E. Engert, Steve Henson]
2219 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2220 difference in days and seconds between two tm or ASN1_TIME structures.
2223 *) Add -rev test option to s_server to just reverse order of characters
2224 received by client and send back to server. Also prints an abbreviated
2225 summary of the connection parameters.
2228 *) New option -brief for s_client and s_server to print out a brief summary
2229 of connection parameters.
2232 *) Add callbacks for arbitrary TLS extensions.
2233 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2235 *) New option -crl_download in several openssl utilities to download CRLs
2236 from CRLDP extension in certificates.
2239 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2242 *) New function X509_CRL_diff to generate a delta CRL from the difference
2243 of two full CRLs. Add support to "crl" utility.
2246 *) New functions to set lookup_crls function and to retrieve
2247 X509_STORE from X509_STORE_CTX.
2250 *) Print out deprecated issuer and subject unique ID fields in
2254 *) Extend OCSP I/O functions so they can be used for simple general purpose
2255 HTTP as well as OCSP. New wrapper function which can be used to download
2256 CRLs using the OCSP API.
2259 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2262 *) SSL_CONF* functions. These provide a common framework for application
2263 configuration using configuration files or command lines.
2266 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2267 message callback and prints the results. Needs compile time option
2268 "enable-ssl-trace". New options to s_client and s_server to enable
2272 *) New ctrl and macro to retrieve supported points extensions.
2273 Print out extension in s_server and s_client.
2276 *) New functions to retrieve certificate signature and signature
2280 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2284 *) New Suite B modes for TLS code. These use and enforce the requirements
2285 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2286 only use Suite B curves. The Suite B modes can be set by using the
2287 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2290 *) New chain verification flags for Suite B levels of security. Check
2291 algorithms are acceptable when flags are set in X509_verify_cert.
2294 *) Make tls1_check_chain return a set of flags indicating checks passed
2295 by a certificate chain. Add additional tests to handle client
2296 certificates: checks for matching certificate type and issuer name
2300 *) If an attempt is made to use a signature algorithm not in the peer
2301 preference list abort the handshake. If client has no suitable
2302 signature algorithms in response to a certificate request do not
2303 use the certificate.
2306 *) If server EC tmp key is not in client preference list abort handshake.
2309 *) Add support for certificate stores in CERT structure. This makes it
2310 possible to have different stores per SSL structure or one store in
2311 the parent SSL_CTX. Include distinct stores for certificate chain
2312 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2313 to build and store a certificate chain in CERT structure: returning
2314 an error if the chain cannot be built: this will allow applications
2315 to test if a chain is correctly configured.
2317 Note: if the CERT based stores are not set then the parent SSL_CTX
2318 store is used to retain compatibility with existing behaviour.
2322 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2323 mask based on the current session, check mask when sending client
2324 hello and checking the requested ciphersuite.
2327 *) New ctrls to retrieve and set certificate types in a certificate
2328 request message. Print out received values in s_client. If certificate
2329 types is not set with custom values set sensible values based on
2330 supported signature algorithms.
2333 *) Support for distinct client and server supported signature algorithms.
2336 *) Add certificate callback. If set this is called whenever a certificate
2337 is required by client or server. An application can decide which
2338 certificate chain to present based on arbitrary criteria: for example
2339 supported signature algorithms. Add very simple example to s_server.
2340 This fixes many of the problems and restrictions of the existing client
2341 certificate callback: for example you can now clear an existing
2342 certificate and specify the whole chain.
2345 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2346 the certificate can be used for (if anything). Set valid_flags field
2347 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2348 to have similar checks in it.
2350 Add new "cert_flags" field to CERT structure and include a "strict mode".
2351 This enforces some TLS certificate requirements (such as only permitting
2352 certificate signature algorithms contained in the supported algorithms
2353 extension) which some implementations ignore: this option should be used
2354 with caution as it could cause interoperability issues.
2357 *) Update and tidy signature algorithm extension processing. Work out
2358 shared signature algorithms based on preferences and peer algorithms
2359 and print them out in s_client and s_server. Abort handshake if no
2360 shared signature algorithms.
2363 *) Add new functions to allow customised supported signature algorithms
2364 for SSL and SSL_CTX structures. Add options to s_client and s_server
2368 *) New function SSL_certs_clear() to delete all references to certificates
2369 from an SSL structure. Before this once a certificate had been added
2370 it couldn't be removed.
2373 *) Integrate hostname, email address and IP address checking with certificate
2374 verification. New verify options supporting checking in openssl utility.
2377 *) Fixes and wildcard matching support to hostname and email checking
2378 functions. Add manual page.
2379 [Florian Weimer (Red Hat Product Security Team)]
2381 *) New functions to check a hostname email or IP address against a
2382 certificate. Add options x509 utility to print results of checks against
2386 *) Fix OCSP checking.
2387 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2389 *) Initial experimental support for explicitly trusted non-root CAs.
2390 OpenSSL still tries to build a complete chain to a root but if an
2391 intermediate CA has a trust setting included that is used. The first
2392 setting is used: whether to trust (e.g., -addtrust option to the x509
2396 *) Add -trusted_first option which attempts to find certificates in the
2397 trusted store even if an untrusted chain is also supplied.
2400 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2401 platform support for Linux and Android.
2404 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2407 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2408 When in FIPS mode the approved implementations are used as normal,
2409 when not in FIPS mode the internal unapproved versions are used instead.
2410 This means that the FIPS capable OpenSSL isn't forced to use the
2411 (often lower performance) FIPS implementations outside FIPS mode.
2414 *) Transparently support X9.42 DH parameters when calling
2415 PEM_read_bio_DHparameters. This means existing applications can handle
2416 the new parameter format automatically.
2419 *) Initial experimental support for X9.42 DH parameter format: mainly
2420 to support use of 'q' parameter for RFC5114 parameters.
2423 *) Add DH parameters from RFC5114 including test data to dhtest.
2426 *) Support for automatic EC temporary key parameter selection. If enabled
2427 the most preferred EC parameters are automatically used instead of
2428 hardcoded fixed parameters. Now a server just has to call:
2429 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2430 support ECDH and use the most appropriate parameters.
2433 *) Enhance and tidy EC curve and point format TLS extension code. Use
2434 static structures instead of allocation if default values are used.
2435 New ctrls to set curves we wish to support and to retrieve shared curves.
2436 Print out shared curves in s_server. New options to s_server and s_client
2437 to set list of supported curves.
2440 *) New ctrls to retrieve supported signature algorithms and
2441 supported curve values as an array of NIDs. Extend openssl utility
2442 to print out received values.
2445 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2446 between NIDs and the more common NIST names such as "P-256". Enhance
2447 ecparam utility and ECC method to recognise the NIST names for curves.
2450 *) Enhance SSL/TLS certificate chain handling to support different
2451 chains for each certificate instead of one chain in the parent SSL_CTX.
2454 *) Support for fixed DH ciphersuite client authentication: where both
2455 server and client use DH certificates with common parameters.
2458 *) Support for fixed DH ciphersuites: those requiring DH server
2462 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2464 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2465 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2466 X509_CINF_get_signature were reverted post internal team review.
2468 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2470 *) Build fixes for the Windows and OpenVMS platforms
2471 [Matt Caswell and Richard Levitte]
2473 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2475 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2476 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2477 dereference. This could lead to a Denial Of Service attack. Thanks to
2478 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2482 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2483 dtls1_buffer_record function under certain conditions. In particular this
2484 could occur if an attacker sent repeated DTLS records with the same
2485 sequence number but for the next epoch. The memory leak could be exploited
2486 by an attacker in a Denial of Service attack through memory exhaustion.
2487 Thanks to Chris Mueller for reporting this issue.
2491 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2492 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2493 method would be set to NULL which could later result in a NULL pointer
2494 dereference. Thanks to Frank Schmirler for reporting this issue.
2498 *) Abort handshake if server key exchange message is omitted for ephemeral
2501 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2502 reporting this issue.
2506 *) Remove non-export ephemeral RSA code on client and server. This code
2507 violated the TLS standard by allowing the use of temporary RSA keys in
2508 non-export ciphersuites and could be used by a server to effectively
2509 downgrade the RSA key length used to a value smaller than the server
2510 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2511 INRIA or reporting this issue.
2515 *) Fixed issue where DH client certificates are accepted without verification.
2516 An OpenSSL server will accept a DH certificate for client authentication
2517 without the certificate verify message. This effectively allows a client to
2518 authenticate without the use of a private key. This only affects servers
2519 which trust a client certificate authority which issues certificates
2520 containing DH keys: these are extremely rare and hardly ever encountered.
2521 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2526 *) Ensure that the session ID context of an SSL is updated when its
2527 SSL_CTX is updated via SSL_set_SSL_CTX.
2529 The session ID context is typically set from the parent SSL_CTX,
2530 and can vary with the CTX.
2533 *) Fix various certificate fingerprint issues.
2535 By using non-DER or invalid encodings outside the signed portion of a
2536 certificate the fingerprint can be changed without breaking the signature.
2537 Although no details of the signed portion of the certificate can be changed
2538 this can cause problems with some applications: e.g. those using the
2539 certificate fingerprint for blacklists.
2541 1. Reject signatures with non zero unused bits.
2543 If the BIT STRING containing the signature has non zero unused bits reject
2544 the signature. All current signature algorithms require zero unused bits.
2546 2. Check certificate algorithm consistency.
2548 Check the AlgorithmIdentifier inside TBS matches the one in the
2549 certificate signature. NB: this will result in signature failure
2550 errors for some broken certificates.
2552 Thanks to Konrad Kraszewski from Google for reporting this issue.
2554 3. Check DSA/ECDSA signatures use DER.
2556 Re-encode DSA/ECDSA signatures and compare with the original received
2557 signature. Return an error if there is a mismatch.
2559 This will reject various cases including garbage after signature
2560 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2561 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2562 (negative or with leading zeroes).
2564 Further analysis was conducted and fixes were developed by Stephen Henson
2565 of the OpenSSL core team.
2570 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2571 results on some platforms, including x86_64. This bug occurs at random
2572 with a very low probability, and is not known to be exploitable in any
2573 way, though its exact impact is difficult to determine. Thanks to Pieter
2574 Wuille (Blockstream) who reported this issue and also suggested an initial
2575 fix. Further analysis was conducted by the OpenSSL development team and
2576 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2577 the OpenSSL core team.
2581 *) Do not resume sessions on the server if the negotiated protocol
2582 version does not match the session's version. Resuming with a different
2583 version, while not strictly forbidden by the RFC, is of questionable
2584 sanity and breaks all known clients.
2585 [David Benjamin, Emilia Käsper]
2587 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2588 early CCS messages during renegotiation. (Note that because
2589 renegotiation is encrypted, this early CCS was not exploitable.)
2592 *) Tighten client-side session ticket handling during renegotiation:
2593 ensure that the client only accepts a session ticket if the server sends
2594 the extension anew in the ServerHello. Previously, a TLS client would
2595 reuse the old extension state and thus accept a session ticket if one was
2596 announced in the initial ServerHello.
2598 Similarly, ensure that the client requires a session ticket if one
2599 was advertised in the ServerHello. Previously, a TLS client would
2600 ignore a missing NewSessionTicket message.
2603 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2605 *) SRTP Memory Leak.
2607 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2608 sends a carefully crafted handshake message, to cause OpenSSL to fail
2609 to free up to 64k of memory causing a memory leak. This could be
2610 exploited in a Denial Of Service attack. This issue affects OpenSSL
2611 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2612 whether SRTP is used or configured. Implementations of OpenSSL that
2613 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2615 The fix was developed by the OpenSSL team.
2619 *) Session Ticket Memory Leak.
2621 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2622 integrity of that ticket is first verified. In the event of a session
2623 ticket integrity check failing, OpenSSL will fail to free memory
2624 causing a memory leak. By sending a large number of invalid session
2625 tickets an attacker could exploit this issue in a Denial Of Service
2630 *) Build option no-ssl3 is incomplete.
2632 When OpenSSL is configured with "no-ssl3" as a build option, servers
2633 could accept and complete a SSL 3.0 handshake, and clients could be
2634 configured to send them.
2636 [Akamai and the OpenSSL team]
2638 *) Add support for TLS_FALLBACK_SCSV.
2639 Client applications doing fallback retries should call
2640 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2642 [Adam Langley, Bodo Moeller]
2644 *) Add additional DigestInfo checks.
2646 Re-encode DigestInto in DER and check against the original when
2647 verifying RSA signature: this will reject any improperly encoded
2648 DigestInfo structures.
2650 Note: this is a precautionary measure and no attacks are currently known.
2654 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2656 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2657 SRP code can be overrun an internal buffer. Add sanity check that
2658 g, A, B < N to SRP code.
2660 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2661 Group for discovering this issue.
2665 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2666 TLS 1.0 instead of higher protocol versions when the ClientHello message
2667 is badly fragmented. This allows a man-in-the-middle attacker to force a
2668 downgrade to TLS 1.0 even if both the server and the client support a
2669 higher protocol version, by modifying the client's TLS records.
2671 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2672 researching this issue.
2676 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2677 to a denial of service attack. A malicious server can crash the client
2678 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2679 ciphersuite and sending carefully crafted handshake messages.
2681 Thanks to Felix Gröbert (Google) for discovering and researching this
2686 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2687 to leak memory. This can be exploited through a Denial of Service attack.
2688 Thanks to Adam Langley for discovering and researching this issue.
2692 *) An attacker can force openssl to consume large amounts of memory whilst
2693 processing DTLS handshake messages. This can be exploited through a
2694 Denial of Service attack.
2695 Thanks to Adam Langley for discovering and researching this issue.
2699 *) An attacker can force an error condition which causes openssl to crash
2700 whilst processing DTLS packets due to memory being freed twice. This
2701 can be exploited through a Denial of Service attack.
2702 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2707 *) If a multithreaded client connects to a malicious server using a resumed
2708 session and the server sends an ec point format extension it could write
2709 up to 255 bytes to freed memory.
2711 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2716 *) A malicious server can crash an OpenSSL client with a null pointer
2717 dereference (read) by specifying an SRP ciphersuite even though it was not
2718 properly negotiated with the client. This can be exploited through a
2719 Denial of Service attack.
2721 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2722 discovering and researching this issue.
2726 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2727 X509_name_oneline, X509_name_print_ex et al. to leak some information
2728 from the stack. Applications may be affected if they echo pretty printing
2729 output to the attacker.
2731 Thanks to Ivan Fratric (Google) for discovering this issue.
2733 [Emilia Käsper, and Steve Henson]
2735 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2736 for corner cases. (Certain input points at infinity could lead to
2737 bogus results, with non-infinity inputs mapped to infinity too.)
2740 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2742 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2743 handshake can force the use of weak keying material in OpenSSL
2744 SSL/TLS clients and servers.
2746 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2747 researching this issue. (CVE-2014-0224)
2748 [KIKUCHI Masashi, Steve Henson]
2750 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2751 OpenSSL DTLS client the code can be made to recurse eventually crashing
2754 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2756 [Imre Rad, Steve Henson]
2758 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2759 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2760 client or server. This is potentially exploitable to run arbitrary
2761 code on a vulnerable client or server.
2763 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2764 [Jüri Aedla, Steve Henson]
2766 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2767 are subject to a denial of service attack.
2769 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2770 this issue. (CVE-2014-3470)
2771 [Felix Gröbert, Ivan Fratric, Steve Henson]
2773 *) Harmonize version and its documentation. -f flag is used to display
2775 [mancha <mancha1@zoho.com>]
2777 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2778 in i2d_ECPrivateKey.
2779 [mancha <mancha1@zoho.com>]
2781 *) Fix some double frees. These are not thought to be exploitable.
2782 [mancha <mancha1@zoho.com>]
2784 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2786 *) A missing bounds check in the handling of the TLS heartbeat extension
2787 can be used to reveal up to 64k of memory to a connected client or
2790 Thanks for Neel Mehta of Google Security for discovering this bug and to
2791 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2792 preparing the fix (CVE-2014-0160)
2793 [Adam Langley, Bodo Moeller]
2795 *) Fix for the attack described in the paper "Recovering OpenSSL
2796 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2797 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2798 http://eprint.iacr.org/2014/140
2800 Thanks to Yuval Yarom and Naomi Benger for discovering this
2801 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2802 [Yuval Yarom and Naomi Benger]
2804 *) TLS pad extension: draft-agl-tls-padding-03
2806 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2807 TLS client Hello record length value would otherwise be > 255 and
2808 less that 512 pad with a dummy extension containing zeroes so it
2809 is at least 512 bytes long.
2811 [Adam Langley, Steve Henson]
2813 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2815 *) Fix for TLS record tampering bug. A carefully crafted invalid
2816 handshake could crash OpenSSL with a NULL pointer exception.
2817 Thanks to Anton Johansson for reporting this issues.
2820 *) Keep original DTLS digest and encryption contexts in retransmission
2821 structures so we can use the previous session parameters if they need
2822 to be resent. (CVE-2013-6450)
2825 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2826 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2827 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2828 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2829 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2830 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2831 [Rob Stradling, Adam Langley]
2833 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2835 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2836 supporting platforms or when small records were transferred.
2837 [Andy Polyakov, Steve Henson]
2839 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2841 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2843 This addresses the flaw in CBC record processing discovered by
2844 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2845 at: http://www.isg.rhul.ac.uk/tls/
2847 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2848 Security Group at Royal Holloway, University of London
2849 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2850 Emilia Käsper for the initial patch.
2852 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2854 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2855 ciphersuites which can be exploited in a denial of service attack.
2856 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2857 and detecting this bug and to Wolfgang Ettlinger
2858 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2862 *) Return an error when checking OCSP signatures when key is NULL.
2863 This fixes a DoS attack. (CVE-2013-0166)
2866 *) Make openssl verify return errors.
2867 [Chris Palmer <palmer@google.com> and Ben Laurie]
2869 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2870 the right response is stapled. Also change SSL_get_certificate()
2871 so it returns the certificate actually sent.
2872 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2873 [Rob Stradling <rob.stradling@comodo.com>]
2875 *) Fix possible deadlock when decoding public keys.
2878 *) Don't use TLS 1.0 record version number in initial client hello
2882 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2884 *) Sanity check record length before skipping explicit IV in TLS
2885 1.2, 1.1 and DTLS to fix DoS attack.
2887 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2888 fuzzing as a service testing platform.
2892 *) Initialise tkeylen properly when encrypting CMS messages.
2893 Thanks to Solar Designer of Openwall for reporting this issue.
2896 *) In FIPS mode don't try to use composite ciphers as they are not
2900 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2902 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2903 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2904 mean any application compiled against OpenSSL 1.0.0 headers setting
2905 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
2906 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2907 0x10000000L Any application which was previously compiled against
2908 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2909 will need to be recompiled as a result. Letting be results in
2910 inability to disable specifically TLS 1.1 and in client context,
2911 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2914 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
2915 disable just protocol X, but all protocols above X *if* there are
2916 protocols *below* X still enabled. In more practical terms it means
2917 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2918 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2919 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2923 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2925 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2926 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2927 in CRYPTO_realloc_clean.
2929 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2930 issue and to Adam Langley <agl@chromium.org> for fixing it.
2932 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2934 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2937 *) Workarounds for some broken servers that "hang" if a client hello
2938 record length exceeds 255 bytes.
2940 1. Do not use record version number > TLS 1.0 in initial client
2941 hello: some (but not all) hanging servers will now work.
2942 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2943 the number of ciphers sent in the client hello. This should be
2944 set to an even number, such as 50, for example by passing:
2945 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2946 Most broken servers should now work.
2947 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2948 TLS 1.2 client support entirely.
2951 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2954 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2956 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2957 STRING form instead of a DigestInfo.
2960 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2961 and the RSA_sign/RSA_verify functions. This was made more apparent when
2962 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2963 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2964 the correct format in RSA_verify so both forms transparently work.
2967 *) Some servers which support TLS 1.0 can choke if we initially indicate
2968 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2969 encrypted premaster secret. As a workaround use the maximum permitted
2970 client version in client hello, this should keep such servers happy
2971 and still work with previous versions of OpenSSL.
2974 *) Add support for TLS/DTLS heartbeats.
2975 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2977 *) Add support for SCTP.
2978 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2980 *) Improved PRNG seeding for VOS.
2981 [Paul Green <Paul.Green@stratus.com>]
2983 *) Extensive assembler packs updates, most notably:
2985 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2986 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2987 - x86_64: bit-sliced AES implementation;
2988 - ARM: NEON support, contemporary platforms optimizations;
2989 - s390x: z196 support;
2990 - *: GHASH and GF(2^m) multiplication implementations;
2994 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2995 (removal of unnecessary code)
2996 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2998 *) Add TLS key material exporter from RFC 5705.
3001 *) Add DTLS-SRTP negotiation from RFC 5764.
3004 *) Add Next Protocol Negotiation,
3005 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3006 disabled with a no-npn flag to config or Configure. Code donated
3008 [Adam Langley <agl@google.com> and Ben Laurie]
3010 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3011 NIST-P256, NIST-P521, with constant-time single point multiplication on
3012 typical inputs. Compiler support for the nonstandard type __uint128_t is
3013 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3014 Code made available under Apache License version 2.0.
3016 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3017 line to include this in your build of OpenSSL, and run "make depend" (or
3018 "make update"). This enables the following EC_METHODs:
3020 EC_GFp_nistp224_method()
3021 EC_GFp_nistp256_method()
3022 EC_GFp_nistp521_method()
3024 EC_GROUP_new_by_curve_name() will automatically use these (while
3025 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3027 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3029 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3030 all platforms. Move ssize_t definition from e_os.h to the public
3031 header file e_os2.h as it now appears in public header file cms.h
3034 *) New -sigopt option to the ca, req and x509 utilities. Additional
3035 signature parameters can be passed using this option and in
3039 *) Add RSA PSS signing function. This will generate and set the
3040 appropriate AlgorithmIdentifiers for PSS based on those in the
3041 corresponding EVP_MD_CTX structure. No application support yet.
3044 *) Support for companion algorithm specific ASN1 signing routines.
3045 New function ASN1_item_sign_ctx() signs a pre-initialised
3046 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3047 the appropriate parameters.
3050 *) Add new algorithm specific ASN1 verification initialisation function
3051 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3052 handling will be the same no matter what EVP_PKEY_METHOD is used.
3053 Add a PSS handler to support verification of PSS signatures: checked
3054 against a number of sample certificates.
3057 *) Add signature printing for PSS. Add PSS OIDs.
3058 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3060 *) Add algorithm specific signature printing. An individual ASN1 method
3061 can now print out signatures instead of the standard hex dump.
3063 More complex signatures (e.g. PSS) can print out more meaningful
3064 information. Include DSA version that prints out the signature
3068 *) Password based recipient info support for CMS library: implementing
3072 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3073 neatly separates the code into cipher and PBE sections and is required
3074 for some algorithms that split PBES2 into separate pieces (such as
3075 password based CMS).
3078 *) Session-handling fixes:
3079 - Fix handling of connections that are resuming with a session ID,
3080 but also support Session Tickets.
3081 - Fix a bug that suppressed issuing of a new ticket if the client
3082 presented a ticket with an expired session.
3083 - Try to set the ticket lifetime hint to something reasonable.
3084 - Make tickets shorter by excluding irrelevant information.
3085 - On the client side, don't ignore renewed tickets.
3086 [Adam Langley, Bodo Moeller (Google)]
3088 *) Fix PSK session representation.
3091 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3093 This work was sponsored by Intel.
3096 *) Add GCM support to TLS library. Some custom code is needed to split
3097 the IV between the fixed (from PRF) and explicit (from TLS record)
3098 portions. This adds all GCM ciphersuites supported by RFC5288 and
3099 RFC5289. Generalise some AES* cipherstrings to include GCM and
3100 add a special AESGCM string for GCM only.
3103 *) Expand range of ctrls for AES GCM. Permit setting invocation
3104 field on decrypt and retrieval of invocation field only on encrypt.
3107 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3108 As required by RFC5289 these ciphersuites cannot be used if for
3109 versions of TLS earlier than 1.2.
3112 *) For FIPS capable OpenSSL interpret a NULL default public key method
3113 as unset and return the appropriate default but do *not* set the default.
3114 This means we can return the appropriate method in applications that
3115 switch between FIPS and non-FIPS modes.
3118 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3119 ENGINE is used then we cannot handle that in the FIPS module so we
3120 keep original code iff non-FIPS operations are allowed.
3123 *) Add -attime option to openssl utilities.
3124 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3126 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3129 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3130 FIPS EC methods unconditionally for now.
3133 *) New build option no-ec2m to disable characteristic 2 code.
3136 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3137 all cases can be covered as some introduce binary incompatibilities.
3140 *) Redirect RSA operations to FIPS module including keygen,
3141 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3144 *) Add similar low level API blocking to ciphers.
3147 *) Low level digest APIs are not approved in FIPS mode: any attempt
3148 to use these will cause a fatal error. Applications that *really* want
3149 to use them can use the private_* version instead.
3152 *) Redirect cipher operations to FIPS module for FIPS builds.
3155 *) Redirect digest operations to FIPS module for FIPS builds.
3158 *) Update build system to add "fips" flag which will link in fipscanister.o
3159 for static and shared library builds embedding a signature if needed.
3162 *) Output TLS supported curves in preference order instead of numerical
3163 order. This is currently hardcoded for the highest order curves first.
3164 This should be configurable so applications can judge speed vs strength.
3167 *) Add TLS v1.2 server support for client authentication.
3170 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3174 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3175 FIPS modules versions.
3178 *) Add TLS v1.2 client side support for client authentication. Keep cache
3179 of handshake records longer as we don't know the hash algorithm to use
3180 until after the certificate request message is received.
3183 *) Initial TLS v1.2 client support. Add a default signature algorithms
3184 extension including all the algorithms we support. Parse new signature
3185 format in client key exchange. Relax some ECC signing restrictions for
3186 TLS v1.2 as indicated in RFC5246.
3189 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3190 to new signature format when needed using client digest preference.
3191 All server ciphersuites should now work correctly in TLS v1.2. No client
3192 support yet and no support for client certificates.
3195 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3196 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3197 ciphersuites. At present only RSA key exchange ciphersuites work with
3198 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3199 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3200 and version checking.
3203 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3204 with this defined it will not be affected by any changes to ssl internal
3205 structures. Add several utility functions to allow openssl application
3206 to work with OPENSSL_NO_SSL_INTERN defined.
3210 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3212 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3215 *) Permit abbreviated handshakes when renegotiating using the function
3216 SSL_renegotiate_abbreviated().
3217 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3219 *) Add call to ENGINE_register_all_complete() to
3220 ENGINE_load_builtin_engines(), so some implementations get used
3221 automatically instead of needing explicit application support.
3224 *) Add support for TLS key exporter as described in RFC5705.
3225 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3227 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3228 a few changes are required:
3230 Add SSL_OP_NO_TLSv1_1 flag.
3231 Add TLSv1_1 methods.
3232 Update version checking logic to handle version 1.1.
3233 Add explicit IV handling (ported from DTLS code).
3234 Add command line options to s_client/s_server.
3237 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3239 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3240 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3241 content decryption and always return the same error. Note: this attack
3242 needs on average 2^20 messages so it only affects automated senders. The
3243 old behaviour can be re-enabled in the CMS code by setting the
3244 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3245 an MMA defence is not necessary.
3246 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3247 this issue. (CVE-2012-0884)
3250 *) Fix CVE-2011-4619: make sure we really are receiving a
3251 client hello before rejecting multiple SGC restarts. Thanks to
3252 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3255 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3257 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3258 Thanks to Antonio Martin, Enterprise Secure Access Research and
3259 Development, Cisco Systems, Inc. for discovering this bug and
3260 preparing a fix. (CVE-2012-0050)
3263 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3265 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3266 of the Vaudenay padding oracle attack on CBC mode encryption
3267 which enables an efficient plaintext recovery attack against
3268 the OpenSSL implementation of DTLS. Their attack exploits timing
3269 differences arising during decryption processing. A research
3270 paper describing this attack can be found at:
3271 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3272 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3273 Security Group at Royal Holloway, University of London
3274 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3275 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3276 for preparing the fix. (CVE-2011-4108)
3277 [Robin Seggelmann, Michael Tuexen]
3279 *) Clear bytes used for block padding of SSL 3.0 records.
3281 [Adam Langley (Google)]
3283 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3284 Kadianakis <desnacked@gmail.com> for discovering this issue and
3285 Adam Langley for preparing the fix. (CVE-2011-4619)
3286 [Adam Langley (Google)]
3288 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3289 [Andrey Kulikov <amdeich@gmail.com>]
3291 *) Prevent malformed RFC3779 data triggering an assertion failure.
3292 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3293 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3294 [Rob Austein <sra@hactrn.net>]
3296 *) Improved PRNG seeding for VOS.
3297 [Paul Green <Paul.Green@stratus.com>]
3299 *) Fix ssl_ciph.c set-up race.
3300 [Adam Langley (Google)]
3302 *) Fix spurious failures in ecdsatest.c.
3303 [Emilia Käsper (Google)]
3305 *) Fix the BIO_f_buffer() implementation (which was mixing different
3306 interpretations of the '..._len' fields).
3307 [Adam Langley (Google)]
3309 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3310 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3311 threads won't reuse the same blinding coefficients.
3313 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3314 lock to call BN_BLINDING_invert_ex, and avoids one use of
3315 BN_BLINDING_update for each BN_BLINDING structure (previously,
3316 the last update always remained unused).
3317 [Emilia Käsper (Google)]
3319 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3320 [Bob Buckholz (Google)]
3322 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3324 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3325 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3326 [Kaspar Brand <ossl@velox.ch>]
3328 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3329 for multi-threaded use of ECDH. (CVE-2011-3210)
3330 [Adam Langley (Google)]
3332 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3335 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3336 signature public key algorithm by using OID xref utilities instead.
3337 Before this you could only use some ECC ciphersuites with SHA1 only.
3340 *) Add protection against ECDSA timing attacks as mentioned in the paper
3341 by Billy Bob Brumley and Nicola Tuveri, see:
3343 http://eprint.iacr.org/2011/232.pdf
3345 [Billy Bob Brumley and Nicola Tuveri]
3347 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3349 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3350 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3352 *) Fix bug in string printing code: if *any* escaping is enabled we must
3353 escape the escape character (backslash) or the resulting string is
3357 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3359 *) Disable code workaround for ancient and obsolete Netscape browsers
3360 and servers: an attacker can use it in a ciphersuite downgrade attack.
3361 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3364 *) Fixed J-PAKE implementation error, originally discovered by
3365 Sebastien Martini, further info and confirmation from Stefan
3366 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3369 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3371 *) Fix extension code to avoid race conditions which can result in a buffer
3372 overrun vulnerability: resumed sessions must not be modified as they can
3373 be shared by multiple threads. CVE-2010-3864
3376 *) Fix WIN32 build system to correctly link an ENGINE directory into
3380 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3382 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3384 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3386 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3388 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3389 context. The operation can be customised via the ctrl mechanism in
3390 case ENGINEs want to include additional functionality.
3393 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3396 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3397 output hashes compatible with older versions of OpenSSL.
3398 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3400 *) Fix compression algorithm handling: if resuming a session use the
3401 compression algorithm of the resumed session instead of determining
3402 it from client hello again. Don't allow server to change algorithm.
3405 *) Add load_crls() function to apps tidying load_certs() too. Add option
3406 to verify utility to allow additional CRLs to be included.
3409 *) Update OCSP request code to permit adding custom headers to the request:
3410 some responders need this.
3413 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3415 [Julia Lawall <julia@diku.dk>]
3417 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3418 needlessly dereferenced structures, used obsolete functions and
3419 didn't handle all updated verify codes correctly.
3422 *) Disable MD2 in the default configuration.
3425 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3426 indicate the initial BIO being pushed or popped. This makes it possible
3427 to determine whether the BIO is the one explicitly called or as a result
3428 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3429 it handles reference counts correctly and doesn't zero out the I/O bio
3430 when it is not being explicitly popped. WARNING: applications which
3431 included workarounds for the old buggy behaviour will need to be modified
3432 or they could free up already freed BIOs.
3435 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3436 renaming to all platforms (within the 0.9.8 branch, this was
3437 done conditionally on Netware platforms to avoid a name clash).
3438 [Guenter <lists@gknw.net>]
3440 *) Add ECDHE and PSK support to DTLS.
3441 [Michael Tuexen <tuexen@fh-muenster.de>]
3443 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3447 *) Add "missing" function EVP_MD_flags() (without this the only way to
3448 retrieve a digest flags is by accessing the structure directly. Update
3449 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3450 or cipher is registered as in the "from" argument. Print out all
3451 registered digests in the dgst usage message instead of manually
3452 attempting to work them out.
3455 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3456 this allows the use of compression and extensions. Change default cipher
3457 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3458 by default unless an application cipher string requests it.
3461 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3462 key ids to find matching certificates and keys but some PKCS#12 files
3463 don't follow the (somewhat unwritten) rules and this strategy fails.
3464 Now just gather all certificates together and the first private key
3465 then look for the first certificate that matches the key.
3468 *) Support use of registered digest and cipher names for dgst and cipher
3469 commands instead of having to add each one as a special case. So now
3476 openssl dgst -sha256 foo
3478 and this works for ENGINE based algorithms too.
3482 *) Update Gost ENGINE to support parameter files.
3483 [Victor B. Wagner <vitus@cryptocom.ru>]
3485 *) Support GeneralizedTime in ca utility.
3486 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3488 *) Enhance the hash format used for certificate directory links. The new
3489 form uses the canonical encoding (meaning equivalent names will work
3490 even if they aren't identical) and uses SHA1 instead of MD5. This form
3491 is incompatible with the older format and as a result c_rehash should
3492 be used to rebuild symbolic links.
3495 *) Make PKCS#8 the default write format for private keys, replacing the
3496 traditional format. This form is standardised, more secure and doesn't
3497 include an implicit MD5 dependency.
3500 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3501 committed to OpenSSL should pass this lot as a minimum.
3504 *) Add session ticket override functionality for use by EAP-FAST.
3505 [Jouni Malinen <j@w1.fi>]
3507 *) Modify HMAC functions to return a value. Since these can be implemented
3508 in an ENGINE errors can occur.
3511 *) Type-checked OBJ_bsearch_ex.
3514 *) Type-checked OBJ_bsearch. Also some constification necessitated
3515 by type-checking. Still to come: TXT_DB, bsearch(?),
3516 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3520 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3521 seconds to a tm structure directly, instead of going through OS
3522 specific date routines. This avoids any issues with OS routines such
3523 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3524 and X509_time_adj_ex() to cover the extended range. The existing
3525 X509_time_adj() is still usable and will no longer have any date issues.
3528 *) Delta CRL support. New use deltas option which will attempt to locate
3529 and search any appropriate delta CRLs available.
3531 This work was sponsored by Google.
3534 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3535 code and add additional score elements. Validate alternate CRL paths
3536 as part of the CRL checking and indicate a new error "CRL path validation
3537 error" in this case. Applications wanting additional details can use
3538 the verify callback and check the new "parent" field. If this is not
3539 NULL CRL path validation is taking place. Existing applications won't
3540 see this because it requires extended CRL support which is off by
3543 This work was sponsored by Google.
3546 *) Support for freshest CRL extension.
3548 This work was sponsored by Google.
3551 *) Initial indirect CRL support. Currently only supported in the CRLs
3552 passed directly and not via lookup. Process certificate issuer
3553 CRL entry extension and lookup CRL entries by bother issuer name
3554 and serial number. Check and process CRL issuer entry in IDP extension.
3556 This work was sponsored by Google.
3559 *) Add support for distinct certificate and CRL paths. The CRL issuer
3560 certificate is validated separately in this case. Only enabled if
3561 an extended CRL support flag is set: this flag will enable additional
3562 CRL functionality in future.
3564 This work was sponsored by Google.
3567 *) Add support for policy mappings extension.
3569 This work was sponsored by Google.
3572 *) Fixes to pathlength constraint, self issued certificate handling,
3573 policy processing to align with RFC3280 and PKITS tests.
3575 This work was sponsored by Google.
3578 *) Support for name constraints certificate extension. DN, email, DNS
3579 and URI types are currently supported.
3581 This work was sponsored by Google.
3584 *) To cater for systems that provide a pointer-based thread ID rather
3585 than numeric, deprecate the current numeric thread ID mechanism and
3586 replace it with a structure and associated callback type. This
3587 mechanism allows a numeric "hash" to be extracted from a thread ID in
3588 either case, and on platforms where pointers are larger than 'long',
3589 mixing is done to help ensure the numeric 'hash' is usable even if it
3590 can't be guaranteed unique. The default mechanism is to use "&errno"
3591 as a pointer-based thread ID to distinguish between threads.
3593 Applications that want to provide their own thread IDs should now use
3594 CRYPTO_THREADID_set_callback() to register a callback that will call
3595 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3597 Note that ERR_remove_state() is now deprecated, because it is tied
3598 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3599 to free the current thread's error state should be replaced by
3600 ERR_remove_thread_state(NULL).
3602 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3603 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3604 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3605 application was previously providing a numeric thread callback that
3606 was inappropriate for distinguishing threads, then uniqueness might
3607 have been obtained with &errno that happened immediately in the
3608 intermediate development versions of OpenSSL; this is no longer the
3609 case, the numeric thread callback will now override the automatic use
3611 [Geoff Thorpe, with help from Bodo Moeller]
3613 *) Initial support for different CRL issuing certificates. This covers a
3614 simple case where the self issued certificates in the chain exist and
3615 the real CRL issuer is higher in the existing chain.
3617 This work was sponsored by Google.
3620 *) Removed effectively defunct crypto/store from the build.
3623 *) Revamp of STACK to provide stronger type-checking. Still to come:
3624 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3625 ASN1_STRING, CONF_VALUE.
3628 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3629 RAM on SSL connections. This option can save about 34k per idle SSL.
3632 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3633 STACK, TXT_DB, bsearch, qsort.
3636 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3637 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3638 support for data, signedData, compressedData, digestedData and
3639 encryptedData, envelopedData types included. Scripts to check against
projects / openssl.git / blob