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 *) Get rid of Makefile.shared, and in the process, make the processing
13 of certain files (rc.obj, or the .def/.map/.opt files produced from
14 the ordinal files) more visible and hopefully easier to trace and
15 debug (or make silent).
18 *) Make it possible to have environment variable assignments as
19 arguments to config / Configure.
22 *) Add multi-prime RSA (RFC 8017) support.
25 *) Add SM3 implemented according to GB/T 32905-2016
26 [ Jack Lloyd <jack.lloyd@ribose.com>,
27 Ronald Tse <ronald.tse@ribose.com>,
28 Erick Borsboom <erick.borsboom@ribose.com> ]
30 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
31 as documented in RFC6066.
32 Based on a patch from Tomasz Moń
33 [Filipe Raimundo da Silva]
35 *) Add SM4 implemented according to GB/T 32907-2016.
36 [ Jack Lloyd <jack.lloyd@ribose.com>,
37 Ronald Tse <ronald.tse@ribose.com>,
38 Erick Borsboom <erick.borsboom@ribose.com> ]
40 *) Reimplement -newreq-nodes and ERR_error_string_n; the
41 original author does not agree with the license change.
44 *) Add ARIA AEAD TLS support.
47 *) Some macro definitions to support VS6 have been removed. Visual
48 Studio 6 has not worked since 1.1.0
51 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
52 without clearing the errors.
55 *) Add "atfork" functions. If building on a system that without
56 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
57 requirements. The RAND facility now uses/requires this.
63 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
64 not possible to disable entirely. However, it's still possible to
65 disable the console reading UI method, UI_OpenSSL() (use UI_null()
68 To disable, configure with 'no-ui-console'. 'no-ui' is still
69 possible to use as an alias. Check at compile time with the
70 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
71 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
74 *) Add a STORE module, which implements a uniform and URI based reader of
75 stores that can contain keys, certificates, CRLs and numerous other
76 objects. The main API is loosely based on a few stdio functions,
77 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
78 OSSL_STORE_error and OSSL_STORE_close.
79 The implementation uses backends called "loaders" to implement arbitrary
80 URI schemes. There is one built in "loader" for the 'file' scheme.
83 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
84 then adjusted to work on FreeBSD 8.4 as well.
85 Enable by configuring with 'enable-devcryptoeng'. This is done by default
86 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
89 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
90 util/mkerr.pl, which is adapted to allow those prefixes, leading to
91 error code calls like this:
93 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
95 With this change, we claim the namespaces OSSL and OPENSSL in a manner
96 that can be encoded in C. For the foreseeable future, this will only
98 [Richard Levitte and Tim Hudson]
100 *) Removed BSD cryptodev engine.
103 *) Add a build target 'build_all_generated', to build all generated files
104 and only that. This can be used to prepare everything that requires
105 things like perl for a system that lacks perl and then move everything
106 to that system and do the rest of the build there.
109 *) In the UI interface, make it possible to duplicate the user data. This
110 can be used by engines that need to retain the data for a longer time
111 than just the call where this user data is passed.
114 *) Ignore the '-named_curve auto' value for compatibility of applications
116 [Tomas Mraz <tmraz@fedoraproject.org>]
118 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
119 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
120 alerts across multiple records (some of which could be empty). In practice
121 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
122 prohibts this altogether and other libraries (BoringSSL, NSS) do not
123 support this at all. Supporting it adds significant complexity to the
124 record layer, and its removal is unlikely to cause inter-operability
128 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
129 with Z. These are meant to replace LONG and ZLONG and to be size safe.
130 The use of LONG and ZLONG is discouraged and scheduled for deprecation
134 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
135 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
136 [Richard Levitte, Andy Polyakov]
138 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
142 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
143 platform rather than 'mingw'.
146 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
147 success if they are asked to add an object which already exists
148 in the store. This change cascades to other functions which load
149 certificates and CRLs.
152 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
153 facilitate stack unwinding even from assembly subroutines.
156 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
157 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
160 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
161 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
162 which is the minimum version we support.
165 *) Certificate time validation (X509_cmp_time) enforces stricter
166 compliance with RFC 5280. Fractional seconds and timezone offsets
167 are no longer allowed.
170 *) Add support for ARIA
173 *) s_client will now send the Server Name Indication (SNI) extension by
174 default unless the new "-noservername" option is used. The server name is
175 based on the host provided to the "-connect" option unless overridden by
179 *) Add support for SipHash
182 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
183 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
184 prevent issues where no progress is being made and the peer continually
185 sends unrecognised record types, using up resources processing them.
188 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
189 using the algorithm defined in
190 https://www.akkadia.org/drepper/SHA-crypt.txt
193 *) Heartbeat support has been removed; the ABI is changed for now.
194 [Richard Levitte, Rich Salz]
196 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
199 *) The RSA "null" method, which was partially supported to avoid patent
200 issues, has been replaced to always returns NULL.
203 Changes between 1.1.0g and 1.1.0h [xx XXX xxxx]
205 *) rsaz_1024_mul_avx2 overflow bug on x86_64
207 There is an overflow bug in the AVX2 Montgomery multiplication procedure
208 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
209 Analysis suggests that attacks against RSA and DSA as a result of this
210 defect would be very difficult to perform and are not believed likely.
211 Attacks against DH1024 are considered just feasible, because most of the
212 work necessary to deduce information about a private key may be performed
213 offline. The amount of resources required for such an attack would be
214 significant. However, for an attack on TLS to be meaningful, the server
215 would have to share the DH1024 private key among multiple clients, which is
216 no longer an option since CVE-2016-0701.
218 This only affects processors that support the AVX2 but not ADX extensions
219 like Intel Haswell (4th generation).
221 This issue was reported to OpenSSL by David Benjamin (Google). The issue
222 was originally found via the OSS-Fuzz project.
226 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
228 *) bn_sqrx8x_internal carry bug on x86_64
230 There is a carry propagating bug in the x86_64 Montgomery squaring
231 procedure. No EC algorithms are affected. Analysis suggests that attacks
232 against RSA and DSA as a result of this defect would be very difficult to
233 perform and are not believed likely. Attacks against DH are considered just
234 feasible (although very difficult) because most of the work necessary to
235 deduce information about a private key may be performed offline. The amount
236 of resources required for such an attack would be very significant and
237 likely only accessible to a limited number of attackers. An attacker would
238 additionally need online access to an unpatched system using the target
239 private key in a scenario with persistent DH parameters and a private
240 key that is shared between multiple clients.
242 This only affects processors that support the BMI1, BMI2 and ADX extensions
243 like Intel Broadwell (5th generation) and later or AMD Ryzen.
245 This issue was reported to OpenSSL by the OSS-Fuzz project.
249 *) Malformed X.509 IPAddressFamily could cause OOB read
251 If an X.509 certificate has a malformed IPAddressFamily extension,
252 OpenSSL could do a one-byte buffer overread. The most likely result
253 would be an erroneous display of the certificate in text format.
255 This issue was reported to OpenSSL by the OSS-Fuzz project.
259 Changes between 1.1.0e and 1.1.0f [25 May 2017]
261 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
262 platform rather than 'mingw'.
265 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
266 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
267 which is the minimum version we support.
270 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
272 *) Encrypt-Then-Mac renegotiation crash
274 During a renegotiation handshake if the Encrypt-Then-Mac extension is
275 negotiated where it was not in the original handshake (or vice-versa) then
276 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
277 and servers are affected.
279 This issue was reported to OpenSSL by Joe Orton (Red Hat).
283 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
285 *) Truncated packet could crash via OOB read
287 If one side of an SSL/TLS path is running on a 32-bit host and a specific
288 cipher is being used, then a truncated packet can cause that host to
289 perform an out-of-bounds read, usually resulting in a crash.
291 This issue was reported to OpenSSL by Robert Święcki of Google.
295 *) Bad (EC)DHE parameters cause a client crash
297 If a malicious server supplies bad parameters for a DHE or ECDHE key
298 exchange then this can result in the client attempting to dereference a
299 NULL pointer leading to a client crash. This could be exploited in a Denial
302 This issue was reported to OpenSSL by Guido Vranken.
306 *) BN_mod_exp may produce incorrect results on x86_64
308 There is a carry propagating bug in the x86_64 Montgomery squaring
309 procedure. No EC algorithms are affected. Analysis suggests that attacks
310 against RSA and DSA as a result of this defect would be very difficult to
311 perform and are not believed likely. Attacks against DH are considered just
312 feasible (although very difficult) because most of the work necessary to
313 deduce information about a private key may be performed offline. The amount
314 of resources required for such an attack would be very significant and
315 likely only accessible to a limited number of attackers. An attacker would
316 additionally need online access to an unpatched system using the target
317 private key in a scenario with persistent DH parameters and a private
318 key that is shared between multiple clients. For example this can occur by
319 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
320 similar to CVE-2015-3193 but must be treated as a separate problem.
322 This issue was reported to OpenSSL by the OSS-Fuzz project.
326 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
328 *) ChaCha20/Poly1305 heap-buffer-overflow
330 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
331 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
332 crash. This issue is not considered to be exploitable beyond a DoS.
334 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
338 *) CMS Null dereference
340 Applications parsing invalid CMS structures can crash with a NULL pointer
341 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
342 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
343 structure callback if an attempt is made to free certain invalid encodings.
344 Only CHOICE structures using a callback which do not handle NULL value are
347 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
351 *) Montgomery multiplication may produce incorrect results
353 There is a carry propagating bug in the Broadwell-specific Montgomery
354 multiplication procedure that handles input lengths divisible by, but
355 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
356 and DH private keys are impossible. This is because the subroutine in
357 question is not used in operations with the private key itself and an input
358 of the attacker's direct choice. Otherwise the bug can manifest itself as
359 transient authentication and key negotiation failures or reproducible
360 erroneous outcome of public-key operations with specially crafted input.
361 Among EC algorithms only Brainpool P-512 curves are affected and one
362 presumably can attack ECDH key negotiation. Impact was not analyzed in
363 detail, because pre-requisites for attack are considered unlikely. Namely
364 multiple clients have to choose the curve in question and the server has to
365 share the private key among them, neither of which is default behaviour.
366 Even then only clients that chose the curve will be affected.
368 This issue was publicly reported as transient failures and was not
369 initially recognized as a security issue. Thanks to Richard Morgan for
370 providing reproducible case.
374 *) Removed automatic addition of RPATH in shared libraries and executables,
375 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
378 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
380 *) Fix Use After Free for large message sizes
382 The patch applied to address CVE-2016-6307 resulted in an issue where if a
383 message larger than approx 16k is received then the underlying buffer to
384 store the incoming message is reallocated and moved. Unfortunately a
385 dangling pointer to the old location is left which results in an attempt to
386 write to the previously freed location. This is likely to result in a
387 crash, however it could potentially lead to execution of arbitrary code.
389 This issue only affects OpenSSL 1.1.0a.
391 This issue was reported to OpenSSL by Robert Święcki.
395 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
397 *) OCSP Status Request extension unbounded memory growth
399 A malicious client can send an excessively large OCSP Status Request
400 extension. If that client continually requests renegotiation, sending a
401 large OCSP Status Request extension each time, then there will be unbounded
402 memory growth on the server. This will eventually lead to a Denial Of
403 Service attack through memory exhaustion. Servers with a default
404 configuration are vulnerable even if they do not support OCSP. Builds using
405 the "no-ocsp" build time option are not affected.
407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
411 *) SSL_peek() hang on empty record
413 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
414 sends an empty record. This could be exploited by a malicious peer in a
415 Denial Of Service attack.
417 This issue was reported to OpenSSL by Alex Gaynor.
421 *) Excessive allocation of memory in tls_get_message_header() and
422 dtls1_preprocess_fragment()
424 A (D)TLS message includes 3 bytes for its length in the header for the
425 message. This would allow for messages up to 16Mb in length. Messages of
426 this length are excessive and OpenSSL includes a check to ensure that a
427 peer is sending reasonably sized messages in order to avoid too much memory
428 being consumed to service a connection. A flaw in the logic of version
429 1.1.0 means that memory for the message is allocated too early, prior to
430 the excessive message length check. Due to way memory is allocated in
431 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
432 to service a connection. This could lead to a Denial of Service through
433 memory exhaustion. However, the excessive message length check still takes
434 place, and this would cause the connection to immediately fail. Assuming
435 that the application calls SSL_free() on the failed connection in a timely
436 manner then the 21Mb of allocated memory will then be immediately freed
437 again. Therefore the excessive memory allocation will be transitory in
438 nature. This then means that there is only a security impact if:
440 1) The application does not call SSL_free() in a timely manner in the event
441 that the connection fails
443 2) The application is working in a constrained environment where there is
444 very little free memory
446 3) The attacker initiates multiple connection attempts such that there are
447 multiple connections in a state where memory has been allocated for the
448 connection; SSL_free() has not yet been called; and there is insufficient
449 memory to service the multiple requests.
451 Except in the instance of (1) above any Denial Of Service is likely to be
452 transitory because as soon as the connection fails the memory is
453 subsequently freed again in the SSL_free() call. However there is an
454 increased risk during this period of application crashes due to the lack of
455 memory - which would then mean a more serious Denial of Service.
457 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
458 (CVE-2016-6307 and CVE-2016-6308)
461 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
462 had to be removed. Primary reason is that vendor assembler can't
463 assemble our modules with -KPIC flag. As result it, assembly
464 support, was not even available as option. But its lack means
465 lack of side-channel resistant code, which is incompatible with
466 security by todays standards. Fortunately gcc is readily available
467 prepackaged option, which we firmly point at...
470 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
472 *) Windows command-line tool supports UTF-8 opt-in option for arguments
473 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
474 (to any value) allows Windows user to access PKCS#12 file generated
475 with Windows CryptoAPI and protected with non-ASCII password, as well
476 as files generated under UTF-8 locale on Linux also protected with
480 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
481 have been disabled by default and removed from DEFAULT, just like RC4.
482 See the RC4 item below to re-enable both.
485 *) The method for finding the storage location for the Windows RAND seed file
486 has changed. First we check %RANDFILE%. If that is not set then we check
487 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
488 all else fails we fall back to C:\.
491 *) The EVP_EncryptUpdate() function has had its return type changed from void
492 to int. A return of 0 indicates and error while a return of 1 indicates
496 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
497 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
498 off the constant time implementation for RSA, DSA and DH have been made
499 no-ops and deprecated.
502 *) Windows RAND implementation was simplified to only get entropy by
503 calling CryptGenRandom(). Various other RAND-related tickets
505 [Joseph Wylie Yandle, Rich Salz]
507 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
508 and OPENSSL_LH_, respectively. The old names are available
509 with API compatibility. They new names are now completely documented.
512 *) Unify TYPE_up_ref(obj) methods signature.
513 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
514 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
515 int (instead of void) like all others TYPE_up_ref() methods.
516 So now these methods also check the return value of CRYPTO_atomic_add(),
517 and the validity of object reference counter.
518 [fdasilvayy@gmail.com]
520 *) With Windows Visual Studio builds, the .pdb files are installed
521 alongside the installed libraries and executables. For a static
522 library installation, ossl_static.pdb is the associate compiler
523 generated .pdb file to be used when linking programs.
526 *) Remove openssl.spec. Packaging files belong with the packagers.
529 *) Automatic Darwin/OSX configuration has had a refresh, it will now
530 recognise x86_64 architectures automatically. You can still decide
531 to build for a different bitness with the environment variable
532 KERNEL_BITS (can be 32 or 64), for example:
534 KERNEL_BITS=32 ./config
538 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
539 256 bit AES and HMAC with SHA256.
542 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
545 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
548 *) To enable users to have their own config files and build file templates,
549 Configure looks in the directory indicated by the environment variable
550 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
551 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
552 name and is used as is.
555 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
556 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
557 X509_CERT_FILE_CTX was removed.
560 *) "shared" builds are now the default. To create only static libraries use
561 the "no-shared" Configure option.
564 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
565 All of these option have not worked for some while and are fundamental
569 *) Make various cleanup routines no-ops and mark them as deprecated. Most
570 global cleanup functions are no longer required because they are handled
571 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
572 Explicitly de-initing can cause problems (e.g. where a library that uses
573 OpenSSL de-inits, but an application is still using it). The affected
574 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
575 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
576 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
580 *) --strict-warnings no longer enables runtime debugging options
581 such as REF_DEBUG. Instead, debug options are automatically
582 enabled with '--debug' builds.
583 [Andy Polyakov, Emilia Käsper]
585 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
586 have been moved out of the public header files. New functions for managing
587 these have been added.
590 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
591 objects have been moved out of the public header files. New
592 functions for managing these have been added.
595 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
596 have been moved out of the public header files. New functions for managing
597 these have been added.
600 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
601 moved out of the public header files. New functions for managing these
605 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
608 *) Removed the mk1mf build scripts.
611 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
612 it is always safe to #include a header now.
615 *) Removed the aged BC-32 config and all its supporting scripts
618 *) Removed support for Ultrix, Netware, and OS/2.
621 *) Add support for HKDF.
624 *) Add support for blake2b and blake2s
627 *) Added support for "pipelining". Ciphers that have the
628 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
629 encryptions/decryptions simultaneously. There are currently no built-in
630 ciphers with this property but the expectation is that engines will be able
631 to offer it to significantly improve throughput. Support has been extended
632 into libssl so that multiple records for a single connection can be
633 processed in one go (for >=TLS 1.1).
636 *) Added the AFALG engine. This is an async capable engine which is able to
637 offload work to the Linux kernel. In this initial version it only supports
638 AES128-CBC. The kernel must be version 4.1.0 or greater.
641 *) OpenSSL now uses a new threading API. It is no longer necessary to
642 set locking callbacks to use OpenSSL in a multi-threaded environment. There
643 are two supported threading models: pthreads and windows threads. It is
644 also possible to configure OpenSSL at compile time for "no-threads". The
645 old threading API should no longer be used. The functions have been
646 replaced with "no-op" compatibility macros.
647 [Alessandro Ghedini, Matt Caswell]
649 *) Modify behavior of ALPN to invoke callback after SNI/servername
650 callback, such that updates to the SSL_CTX affect ALPN.
653 *) Add SSL_CIPHER queries for authentication and key-exchange.
656 *) Changes to the DEFAULT cipherlist:
657 - Prefer (EC)DHE handshakes over plain RSA.
658 - Prefer AEAD ciphers over legacy ciphers.
659 - Prefer ECDSA over RSA when both certificates are available.
660 - Prefer TLSv1.2 ciphers/PRF.
661 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
665 *) Change the ECC default curve list to be this, in order: x25519,
666 secp256r1, secp521r1, secp384r1.
669 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
670 disabled by default. They can be re-enabled using the
671 enable-weak-ssl-ciphers option to Configure.
674 *) If the server has ALPN configured, but supports no protocols that the
675 client advertises, send a fatal "no_application_protocol" alert.
676 This behaviour is SHALL in RFC 7301, though it isn't universally
677 implemented by other servers.
680 *) Add X25519 support.
681 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
682 for public and private key encoding using the format documented in
683 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
684 key generation and key derivation.
686 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
690 *) Deprecate SRP_VBASE_get_by_user.
691 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
692 In order to fix an unavoidable memory leak (CVE-2016-0798),
693 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
694 seed, even if the seed is configured.
696 Users should use SRP_VBASE_get1_by_user instead. Note that in
697 SRP_VBASE_get1_by_user, caller must free the returned value. Note
698 also that even though configuring the SRP seed attempts to hide
699 invalid usernames by continuing the handshake with fake
700 credentials, this behaviour is not constant time and no strong
701 guarantees are made that the handshake is indistinguishable from
702 that of a valid user.
705 *) Configuration change; it's now possible to build dynamic engines
706 without having to build shared libraries and vice versa. This
707 only applies to the engines in engines/, those in crypto/engine/
708 will always be built into libcrypto (i.e. "static").
710 Building dynamic engines is enabled by default; to disable, use
711 the configuration option "disable-dynamic-engine".
713 The only requirements for building dynamic engines are the
714 presence of the DSO module and building with position independent
715 code, so they will also automatically be disabled if configuring
716 with "disable-dso" or "disable-pic".
718 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
719 are also taken away from openssl/opensslconf.h, as they are
723 *) Configuration change; if there is a known flag to compile
724 position independent code, it will always be applied on the
725 libcrypto and libssl object files, and never on the application
726 object files. This means other libraries that use routines from
727 libcrypto / libssl can be made into shared libraries regardless
728 of how OpenSSL was configured.
730 If this isn't desirable, the configuration options "disable-pic"
731 or "no-pic" can be used to disable the use of PIC. This will
732 also disable building shared libraries and dynamic engines.
735 *) Removed JPAKE code. It was experimental and has no wide use.
738 *) The INSTALL_PREFIX Makefile variable has been renamed to
739 DESTDIR. That makes for less confusion on what this variable
740 is for. Also, the configuration option --install_prefix is
744 *) Heartbeat for TLS has been removed and is disabled by default
745 for DTLS; configure with enable-heartbeats. Code that uses the
746 old #define's might need to be updated.
747 [Emilia Käsper, Rich Salz]
749 *) Rename REF_CHECK to REF_DEBUG.
752 *) New "unified" build system
754 The "unified" build system is aimed to be a common system for all
755 platforms we support. With it comes new support for VMS.
757 This system builds supports building in a different directory tree
758 than the source tree. It produces one Makefile (for unix family
759 or lookalikes), or one descrip.mms (for VMS).
761 The source of information to make the Makefile / descrip.mms is
762 small files called 'build.info', holding the necessary
763 information for each directory with source to compile, and a
764 template in Configurations, like unix-Makefile.tmpl or
767 With this change, the library names were also renamed on Windows
768 and on VMS. They now have names that are closer to the standard
769 on Unix, and include the major version number, and in certain
770 cases, the architecture they are built for. See "Notes on shared
771 libraries" in INSTALL.
773 We rely heavily on the perl module Text::Template.
776 *) Added support for auto-initialisation and de-initialisation of the library.
777 OpenSSL no longer requires explicit init or deinit routines to be called,
778 except in certain circumstances. See the OPENSSL_init_crypto() and
779 OPENSSL_init_ssl() man pages for further information.
782 *) The arguments to the DTLSv1_listen function have changed. Specifically the
783 "peer" argument is now expected to be a BIO_ADDR object.
785 *) Rewrite of BIO networking library. The BIO library lacked consistent
786 support of IPv6, and adding it required some more extensive
787 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
788 which hold all types of addresses and chains of address information.
789 It also introduces a new API, with functions like BIO_socket,
790 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
791 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
792 have been adapted accordingly.
795 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
799 *) CRIME protection: disable compression by default, even if OpenSSL is
800 compiled with zlib enabled. Applications can still enable compression
801 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
802 using the SSL_CONF library to configure compression.
805 *) The signature of the session callback configured with
806 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
807 was explicitly marked as 'const unsigned char*' instead of
811 *) Always DPURIFY. Remove the use of uninitialized memory in the
812 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
815 *) Removed many obsolete configuration items, including
816 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
817 MD2_CHAR, MD2_INT, MD2_LONG
819 IDEA_SHORT, IDEA_LONG
820 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
821 [Rich Salz, with advice from Andy Polyakov]
823 *) Many BN internals have been moved to an internal header file.
824 [Rich Salz with help from Andy Polyakov]
826 *) Configuration and writing out the results from it has changed.
827 Files such as Makefile include/openssl/opensslconf.h and are now
828 produced through general templates, such as Makefile.in and
829 crypto/opensslconf.h.in and some help from the perl module
832 Also, the center of configuration information is no longer
833 Makefile. Instead, Configure produces a perl module in
834 configdata.pm which holds most of the config data (in the hash
835 table %config), the target data that comes from the target
836 configuration in one of the Configurations/*.conf files (in
840 *) To clarify their intended purposes, the Configure options
841 --prefix and --openssldir change their semantics, and become more
842 straightforward and less interdependent.
844 --prefix shall be used exclusively to give the location INSTALLTOP
845 where programs, scripts, libraries, include files and manuals are
846 going to be installed. The default is now /usr/local.
848 --openssldir shall be used exclusively to give the default
849 location OPENSSLDIR where certificates, private keys, CRLs are
850 managed. This is also where the default openssl.cnf gets
852 If the directory given with this option is a relative path, the
853 values of both the --prefix value and the --openssldir value will
854 be combined to become OPENSSLDIR.
855 The default for --openssldir is INSTALLTOP/ssl.
857 Anyone who uses --openssldir to specify where OpenSSL is to be
858 installed MUST change to use --prefix instead.
861 *) The GOST engine was out of date and therefore it has been removed. An up
862 to date GOST engine is now being maintained in an external repository.
863 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
864 support for GOST ciphersuites (these are only activated if a GOST engine
868 *) EGD is no longer supported by default; use enable-egd when
870 [Ben Kaduk and Rich Salz]
872 *) The distribution now has Makefile.in files, which are used to
873 create Makefile's when Configure is run. *Configure must be run
874 before trying to build now.*
877 *) The return value for SSL_CIPHER_description() for error conditions
881 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
883 Obtaining and performing DNSSEC validation of TLSA records is
884 the application's responsibility. The application provides
885 the TLSA records of its choice to OpenSSL, and these are then
886 used to authenticate the peer.
888 The TLSA records need not even come from DNS. They can, for
889 example, be used to implement local end-entity certificate or
890 trust-anchor "pinning", where the "pin" data takes the form
891 of TLSA records, which can augment or replace verification
892 based on the usual WebPKI public certification authorities.
895 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
896 continues to support deprecated interfaces in default builds.
897 However, applications are strongly advised to compile their
898 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
899 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
900 or the 1.1.0 releases.
902 In environments in which all applications have been ported to
903 not use any deprecated interfaces OpenSSL's Configure script
904 should be used with the --api=1.1.0 option to entirely remove
905 support for the deprecated features from the library and
906 unconditionally disable them in the installed headers.
907 Essentially the same effect can be achieved with the "no-deprecated"
908 argument to Configure, except that this will always restrict
909 the build to just the latest API, rather than a fixed API
912 As applications are ported to future revisions of the API,
913 they should update their compile-time OPENSSL_API_COMPAT define
914 accordingly, but in most cases should be able to continue to
915 compile with later releases.
917 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
918 0x10000000L and 0x00908000L, respectively. However those
919 versions did not support the OPENSSL_API_COMPAT feature, and
920 so applications are not typically tested for explicit support
921 of just the undeprecated features of either release.
924 *) Add support for setting the minimum and maximum supported protocol.
925 It can bet set via the SSL_set_min_proto_version() and
926 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
927 MaxProtocol. It's recommended to use the new APIs to disable
928 protocols instead of disabling individual protocols using
929 SSL_set_options() or SSL_CONF's Protocol. This change also
930 removes support for disabling TLS 1.2 in the OpenSSL TLS
931 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
934 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
937 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
938 and integrates ECDSA and ECDH functionality into EC. Implementations can
939 now redirect key generation and no longer need to convert to or from
942 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
943 include the ec.h header file instead.
946 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
947 ciphers who are no longer supported and drops support the ephemeral RSA key
948 exchange. The LOW ciphers currently doesn't have any ciphers in it.
951 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
952 opaque. For HMAC_CTX, the following constructors and destructors
955 HMAC_CTX *HMAC_CTX_new(void);
956 void HMAC_CTX_free(HMAC_CTX *ctx);
958 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
959 destroy such methods has been added. See EVP_MD_meth_new(3) and
960 EVP_CIPHER_meth_new(3) for documentation.
963 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
964 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
965 EVP_MD_CTX_reset() should be called instead to reinitialise
966 an already created structure.
967 2) For consistency with the majority of our object creators and
968 destructors, EVP_MD_CTX_(create|destroy) were renamed to
969 EVP_MD_CTX_(new|free). The old names are retained as macros
970 for deprecated builds.
973 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
974 cryptographic operations to be performed asynchronously as long as an
975 asynchronous capable engine is used. See the ASYNC_start_job() man page for
976 further details. Libssl has also had this capability integrated with the
977 introduction of the new mode SSL_MODE_ASYNC and associated error
978 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
979 pages. This work was developed in partnership with Intel Corp.
982 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
983 always enabled now. If you want to disable the support you should
984 exclude it using the list of supported ciphers. This also means that the
985 "-no_ecdhe" option has been removed from s_server.
988 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
989 SSL_{CTX_}set1_curves() which can set a list.
992 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
993 curve you want to support using SSL_{CTX_}set1_curves().
996 *) State machine rewrite. The state machine code has been significantly
997 refactored in order to remove much duplication of code and solve issues
998 with the old code (see ssl/statem/README for further details). This change
999 does have some associated API changes. Notably the SSL_state() function
1000 has been removed and replaced by SSL_get_state which now returns an
1001 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1002 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1006 *) All instances of the string "ssleay" in the public API were replaced
1007 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1008 Some error codes related to internal RSA_eay API's were renamed.
1011 *) The demo files in crypto/threads were moved to demo/threads.
1014 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1016 [Matt Caswell, Rich Salz]
1018 *) New ASN.1 embed macro.
1020 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1021 structure is not allocated: it is part of the parent. That is instead of
1029 This reduces memory fragmentation and make it impossible to accidentally
1030 set a mandatory field to NULL.
1032 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1033 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1034 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1038 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1041 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1042 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1043 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1044 DES and RC4 ciphersuites.
1047 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1048 This changes the decoding behaviour for some invalid messages,
1049 though the change is mostly in the more lenient direction, and
1050 legacy behaviour is preserved as much as possible.
1053 *) Fix no-stdio build.
1054 [ David Woodhouse <David.Woodhouse@intel.com> and also
1055 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1057 *) New testing framework
1058 The testing framework has been largely rewritten and is now using
1059 perl and the perl modules Test::Harness and an extended variant of
1060 Test::More called OpenSSL::Test to do its work. All test scripts in
1061 test/ have been rewritten into test recipes, and all direct calls to
1062 executables in test/Makefile have become individual recipes using the
1063 simplified testing OpenSSL::Test::Simple.
1065 For documentation on our testing modules, do:
1067 perldoc test/testlib/OpenSSL/Test/Simple.pm
1068 perldoc test/testlib/OpenSSL/Test.pm
1072 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1073 are used; the latter aborts on memory leaks (usually checked on exit).
1074 Some undocumented "set malloc, etc., hooks" functions were removed
1075 and others were changed. All are now documented.
1078 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1080 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1082 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1083 from RFC4279, RFC4785, RFC5487, RFC5489.
1085 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1086 original RSA_PSK patch.
1089 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1090 era flag was never set throughout the codebase (only read). Also removed
1091 SSL3_FLAGS_POP_BUFFER which was only used if
1092 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1095 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1096 to be "oneline" instead of "compat".
1099 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1100 not aware of clients that still exhibit this bug, and the workaround
1101 hasn't been working properly for a while.
1104 *) The return type of BIO_number_read() and BIO_number_written() as well as
1105 the corresponding num_read and num_write members in the BIO structure has
1106 changed from unsigned long to uint64_t. On platforms where an unsigned
1107 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1111 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1112 OpenSSL without support for them. It also means that maintaining
1113 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1114 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1117 *) Removed support for the two export grade static DH ciphersuites
1118 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1119 were newly added (along with a number of other static DH ciphersuites) to
1120 1.0.2. However the two export ones have *never* worked since they were
1121 introduced. It seems strange in any case to be adding new export
1122 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1125 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1126 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1127 and turned into macros which simply call the new preferred function names
1128 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1129 should use the new names instead. Also as part of this change the ssl23.h
1130 header file has been removed.
1133 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1134 code and the associated standard is no longer considered fit-for-purpose.
1137 *) RT2547 was closed. When generating a private key, try to make the
1138 output file readable only by the owner. This behavior change might
1139 be noticeable when interacting with other software.
1141 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1145 *) Added HTTP GET support to the ocsp command.
1148 *) Changed default digest for the dgst and enc commands from MD5 to
1152 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1155 *) Added support for TLS extended master secret from
1156 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1157 initial patch which was a great help during development.
1160 *) All libssl internal structures have been removed from the public header
1161 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1162 now redundant). Users should not attempt to access internal structures
1163 directly. Instead they should use the provided API functions.
1166 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1167 Access to deprecated functions can be re-enabled by running config with
1168 "enable-deprecated". In addition applications wishing to use deprecated
1169 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1170 will, by default, disable some transitive includes that previously existed
1171 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1174 *) Added support for OCB mode. OpenSSL has been granted a patent license
1175 compatible with the OpenSSL license for use of OCB. Details are available
1176 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1177 for OCB can be removed by calling config with no-ocb.
1180 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1181 compatible client hello.
1184 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1185 done while fixing the error code for the key-too-small case.
1186 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1188 *) CA.sh has been removed; use CA.pl instead.
1191 *) Removed old DES API.
1194 *) Remove various unsupported platforms:
1200 Sinix/ReliantUNIX RM400
1205 16-bit platforms such as WIN16
1208 *) Clean up OPENSSL_NO_xxx #define's
1209 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1210 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1211 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1212 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1213 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1214 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1215 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1216 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1217 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1218 Remove MS_STATIC; it's a relic from platforms <32 bits.
1221 *) Cleaned up dead code
1222 Remove all but one '#ifdef undef' which is to be looked at.
1225 *) Clean up calling of xxx_free routines.
1226 Just like free(), fix most of the xxx_free routines to accept
1227 NULL. Remove the non-null checks from callers. Save much code.
1230 *) Add secure heap for storage of private keys (when possible).
1231 Add BIO_s_secmem(), CBIGNUM, etc.
1232 Contributed by Akamai Technologies under our Corporate CLA.
1235 *) Experimental support for a new, fast, unbiased prime candidate generator,
1236 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1237 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1239 *) New output format NSS in the sess_id command line tool. This allows
1240 exporting the session id and the master key in NSS keylog format.
1241 [Martin Kaiser <martin@kaiser.cx>]
1243 *) Harmonize version and its documentation. -f flag is used to display
1245 [mancha <mancha1@zoho.com>]
1247 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1248 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1249 [mancha <mancha1@zoho.com>]
1251 *) Fix some double frees. These are not thought to be exploitable.
1252 [mancha <mancha1@zoho.com>]
1254 *) A missing bounds check in the handling of the TLS heartbeat extension
1255 can be used to reveal up to 64k of memory to a connected client or
1258 Thanks for Neel Mehta of Google Security for discovering this bug and to
1259 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1260 preparing the fix (CVE-2014-0160)
1261 [Adam Langley, Bodo Moeller]
1263 *) Fix for the attack described in the paper "Recovering OpenSSL
1264 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1265 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1266 http://eprint.iacr.org/2014/140
1268 Thanks to Yuval Yarom and Naomi Benger for discovering this
1269 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1270 [Yuval Yarom and Naomi Benger]
1272 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1273 this fixes a limitation in previous versions of OpenSSL.
1276 *) Experimental encrypt-then-mac support.
1278 Experimental support for encrypt then mac from
1279 draft-gutmann-tls-encrypt-then-mac-02.txt
1281 To enable it set the appropriate extension number (0x42 for the test
1282 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1284 For non-compliant peers (i.e. just about everything) this should have no
1287 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1291 *) Add EVP support for key wrapping algorithms, to avoid problems with
1292 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1293 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1294 algorithms and include tests cases.
1297 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1301 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1302 MGF1 digest and OAEP label.
1305 *) Make openssl verify return errors.
1306 [Chris Palmer <palmer@google.com> and Ben Laurie]
1308 *) New function ASN1_TIME_diff to calculate the difference between two
1309 ASN1_TIME structures or one structure and the current time.
1312 *) Update fips_test_suite to support multiple command line options. New
1313 test to induce all self test errors in sequence and check expected
1317 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1318 sign or verify all in one operation.
1321 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1322 test programs and fips_test_suite. Includes functionality to parse
1323 the minimal script output of fipsalgest.pl directly.
1326 *) Add authorisation parameter to FIPS_module_mode_set().
1329 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1332 *) Use separate DRBG fields for internal and external flags. New function
1333 FIPS_drbg_health_check() to perform on demand health checking. Add
1334 generation tests to fips_test_suite with reduced health check interval to
1335 demonstrate periodic health checking. Add "nodh" option to
1336 fips_test_suite to skip very slow DH test.
1339 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1343 *) More extensive health check for DRBG checking many more failure modes.
1344 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1345 combination: call this in fips_test_suite.
1348 *) Add support for canonical generation of DSA parameter 'g'. See
1351 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1352 POST to handle HMAC cases.
1355 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1356 to return numerical and string versions of the FIPS module number.
1359 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1360 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1361 outside the validated module in the FIPS capable OpenSSL.
1364 *) Minor change to DRBG entropy callback semantics. In some cases
1365 there is no multiple of the block length between min_len and
1366 max_len. Allow the callback to return more than max_len bytes
1367 of entropy but discard any extra: it is the callback's responsibility
1368 to ensure that the extra data discarded does not impact the
1369 requested amount of entropy.
1372 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1373 information in FIPS186-3, SP800-57 and SP800-131A.
1376 *) CCM support via EVP. Interface is very similar to GCM case except we
1377 must supply all data in one chunk (i.e. no update, final) and the
1378 message length must be supplied if AAD is used. Add algorithm test
1382 *) Initial version of POST overhaul. Add POST callback to allow the status
1383 of POST to be monitored and/or failures induced. Modify fips_test_suite
1384 to use callback. Always run all selftests even if one fails.
1387 *) XTS support including algorithm test driver in the fips_gcmtest program.
1388 Note: this does increase the maximum key length from 32 to 64 bytes but
1389 there should be no binary compatibility issues as existing applications
1390 will never use XTS mode.
1393 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1394 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1395 performs algorithm blocking for unapproved PRNG types. Also do not
1396 set PRNG type in FIPS_mode_set(): leave this to the application.
1397 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1398 the standard OpenSSL PRNG: set additional data to a date time vector.
1401 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1402 This shouldn't present any incompatibility problems because applications
1403 shouldn't be using these directly and any that are will need to rethink
1404 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1407 *) Extensive self tests and health checking required by SP800-90 DRBG.
1408 Remove strength parameter from FIPS_drbg_instantiate and always
1409 instantiate at maximum supported strength.
1412 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1415 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1418 *) New function DH_compute_key_padded() to compute a DH key and pad with
1419 leading zeroes if needed: this complies with SP800-56A et al.
1422 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1423 anything, incomplete, subject to change and largely untested at present.
1426 *) Modify fipscanisteronly build option to only build the necessary object
1427 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1430 *) Add experimental option FIPSSYMS to give all symbols in
1431 fipscanister.o and FIPS or fips prefix. This will avoid
1432 conflicts with future versions of OpenSSL. Add perl script
1433 util/fipsas.pl to preprocess assembly language source files
1434 and rename any affected symbols.
1437 *) Add selftest checks and algorithm block of non-fips algorithms in
1438 FIPS mode. Remove DES2 from selftests.
1441 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1442 return internal method without any ENGINE dependencies. Add new
1443 tiny fips sign and verify functions.
1446 *) New build option no-ec2m to disable characteristic 2 code.
1449 *) New build option "fipscanisteronly". This only builds fipscanister.o
1450 and (currently) associated fips utilities. Uses the file Makefile.fips
1451 instead of Makefile.org as the prototype.
1454 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1455 Update fips_gcmtest to use IV generator.
1458 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1459 setting output buffer to NULL. The *Final function must be
1460 called although it will not retrieve any additional data. The tag
1461 can be set or retrieved with a ctrl. The IV length is by default 12
1462 bytes (96 bits) but can be set to an alternative value. If the IV
1463 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1467 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1468 underlying do_cipher function handles all cipher semantics itself
1469 including padding and finalisation. This is useful if (for example)
1470 an ENGINE cipher handles block padding itself. The behaviour of
1471 do_cipher is subtly changed if this flag is set: the return value
1472 is the number of characters written to the output buffer (zero is
1473 no longer an error code) or a negative error code. Also if the
1474 input buffer is NULL and length 0 finalisation should be performed.
1477 *) If a candidate issuer certificate is already part of the constructed
1478 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1481 *) Improve forward-security support: add functions
1483 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1484 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1486 for use by SSL/TLS servers; the callback function will be called whenever a
1487 new session is created, and gets to decide whether the session may be
1488 cached to make it resumable (return 0) or not (return 1). (As by the
1489 SSL/TLS protocol specifications, the session_id sent by the server will be
1490 empty to indicate that the session is not resumable; also, the server will
1491 not generate RFC 4507 (RFC 5077) session tickets.)
1493 A simple reasonable callback implementation is to return is_forward_secure.
1494 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1495 by the SSL/TLS server library, indicating whether it can provide forward
1497 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1499 *) New -verify_name option in command line utilities to set verification
1503 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1504 Add CMAC pkey methods.
1507 *) Experimental renegotiation in s_server -www mode. If the client
1508 browses /reneg connection is renegotiated. If /renegcert it is
1509 renegotiated requesting a certificate.
1512 *) Add an "external" session cache for debugging purposes to s_server. This
1513 should help trace issues which normally are only apparent in deployed
1514 multi-process servers.
1517 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1518 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1519 BIO_set_cipher() and some obscure PEM functions were changed so they
1520 can now return an error. The RAND changes required a change to the
1521 RAND_METHOD structure.
1524 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1525 a gcc attribute to warn if the result of a function is ignored. This
1526 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1527 whose return value is often ignored.
1530 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1531 These allow SCTs (signed certificate timestamps) to be requested and
1532 validated when establishing a connection.
1533 [Rob Percival <robpercival@google.com>]
1535 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1537 *) Prevent padding oracle in AES-NI CBC MAC check
1539 A MITM attacker can use a padding oracle attack to decrypt traffic
1540 when the connection uses an AES CBC cipher and the server support
1543 This issue was introduced as part of the fix for Lucky 13 padding
1544 attack (CVE-2013-0169). The padding check was rewritten to be in
1545 constant time by making sure that always the same bytes are read and
1546 compared against either the MAC or padding bytes. But it no longer
1547 checked that there was enough data to have both the MAC and padding
1550 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1554 *) Fix EVP_EncodeUpdate overflow
1556 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1557 Base64 encoding of binary data. If an attacker is able to supply very large
1558 amounts of input data then a length check can overflow resulting in a heap
1561 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1562 the PEM_write_bio* family of functions. These are mainly used within the
1563 OpenSSL command line applications, so any application which processes data
1564 from an untrusted source and outputs it as a PEM file should be considered
1565 vulnerable to this issue. User applications that call these APIs directly
1566 with large amounts of untrusted data may also be vulnerable.
1568 This issue was reported by Guido Vranken.
1572 *) Fix EVP_EncryptUpdate overflow
1574 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1575 is able to supply very large amounts of input data after a previous call to
1576 EVP_EncryptUpdate() with a partial block then a length check can overflow
1577 resulting in a heap corruption. Following an analysis of all OpenSSL
1578 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1579 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1580 the first called function after an EVP_EncryptInit(), and therefore that
1581 specific call must be safe. The second form is where the length passed to
1582 EVP_EncryptUpdate() can be seen from the code to be some small value and
1583 therefore there is no possibility of an overflow. Since all instances are
1584 one of these two forms, it is believed that there can be no overflows in
1585 internal code due to this problem. It should be noted that
1586 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1587 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1588 of these calls have also been analysed too and it is believed there are no
1589 instances in internal usage where an overflow could occur.
1591 This issue was reported by Guido Vranken.
1595 *) Prevent ASN.1 BIO excessive memory allocation
1597 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1598 a short invalid encoding can cause allocation of large amounts of memory
1599 potentially consuming excessive resources or exhausting memory.
1601 Any application parsing untrusted data through d2i BIO functions is
1602 affected. The memory based functions such as d2i_X509() are *not* affected.
1603 Since the memory based functions are used by the TLS library, TLS
1604 applications are not affected.
1606 This issue was reported by Brian Carpenter.
1612 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1613 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1614 in arbitrary stack data being returned in the buffer.
1616 This issue was reported by Guido Vranken.
1620 *) Modify behavior of ALPN to invoke callback after SNI/servername
1621 callback, such that updates to the SSL_CTX affect ALPN.
1624 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1628 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1629 methods are enabled and ssl2 is disabled the methods return NULL.
1632 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1634 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1635 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1636 provide any "EXPORT" or "LOW" strength ciphers.
1639 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1640 is by default disabled at build-time. Builds that are not configured with
1641 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1642 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1643 will need to explicitly call either of:
1645 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1647 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1649 as appropriate. Even if either of those is used, or the application
1650 explicitly uses the version-specific SSLv2_method() or its client and
1651 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1652 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1653 ciphers, and SSLv2 56-bit DES are no longer available.
1657 *) Fix a double-free in DSA code
1659 A double free bug was discovered when OpenSSL parses malformed DSA private
1660 keys and could lead to a DoS attack or memory corruption for applications
1661 that receive DSA private keys from untrusted sources. This scenario is
1664 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1669 *) Disable SRP fake user seed to address a server memory leak.
1671 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1673 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1674 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1675 was changed to ignore the "fake user" SRP seed, even if the seed
1678 Users should use SRP_VBASE_get1_by_user instead. Note that in
1679 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1680 also that even though configuring the SRP seed attempts to hide
1681 invalid usernames by continuing the handshake with fake
1682 credentials, this behaviour is not constant time and no strong
1683 guarantees are made that the handshake is indistinguishable from
1684 that of a valid user.
1688 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1690 In the BN_hex2bn function the number of hex digits is calculated using an
1691 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1692 large values of |i| this can result in |bn_expand| not allocating any
1693 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1694 field as NULL leading to a subsequent NULL ptr deref. For very large values
1695 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1696 In this case memory is allocated to the internal BIGNUM data field, but it
1697 is insufficiently sized leading to heap corruption. A similar issue exists
1698 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1699 is ever called by user applications with very large untrusted hex/dec data.
1700 This is anticipated to be a rare occurrence.
1702 All OpenSSL internal usage of these functions use data that is not expected
1703 to be untrusted, e.g. config file data or application command line
1704 arguments. If user developed applications generate config file data based
1705 on untrusted data then it is possible that this could also lead to security
1706 consequences. This is also anticipated to be rare.
1708 This issue was reported to OpenSSL by Guido Vranken.
1712 *) Fix memory issues in BIO_*printf functions
1714 The internal |fmtstr| function used in processing a "%s" format string in
1715 the BIO_*printf functions could overflow while calculating the length of a
1716 string and cause an OOB read when printing very long strings.
1718 Additionally the internal |doapr_outch| function can attempt to write to an
1719 OOB memory location (at an offset from the NULL pointer) in the event of a
1720 memory allocation failure. In 1.0.2 and below this could be caused where
1721 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1722 could be in processing a very long "%s" format string. Memory leaks can
1725 The first issue may mask the second issue dependent on compiler behaviour.
1726 These problems could enable attacks where large amounts of untrusted data
1727 is passed to the BIO_*printf functions. If applications use these functions
1728 in this way then they could be vulnerable. OpenSSL itself uses these
1729 functions when printing out human-readable dumps of ASN.1 data. Therefore
1730 applications that print this data could be vulnerable if the data is from
1731 untrusted sources. OpenSSL command line applications could also be
1732 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1733 as command line arguments.
1735 Libssl is not considered directly vulnerable. Additionally certificates etc
1736 received via remote connections via libssl are also unlikely to be able to
1737 trigger these issues because of message size limits enforced within libssl.
1739 This issue was reported to OpenSSL Guido Vranken.
1743 *) Side channel attack on modular exponentiation
1745 A side-channel attack was found which makes use of cache-bank conflicts on
1746 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1747 of RSA keys. The ability to exploit this issue is limited as it relies on
1748 an attacker who has control of code in a thread running on the same
1749 hyper-threaded core as the victim thread which is performing decryptions.
1751 This issue was reported to OpenSSL by Yuval Yarom, The University of
1752 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1753 Nadia Heninger, University of Pennsylvania with more information at
1754 http://cachebleed.info.
1758 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1759 if no keysize is specified with default_bits. This fixes an
1760 omission in an earlier change that changed all RSA/DSA key generation
1761 apps to use 2048 bits by default.
1764 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1765 *) DH small subgroups
1767 Historically OpenSSL only ever generated DH parameters based on "safe"
1768 primes. More recently (in version 1.0.2) support was provided for
1769 generating X9.42 style parameter files such as those required for RFC 5114
1770 support. The primes used in such files may not be "safe". Where an
1771 application is using DH configured with parameters based on primes that are
1772 not "safe" then an attacker could use this fact to find a peer's private
1773 DH exponent. This attack requires that the attacker complete multiple
1774 handshakes in which the peer uses the same private DH exponent. For example
1775 this could be used to discover a TLS server's private DH exponent if it's
1776 reusing the private DH exponent or it's using a static DH ciphersuite.
1778 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1779 TLS. It is not on by default. If the option is not set then the server
1780 reuses the same private DH exponent for the life of the server process and
1781 would be vulnerable to this attack. It is believed that many popular
1782 applications do set this option and would therefore not be at risk.
1784 The fix for this issue adds an additional check where a "q" parameter is
1785 available (as is the case in X9.42 based parameters). This detects the
1786 only known attack, and is the only possible defense for static DH
1787 ciphersuites. This could have some performance impact.
1789 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1790 default and cannot be disabled. This could have some performance impact.
1792 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1796 *) SSLv2 doesn't block disabled ciphers
1798 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1799 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1800 been disabled, provided that the SSLv2 protocol was not also disabled via
1803 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1804 and Sebastian Schinzel.
1808 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1810 *) BN_mod_exp may produce incorrect results on x86_64
1812 There is a carry propagating bug in the x86_64 Montgomery squaring
1813 procedure. No EC algorithms are affected. Analysis suggests that attacks
1814 against RSA and DSA as a result of this defect would be very difficult to
1815 perform and are not believed likely. Attacks against DH are considered just
1816 feasible (although very difficult) because most of the work necessary to
1817 deduce information about a private key may be performed offline. The amount
1818 of resources required for such an attack would be very significant and
1819 likely only accessible to a limited number of attackers. An attacker would
1820 additionally need online access to an unpatched system using the target
1821 private key in a scenario with persistent DH parameters and a private
1822 key that is shared between multiple clients. For example this can occur by
1823 default in OpenSSL DHE based SSL/TLS ciphersuites.
1825 This issue was reported to OpenSSL by Hanno Böck.
1829 *) Certificate verify crash with missing PSS parameter
1831 The signature verification routines will crash with a NULL pointer
1832 dereference if presented with an ASN.1 signature using the RSA PSS
1833 algorithm and absent mask generation function parameter. Since these
1834 routines are used to verify certificate signature algorithms this can be
1835 used to crash any certificate verification operation and exploited in a
1836 DoS attack. Any application which performs certificate verification is
1837 vulnerable including OpenSSL clients and servers which enable client
1840 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1844 *) X509_ATTRIBUTE memory leak
1846 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1847 memory. This structure is used by the PKCS#7 and CMS routines so any
1848 application which reads PKCS#7 or CMS data from untrusted sources is
1849 affected. SSL/TLS is not affected.
1851 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1856 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1857 This changes the decoding behaviour for some invalid messages,
1858 though the change is mostly in the more lenient direction, and
1859 legacy behaviour is preserved as much as possible.
1862 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1864 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1866 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1868 *) Alternate chains certificate forgery
1870 During certificate verification, OpenSSL will attempt to find an
1871 alternative certificate chain if the first attempt to build such a chain
1872 fails. An error in the implementation of this logic can mean that an
1873 attacker could cause certain checks on untrusted certificates to be
1874 bypassed, such as the CA flag, enabling them to use a valid leaf
1875 certificate to act as a CA and "issue" an invalid certificate.
1877 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1881 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1883 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1884 incompatibility in the handling of HMAC. The previous ABI has now been
1888 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1890 *) Malformed ECParameters causes infinite loop
1892 When processing an ECParameters structure OpenSSL enters an infinite loop
1893 if the curve specified is over a specially malformed binary polynomial
1896 This can be used to perform denial of service against any
1897 system which processes public keys, certificate requests or
1898 certificates. This includes TLS clients and TLS servers with
1899 client authentication enabled.
1901 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1905 *) Exploitable out-of-bounds read in X509_cmp_time
1907 X509_cmp_time does not properly check the length of the ASN1_TIME
1908 string and can read a few bytes out of bounds. In addition,
1909 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1912 An attacker can use this to craft malformed certificates and CRLs of
1913 various sizes and potentially cause a segmentation fault, resulting in
1914 a DoS on applications that verify certificates or CRLs. TLS clients
1915 that verify CRLs are affected. TLS clients and servers with client
1916 authentication enabled may be affected if they use custom verification
1919 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1920 independently by Hanno Böck.
1924 *) PKCS7 crash with missing EnvelopedContent
1926 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1927 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1928 with missing content and trigger a NULL pointer dereference on parsing.
1930 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1931 structures from untrusted sources are affected. OpenSSL clients and
1932 servers are not affected.
1934 This issue was reported to OpenSSL by Michal Zalewski (Google).
1938 *) CMS verify infinite loop with unknown hash function
1940 When verifying a signedData message the CMS code can enter an infinite loop
1941 if presented with an unknown hash function OID. This can be used to perform
1942 denial of service against any system which verifies signedData messages using
1944 This issue was reported to OpenSSL by Johannes Bauer.
1948 *) Race condition handling NewSessionTicket
1950 If a NewSessionTicket is received by a multi-threaded client when attempting to
1951 reuse a previous ticket then a race condition can occur potentially leading to
1952 a double free of the ticket data.
1956 *) Only support 256-bit or stronger elliptic curves with the
1957 'ecdh_auto' setting (server) or by default (client). Of supported
1958 curves, prefer P-256 (both).
1961 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1963 *) ClientHello sigalgs DoS fix
1965 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1966 invalid signature algorithms extension a NULL pointer dereference will
1967 occur. This can be exploited in a DoS attack against the server.
1969 This issue was was reported to OpenSSL by David Ramos of Stanford
1972 [Stephen Henson and Matt Caswell]
1974 *) Multiblock corrupted pointer fix
1976 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1977 feature only applies on 64 bit x86 architecture platforms that support AES
1978 NI instructions. A defect in the implementation of "multiblock" can cause
1979 OpenSSL's internal write buffer to become incorrectly set to NULL when
1980 using non-blocking IO. Typically, when the user application is using a
1981 socket BIO for writing, this will only result in a failed connection.
1982 However if some other BIO is used then it is likely that a segmentation
1983 fault will be triggered, thus enabling a potential DoS attack.
1985 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1989 *) Segmentation fault in DTLSv1_listen fix
1991 The DTLSv1_listen function is intended to be stateless and processes the
1992 initial ClientHello from many peers. It is common for user code to loop
1993 over the call to DTLSv1_listen until a valid ClientHello is received with
1994 an associated cookie. A defect in the implementation of DTLSv1_listen means
1995 that state is preserved in the SSL object from one invocation to the next
1996 that can lead to a segmentation fault. Errors processing the initial
1997 ClientHello can trigger this scenario. An example of such an error could be
1998 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2001 This issue was reported to OpenSSL by Per Allansson.
2005 *) Segmentation fault in ASN1_TYPE_cmp fix
2007 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2008 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2009 certificate signature algorithm consistency this can be used to crash any
2010 certificate verification operation and exploited in a DoS attack. Any
2011 application which performs certificate verification is vulnerable including
2012 OpenSSL clients and servers which enable client authentication.
2016 *) Segmentation fault for invalid PSS parameters fix
2018 The signature verification routines will crash with a NULL pointer
2019 dereference if presented with an ASN.1 signature using the RSA PSS
2020 algorithm and invalid parameters. Since these routines are used to verify
2021 certificate signature algorithms this can be used to crash any
2022 certificate verification operation and exploited in a DoS attack. Any
2023 application which performs certificate verification is vulnerable including
2024 OpenSSL clients and servers which enable client authentication.
2026 This issue was was reported to OpenSSL by Brian Carpenter.
2030 *) ASN.1 structure reuse memory corruption fix
2032 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2033 memory corruption via an invalid write. Such reuse is and has been
2034 strongly discouraged and is believed to be rare.
2036 Applications that parse structures containing CHOICE or ANY DEFINED BY
2037 components may be affected. Certificate parsing (d2i_X509 and related
2038 functions) are however not affected. OpenSSL clients and servers are
2043 *) PKCS7 NULL pointer dereferences fix
2045 The PKCS#7 parsing code does not handle missing outer ContentInfo
2046 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2047 missing content and trigger a NULL pointer dereference on parsing.
2049 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2050 otherwise parse PKCS#7 structures from untrusted sources are
2051 affected. OpenSSL clients and servers are not affected.
2053 This issue was reported to OpenSSL by Michal Zalewski (Google).
2057 *) DoS via reachable assert in SSLv2 servers fix
2059 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2060 servers that both support SSLv2 and enable export cipher suites by sending
2061 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2063 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2064 (OpenSSL development team).
2068 *) Empty CKE with client auth and DHE fix
2070 If client auth is used then a server can seg fault in the event of a DHE
2071 ciphersuite being selected and a zero length ClientKeyExchange message
2072 being sent by the client. This could be exploited in a DoS attack.
2076 *) Handshake with unseeded PRNG fix
2078 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2079 with an unseeded PRNG. The conditions are:
2080 - The client is on a platform where the PRNG has not been seeded
2081 automatically, and the user has not seeded manually
2082 - A protocol specific client method version has been used (i.e. not
2083 SSL_client_methodv23)
2084 - A ciphersuite is used that does not require additional random data from
2085 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2087 If the handshake succeeds then the client random that has been used will
2088 have been generated from a PRNG with insufficient entropy and therefore the
2089 output may be predictable.
2091 For example using the following command with an unseeded openssl will
2092 succeed on an unpatched platform:
2094 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2098 *) Use After Free following d2i_ECPrivatekey error fix
2100 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2101 could cause a use after free condition. This, in turn, could cause a double
2102 free in several private key parsing functions (such as d2i_PrivateKey
2103 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2104 for applications that receive EC private keys from untrusted
2105 sources. This scenario is considered rare.
2107 This issue was discovered by the BoringSSL project and fixed in their
2112 *) X509_to_X509_REQ NULL pointer deref fix
2114 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2115 the certificate key is invalid. This function is rarely used in practice.
2117 This issue was discovered by Brian Carpenter.
2121 *) Removed the export ciphers from the DEFAULT ciphers
2124 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2126 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2127 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2128 So far those who have to target multiple platforms would compromise
2129 and argue that binary targeting say ARMv5 would still execute on
2130 ARMv8. "Universal" build resolves this compromise by providing
2131 near-optimal performance even on newer platforms.
2134 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2135 (other platforms pending).
2136 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2138 *) Add support for the SignedCertificateTimestampList certificate and
2139 OCSP response extensions from RFC6962.
2142 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2143 for corner cases. (Certain input points at infinity could lead to
2144 bogus results, with non-infinity inputs mapped to infinity too.)
2147 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2148 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2149 common cases are optimized and there still is room for further
2150 improvements. Vector Permutation AES for Altivec is also added.
2153 *) Add support for little-endian ppc64 Linux target.
2154 [Marcelo Cerri (IBM)]
2156 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2157 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2158 are optimized and there still is room for further improvements.
2159 Both 32- and 64-bit modes are supported.
2160 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2162 *) Improved ARMv7 NEON support.
2165 *) Support for SPARC Architecture 2011 crypto extensions, first
2166 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2167 SHA256/512, MD5, GHASH and modular exponentiation.
2168 [Andy Polyakov, David Miller]
2170 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2172 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2174 *) Support for new and upcoming Intel processors, including AVX2,
2175 BMI and SHA ISA extensions. This includes additional "stitched"
2176 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2179 This work was sponsored by Intel Corp.
2182 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2183 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2184 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2187 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2188 this fixes a limitation in previous versions of OpenSSL.
2191 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2192 MGF1 digest and OAEP label.
2195 *) Add EVP support for key wrapping algorithms, to avoid problems with
2196 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2197 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2198 algorithms and include tests cases.
2201 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2203 [Douglas E. Engert, Steve Henson]
2205 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2206 difference in days and seconds between two tm or ASN1_TIME structures.
2209 *) Add -rev test option to s_server to just reverse order of characters
2210 received by client and send back to server. Also prints an abbreviated
2211 summary of the connection parameters.
2214 *) New option -brief for s_client and s_server to print out a brief summary
2215 of connection parameters.
2218 *) Add callbacks for arbitrary TLS extensions.
2219 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2221 *) New option -crl_download in several openssl utilities to download CRLs
2222 from CRLDP extension in certificates.
2225 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2228 *) New function X509_CRL_diff to generate a delta CRL from the difference
2229 of two full CRLs. Add support to "crl" utility.
2232 *) New functions to set lookup_crls function and to retrieve
2233 X509_STORE from X509_STORE_CTX.
2236 *) Print out deprecated issuer and subject unique ID fields in
2240 *) Extend OCSP I/O functions so they can be used for simple general purpose
2241 HTTP as well as OCSP. New wrapper function which can be used to download
2242 CRLs using the OCSP API.
2245 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2248 *) SSL_CONF* functions. These provide a common framework for application
2249 configuration using configuration files or command lines.
2252 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2253 message callback and prints the results. Needs compile time option
2254 "enable-ssl-trace". New options to s_client and s_server to enable
2258 *) New ctrl and macro to retrieve supported points extensions.
2259 Print out extension in s_server and s_client.
2262 *) New functions to retrieve certificate signature and signature
2266 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2270 *) New Suite B modes for TLS code. These use and enforce the requirements
2271 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2272 only use Suite B curves. The Suite B modes can be set by using the
2273 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2276 *) New chain verification flags for Suite B levels of security. Check
2277 algorithms are acceptable when flags are set in X509_verify_cert.
2280 *) Make tls1_check_chain return a set of flags indicating checks passed
2281 by a certificate chain. Add additional tests to handle client
2282 certificates: checks for matching certificate type and issuer name
2286 *) If an attempt is made to use a signature algorithm not in the peer
2287 preference list abort the handshake. If client has no suitable
2288 signature algorithms in response to a certificate request do not
2289 use the certificate.
2292 *) If server EC tmp key is not in client preference list abort handshake.
2295 *) Add support for certificate stores in CERT structure. This makes it
2296 possible to have different stores per SSL structure or one store in
2297 the parent SSL_CTX. Include distinct stores for certificate chain
2298 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2299 to build and store a certificate chain in CERT structure: returning
2300 an error if the chain cannot be built: this will allow applications
2301 to test if a chain is correctly configured.
2303 Note: if the CERT based stores are not set then the parent SSL_CTX
2304 store is used to retain compatibility with existing behaviour.
2308 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2309 mask based on the current session, check mask when sending client
2310 hello and checking the requested ciphersuite.
2313 *) New ctrls to retrieve and set certificate types in a certificate
2314 request message. Print out received values in s_client. If certificate
2315 types is not set with custom values set sensible values based on
2316 supported signature algorithms.
2319 *) Support for distinct client and server supported signature algorithms.
2322 *) Add certificate callback. If set this is called whenever a certificate
2323 is required by client or server. An application can decide which
2324 certificate chain to present based on arbitrary criteria: for example
2325 supported signature algorithms. Add very simple example to s_server.
2326 This fixes many of the problems and restrictions of the existing client
2327 certificate callback: for example you can now clear an existing
2328 certificate and specify the whole chain.
2331 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2332 the certificate can be used for (if anything). Set valid_flags field
2333 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2334 to have similar checks in it.
2336 Add new "cert_flags" field to CERT structure and include a "strict mode".
2337 This enforces some TLS certificate requirements (such as only permitting
2338 certificate signature algorithms contained in the supported algorithms
2339 extension) which some implementations ignore: this option should be used
2340 with caution as it could cause interoperability issues.
2343 *) Update and tidy signature algorithm extension processing. Work out
2344 shared signature algorithms based on preferences and peer algorithms
2345 and print them out in s_client and s_server. Abort handshake if no
2346 shared signature algorithms.
2349 *) Add new functions to allow customised supported signature algorithms
2350 for SSL and SSL_CTX structures. Add options to s_client and s_server
2354 *) New function SSL_certs_clear() to delete all references to certificates
2355 from an SSL structure. Before this once a certificate had been added
2356 it couldn't be removed.
2359 *) Integrate hostname, email address and IP address checking with certificate
2360 verification. New verify options supporting checking in openssl utility.
2363 *) Fixes and wildcard matching support to hostname and email checking
2364 functions. Add manual page.
2365 [Florian Weimer (Red Hat Product Security Team)]
2367 *) New functions to check a hostname email or IP address against a
2368 certificate. Add options x509 utility to print results of checks against
2372 *) Fix OCSP checking.
2373 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2375 *) Initial experimental support for explicitly trusted non-root CAs.
2376 OpenSSL still tries to build a complete chain to a root but if an
2377 intermediate CA has a trust setting included that is used. The first
2378 setting is used: whether to trust (e.g., -addtrust option to the x509
2382 *) Add -trusted_first option which attempts to find certificates in the
2383 trusted store even if an untrusted chain is also supplied.
2386 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2387 platform support for Linux and Android.
2390 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2393 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2394 When in FIPS mode the approved implementations are used as normal,
2395 when not in FIPS mode the internal unapproved versions are used instead.
2396 This means that the FIPS capable OpenSSL isn't forced to use the
2397 (often lower performance) FIPS implementations outside FIPS mode.
2400 *) Transparently support X9.42 DH parameters when calling
2401 PEM_read_bio_DHparameters. This means existing applications can handle
2402 the new parameter format automatically.
2405 *) Initial experimental support for X9.42 DH parameter format: mainly
2406 to support use of 'q' parameter for RFC5114 parameters.
2409 *) Add DH parameters from RFC5114 including test data to dhtest.
2412 *) Support for automatic EC temporary key parameter selection. If enabled
2413 the most preferred EC parameters are automatically used instead of
2414 hardcoded fixed parameters. Now a server just has to call:
2415 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2416 support ECDH and use the most appropriate parameters.
2419 *) Enhance and tidy EC curve and point format TLS extension code. Use
2420 static structures instead of allocation if default values are used.
2421 New ctrls to set curves we wish to support and to retrieve shared curves.
2422 Print out shared curves in s_server. New options to s_server and s_client
2423 to set list of supported curves.
2426 *) New ctrls to retrieve supported signature algorithms and
2427 supported curve values as an array of NIDs. Extend openssl utility
2428 to print out received values.
2431 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2432 between NIDs and the more common NIST names such as "P-256". Enhance
2433 ecparam utility and ECC method to recognise the NIST names for curves.
2436 *) Enhance SSL/TLS certificate chain handling to support different
2437 chains for each certificate instead of one chain in the parent SSL_CTX.
2440 *) Support for fixed DH ciphersuite client authentication: where both
2441 server and client use DH certificates with common parameters.
2444 *) Support for fixed DH ciphersuites: those requiring DH server
2448 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2450 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2451 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2452 X509_CINF_get_signature were reverted post internal team review.
2454 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2456 *) Build fixes for the Windows and OpenVMS platforms
2457 [Matt Caswell and Richard Levitte]
2459 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2461 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2462 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2463 dereference. This could lead to a Denial Of Service attack. Thanks to
2464 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2468 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2469 dtls1_buffer_record function under certain conditions. In particular this
2470 could occur if an attacker sent repeated DTLS records with the same
2471 sequence number but for the next epoch. The memory leak could be exploited
2472 by an attacker in a Denial of Service attack through memory exhaustion.
2473 Thanks to Chris Mueller for reporting this issue.
2477 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2478 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2479 method would be set to NULL which could later result in a NULL pointer
2480 dereference. Thanks to Frank Schmirler for reporting this issue.
2484 *) Abort handshake if server key exchange message is omitted for ephemeral
2487 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2488 reporting this issue.
2492 *) Remove non-export ephemeral RSA code on client and server. This code
2493 violated the TLS standard by allowing the use of temporary RSA keys in
2494 non-export ciphersuites and could be used by a server to effectively
2495 downgrade the RSA key length used to a value smaller than the server
2496 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2497 INRIA or reporting this issue.
2501 *) Fixed issue where DH client certificates are accepted without verification.
2502 An OpenSSL server will accept a DH certificate for client authentication
2503 without the certificate verify message. This effectively allows a client to
2504 authenticate without the use of a private key. This only affects servers
2505 which trust a client certificate authority which issues certificates
2506 containing DH keys: these are extremely rare and hardly ever encountered.
2507 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2512 *) Ensure that the session ID context of an SSL is updated when its
2513 SSL_CTX is updated via SSL_set_SSL_CTX.
2515 The session ID context is typically set from the parent SSL_CTX,
2516 and can vary with the CTX.
2519 *) Fix various certificate fingerprint issues.
2521 By using non-DER or invalid encodings outside the signed portion of a
2522 certificate the fingerprint can be changed without breaking the signature.
2523 Although no details of the signed portion of the certificate can be changed
2524 this can cause problems with some applications: e.g. those using the
2525 certificate fingerprint for blacklists.
2527 1. Reject signatures with non zero unused bits.
2529 If the BIT STRING containing the signature has non zero unused bits reject
2530 the signature. All current signature algorithms require zero unused bits.
2532 2. Check certificate algorithm consistency.
2534 Check the AlgorithmIdentifier inside TBS matches the one in the
2535 certificate signature. NB: this will result in signature failure
2536 errors for some broken certificates.
2538 Thanks to Konrad Kraszewski from Google for reporting this issue.
2540 3. Check DSA/ECDSA signatures use DER.
2542 Re-encode DSA/ECDSA signatures and compare with the original received
2543 signature. Return an error if there is a mismatch.
2545 This will reject various cases including garbage after signature
2546 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2547 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2548 (negative or with leading zeroes).
2550 Further analysis was conducted and fixes were developed by Stephen Henson
2551 of the OpenSSL core team.
2556 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2557 results on some platforms, including x86_64. This bug occurs at random
2558 with a very low probability, and is not known to be exploitable in any
2559 way, though its exact impact is difficult to determine. Thanks to Pieter
2560 Wuille (Blockstream) who reported this issue and also suggested an initial
2561 fix. Further analysis was conducted by the OpenSSL development team and
2562 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2563 the OpenSSL core team.
2567 *) Do not resume sessions on the server if the negotiated protocol
2568 version does not match the session's version. Resuming with a different
2569 version, while not strictly forbidden by the RFC, is of questionable
2570 sanity and breaks all known clients.
2571 [David Benjamin, Emilia Käsper]
2573 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2574 early CCS messages during renegotiation. (Note that because
2575 renegotiation is encrypted, this early CCS was not exploitable.)
2578 *) Tighten client-side session ticket handling during renegotiation:
2579 ensure that the client only accepts a session ticket if the server sends
2580 the extension anew in the ServerHello. Previously, a TLS client would
2581 reuse the old extension state and thus accept a session ticket if one was
2582 announced in the initial ServerHello.
2584 Similarly, ensure that the client requires a session ticket if one
2585 was advertised in the ServerHello. Previously, a TLS client would
2586 ignore a missing NewSessionTicket message.
2589 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2591 *) SRTP Memory Leak.
2593 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2594 sends a carefully crafted handshake message, to cause OpenSSL to fail
2595 to free up to 64k of memory causing a memory leak. This could be
2596 exploited in a Denial Of Service attack. This issue affects OpenSSL
2597 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2598 whether SRTP is used or configured. Implementations of OpenSSL that
2599 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2601 The fix was developed by the OpenSSL team.
2605 *) Session Ticket Memory Leak.
2607 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2608 integrity of that ticket is first verified. In the event of a session
2609 ticket integrity check failing, OpenSSL will fail to free memory
2610 causing a memory leak. By sending a large number of invalid session
2611 tickets an attacker could exploit this issue in a Denial Of Service
2616 *) Build option no-ssl3 is incomplete.
2618 When OpenSSL is configured with "no-ssl3" as a build option, servers
2619 could accept and complete a SSL 3.0 handshake, and clients could be
2620 configured to send them.
2622 [Akamai and the OpenSSL team]
2624 *) Add support for TLS_FALLBACK_SCSV.
2625 Client applications doing fallback retries should call
2626 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2628 [Adam Langley, Bodo Moeller]
2630 *) Add additional DigestInfo checks.
2632 Re-encode DigestInto in DER and check against the original when
2633 verifying RSA signature: this will reject any improperly encoded
2634 DigestInfo structures.
2636 Note: this is a precautionary measure and no attacks are currently known.
2640 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2642 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2643 SRP code can be overrun an internal buffer. Add sanity check that
2644 g, A, B < N to SRP code.
2646 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2647 Group for discovering this issue.
2651 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2652 TLS 1.0 instead of higher protocol versions when the ClientHello message
2653 is badly fragmented. This allows a man-in-the-middle attacker to force a
2654 downgrade to TLS 1.0 even if both the server and the client support a
2655 higher protocol version, by modifying the client's TLS records.
2657 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2658 researching this issue.
2662 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2663 to a denial of service attack. A malicious server can crash the client
2664 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2665 ciphersuite and sending carefully crafted handshake messages.
2667 Thanks to Felix Gröbert (Google) for discovering and researching this
2672 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2673 to leak memory. This can be exploited through a Denial of Service attack.
2674 Thanks to Adam Langley for discovering and researching this issue.
2678 *) An attacker can force openssl to consume large amounts of memory whilst
2679 processing DTLS handshake messages. This can be exploited through a
2680 Denial of Service attack.
2681 Thanks to Adam Langley for discovering and researching this issue.
2685 *) An attacker can force an error condition which causes openssl to crash
2686 whilst processing DTLS packets due to memory being freed twice. This
2687 can be exploited through a Denial of Service attack.
2688 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2693 *) If a multithreaded client connects to a malicious server using a resumed
2694 session and the server sends an ec point format extension it could write
2695 up to 255 bytes to freed memory.
2697 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2702 *) A malicious server can crash an OpenSSL client with a null pointer
2703 dereference (read) by specifying an SRP ciphersuite even though it was not
2704 properly negotiated with the client. This can be exploited through a
2705 Denial of Service attack.
2707 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2708 discovering and researching this issue.
2712 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2713 X509_name_oneline, X509_name_print_ex et al. to leak some information
2714 from the stack. Applications may be affected if they echo pretty printing
2715 output to the attacker.
2717 Thanks to Ivan Fratric (Google) for discovering this issue.
2719 [Emilia Käsper, and Steve Henson]
2721 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2722 for corner cases. (Certain input points at infinity could lead to
2723 bogus results, with non-infinity inputs mapped to infinity too.)
2726 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2728 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2729 handshake can force the use of weak keying material in OpenSSL
2730 SSL/TLS clients and servers.
2732 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2733 researching this issue. (CVE-2014-0224)
2734 [KIKUCHI Masashi, Steve Henson]
2736 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2737 OpenSSL DTLS client the code can be made to recurse eventually crashing
2740 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2742 [Imre Rad, Steve Henson]
2744 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2745 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2746 client or server. This is potentially exploitable to run arbitrary
2747 code on a vulnerable client or server.
2749 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2750 [Jüri Aedla, Steve Henson]
2752 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2753 are subject to a denial of service attack.
2755 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2756 this issue. (CVE-2014-3470)
2757 [Felix Gröbert, Ivan Fratric, Steve Henson]
2759 *) Harmonize version and its documentation. -f flag is used to display
2761 [mancha <mancha1@zoho.com>]
2763 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2764 in i2d_ECPrivateKey.
2765 [mancha <mancha1@zoho.com>]
2767 *) Fix some double frees. These are not thought to be exploitable.
2768 [mancha <mancha1@zoho.com>]
2770 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2772 *) A missing bounds check in the handling of the TLS heartbeat extension
2773 can be used to reveal up to 64k of memory to a connected client or
2776 Thanks for Neel Mehta of Google Security for discovering this bug and to
2777 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2778 preparing the fix (CVE-2014-0160)
2779 [Adam Langley, Bodo Moeller]
2781 *) Fix for the attack described in the paper "Recovering OpenSSL
2782 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2783 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2784 http://eprint.iacr.org/2014/140
2786 Thanks to Yuval Yarom and Naomi Benger for discovering this
2787 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2788 [Yuval Yarom and Naomi Benger]
2790 *) TLS pad extension: draft-agl-tls-padding-03
2792 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2793 TLS client Hello record length value would otherwise be > 255 and
2794 less that 512 pad with a dummy extension containing zeroes so it
2795 is at least 512 bytes long.
2797 [Adam Langley, Steve Henson]
2799 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2801 *) Fix for TLS record tampering bug. A carefully crafted invalid
2802 handshake could crash OpenSSL with a NULL pointer exception.
2803 Thanks to Anton Johansson for reporting this issues.
2806 *) Keep original DTLS digest and encryption contexts in retransmission
2807 structures so we can use the previous session parameters if they need
2808 to be resent. (CVE-2013-6450)
2811 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2812 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2813 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2814 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2815 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2816 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2817 [Rob Stradling, Adam Langley]
2819 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2821 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2822 supporting platforms or when small records were transferred.
2823 [Andy Polyakov, Steve Henson]
2825 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2827 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2829 This addresses the flaw in CBC record processing discovered by
2830 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2831 at: http://www.isg.rhul.ac.uk/tls/
2833 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2834 Security Group at Royal Holloway, University of London
2835 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2836 Emilia Käsper for the initial patch.
2838 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2840 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2841 ciphersuites which can be exploited in a denial of service attack.
2842 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2843 and detecting this bug and to Wolfgang Ettlinger
2844 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2848 *) Return an error when checking OCSP signatures when key is NULL.
2849 This fixes a DoS attack. (CVE-2013-0166)
2852 *) Make openssl verify return errors.
2853 [Chris Palmer <palmer@google.com> and Ben Laurie]
2855 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2856 the right response is stapled. Also change SSL_get_certificate()
2857 so it returns the certificate actually sent.
2858 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2859 [Rob Stradling <rob.stradling@comodo.com>]
2861 *) Fix possible deadlock when decoding public keys.
2864 *) Don't use TLS 1.0 record version number in initial client hello
2868 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2870 *) Sanity check record length before skipping explicit IV in TLS
2871 1.2, 1.1 and DTLS to fix DoS attack.
2873 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2874 fuzzing as a service testing platform.
2878 *) Initialise tkeylen properly when encrypting CMS messages.
2879 Thanks to Solar Designer of Openwall for reporting this issue.
2882 *) In FIPS mode don't try to use composite ciphers as they are not
2886 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2888 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2889 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2890 mean any application compiled against OpenSSL 1.0.0 headers setting
2891 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
2892 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2893 0x10000000L Any application which was previously compiled against
2894 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2895 will need to be recompiled as a result. Letting be results in
2896 inability to disable specifically TLS 1.1 and in client context,
2897 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2900 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
2901 disable just protocol X, but all protocols above X *if* there are
2902 protocols *below* X still enabled. In more practical terms it means
2903 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2904 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2905 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2909 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2911 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2912 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2913 in CRYPTO_realloc_clean.
2915 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2916 issue and to Adam Langley <agl@chromium.org> for fixing it.
2918 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2920 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2923 *) Workarounds for some broken servers that "hang" if a client hello
2924 record length exceeds 255 bytes.
2926 1. Do not use record version number > TLS 1.0 in initial client
2927 hello: some (but not all) hanging servers will now work.
2928 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2929 the number of ciphers sent in the client hello. This should be
2930 set to an even number, such as 50, for example by passing:
2931 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2932 Most broken servers should now work.
2933 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2934 TLS 1.2 client support entirely.
2937 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2940 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2942 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2943 STRING form instead of a DigestInfo.
2946 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2947 and the RSA_sign/RSA_verify functions. This was made more apparent when
2948 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2949 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2950 the correct format in RSA_verify so both forms transparently work.
2953 *) Some servers which support TLS 1.0 can choke if we initially indicate
2954 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2955 encrypted premaster secret. As a workaround use the maximum permitted
2956 client version in client hello, this should keep such servers happy
2957 and still work with previous versions of OpenSSL.
2960 *) Add support for TLS/DTLS heartbeats.
2961 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2963 *) Add support for SCTP.
2964 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2966 *) Improved PRNG seeding for VOS.
2967 [Paul Green <Paul.Green@stratus.com>]
2969 *) Extensive assembler packs updates, most notably:
2971 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2972 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2973 - x86_64: bit-sliced AES implementation;
2974 - ARM: NEON support, contemporary platforms optimizations;
2975 - s390x: z196 support;
2976 - *: GHASH and GF(2^m) multiplication implementations;
2980 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2981 (removal of unnecessary code)
2982 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2984 *) Add TLS key material exporter from RFC 5705.
2987 *) Add DTLS-SRTP negotiation from RFC 5764.
2990 *) Add Next Protocol Negotiation,
2991 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2992 disabled with a no-npn flag to config or Configure. Code donated
2994 [Adam Langley <agl@google.com> and Ben Laurie]
2996 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2997 NIST-P256, NIST-P521, with constant-time single point multiplication on
2998 typical inputs. Compiler support for the nonstandard type __uint128_t is
2999 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3000 Code made available under Apache License version 2.0.
3002 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3003 line to include this in your build of OpenSSL, and run "make depend" (or
3004 "make update"). This enables the following EC_METHODs:
3006 EC_GFp_nistp224_method()
3007 EC_GFp_nistp256_method()
3008 EC_GFp_nistp521_method()
3010 EC_GROUP_new_by_curve_name() will automatically use these (while
3011 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3013 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3015 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3016 all platforms. Move ssize_t definition from e_os.h to the public
3017 header file e_os2.h as it now appears in public header file cms.h
3020 *) New -sigopt option to the ca, req and x509 utilities. Additional
3021 signature parameters can be passed using this option and in
3025 *) Add RSA PSS signing function. This will generate and set the
3026 appropriate AlgorithmIdentifiers for PSS based on those in the
3027 corresponding EVP_MD_CTX structure. No application support yet.
3030 *) Support for companion algorithm specific ASN1 signing routines.
3031 New function ASN1_item_sign_ctx() signs a pre-initialised
3032 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3033 the appropriate parameters.
3036 *) Add new algorithm specific ASN1 verification initialisation function
3037 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3038 handling will be the same no matter what EVP_PKEY_METHOD is used.
3039 Add a PSS handler to support verification of PSS signatures: checked
3040 against a number of sample certificates.
3043 *) Add signature printing for PSS. Add PSS OIDs.
3044 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3046 *) Add algorithm specific signature printing. An individual ASN1 method
3047 can now print out signatures instead of the standard hex dump.
3049 More complex signatures (e.g. PSS) can print out more meaningful
3050 information. Include DSA version that prints out the signature
3054 *) Password based recipient info support for CMS library: implementing
3058 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3059 neatly separates the code into cipher and PBE sections and is required
3060 for some algorithms that split PBES2 into separate pieces (such as
3061 password based CMS).
3064 *) Session-handling fixes:
3065 - Fix handling of connections that are resuming with a session ID,
3066 but also support Session Tickets.
3067 - Fix a bug that suppressed issuing of a new ticket if the client
3068 presented a ticket with an expired session.
3069 - Try to set the ticket lifetime hint to something reasonable.
3070 - Make tickets shorter by excluding irrelevant information.
3071 - On the client side, don't ignore renewed tickets.
3072 [Adam Langley, Bodo Moeller (Google)]
3074 *) Fix PSK session representation.
3077 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3079 This work was sponsored by Intel.
3082 *) Add GCM support to TLS library. Some custom code is needed to split
3083 the IV between the fixed (from PRF) and explicit (from TLS record)
3084 portions. This adds all GCM ciphersuites supported by RFC5288 and
3085 RFC5289. Generalise some AES* cipherstrings to include GCM and
3086 add a special AESGCM string for GCM only.
3089 *) Expand range of ctrls for AES GCM. Permit setting invocation
3090 field on decrypt and retrieval of invocation field only on encrypt.
3093 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3094 As required by RFC5289 these ciphersuites cannot be used if for
3095 versions of TLS earlier than 1.2.
3098 *) For FIPS capable OpenSSL interpret a NULL default public key method
3099 as unset and return the appropriate default but do *not* set the default.
3100 This means we can return the appropriate method in applications that
3101 switch between FIPS and non-FIPS modes.
3104 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3105 ENGINE is used then we cannot handle that in the FIPS module so we
3106 keep original code iff non-FIPS operations are allowed.
3109 *) Add -attime option to openssl utilities.
3110 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3112 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3115 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3116 FIPS EC methods unconditionally for now.
3119 *) New build option no-ec2m to disable characteristic 2 code.
3122 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3123 all cases can be covered as some introduce binary incompatibilities.
3126 *) Redirect RSA operations to FIPS module including keygen,
3127 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3130 *) Add similar low level API blocking to ciphers.
3133 *) Low level digest APIs are not approved in FIPS mode: any attempt
3134 to use these will cause a fatal error. Applications that *really* want
3135 to use them can use the private_* version instead.
3138 *) Redirect cipher operations to FIPS module for FIPS builds.
3141 *) Redirect digest operations to FIPS module for FIPS builds.
3144 *) Update build system to add "fips" flag which will link in fipscanister.o
3145 for static and shared library builds embedding a signature if needed.
3148 *) Output TLS supported curves in preference order instead of numerical
3149 order. This is currently hardcoded for the highest order curves first.
3150 This should be configurable so applications can judge speed vs strength.
3153 *) Add TLS v1.2 server support for client authentication.
3156 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3160 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3161 FIPS modules versions.
3164 *) Add TLS v1.2 client side support for client authentication. Keep cache
3165 of handshake records longer as we don't know the hash algorithm to use
3166 until after the certificate request message is received.
3169 *) Initial TLS v1.2 client support. Add a default signature algorithms
3170 extension including all the algorithms we support. Parse new signature
3171 format in client key exchange. Relax some ECC signing restrictions for
3172 TLS v1.2 as indicated in RFC5246.
3175 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3176 to new signature format when needed using client digest preference.
3177 All server ciphersuites should now work correctly in TLS v1.2. No client
3178 support yet and no support for client certificates.
3181 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3182 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3183 ciphersuites. At present only RSA key exchange ciphersuites work with
3184 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3185 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3186 and version checking.
3189 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3190 with this defined it will not be affected by any changes to ssl internal
3191 structures. Add several utility functions to allow openssl application
3192 to work with OPENSSL_NO_SSL_INTERN defined.
3196 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
3198 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3201 *) Permit abbreviated handshakes when renegotiating using the function
3202 SSL_renegotiate_abbreviated().
3203 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3205 *) Add call to ENGINE_register_all_complete() to
3206 ENGINE_load_builtin_engines(), so some implementations get used
3207 automatically instead of needing explicit application support.
3210 *) Add support for TLS key exporter as described in RFC5705.
3211 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3213 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3214 a few changes are required:
3216 Add SSL_OP_NO_TLSv1_1 flag.
3217 Add TLSv1_1 methods.
3218 Update version checking logic to handle version 1.1.
3219 Add explicit IV handling (ported from DTLS code).
3220 Add command line options to s_client/s_server.
3223 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3225 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3226 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3227 content decryption and always return the same error. Note: this attack
3228 needs on average 2^20 messages so it only affects automated senders. The
3229 old behaviour can be re-enabled in the CMS code by setting the
3230 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3231 an MMA defence is not necessary.
3232 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3233 this issue. (CVE-2012-0884)
3236 *) Fix CVE-2011-4619: make sure we really are receiving a
3237 client hello before rejecting multiple SGC restarts. Thanks to
3238 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3241 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3243 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3244 Thanks to Antonio Martin, Enterprise Secure Access Research and
3245 Development, Cisco Systems, Inc. for discovering this bug and
3246 preparing a fix. (CVE-2012-0050)
3249 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3251 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3252 of the Vaudenay padding oracle attack on CBC mode encryption
3253 which enables an efficient plaintext recovery attack against
3254 the OpenSSL implementation of DTLS. Their attack exploits timing
3255 differences arising during decryption processing. A research
3256 paper describing this attack can be found at:
3257 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3258 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3259 Security Group at Royal Holloway, University of London
3260 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3261 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3262 for preparing the fix. (CVE-2011-4108)
3263 [Robin Seggelmann, Michael Tuexen]
3265 *) Clear bytes used for block padding of SSL 3.0 records.
3267 [Adam Langley (Google)]
3269 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3270 Kadianakis <desnacked@gmail.com> for discovering this issue and
3271 Adam Langley for preparing the fix. (CVE-2011-4619)
3272 [Adam Langley (Google)]
3274 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3275 [Andrey Kulikov <amdeich@gmail.com>]
3277 *) Prevent malformed RFC3779 data triggering an assertion failure.
3278 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3279 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3280 [Rob Austein <sra@hactrn.net>]
3282 *) Improved PRNG seeding for VOS.
3283 [Paul Green <Paul.Green@stratus.com>]
3285 *) Fix ssl_ciph.c set-up race.
3286 [Adam Langley (Google)]
3288 *) Fix spurious failures in ecdsatest.c.
3289 [Emilia Käsper (Google)]
3291 *) Fix the BIO_f_buffer() implementation (which was mixing different
3292 interpretations of the '..._len' fields).
3293 [Adam Langley (Google)]
3295 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3296 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3297 threads won't reuse the same blinding coefficients.
3299 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3300 lock to call BN_BLINDING_invert_ex, and avoids one use of
3301 BN_BLINDING_update for each BN_BLINDING structure (previously,
3302 the last update always remained unused).
3303 [Emilia Käsper (Google)]
3305 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3306 [Bob Buckholz (Google)]
3308 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3310 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3311 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3312 [Kaspar Brand <ossl@velox.ch>]
3314 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3315 for multi-threaded use of ECDH. (CVE-2011-3210)
3316 [Adam Langley (Google)]
3318 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3321 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3322 signature public key algorithm by using OID xref utilities instead.
3323 Before this you could only use some ECC ciphersuites with SHA1 only.
3326 *) Add protection against ECDSA timing attacks as mentioned in the paper
3327 by Billy Bob Brumley and Nicola Tuveri, see:
3329 http://eprint.iacr.org/2011/232.pdf
3331 [Billy Bob Brumley and Nicola Tuveri]
3333 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3335 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3336 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3338 *) Fix bug in string printing code: if *any* escaping is enabled we must
3339 escape the escape character (backslash) or the resulting string is
3343 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3345 *) Disable code workaround for ancient and obsolete Netscape browsers
3346 and servers: an attacker can use it in a ciphersuite downgrade attack.
3347 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3350 *) Fixed J-PAKE implementation error, originally discovered by
3351 Sebastien Martini, further info and confirmation from Stefan
3352 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3355 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3357 *) Fix extension code to avoid race conditions which can result in a buffer
3358 overrun vulnerability: resumed sessions must not be modified as they can
3359 be shared by multiple threads. CVE-2010-3864
3362 *) Fix WIN32 build system to correctly link an ENGINE directory into
3366 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3368 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3370 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3372 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3374 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3375 context. The operation can be customised via the ctrl mechanism in
3376 case ENGINEs want to include additional functionality.
3379 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3382 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3383 output hashes compatible with older versions of OpenSSL.
3384 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3386 *) Fix compression algorithm handling: if resuming a session use the
3387 compression algorithm of the resumed session instead of determining
3388 it from client hello again. Don't allow server to change algorithm.
3391 *) Add load_crls() function to apps tidying load_certs() too. Add option
3392 to verify utility to allow additional CRLs to be included.
3395 *) Update OCSP request code to permit adding custom headers to the request:
3396 some responders need this.
3399 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3401 [Julia Lawall <julia@diku.dk>]
3403 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3404 needlessly dereferenced structures, used obsolete functions and
3405 didn't handle all updated verify codes correctly.
3408 *) Disable MD2 in the default configuration.
3411 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3412 indicate the initial BIO being pushed or popped. This makes it possible
3413 to determine whether the BIO is the one explicitly called or as a result
3414 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3415 it handles reference counts correctly and doesn't zero out the I/O bio
3416 when it is not being explicitly popped. WARNING: applications which
3417 included workarounds for the old buggy behaviour will need to be modified
3418 or they could free up already freed BIOs.
3421 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3422 renaming to all platforms (within the 0.9.8 branch, this was
3423 done conditionally on Netware platforms to avoid a name clash).
3424 [Guenter <lists@gknw.net>]
3426 *) Add ECDHE and PSK support to DTLS.
3427 [Michael Tuexen <tuexen@fh-muenster.de>]
3429 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3433 *) Add "missing" function EVP_MD_flags() (without this the only way to
3434 retrieve a digest flags is by accessing the structure directly. Update
3435 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3436 or cipher is registered as in the "from" argument. Print out all
3437 registered digests in the dgst usage message instead of manually
3438 attempting to work them out.
3441 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3442 this allows the use of compression and extensions. Change default cipher
3443 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3444 by default unless an application cipher string requests it.
3447 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3448 key ids to find matching certificates and keys but some PKCS#12 files
3449 don't follow the (somewhat unwritten) rules and this strategy fails.
3450 Now just gather all certificates together and the first private key
3451 then look for the first certificate that matches the key.
3454 *) Support use of registered digest and cipher names for dgst and cipher
3455 commands instead of having to add each one as a special case. So now
3462 openssl dgst -sha256 foo
3464 and this works for ENGINE based algorithms too.
3468 *) Update Gost ENGINE to support parameter files.
3469 [Victor B. Wagner <vitus@cryptocom.ru>]
3471 *) Support GeneralizedTime in ca utility.
3472 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3474 *) Enhance the hash format used for certificate directory links. The new
3475 form uses the canonical encoding (meaning equivalent names will work
3476 even if they aren't identical) and uses SHA1 instead of MD5. This form
3477 is incompatible with the older format and as a result c_rehash should
3478 be used to rebuild symbolic links.
3481 *) Make PKCS#8 the default write format for private keys, replacing the
3482 traditional format. This form is standardised, more secure and doesn't
3483 include an implicit MD5 dependency.
3486 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3487 committed to OpenSSL should pass this lot as a minimum.
3490 *) Add session ticket override functionality for use by EAP-FAST.
3491 [Jouni Malinen <j@w1.fi>]
3493 *) Modify HMAC functions to return a value. Since these can be implemented
3494 in an ENGINE errors can occur.
3497 *) Type-checked OBJ_bsearch_ex.
3500 *) Type-checked OBJ_bsearch. Also some constification necessitated
3501 by type-checking. Still to come: TXT_DB, bsearch(?),
3502 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3506 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3507 seconds to a tm structure directly, instead of going through OS
3508 specific date routines. This avoids any issues with OS routines such
3509 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3510 and X509_time_adj_ex() to cover the extended range. The existing
3511 X509_time_adj() is still usable and will no longer have any date issues.
3514 *) Delta CRL support. New use deltas option which will attempt to locate
3515 and search any appropriate delta CRLs available.
3517 This work was sponsored by Google.
3520 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3521 code and add additional score elements. Validate alternate CRL paths
3522 as part of the CRL checking and indicate a new error "CRL path validation
3523 error" in this case. Applications wanting additional details can use
3524 the verify callback and check the new "parent" field. If this is not
3525 NULL CRL path validation is taking place. Existing applications won't
3526 see this because it requires extended CRL support which is off by
3529 This work was sponsored by Google.
3532 *) Support for freshest CRL extension.
3534 This work was sponsored by Google.
3537 *) Initial indirect CRL support. Currently only supported in the CRLs
3538 passed directly and not via lookup. Process certificate issuer
3539 CRL entry extension and lookup CRL entries by bother issuer name
3540 and serial number. Check and process CRL issuer entry in IDP extension.
3542 This work was sponsored by Google.
3545 *) Add support for distinct certificate and CRL paths. The CRL issuer
3546 certificate is validated separately in this case. Only enabled if
3547 an extended CRL support flag is set: this flag will enable additional
3548 CRL functionality in future.
3550 This work was sponsored by Google.
3553 *) Add support for policy mappings extension.
3555 This work was sponsored by Google.
3558 *) Fixes to pathlength constraint, self issued certificate handling,
3559 policy processing to align with RFC3280 and PKITS tests.
3561 This work was sponsored by Google.
3564 *) Support for name constraints certificate extension. DN, email, DNS
3565 and URI types are currently supported.
3567 This work was sponsored by Google.
3570 *) To cater for systems that provide a pointer-based thread ID rather
3571 than numeric, deprecate the current numeric thread ID mechanism and
3572 replace it with a structure and associated callback type. This
3573 mechanism allows a numeric "hash" to be extracted from a thread ID in
3574 either case, and on platforms where pointers are larger than 'long',
3575 mixing is done to help ensure the numeric 'hash' is usable even if it
3576 can't be guaranteed unique. The default mechanism is to use "&errno"
3577 as a pointer-based thread ID to distinguish between threads.
3579 Applications that want to provide their own thread IDs should now use
3580 CRYPTO_THREADID_set_callback() to register a callback that will call
3581 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3583 Note that ERR_remove_state() is now deprecated, because it is tied
3584 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3585 to free the current thread's error state should be replaced by
3586 ERR_remove_thread_state(NULL).
3588 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3589 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3590 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3591 application was previously providing a numeric thread callback that
3592 was inappropriate for distinguishing threads, then uniqueness might
3593 have been obtained with &errno that happened immediately in the
3594 intermediate development versions of OpenSSL; this is no longer the
3595 case, the numeric thread callback will now override the automatic use
3597 [Geoff Thorpe, with help from Bodo Moeller]
3599 *) Initial support for different CRL issuing certificates. This covers a
3600 simple case where the self issued certificates in the chain exist and
3601 the real CRL issuer is higher in the existing chain.
3603 This work was sponsored by Google.
3606 *) Removed effectively defunct crypto/store from the build.
3609 *) Revamp of STACK to provide stronger type-checking. Still to come:
3610 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3611 ASN1_STRING, CONF_VALUE.
3614 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3615 RAM on SSL connections. This option can save about 34k per idle SSL.
3618 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3619 STACK, TXT_DB, bsearch, qsort.
3622 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3623 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3624 support for data, signedData, compressedData, digestedData and
3625 encryptedData, envelopedData types included. Scripts to check against
3626 RFC4134 examples draft and interop and consistency checks of many
3627 content types and variants.
3630 *) Add options to enc utility to support use of zlib compression BIO.
3633 *) Extend mk1mf to support importing of options and assembly language
3634 files from Configure script, currently only included in VC-WIN32.
projects / openssl.git / blob