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 1.1.0h and 1.1.1 [xx XXX xxxx]
11 *) Add coordinate blinding for EC_POINT and implement projective
12 coordinate blinding for generic prime curves as a countermeasure to
13 chosen point SCA attacks.
14 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
16 *) Add blinding to an ECDSA signature to protect against side channel attacks
17 discovered by Keegan Ryan (NCC Group).
20 *) Enforce checking in the pkeyutl command line app to ensure that the input
21 length does not exceed the maximum supported digest length when performing
22 a sign, verify or verifyrecover operation.
25 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
26 I/O in combination with something like select() or poll() will hang. This
27 can be turned off again using SSL_CTX_clear_mode().
28 Many applications do not properly handle non-application data records, and
29 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
30 around the problems in those applications, but can also break some.
31 It's recommended to read the manpages about SSL_read(), SSL_write(),
32 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
33 SSL_CTX_set_read_ahead() again.
36 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
37 now allow empty (zero character) pass phrases.
40 *) Apply blinding to binary field modular inversion and remove patent
41 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
44 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
45 binary and prime elliptic curves.
48 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
49 constant time fixed point multiplication.
52 *) Revise elliptic curve scalar multiplication with timing attack
53 defenses: ec_wNAF_mul redirects to a constant time implementation
54 when computing fixed point and variable point multiplication (which
55 in OpenSSL are mostly used with secret scalars in keygen, sign,
56 ECDH derive operations).
57 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
60 *) Updated CONTRIBUTING
63 *) Updated DRBG / RAND to request nonce and additional low entropy
64 randomness from the system.
67 *) Updated 'openssl rehash' to use OpenSSL consistent default.
70 *) Moved the load of the ssl_conf module to libcrypto, which helps
71 loading engines that libssl uses before libssl is initialised.
74 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
77 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
78 [Ingo Schwarze, Rich Salz]
80 *) Added output of accepting IP address and port for 'openssl s_server'
83 *) Added a new API for TLSv1.3 ciphersuites:
84 SSL_CTX_set_ciphersuites()
85 SSL_set_ciphersuites()
88 *) Memory allocation failures consistenly add an error to the error
92 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
93 in libcrypto when run as setuid/setgid.
96 *) Load any config file by default when libssl is used.
99 *) Added new public header file <openssl/rand_drbg.h> and documentation
100 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
101 [Matthias St. Pierre]
103 *) QNX support removed (cannot find contributors to get their approval
104 for the license change).
107 *) TLSv1.3 replay protection for early data has been implemented. See the
108 SSL_read_early_data() man page for further details.
111 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
112 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
113 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
114 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
115 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
116 configuration has been separated out. See the ciphers man page or the
117 SSL_CTX_set_ciphersuites() man page for more information.
120 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
121 in responder mode now supports the new "-multi" option, which
122 spawns the specified number of child processes to handle OCSP
123 requests. The "-timeout" option now also limits the OCSP
124 responder's patience to wait to receive the full client request
125 on a newly accepted connection. Child processes are respawned
126 as needed, and the CA index file is automatically reloaded
127 when changed. This makes it possible to run the "ocsp" responder
128 as a long-running service, making the OpenSSL CA somewhat more
129 feature-complete. In this mode, most diagnostic messages logged
130 after entering the event loop are logged via syslog(3) rather than
134 *) Added support for X448 and Ed448. Heavily based on original work by
138 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
139 objects loaded. This adds the functions OSSL_STORE_expect() and
140 OSSL_STORE_find() as well as needed tools to construct searches and
141 get the search data out of them.
144 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
145 version of OpenSSL should review their configuration settings to ensure
146 that they are still appropriate for TLSv1.3. For further information see:
147 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
149 NOTE: In this pre-release of OpenSSL a draft version of the
150 TLSv1.3 standard has been implemented. Implementations of different draft
151 versions of the standard do not inter-operate, and this version will not
152 inter-operate with an implementation of the final standard when it is
153 eventually published. Different pre-release versions may implement
154 different versions of the draft. The final version of OpenSSL 1.1.1 will
155 implement the final version of the standard.
156 TODO(TLS1.3): Remove the above note before final release
159 *) Grand redesign of the OpenSSL random generator
161 The default RAND method now utilizes an AES-CTR DRBG according to
162 NIST standard SP 800-90Ar1. The new random generator is essentially
163 a port of the default random generator from the OpenSSL FIPS 2.0
164 object module. It is a hybrid deterministic random bit generator
165 using an AES-CTR bit stream and which seeds and reseeds itself
166 automatically using trusted system entropy sources.
168 Some of its new features are:
169 o Support for multiple DRBG instances with seed chaining.
170 o Add a public DRBG instance for the default RAND method.
171 o Add a dedicated DRBG instance for generating long term private keys.
172 o Make the DRBG instances fork-safe.
173 o Keep all global DRBG instances on the secure heap if it is enabled.
174 o Add a DRBG instance to every SSL instance for lock free operation
175 and to increase unpredictability.
176 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
178 *) Changed Configure so it only says what it does and doesn't dump
179 so much data. Instead, ./configdata.pm should be used as a script
180 to display all sorts of configuration data.
183 *) Added processing of "make variables" to Configure.
186 *) Added SHA512/224 and SHA512/256 algorithm support.
189 *) The last traces of Netware support, first removed in 1.1.0, have
193 *) Get rid of Makefile.shared, and in the process, make the processing
194 of certain files (rc.obj, or the .def/.map/.opt files produced from
195 the ordinal files) more visible and hopefully easier to trace and
196 debug (or make silent).
199 *) Make it possible to have environment variable assignments as
200 arguments to config / Configure.
203 *) Add multi-prime RSA (RFC 8017) support.
206 *) Add SM3 implemented according to GB/T 32905-2016
207 [ Jack Lloyd <jack.lloyd@ribose.com>,
208 Ronald Tse <ronald.tse@ribose.com>,
209 Erick Borsboom <erick.borsboom@ribose.com> ]
211 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
212 as documented in RFC6066.
213 Based on a patch from Tomasz Moń
214 [Filipe Raimundo da Silva]
216 *) Add SM4 implemented according to GB/T 32907-2016.
217 [ Jack Lloyd <jack.lloyd@ribose.com>,
218 Ronald Tse <ronald.tse@ribose.com>,
219 Erick Borsboom <erick.borsboom@ribose.com> ]
221 *) Reimplement -newreq-nodes and ERR_error_string_n; the
222 original author does not agree with the license change.
225 *) Add ARIA AEAD TLS support.
228 *) Some macro definitions to support VS6 have been removed. Visual
229 Studio 6 has not worked since 1.1.0
232 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
233 without clearing the errors.
236 *) Add "atfork" functions. If building on a system that without
237 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
238 requirements. The RAND facility now uses/requires this.
244 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
245 not possible to disable entirely. However, it's still possible to
246 disable the console reading UI method, UI_OpenSSL() (use UI_null()
249 To disable, configure with 'no-ui-console'. 'no-ui' is still
250 possible to use as an alias. Check at compile time with the
251 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
252 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
255 *) Add a STORE module, which implements a uniform and URI based reader of
256 stores that can contain keys, certificates, CRLs and numerous other
257 objects. The main API is loosely based on a few stdio functions,
258 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
259 OSSL_STORE_error and OSSL_STORE_close.
260 The implementation uses backends called "loaders" to implement arbitrary
261 URI schemes. There is one built in "loader" for the 'file' scheme.
264 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
265 then adjusted to work on FreeBSD 8.4 as well.
266 Enable by configuring with 'enable-devcryptoeng'. This is done by default
267 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
270 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
271 util/mkerr.pl, which is adapted to allow those prefixes, leading to
272 error code calls like this:
274 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
276 With this change, we claim the namespaces OSSL and OPENSSL in a manner
277 that can be encoded in C. For the foreseeable future, this will only
279 [Richard Levitte and Tim Hudson]
281 *) Removed BSD cryptodev engine.
284 *) Add a build target 'build_all_generated', to build all generated files
285 and only that. This can be used to prepare everything that requires
286 things like perl for a system that lacks perl and then move everything
287 to that system and do the rest of the build there.
290 *) In the UI interface, make it possible to duplicate the user data. This
291 can be used by engines that need to retain the data for a longer time
292 than just the call where this user data is passed.
295 *) Ignore the '-named_curve auto' value for compatibility of applications
297 [Tomas Mraz <tmraz@fedoraproject.org>]
299 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
300 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
301 alerts across multiple records (some of which could be empty). In practice
302 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
303 prohibts this altogether and other libraries (BoringSSL, NSS) do not
304 support this at all. Supporting it adds significant complexity to the
305 record layer, and its removal is unlikely to cause inter-operability
309 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
310 with Z. These are meant to replace LONG and ZLONG and to be size safe.
311 The use of LONG and ZLONG is discouraged and scheduled for deprecation
315 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
316 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
317 [Richard Levitte, Andy Polyakov]
319 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
323 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
324 platform rather than 'mingw'.
327 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
328 success if they are asked to add an object which already exists
329 in the store. This change cascades to other functions which load
330 certificates and CRLs.
333 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
334 facilitate stack unwinding even from assembly subroutines.
337 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
338 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
341 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
342 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
343 which is the minimum version we support.
346 *) Certificate time validation (X509_cmp_time) enforces stricter
347 compliance with RFC 5280. Fractional seconds and timezone offsets
348 are no longer allowed.
351 *) Add support for ARIA
354 *) s_client will now send the Server Name Indication (SNI) extension by
355 default unless the new "-noservername" option is used. The server name is
356 based on the host provided to the "-connect" option unless overridden by
360 *) Add support for SipHash
363 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
364 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
365 prevent issues where no progress is being made and the peer continually
366 sends unrecognised record types, using up resources processing them.
369 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
370 using the algorithm defined in
371 https://www.akkadia.org/drepper/SHA-crypt.txt
374 *) Heartbeat support has been removed; the ABI is changed for now.
375 [Richard Levitte, Rich Salz]
377 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
380 *) The RSA "null" method, which was partially supported to avoid patent
381 issues, has been replaced to always returns NULL.
385 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
387 *) Fixed a text canonicalisation bug in CMS
389 Where a CMS detached signature is used with text content the text goes
390 through a canonicalisation process first prior to signing or verifying a
391 signature. This process strips trailing space at the end of lines, converts
392 line terminators to CRLF and removes additional trailing line terminators
393 at the end of a file. A bug in the canonicalisation process meant that
394 some characters, such as form-feed, were incorrectly treated as whitespace
395 and removed. This is contrary to the specification (RFC5485). This fix
396 could mean that detached text data signed with an earlier version of
397 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
398 signed with a fixed OpenSSL may fail to verify with an earlier version of
399 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
400 and use the "-binary" flag (for the "cms" command line application) or set
401 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
404 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
406 *) Constructed ASN.1 types with a recursive definition could exceed the stack
408 Constructed ASN.1 types with a recursive definition (such as can be found
409 in PKCS7) could eventually exceed the stack given malicious input with
410 excessive recursion. This could result in a Denial Of Service attack. There
411 are no such structures used within SSL/TLS that come from untrusted sources
412 so this is considered safe.
414 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
419 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
421 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
422 effectively reduced to only comparing the least significant bit of each
423 byte. This allows an attacker to forge messages that would be considered as
424 authenticated in an amount of tries lower than that guaranteed by the
425 security claims of the scheme. The module can only be compiled by the
426 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
428 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
433 *) Add a build target 'build_all_generated', to build all generated files
434 and only that. This can be used to prepare everything that requires
435 things like perl for a system that lacks perl and then move everything
436 to that system and do the rest of the build there.
439 *) Backport SSL_OP_NO_RENGOTIATION
441 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
442 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
443 changes this is no longer possible in 1.1.0. Therefore the new
444 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
445 1.1.0 to provide equivalent functionality.
447 Note that if an application built against 1.1.0h headers (or above) is run
448 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
449 accepted but nothing will happen, i.e. renegotiation will not be prevented.
452 *) Removed the OS390-Unix config target. It relied on a script that doesn't
456 *) rsaz_1024_mul_avx2 overflow bug on x86_64
458 There is an overflow bug in the AVX2 Montgomery multiplication procedure
459 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
460 Analysis suggests that attacks against RSA and DSA as a result of this
461 defect would be very difficult to perform and are not believed likely.
462 Attacks against DH1024 are considered just feasible, because most of the
463 work necessary to deduce information about a private key may be performed
464 offline. The amount of resources required for such an attack would be
465 significant. However, for an attack on TLS to be meaningful, the server
466 would have to share the DH1024 private key among multiple clients, which is
467 no longer an option since CVE-2016-0701.
469 This only affects processors that support the AVX2 but not ADX extensions
470 like Intel Haswell (4th generation).
472 This issue was reported to OpenSSL by David Benjamin (Google). The issue
473 was originally found via the OSS-Fuzz project.
477 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
479 *) bn_sqrx8x_internal carry bug on x86_64
481 There is a carry propagating bug in the x86_64 Montgomery squaring
482 procedure. No EC algorithms are affected. Analysis suggests that attacks
483 against RSA and DSA as a result of this defect would be very difficult to
484 perform and are not believed likely. Attacks against DH are considered just
485 feasible (although very difficult) because most of the work necessary to
486 deduce information about a private key may be performed offline. The amount
487 of resources required for such an attack would be very significant and
488 likely only accessible to a limited number of attackers. An attacker would
489 additionally need online access to an unpatched system using the target
490 private key in a scenario with persistent DH parameters and a private
491 key that is shared between multiple clients.
493 This only affects processors that support the BMI1, BMI2 and ADX extensions
494 like Intel Broadwell (5th generation) and later or AMD Ryzen.
496 This issue was reported to OpenSSL by the OSS-Fuzz project.
500 *) Malformed X.509 IPAddressFamily could cause OOB read
502 If an X.509 certificate has a malformed IPAddressFamily extension,
503 OpenSSL could do a one-byte buffer overread. The most likely result
504 would be an erroneous display of the certificate in text format.
506 This issue was reported to OpenSSL by the OSS-Fuzz project.
510 Changes between 1.1.0e and 1.1.0f [25 May 2017]
512 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
513 platform rather than 'mingw'.
516 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
517 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
518 which is the minimum version we support.
521 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
523 *) Encrypt-Then-Mac renegotiation crash
525 During a renegotiation handshake if the Encrypt-Then-Mac extension is
526 negotiated where it was not in the original handshake (or vice-versa) then
527 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
528 and servers are affected.
530 This issue was reported to OpenSSL by Joe Orton (Red Hat).
534 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
536 *) Truncated packet could crash via OOB read
538 If one side of an SSL/TLS path is running on a 32-bit host and a specific
539 cipher is being used, then a truncated packet can cause that host to
540 perform an out-of-bounds read, usually resulting in a crash.
542 This issue was reported to OpenSSL by Robert Święcki of Google.
546 *) Bad (EC)DHE parameters cause a client crash
548 If a malicious server supplies bad parameters for a DHE or ECDHE key
549 exchange then this can result in the client attempting to dereference a
550 NULL pointer leading to a client crash. This could be exploited in a Denial
553 This issue was reported to OpenSSL by Guido Vranken.
557 *) BN_mod_exp may produce incorrect results on x86_64
559 There is a carry propagating bug in the x86_64 Montgomery squaring
560 procedure. No EC algorithms are affected. Analysis suggests that attacks
561 against RSA and DSA as a result of this defect would be very difficult to
562 perform and are not believed likely. Attacks against DH are considered just
563 feasible (although very difficult) because most of the work necessary to
564 deduce information about a private key may be performed offline. The amount
565 of resources required for such an attack would be very significant and
566 likely only accessible to a limited number of attackers. An attacker would
567 additionally need online access to an unpatched system using the target
568 private key in a scenario with persistent DH parameters and a private
569 key that is shared between multiple clients. For example this can occur by
570 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
571 similar to CVE-2015-3193 but must be treated as a separate problem.
573 This issue was reported to OpenSSL by the OSS-Fuzz project.
577 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
579 *) ChaCha20/Poly1305 heap-buffer-overflow
581 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
582 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
583 crash. This issue is not considered to be exploitable beyond a DoS.
585 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
589 *) CMS Null dereference
591 Applications parsing invalid CMS structures can crash with a NULL pointer
592 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
593 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
594 structure callback if an attempt is made to free certain invalid encodings.
595 Only CHOICE structures using a callback which do not handle NULL value are
598 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
602 *) Montgomery multiplication may produce incorrect results
604 There is a carry propagating bug in the Broadwell-specific Montgomery
605 multiplication procedure that handles input lengths divisible by, but
606 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
607 and DH private keys are impossible. This is because the subroutine in
608 question is not used in operations with the private key itself and an input
609 of the attacker's direct choice. Otherwise the bug can manifest itself as
610 transient authentication and key negotiation failures or reproducible
611 erroneous outcome of public-key operations with specially crafted input.
612 Among EC algorithms only Brainpool P-512 curves are affected and one
613 presumably can attack ECDH key negotiation. Impact was not analyzed in
614 detail, because pre-requisites for attack are considered unlikely. Namely
615 multiple clients have to choose the curve in question and the server has to
616 share the private key among them, neither of which is default behaviour.
617 Even then only clients that chose the curve will be affected.
619 This issue was publicly reported as transient failures and was not
620 initially recognized as a security issue. Thanks to Richard Morgan for
621 providing reproducible case.
625 *) Removed automatic addition of RPATH in shared libraries and executables,
626 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
629 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
631 *) Fix Use After Free for large message sizes
633 The patch applied to address CVE-2016-6307 resulted in an issue where if a
634 message larger than approx 16k is received then the underlying buffer to
635 store the incoming message is reallocated and moved. Unfortunately a
636 dangling pointer to the old location is left which results in an attempt to
637 write to the previously freed location. This is likely to result in a
638 crash, however it could potentially lead to execution of arbitrary code.
640 This issue only affects OpenSSL 1.1.0a.
642 This issue was reported to OpenSSL by Robert Święcki.
646 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
648 *) OCSP Status Request extension unbounded memory growth
650 A malicious client can send an excessively large OCSP Status Request
651 extension. If that client continually requests renegotiation, sending a
652 large OCSP Status Request extension each time, then there will be unbounded
653 memory growth on the server. This will eventually lead to a Denial Of
654 Service attack through memory exhaustion. Servers with a default
655 configuration are vulnerable even if they do not support OCSP. Builds using
656 the "no-ocsp" build time option are not affected.
658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
662 *) SSL_peek() hang on empty record
664 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
665 sends an empty record. This could be exploited by a malicious peer in a
666 Denial Of Service attack.
668 This issue was reported to OpenSSL by Alex Gaynor.
672 *) Excessive allocation of memory in tls_get_message_header() and
673 dtls1_preprocess_fragment()
675 A (D)TLS message includes 3 bytes for its length in the header for the
676 message. This would allow for messages up to 16Mb in length. Messages of
677 this length are excessive and OpenSSL includes a check to ensure that a
678 peer is sending reasonably sized messages in order to avoid too much memory
679 being consumed to service a connection. A flaw in the logic of version
680 1.1.0 means that memory for the message is allocated too early, prior to
681 the excessive message length check. Due to way memory is allocated in
682 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
683 to service a connection. This could lead to a Denial of Service through
684 memory exhaustion. However, the excessive message length check still takes
685 place, and this would cause the connection to immediately fail. Assuming
686 that the application calls SSL_free() on the failed connection in a timely
687 manner then the 21Mb of allocated memory will then be immediately freed
688 again. Therefore the excessive memory allocation will be transitory in
689 nature. This then means that there is only a security impact if:
691 1) The application does not call SSL_free() in a timely manner in the event
692 that the connection fails
694 2) The application is working in a constrained environment where there is
695 very little free memory
697 3) The attacker initiates multiple connection attempts such that there are
698 multiple connections in a state where memory has been allocated for the
699 connection; SSL_free() has not yet been called; and there is insufficient
700 memory to service the multiple requests.
702 Except in the instance of (1) above any Denial Of Service is likely to be
703 transitory because as soon as the connection fails the memory is
704 subsequently freed again in the SSL_free() call. However there is an
705 increased risk during this period of application crashes due to the lack of
706 memory - which would then mean a more serious Denial of Service.
708 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
709 (CVE-2016-6307 and CVE-2016-6308)
712 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
713 had to be removed. Primary reason is that vendor assembler can't
714 assemble our modules with -KPIC flag. As result it, assembly
715 support, was not even available as option. But its lack means
716 lack of side-channel resistant code, which is incompatible with
717 security by todays standards. Fortunately gcc is readily available
718 prepackaged option, which we firmly point at...
721 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
723 *) Windows command-line tool supports UTF-8 opt-in option for arguments
724 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
725 (to any value) allows Windows user to access PKCS#12 file generated
726 with Windows CryptoAPI and protected with non-ASCII password, as well
727 as files generated under UTF-8 locale on Linux also protected with
731 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
732 have been disabled by default and removed from DEFAULT, just like RC4.
733 See the RC4 item below to re-enable both.
736 *) The method for finding the storage location for the Windows RAND seed file
737 has changed. First we check %RANDFILE%. If that is not set then we check
738 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
739 all else fails we fall back to C:\.
742 *) The EVP_EncryptUpdate() function has had its return type changed from void
743 to int. A return of 0 indicates and error while a return of 1 indicates
747 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
748 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
749 off the constant time implementation for RSA, DSA and DH have been made
750 no-ops and deprecated.
753 *) Windows RAND implementation was simplified to only get entropy by
754 calling CryptGenRandom(). Various other RAND-related tickets
756 [Joseph Wylie Yandle, Rich Salz]
758 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
759 and OPENSSL_LH_, respectively. The old names are available
760 with API compatibility. They new names are now completely documented.
763 *) Unify TYPE_up_ref(obj) methods signature.
764 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
765 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
766 int (instead of void) like all others TYPE_up_ref() methods.
767 So now these methods also check the return value of CRYPTO_atomic_add(),
768 and the validity of object reference counter.
769 [fdasilvayy@gmail.com]
771 *) With Windows Visual Studio builds, the .pdb files are installed
772 alongside the installed libraries and executables. For a static
773 library installation, ossl_static.pdb is the associate compiler
774 generated .pdb file to be used when linking programs.
777 *) Remove openssl.spec. Packaging files belong with the packagers.
780 *) Automatic Darwin/OSX configuration has had a refresh, it will now
781 recognise x86_64 architectures automatically. You can still decide
782 to build for a different bitness with the environment variable
783 KERNEL_BITS (can be 32 or 64), for example:
785 KERNEL_BITS=32 ./config
789 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
790 256 bit AES and HMAC with SHA256.
793 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
796 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
799 *) To enable users to have their own config files and build file templates,
800 Configure looks in the directory indicated by the environment variable
801 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
802 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
803 name and is used as is.
806 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
807 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
808 X509_CERT_FILE_CTX was removed.
811 *) "shared" builds are now the default. To create only static libraries use
812 the "no-shared" Configure option.
815 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
816 All of these option have not worked for some while and are fundamental
820 *) Make various cleanup routines no-ops and mark them as deprecated. Most
821 global cleanup functions are no longer required because they are handled
822 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
823 Explicitly de-initing can cause problems (e.g. where a library that uses
824 OpenSSL de-inits, but an application is still using it). The affected
825 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
826 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
827 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
831 *) --strict-warnings no longer enables runtime debugging options
832 such as REF_DEBUG. Instead, debug options are automatically
833 enabled with '--debug' builds.
834 [Andy Polyakov, Emilia Käsper]
836 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
837 have been moved out of the public header files. New functions for managing
838 these have been added.
841 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
842 objects have been moved out of the public header files. New
843 functions for managing these have been added.
846 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
847 have been moved out of the public header files. New functions for managing
848 these have been added.
851 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
852 moved out of the public header files. New functions for managing these
856 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
859 *) Removed the mk1mf build scripts.
862 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
863 it is always safe to #include a header now.
866 *) Removed the aged BC-32 config and all its supporting scripts
869 *) Removed support for Ultrix, Netware, and OS/2.
872 *) Add support for HKDF.
875 *) Add support for blake2b and blake2s
878 *) Added support for "pipelining". Ciphers that have the
879 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
880 encryptions/decryptions simultaneously. There are currently no built-in
881 ciphers with this property but the expectation is that engines will be able
882 to offer it to significantly improve throughput. Support has been extended
883 into libssl so that multiple records for a single connection can be
884 processed in one go (for >=TLS 1.1).
887 *) Added the AFALG engine. This is an async capable engine which is able to
888 offload work to the Linux kernel. In this initial version it only supports
889 AES128-CBC. The kernel must be version 4.1.0 or greater.
892 *) OpenSSL now uses a new threading API. It is no longer necessary to
893 set locking callbacks to use OpenSSL in a multi-threaded environment. There
894 are two supported threading models: pthreads and windows threads. It is
895 also possible to configure OpenSSL at compile time for "no-threads". The
896 old threading API should no longer be used. The functions have been
897 replaced with "no-op" compatibility macros.
898 [Alessandro Ghedini, Matt Caswell]
900 *) Modify behavior of ALPN to invoke callback after SNI/servername
901 callback, such that updates to the SSL_CTX affect ALPN.
904 *) Add SSL_CIPHER queries for authentication and key-exchange.
907 *) Changes to the DEFAULT cipherlist:
908 - Prefer (EC)DHE handshakes over plain RSA.
909 - Prefer AEAD ciphers over legacy ciphers.
910 - Prefer ECDSA over RSA when both certificates are available.
911 - Prefer TLSv1.2 ciphers/PRF.
912 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
916 *) Change the ECC default curve list to be this, in order: x25519,
917 secp256r1, secp521r1, secp384r1.
920 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
921 disabled by default. They can be re-enabled using the
922 enable-weak-ssl-ciphers option to Configure.
925 *) If the server has ALPN configured, but supports no protocols that the
926 client advertises, send a fatal "no_application_protocol" alert.
927 This behaviour is SHALL in RFC 7301, though it isn't universally
928 implemented by other servers.
931 *) Add X25519 support.
932 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
933 for public and private key encoding using the format documented in
934 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
935 key generation and key derivation.
937 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
941 *) Deprecate SRP_VBASE_get_by_user.
942 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
943 In order to fix an unavoidable memory leak (CVE-2016-0798),
944 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
945 seed, even if the seed is configured.
947 Users should use SRP_VBASE_get1_by_user instead. Note that in
948 SRP_VBASE_get1_by_user, caller must free the returned value. Note
949 also that even though configuring the SRP seed attempts to hide
950 invalid usernames by continuing the handshake with fake
951 credentials, this behaviour is not constant time and no strong
952 guarantees are made that the handshake is indistinguishable from
953 that of a valid user.
956 *) Configuration change; it's now possible to build dynamic engines
957 without having to build shared libraries and vice versa. This
958 only applies to the engines in engines/, those in crypto/engine/
959 will always be built into libcrypto (i.e. "static").
961 Building dynamic engines is enabled by default; to disable, use
962 the configuration option "disable-dynamic-engine".
964 The only requirements for building dynamic engines are the
965 presence of the DSO module and building with position independent
966 code, so they will also automatically be disabled if configuring
967 with "disable-dso" or "disable-pic".
969 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
970 are also taken away from openssl/opensslconf.h, as they are
974 *) Configuration change; if there is a known flag to compile
975 position independent code, it will always be applied on the
976 libcrypto and libssl object files, and never on the application
977 object files. This means other libraries that use routines from
978 libcrypto / libssl can be made into shared libraries regardless
979 of how OpenSSL was configured.
981 If this isn't desirable, the configuration options "disable-pic"
982 or "no-pic" can be used to disable the use of PIC. This will
983 also disable building shared libraries and dynamic engines.
986 *) Removed JPAKE code. It was experimental and has no wide use.
989 *) The INSTALL_PREFIX Makefile variable has been renamed to
990 DESTDIR. That makes for less confusion on what this variable
991 is for. Also, the configuration option --install_prefix is
995 *) Heartbeat for TLS has been removed and is disabled by default
996 for DTLS; configure with enable-heartbeats. Code that uses the
997 old #define's might need to be updated.
998 [Emilia Käsper, Rich Salz]
1000 *) Rename REF_CHECK to REF_DEBUG.
1003 *) New "unified" build system
1005 The "unified" build system is aimed to be a common system for all
1006 platforms we support. With it comes new support for VMS.
1008 This system builds supports building in a different directory tree
1009 than the source tree. It produces one Makefile (for unix family
1010 or lookalikes), or one descrip.mms (for VMS).
1012 The source of information to make the Makefile / descrip.mms is
1013 small files called 'build.info', holding the necessary
1014 information for each directory with source to compile, and a
1015 template in Configurations, like unix-Makefile.tmpl or
1018 With this change, the library names were also renamed on Windows
1019 and on VMS. They now have names that are closer to the standard
1020 on Unix, and include the major version number, and in certain
1021 cases, the architecture they are built for. See "Notes on shared
1022 libraries" in INSTALL.
1024 We rely heavily on the perl module Text::Template.
1027 *) Added support for auto-initialisation and de-initialisation of the library.
1028 OpenSSL no longer requires explicit init or deinit routines to be called,
1029 except in certain circumstances. See the OPENSSL_init_crypto() and
1030 OPENSSL_init_ssl() man pages for further information.
1033 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1034 "peer" argument is now expected to be a BIO_ADDR object.
1036 *) Rewrite of BIO networking library. The BIO library lacked consistent
1037 support of IPv6, and adding it required some more extensive
1038 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1039 which hold all types of addresses and chains of address information.
1040 It also introduces a new API, with functions like BIO_socket,
1041 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1042 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1043 have been adapted accordingly.
1046 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1050 *) CRIME protection: disable compression by default, even if OpenSSL is
1051 compiled with zlib enabled. Applications can still enable compression
1052 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1053 using the SSL_CONF library to configure compression.
1056 *) The signature of the session callback configured with
1057 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1058 was explicitly marked as 'const unsigned char*' instead of
1062 *) Always DPURIFY. Remove the use of uninitialized memory in the
1063 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1066 *) Removed many obsolete configuration items, including
1067 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1068 MD2_CHAR, MD2_INT, MD2_LONG
1070 IDEA_SHORT, IDEA_LONG
1071 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1072 [Rich Salz, with advice from Andy Polyakov]
1074 *) Many BN internals have been moved to an internal header file.
1075 [Rich Salz with help from Andy Polyakov]
1077 *) Configuration and writing out the results from it has changed.
1078 Files such as Makefile include/openssl/opensslconf.h and are now
1079 produced through general templates, such as Makefile.in and
1080 crypto/opensslconf.h.in and some help from the perl module
1083 Also, the center of configuration information is no longer
1084 Makefile. Instead, Configure produces a perl module in
1085 configdata.pm which holds most of the config data (in the hash
1086 table %config), the target data that comes from the target
1087 configuration in one of the Configurations/*.conf files (in
1091 *) To clarify their intended purposes, the Configure options
1092 --prefix and --openssldir change their semantics, and become more
1093 straightforward and less interdependent.
1095 --prefix shall be used exclusively to give the location INSTALLTOP
1096 where programs, scripts, libraries, include files and manuals are
1097 going to be installed. The default is now /usr/local.
1099 --openssldir shall be used exclusively to give the default
1100 location OPENSSLDIR where certificates, private keys, CRLs are
1101 managed. This is also where the default openssl.cnf gets
1103 If the directory given with this option is a relative path, the
1104 values of both the --prefix value and the --openssldir value will
1105 be combined to become OPENSSLDIR.
1106 The default for --openssldir is INSTALLTOP/ssl.
1108 Anyone who uses --openssldir to specify where OpenSSL is to be
1109 installed MUST change to use --prefix instead.
1112 *) The GOST engine was out of date and therefore it has been removed. An up
1113 to date GOST engine is now being maintained in an external repository.
1114 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1115 support for GOST ciphersuites (these are only activated if a GOST engine
1119 *) EGD is no longer supported by default; use enable-egd when
1121 [Ben Kaduk and Rich Salz]
1123 *) The distribution now has Makefile.in files, which are used to
1124 create Makefile's when Configure is run. *Configure must be run
1125 before trying to build now.*
1128 *) The return value for SSL_CIPHER_description() for error conditions
1132 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1134 Obtaining and performing DNSSEC validation of TLSA records is
1135 the application's responsibility. The application provides
1136 the TLSA records of its choice to OpenSSL, and these are then
1137 used to authenticate the peer.
1139 The TLSA records need not even come from DNS. They can, for
1140 example, be used to implement local end-entity certificate or
1141 trust-anchor "pinning", where the "pin" data takes the form
1142 of TLSA records, which can augment or replace verification
1143 based on the usual WebPKI public certification authorities.
1146 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1147 continues to support deprecated interfaces in default builds.
1148 However, applications are strongly advised to compile their
1149 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1150 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1151 or the 1.1.0 releases.
1153 In environments in which all applications have been ported to
1154 not use any deprecated interfaces OpenSSL's Configure script
1155 should be used with the --api=1.1.0 option to entirely remove
1156 support for the deprecated features from the library and
1157 unconditionally disable them in the installed headers.
1158 Essentially the same effect can be achieved with the "no-deprecated"
1159 argument to Configure, except that this will always restrict
1160 the build to just the latest API, rather than a fixed API
1163 As applications are ported to future revisions of the API,
1164 they should update their compile-time OPENSSL_API_COMPAT define
1165 accordingly, but in most cases should be able to continue to
1166 compile with later releases.
1168 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1169 0x10000000L and 0x00908000L, respectively. However those
1170 versions did not support the OPENSSL_API_COMPAT feature, and
1171 so applications are not typically tested for explicit support
1172 of just the undeprecated features of either release.
1175 *) Add support for setting the minimum and maximum supported protocol.
1176 It can bet set via the SSL_set_min_proto_version() and
1177 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1178 MaxProtocol. It's recommended to use the new APIs to disable
1179 protocols instead of disabling individual protocols using
1180 SSL_set_options() or SSL_CONF's Protocol. This change also
1181 removes support for disabling TLS 1.2 in the OpenSSL TLS
1182 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1185 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1188 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1189 and integrates ECDSA and ECDH functionality into EC. Implementations can
1190 now redirect key generation and no longer need to convert to or from
1193 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1194 include the ec.h header file instead.
1197 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1198 ciphers who are no longer supported and drops support the ephemeral RSA key
1199 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1202 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1203 opaque. For HMAC_CTX, the following constructors and destructors
1206 HMAC_CTX *HMAC_CTX_new(void);
1207 void HMAC_CTX_free(HMAC_CTX *ctx);
1209 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1210 destroy such methods has been added. See EVP_MD_meth_new(3) and
1211 EVP_CIPHER_meth_new(3) for documentation.
1214 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1215 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1216 EVP_MD_CTX_reset() should be called instead to reinitialise
1217 an already created structure.
1218 2) For consistency with the majority of our object creators and
1219 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1220 EVP_MD_CTX_(new|free). The old names are retained as macros
1221 for deprecated builds.
1224 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1225 cryptographic operations to be performed asynchronously as long as an
1226 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1227 further details. Libssl has also had this capability integrated with the
1228 introduction of the new mode SSL_MODE_ASYNC and associated error
1229 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1230 pages. This work was developed in partnership with Intel Corp.
1233 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1234 always enabled now. If you want to disable the support you should
1235 exclude it using the list of supported ciphers. This also means that the
1236 "-no_ecdhe" option has been removed from s_server.
1239 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1240 SSL_{CTX_}set1_curves() which can set a list.
1243 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1244 curve you want to support using SSL_{CTX_}set1_curves().
1247 *) State machine rewrite. The state machine code has been significantly
1248 refactored in order to remove much duplication of code and solve issues
1249 with the old code (see ssl/statem/README for further details). This change
1250 does have some associated API changes. Notably the SSL_state() function
1251 has been removed and replaced by SSL_get_state which now returns an
1252 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1253 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1257 *) All instances of the string "ssleay" in the public API were replaced
1258 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1259 Some error codes related to internal RSA_eay API's were renamed.
1262 *) The demo files in crypto/threads were moved to demo/threads.
1265 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1267 [Matt Caswell, Rich Salz]
1269 *) New ASN.1 embed macro.
1271 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1272 structure is not allocated: it is part of the parent. That is instead of
1280 This reduces memory fragmentation and make it impossible to accidentally
1281 set a mandatory field to NULL.
1283 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1284 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1285 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1289 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1292 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1293 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1294 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1295 DES and RC4 ciphersuites.
1298 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1299 This changes the decoding behaviour for some invalid messages,
1300 though the change is mostly in the more lenient direction, and
1301 legacy behaviour is preserved as much as possible.
1304 *) Fix no-stdio build.
1305 [ David Woodhouse <David.Woodhouse@intel.com> and also
1306 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1308 *) New testing framework
1309 The testing framework has been largely rewritten and is now using
1310 perl and the perl modules Test::Harness and an extended variant of
1311 Test::More called OpenSSL::Test to do its work. All test scripts in
1312 test/ have been rewritten into test recipes, and all direct calls to
1313 executables in test/Makefile have become individual recipes using the
1314 simplified testing OpenSSL::Test::Simple.
1316 For documentation on our testing modules, do:
1318 perldoc test/testlib/OpenSSL/Test/Simple.pm
1319 perldoc test/testlib/OpenSSL/Test.pm
1323 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1324 are used; the latter aborts on memory leaks (usually checked on exit).
1325 Some undocumented "set malloc, etc., hooks" functions were removed
1326 and others were changed. All are now documented.
1329 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1331 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1333 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1334 from RFC4279, RFC4785, RFC5487, RFC5489.
1336 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1337 original RSA_PSK patch.
1340 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1341 era flag was never set throughout the codebase (only read). Also removed
1342 SSL3_FLAGS_POP_BUFFER which was only used if
1343 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1346 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1347 to be "oneline" instead of "compat".
1350 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1351 not aware of clients that still exhibit this bug, and the workaround
1352 hasn't been working properly for a while.
1355 *) The return type of BIO_number_read() and BIO_number_written() as well as
1356 the corresponding num_read and num_write members in the BIO structure has
1357 changed from unsigned long to uint64_t. On platforms where an unsigned
1358 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1362 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1363 OpenSSL without support for them. It also means that maintaining
1364 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1365 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1368 *) Removed support for the two export grade static DH ciphersuites
1369 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1370 were newly added (along with a number of other static DH ciphersuites) to
1371 1.0.2. However the two export ones have *never* worked since they were
1372 introduced. It seems strange in any case to be adding new export
1373 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1376 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1377 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1378 and turned into macros which simply call the new preferred function names
1379 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1380 should use the new names instead. Also as part of this change the ssl23.h
1381 header file has been removed.
1384 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1385 code and the associated standard is no longer considered fit-for-purpose.
1388 *) RT2547 was closed. When generating a private key, try to make the
1389 output file readable only by the owner. This behavior change might
1390 be noticeable when interacting with other software.
1392 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1396 *) Added HTTP GET support to the ocsp command.
1399 *) Changed default digest for the dgst and enc commands from MD5 to
1403 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1406 *) Added support for TLS extended master secret from
1407 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1408 initial patch which was a great help during development.
1411 *) All libssl internal structures have been removed from the public header
1412 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1413 now redundant). Users should not attempt to access internal structures
1414 directly. Instead they should use the provided API functions.
1417 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1418 Access to deprecated functions can be re-enabled by running config with
1419 "enable-deprecated". In addition applications wishing to use deprecated
1420 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1421 will, by default, disable some transitive includes that previously existed
1422 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1425 *) Added support for OCB mode. OpenSSL has been granted a patent license
1426 compatible with the OpenSSL license for use of OCB. Details are available
1427 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1428 for OCB can be removed by calling config with no-ocb.
1431 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1432 compatible client hello.
1435 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1436 done while fixing the error code for the key-too-small case.
1437 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1439 *) CA.sh has been removed; use CA.pl instead.
1442 *) Removed old DES API.
1445 *) Remove various unsupported platforms:
1451 Sinix/ReliantUNIX RM400
1456 16-bit platforms such as WIN16
1459 *) Clean up OPENSSL_NO_xxx #define's
1460 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1461 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1462 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1463 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1464 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1465 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1466 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1467 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1468 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1469 Remove MS_STATIC; it's a relic from platforms <32 bits.
1472 *) Cleaned up dead code
1473 Remove all but one '#ifdef undef' which is to be looked at.
1476 *) Clean up calling of xxx_free routines.
1477 Just like free(), fix most of the xxx_free routines to accept
1478 NULL. Remove the non-null checks from callers. Save much code.
1481 *) Add secure heap for storage of private keys (when possible).
1482 Add BIO_s_secmem(), CBIGNUM, etc.
1483 Contributed by Akamai Technologies under our Corporate CLA.
1486 *) Experimental support for a new, fast, unbiased prime candidate generator,
1487 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1488 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1490 *) New output format NSS in the sess_id command line tool. This allows
1491 exporting the session id and the master key in NSS keylog format.
1492 [Martin Kaiser <martin@kaiser.cx>]
1494 *) Harmonize version and its documentation. -f flag is used to display
1496 [mancha <mancha1@zoho.com>]
1498 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1499 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1500 [mancha <mancha1@zoho.com>]
1502 *) Fix some double frees. These are not thought to be exploitable.
1503 [mancha <mancha1@zoho.com>]
1505 *) A missing bounds check in the handling of the TLS heartbeat extension
1506 can be used to reveal up to 64k of memory to a connected client or
1509 Thanks for Neel Mehta of Google Security for discovering this bug and to
1510 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1511 preparing the fix (CVE-2014-0160)
1512 [Adam Langley, Bodo Moeller]
1514 *) Fix for the attack described in the paper "Recovering OpenSSL
1515 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1516 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1517 http://eprint.iacr.org/2014/140
1519 Thanks to Yuval Yarom and Naomi Benger for discovering this
1520 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1521 [Yuval Yarom and Naomi Benger]
1523 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1524 this fixes a limitation in previous versions of OpenSSL.
1527 *) Experimental encrypt-then-mac support.
1529 Experimental support for encrypt then mac from
1530 draft-gutmann-tls-encrypt-then-mac-02.txt
1532 To enable it set the appropriate extension number (0x42 for the test
1533 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1535 For non-compliant peers (i.e. just about everything) this should have no
1538 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1542 *) Add EVP support for key wrapping algorithms, to avoid problems with
1543 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1544 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1545 algorithms and include tests cases.
1548 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1552 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1553 MGF1 digest and OAEP label.
1556 *) Make openssl verify return errors.
1557 [Chris Palmer <palmer@google.com> and Ben Laurie]
1559 *) New function ASN1_TIME_diff to calculate the difference between two
1560 ASN1_TIME structures or one structure and the current time.
1563 *) Update fips_test_suite to support multiple command line options. New
1564 test to induce all self test errors in sequence and check expected
1568 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1569 sign or verify all in one operation.
1572 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1573 test programs and fips_test_suite. Includes functionality to parse
1574 the minimal script output of fipsalgest.pl directly.
1577 *) Add authorisation parameter to FIPS_module_mode_set().
1580 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1583 *) Use separate DRBG fields for internal and external flags. New function
1584 FIPS_drbg_health_check() to perform on demand health checking. Add
1585 generation tests to fips_test_suite with reduced health check interval to
1586 demonstrate periodic health checking. Add "nodh" option to
1587 fips_test_suite to skip very slow DH test.
1590 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1594 *) More extensive health check for DRBG checking many more failure modes.
1595 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1596 combination: call this in fips_test_suite.
1599 *) Add support for canonical generation of DSA parameter 'g'. See
1602 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1603 POST to handle HMAC cases.
1606 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1607 to return numerical and string versions of the FIPS module number.
1610 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1611 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1612 outside the validated module in the FIPS capable OpenSSL.
1615 *) Minor change to DRBG entropy callback semantics. In some cases
1616 there is no multiple of the block length between min_len and
1617 max_len. Allow the callback to return more than max_len bytes
1618 of entropy but discard any extra: it is the callback's responsibility
1619 to ensure that the extra data discarded does not impact the
1620 requested amount of entropy.
1623 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1624 information in FIPS186-3, SP800-57 and SP800-131A.
1627 *) CCM support via EVP. Interface is very similar to GCM case except we
1628 must supply all data in one chunk (i.e. no update, final) and the
1629 message length must be supplied if AAD is used. Add algorithm test
1633 *) Initial version of POST overhaul. Add POST callback to allow the status
1634 of POST to be monitored and/or failures induced. Modify fips_test_suite
1635 to use callback. Always run all selftests even if one fails.
1638 *) XTS support including algorithm test driver in the fips_gcmtest program.
1639 Note: this does increase the maximum key length from 32 to 64 bytes but
1640 there should be no binary compatibility issues as existing applications
1641 will never use XTS mode.
1644 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1645 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1646 performs algorithm blocking for unapproved PRNG types. Also do not
1647 set PRNG type in FIPS_mode_set(): leave this to the application.
1648 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1649 the standard OpenSSL PRNG: set additional data to a date time vector.
1652 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1653 This shouldn't present any incompatibility problems because applications
1654 shouldn't be using these directly and any that are will need to rethink
1655 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1658 *) Extensive self tests and health checking required by SP800-90 DRBG.
1659 Remove strength parameter from FIPS_drbg_instantiate and always
1660 instantiate at maximum supported strength.
1663 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1666 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1669 *) New function DH_compute_key_padded() to compute a DH key and pad with
1670 leading zeroes if needed: this complies with SP800-56A et al.
1673 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1674 anything, incomplete, subject to change and largely untested at present.
1677 *) Modify fipscanisteronly build option to only build the necessary object
1678 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1681 *) Add experimental option FIPSSYMS to give all symbols in
1682 fipscanister.o and FIPS or fips prefix. This will avoid
1683 conflicts with future versions of OpenSSL. Add perl script
1684 util/fipsas.pl to preprocess assembly language source files
1685 and rename any affected symbols.
1688 *) Add selftest checks and algorithm block of non-fips algorithms in
1689 FIPS mode. Remove DES2 from selftests.
1692 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1693 return internal method without any ENGINE dependencies. Add new
1694 tiny fips sign and verify functions.
1697 *) New build option no-ec2m to disable characteristic 2 code.
1700 *) New build option "fipscanisteronly". This only builds fipscanister.o
1701 and (currently) associated fips utilities. Uses the file Makefile.fips
1702 instead of Makefile.org as the prototype.
1705 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1706 Update fips_gcmtest to use IV generator.
1709 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1710 setting output buffer to NULL. The *Final function must be
1711 called although it will not retrieve any additional data. The tag
1712 can be set or retrieved with a ctrl. The IV length is by default 12
1713 bytes (96 bits) but can be set to an alternative value. If the IV
1714 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1718 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1719 underlying do_cipher function handles all cipher semantics itself
1720 including padding and finalisation. This is useful if (for example)
1721 an ENGINE cipher handles block padding itself. The behaviour of
1722 do_cipher is subtly changed if this flag is set: the return value
1723 is the number of characters written to the output buffer (zero is
1724 no longer an error code) or a negative error code. Also if the
1725 input buffer is NULL and length 0 finalisation should be performed.
1728 *) If a candidate issuer certificate is already part of the constructed
1729 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1732 *) Improve forward-security support: add functions
1734 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1735 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1737 for use by SSL/TLS servers; the callback function will be called whenever a
1738 new session is created, and gets to decide whether the session may be
1739 cached to make it resumable (return 0) or not (return 1). (As by the
1740 SSL/TLS protocol specifications, the session_id sent by the server will be
1741 empty to indicate that the session is not resumable; also, the server will
1742 not generate RFC 4507 (RFC 5077) session tickets.)
1744 A simple reasonable callback implementation is to return is_forward_secure.
1745 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1746 by the SSL/TLS server library, indicating whether it can provide forward
1748 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1750 *) New -verify_name option in command line utilities to set verification
1754 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1755 Add CMAC pkey methods.
1758 *) Experimental renegotiation in s_server -www mode. If the client
1759 browses /reneg connection is renegotiated. If /renegcert it is
1760 renegotiated requesting a certificate.
1763 *) Add an "external" session cache for debugging purposes to s_server. This
1764 should help trace issues which normally are only apparent in deployed
1765 multi-process servers.
1768 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1769 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1770 BIO_set_cipher() and some obscure PEM functions were changed so they
1771 can now return an error. The RAND changes required a change to the
1772 RAND_METHOD structure.
1775 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1776 a gcc attribute to warn if the result of a function is ignored. This
1777 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1778 whose return value is often ignored.
1781 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1782 These allow SCTs (signed certificate timestamps) to be requested and
1783 validated when establishing a connection.
1784 [Rob Percival <robpercival@google.com>]
1786 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1788 *) Prevent padding oracle in AES-NI CBC MAC check
1790 A MITM attacker can use a padding oracle attack to decrypt traffic
1791 when the connection uses an AES CBC cipher and the server support
1794 This issue was introduced as part of the fix for Lucky 13 padding
1795 attack (CVE-2013-0169). The padding check was rewritten to be in
1796 constant time by making sure that always the same bytes are read and
1797 compared against either the MAC or padding bytes. But it no longer
1798 checked that there was enough data to have both the MAC and padding
1801 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1805 *) Fix EVP_EncodeUpdate overflow
1807 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1808 Base64 encoding of binary data. If an attacker is able to supply very large
1809 amounts of input data then a length check can overflow resulting in a heap
1812 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1813 the PEM_write_bio* family of functions. These are mainly used within the
1814 OpenSSL command line applications, so any application which processes data
1815 from an untrusted source and outputs it as a PEM file should be considered
1816 vulnerable to this issue. User applications that call these APIs directly
1817 with large amounts of untrusted data may also be vulnerable.
1819 This issue was reported by Guido Vranken.
1823 *) Fix EVP_EncryptUpdate overflow
1825 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1826 is able to supply very large amounts of input data after a previous call to
1827 EVP_EncryptUpdate() with a partial block then a length check can overflow
1828 resulting in a heap corruption. Following an analysis of all OpenSSL
1829 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1830 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1831 the first called function after an EVP_EncryptInit(), and therefore that
1832 specific call must be safe. The second form is where the length passed to
1833 EVP_EncryptUpdate() can be seen from the code to be some small value and
1834 therefore there is no possibility of an overflow. Since all instances are
1835 one of these two forms, it is believed that there can be no overflows in
1836 internal code due to this problem. It should be noted that
1837 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1838 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1839 of these calls have also been analysed too and it is believed there are no
1840 instances in internal usage where an overflow could occur.
1842 This issue was reported by Guido Vranken.
1846 *) Prevent ASN.1 BIO excessive memory allocation
1848 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1849 a short invalid encoding can cause allocation of large amounts of memory
1850 potentially consuming excessive resources or exhausting memory.
1852 Any application parsing untrusted data through d2i BIO functions is
1853 affected. The memory based functions such as d2i_X509() are *not* affected.
1854 Since the memory based functions are used by the TLS library, TLS
1855 applications are not affected.
1857 This issue was reported by Brian Carpenter.
1863 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1864 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1865 in arbitrary stack data being returned in the buffer.
1867 This issue was reported by Guido Vranken.
1871 *) Modify behavior of ALPN to invoke callback after SNI/servername
1872 callback, such that updates to the SSL_CTX affect ALPN.
1875 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1879 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1880 methods are enabled and ssl2 is disabled the methods return NULL.
1883 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1885 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1886 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1887 provide any "EXPORT" or "LOW" strength ciphers.
1890 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1891 is by default disabled at build-time. Builds that are not configured with
1892 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1893 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1894 will need to explicitly call either of:
1896 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1898 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1900 as appropriate. Even if either of those is used, or the application
1901 explicitly uses the version-specific SSLv2_method() or its client and
1902 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1903 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1904 ciphers, and SSLv2 56-bit DES are no longer available.
1908 *) Fix a double-free in DSA code
1910 A double free bug was discovered when OpenSSL parses malformed DSA private
1911 keys and could lead to a DoS attack or memory corruption for applications
1912 that receive DSA private keys from untrusted sources. This scenario is
1915 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1920 *) Disable SRP fake user seed to address a server memory leak.
1922 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1924 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1925 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1926 was changed to ignore the "fake user" SRP seed, even if the seed
1929 Users should use SRP_VBASE_get1_by_user instead. Note that in
1930 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1931 also that even though configuring the SRP seed attempts to hide
1932 invalid usernames by continuing the handshake with fake
1933 credentials, this behaviour is not constant time and no strong
1934 guarantees are made that the handshake is indistinguishable from
1935 that of a valid user.
1939 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1941 In the BN_hex2bn function the number of hex digits is calculated using an
1942 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1943 large values of |i| this can result in |bn_expand| not allocating any
1944 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1945 field as NULL leading to a subsequent NULL ptr deref. For very large values
1946 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1947 In this case memory is allocated to the internal BIGNUM data field, but it
1948 is insufficiently sized leading to heap corruption. A similar issue exists
1949 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1950 is ever called by user applications with very large untrusted hex/dec data.
1951 This is anticipated to be a rare occurrence.
1953 All OpenSSL internal usage of these functions use data that is not expected
1954 to be untrusted, e.g. config file data or application command line
1955 arguments. If user developed applications generate config file data based
1956 on untrusted data then it is possible that this could also lead to security
1957 consequences. This is also anticipated to be rare.
1959 This issue was reported to OpenSSL by Guido Vranken.
1963 *) Fix memory issues in BIO_*printf functions
1965 The internal |fmtstr| function used in processing a "%s" format string in
1966 the BIO_*printf functions could overflow while calculating the length of a
1967 string and cause an OOB read when printing very long strings.
1969 Additionally the internal |doapr_outch| function can attempt to write to an
1970 OOB memory location (at an offset from the NULL pointer) in the event of a
1971 memory allocation failure. In 1.0.2 and below this could be caused where
1972 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1973 could be in processing a very long "%s" format string. Memory leaks can
1976 The first issue may mask the second issue dependent on compiler behaviour.
1977 These problems could enable attacks where large amounts of untrusted data
1978 is passed to the BIO_*printf functions. If applications use these functions
1979 in this way then they could be vulnerable. OpenSSL itself uses these
1980 functions when printing out human-readable dumps of ASN.1 data. Therefore
1981 applications that print this data could be vulnerable if the data is from
1982 untrusted sources. OpenSSL command line applications could also be
1983 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1984 as command line arguments.
1986 Libssl is not considered directly vulnerable. Additionally certificates etc
1987 received via remote connections via libssl are also unlikely to be able to
1988 trigger these issues because of message size limits enforced within libssl.
1990 This issue was reported to OpenSSL Guido Vranken.
1994 *) Side channel attack on modular exponentiation
1996 A side-channel attack was found which makes use of cache-bank conflicts on
1997 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1998 of RSA keys. The ability to exploit this issue is limited as it relies on
1999 an attacker who has control of code in a thread running on the same
2000 hyper-threaded core as the victim thread which is performing decryptions.
2002 This issue was reported to OpenSSL by Yuval Yarom, The University of
2003 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2004 Nadia Heninger, University of Pennsylvania with more information at
2005 http://cachebleed.info.
2009 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2010 if no keysize is specified with default_bits. This fixes an
2011 omission in an earlier change that changed all RSA/DSA key generation
2012 apps to use 2048 bits by default.
2015 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2016 *) DH small subgroups
2018 Historically OpenSSL only ever generated DH parameters based on "safe"
2019 primes. More recently (in version 1.0.2) support was provided for
2020 generating X9.42 style parameter files such as those required for RFC 5114
2021 support. The primes used in such files may not be "safe". Where an
2022 application is using DH configured with parameters based on primes that are
2023 not "safe" then an attacker could use this fact to find a peer's private
2024 DH exponent. This attack requires that the attacker complete multiple
2025 handshakes in which the peer uses the same private DH exponent. For example
2026 this could be used to discover a TLS server's private DH exponent if it's
2027 reusing the private DH exponent or it's using a static DH ciphersuite.
2029 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2030 TLS. It is not on by default. If the option is not set then the server
2031 reuses the same private DH exponent for the life of the server process and
2032 would be vulnerable to this attack. It is believed that many popular
2033 applications do set this option and would therefore not be at risk.
2035 The fix for this issue adds an additional check where a "q" parameter is
2036 available (as is the case in X9.42 based parameters). This detects the
2037 only known attack, and is the only possible defense for static DH
2038 ciphersuites. This could have some performance impact.
2040 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2041 default and cannot be disabled. This could have some performance impact.
2043 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2047 *) SSLv2 doesn't block disabled ciphers
2049 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2050 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2051 been disabled, provided that the SSLv2 protocol was not also disabled via
2054 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2055 and Sebastian Schinzel.
2059 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2061 *) BN_mod_exp may produce incorrect results on x86_64
2063 There is a carry propagating bug in the x86_64 Montgomery squaring
2064 procedure. No EC algorithms are affected. Analysis suggests that attacks
2065 against RSA and DSA as a result of this defect would be very difficult to
2066 perform and are not believed likely. Attacks against DH are considered just
2067 feasible (although very difficult) because most of the work necessary to
2068 deduce information about a private key may be performed offline. The amount
2069 of resources required for such an attack would be very significant and
2070 likely only accessible to a limited number of attackers. An attacker would
2071 additionally need online access to an unpatched system using the target
2072 private key in a scenario with persistent DH parameters and a private
2073 key that is shared between multiple clients. For example this can occur by
2074 default in OpenSSL DHE based SSL/TLS ciphersuites.
2076 This issue was reported to OpenSSL by Hanno Böck.
2080 *) Certificate verify crash with missing PSS parameter
2082 The signature verification routines will crash with a NULL pointer
2083 dereference if presented with an ASN.1 signature using the RSA PSS
2084 algorithm and absent mask generation function parameter. Since these
2085 routines are used to verify certificate signature algorithms this can be
2086 used to crash any certificate verification operation and exploited in a
2087 DoS attack. Any application which performs certificate verification is
2088 vulnerable including OpenSSL clients and servers which enable client
2091 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2095 *) X509_ATTRIBUTE memory leak
2097 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2098 memory. This structure is used by the PKCS#7 and CMS routines so any
2099 application which reads PKCS#7 or CMS data from untrusted sources is
2100 affected. SSL/TLS is not affected.
2102 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2107 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2108 This changes the decoding behaviour for some invalid messages,
2109 though the change is mostly in the more lenient direction, and
2110 legacy behaviour is preserved as much as possible.
2113 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2115 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2117 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2119 *) Alternate chains certificate forgery
2121 During certificate verification, OpenSSL will attempt to find an
2122 alternative certificate chain if the first attempt to build such a chain
2123 fails. An error in the implementation of this logic can mean that an
2124 attacker could cause certain checks on untrusted certificates to be
2125 bypassed, such as the CA flag, enabling them to use a valid leaf
2126 certificate to act as a CA and "issue" an invalid certificate.
2128 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2132 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2134 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2135 incompatibility in the handling of HMAC. The previous ABI has now been
2139 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2141 *) Malformed ECParameters causes infinite loop
2143 When processing an ECParameters structure OpenSSL enters an infinite loop
2144 if the curve specified is over a specially malformed binary polynomial
2147 This can be used to perform denial of service against any
2148 system which processes public keys, certificate requests or
2149 certificates. This includes TLS clients and TLS servers with
2150 client authentication enabled.
2152 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2156 *) Exploitable out-of-bounds read in X509_cmp_time
2158 X509_cmp_time does not properly check the length of the ASN1_TIME
2159 string and can read a few bytes out of bounds. In addition,
2160 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2163 An attacker can use this to craft malformed certificates and CRLs of
2164 various sizes and potentially cause a segmentation fault, resulting in
2165 a DoS on applications that verify certificates or CRLs. TLS clients
2166 that verify CRLs are affected. TLS clients and servers with client
2167 authentication enabled may be affected if they use custom verification
2170 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2171 independently by Hanno Böck.
2175 *) PKCS7 crash with missing EnvelopedContent
2177 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2178 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2179 with missing content and trigger a NULL pointer dereference on parsing.
2181 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2182 structures from untrusted sources are affected. OpenSSL clients and
2183 servers are not affected.
2185 This issue was reported to OpenSSL by Michal Zalewski (Google).
2189 *) CMS verify infinite loop with unknown hash function
2191 When verifying a signedData message the CMS code can enter an infinite loop
2192 if presented with an unknown hash function OID. This can be used to perform
2193 denial of service against any system which verifies signedData messages using
2195 This issue was reported to OpenSSL by Johannes Bauer.
2199 *) Race condition handling NewSessionTicket
2201 If a NewSessionTicket is received by a multi-threaded client when attempting to
2202 reuse a previous ticket then a race condition can occur potentially leading to
2203 a double free of the ticket data.
2207 *) Only support 256-bit or stronger elliptic curves with the
2208 'ecdh_auto' setting (server) or by default (client). Of supported
2209 curves, prefer P-256 (both).
2212 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2214 *) ClientHello sigalgs DoS fix
2216 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2217 invalid signature algorithms extension a NULL pointer dereference will
2218 occur. This can be exploited in a DoS attack against the server.
2220 This issue was was reported to OpenSSL by David Ramos of Stanford
2223 [Stephen Henson and Matt Caswell]
2225 *) Multiblock corrupted pointer fix
2227 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2228 feature only applies on 64 bit x86 architecture platforms that support AES
2229 NI instructions. A defect in the implementation of "multiblock" can cause
2230 OpenSSL's internal write buffer to become incorrectly set to NULL when
2231 using non-blocking IO. Typically, when the user application is using a
2232 socket BIO for writing, this will only result in a failed connection.
2233 However if some other BIO is used then it is likely that a segmentation
2234 fault will be triggered, thus enabling a potential DoS attack.
2236 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2240 *) Segmentation fault in DTLSv1_listen fix
2242 The DTLSv1_listen function is intended to be stateless and processes the
2243 initial ClientHello from many peers. It is common for user code to loop
2244 over the call to DTLSv1_listen until a valid ClientHello is received with
2245 an associated cookie. A defect in the implementation of DTLSv1_listen means
2246 that state is preserved in the SSL object from one invocation to the next
2247 that can lead to a segmentation fault. Errors processing the initial
2248 ClientHello can trigger this scenario. An example of such an error could be
2249 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2252 This issue was reported to OpenSSL by Per Allansson.
2256 *) Segmentation fault in ASN1_TYPE_cmp fix
2258 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2259 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2260 certificate signature algorithm consistency this can be used to crash any
2261 certificate verification operation and exploited in a DoS attack. Any
2262 application which performs certificate verification is vulnerable including
2263 OpenSSL clients and servers which enable client authentication.
2267 *) Segmentation fault for invalid PSS parameters fix
2269 The signature verification routines will crash with a NULL pointer
2270 dereference if presented with an ASN.1 signature using the RSA PSS
2271 algorithm and invalid parameters. Since these routines are used to verify
2272 certificate signature algorithms this can be used to crash any
2273 certificate verification operation and exploited in a DoS attack. Any
2274 application which performs certificate verification is vulnerable including
2275 OpenSSL clients and servers which enable client authentication.
2277 This issue was was reported to OpenSSL by Brian Carpenter.
2281 *) ASN.1 structure reuse memory corruption fix
2283 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2284 memory corruption via an invalid write. Such reuse is and has been
2285 strongly discouraged and is believed to be rare.
2287 Applications that parse structures containing CHOICE or ANY DEFINED BY
2288 components may be affected. Certificate parsing (d2i_X509 and related
2289 functions) are however not affected. OpenSSL clients and servers are
2294 *) PKCS7 NULL pointer dereferences fix
2296 The PKCS#7 parsing code does not handle missing outer ContentInfo
2297 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2298 missing content and trigger a NULL pointer dereference on parsing.
2300 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2301 otherwise parse PKCS#7 structures from untrusted sources are
2302 affected. OpenSSL clients and servers are not affected.
2304 This issue was reported to OpenSSL by Michal Zalewski (Google).
2308 *) DoS via reachable assert in SSLv2 servers fix
2310 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2311 servers that both support SSLv2 and enable export cipher suites by sending
2312 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2314 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2315 (OpenSSL development team).
2319 *) Empty CKE with client auth and DHE fix
2321 If client auth is used then a server can seg fault in the event of a DHE
2322 ciphersuite being selected and a zero length ClientKeyExchange message
2323 being sent by the client. This could be exploited in a DoS attack.
2327 *) Handshake with unseeded PRNG fix
2329 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2330 with an unseeded PRNG. The conditions are:
2331 - The client is on a platform where the PRNG has not been seeded
2332 automatically, and the user has not seeded manually
2333 - A protocol specific client method version has been used (i.e. not
2334 SSL_client_methodv23)
2335 - A ciphersuite is used that does not require additional random data from
2336 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2338 If the handshake succeeds then the client random that has been used will
2339 have been generated from a PRNG with insufficient entropy and therefore the
2340 output may be predictable.
2342 For example using the following command with an unseeded openssl will
2343 succeed on an unpatched platform:
2345 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2349 *) Use After Free following d2i_ECPrivatekey error fix
2351 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2352 could cause a use after free condition. This, in turn, could cause a double
2353 free in several private key parsing functions (such as d2i_PrivateKey
2354 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2355 for applications that receive EC private keys from untrusted
2356 sources. This scenario is considered rare.
2358 This issue was discovered by the BoringSSL project and fixed in their
2363 *) X509_to_X509_REQ NULL pointer deref fix
2365 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2366 the certificate key is invalid. This function is rarely used in practice.
2368 This issue was discovered by Brian Carpenter.
2372 *) Removed the export ciphers from the DEFAULT ciphers
2375 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2377 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2378 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2379 So far those who have to target multiple platforms would compromise
2380 and argue that binary targeting say ARMv5 would still execute on
2381 ARMv8. "Universal" build resolves this compromise by providing
2382 near-optimal performance even on newer platforms.
2385 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2386 (other platforms pending).
2387 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2389 *) Add support for the SignedCertificateTimestampList certificate and
2390 OCSP response extensions from RFC6962.
2393 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2394 for corner cases. (Certain input points at infinity could lead to
2395 bogus results, with non-infinity inputs mapped to infinity too.)
2398 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2399 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2400 common cases are optimized and there still is room for further
2401 improvements. Vector Permutation AES for Altivec is also added.
2404 *) Add support for little-endian ppc64 Linux target.
2405 [Marcelo Cerri (IBM)]
2407 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2408 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2409 are optimized and there still is room for further improvements.
2410 Both 32- and 64-bit modes are supported.
2411 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2413 *) Improved ARMv7 NEON support.
2416 *) Support for SPARC Architecture 2011 crypto extensions, first
2417 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2418 SHA256/512, MD5, GHASH and modular exponentiation.
2419 [Andy Polyakov, David Miller]
2421 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2423 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2425 *) Support for new and upcoming Intel processors, including AVX2,
2426 BMI and SHA ISA extensions. This includes additional "stitched"
2427 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2430 This work was sponsored by Intel Corp.
2433 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2434 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2435 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2438 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2439 this fixes a limitation in previous versions of OpenSSL.
2442 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2443 MGF1 digest and OAEP label.
2446 *) Add EVP support for key wrapping algorithms, to avoid problems with
2447 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2448 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2449 algorithms and include tests cases.
2452 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2454 [Douglas E. Engert, Steve Henson]
2456 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2457 difference in days and seconds between two tm or ASN1_TIME structures.
2460 *) Add -rev test option to s_server to just reverse order of characters
2461 received by client and send back to server. Also prints an abbreviated
2462 summary of the connection parameters.
2465 *) New option -brief for s_client and s_server to print out a brief summary
2466 of connection parameters.
2469 *) Add callbacks for arbitrary TLS extensions.
2470 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2472 *) New option -crl_download in several openssl utilities to download CRLs
2473 from CRLDP extension in certificates.
2476 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2479 *) New function X509_CRL_diff to generate a delta CRL from the difference
2480 of two full CRLs. Add support to "crl" utility.
2483 *) New functions to set lookup_crls function and to retrieve
2484 X509_STORE from X509_STORE_CTX.
2487 *) Print out deprecated issuer and subject unique ID fields in
2491 *) Extend OCSP I/O functions so they can be used for simple general purpose
2492 HTTP as well as OCSP. New wrapper function which can be used to download
2493 CRLs using the OCSP API.
2496 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2499 *) SSL_CONF* functions. These provide a common framework for application
2500 configuration using configuration files or command lines.
2503 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2504 message callback and prints the results. Needs compile time option
2505 "enable-ssl-trace". New options to s_client and s_server to enable
2509 *) New ctrl and macro to retrieve supported points extensions.
2510 Print out extension in s_server and s_client.
2513 *) New functions to retrieve certificate signature and signature
2517 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2521 *) New Suite B modes for TLS code. These use and enforce the requirements
2522 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2523 only use Suite B curves. The Suite B modes can be set by using the
2524 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2527 *) New chain verification flags for Suite B levels of security. Check
2528 algorithms are acceptable when flags are set in X509_verify_cert.
2531 *) Make tls1_check_chain return a set of flags indicating checks passed
2532 by a certificate chain. Add additional tests to handle client
2533 certificates: checks for matching certificate type and issuer name
2537 *) If an attempt is made to use a signature algorithm not in the peer
2538 preference list abort the handshake. If client has no suitable
2539 signature algorithms in response to a certificate request do not
2540 use the certificate.
2543 *) If server EC tmp key is not in client preference list abort handshake.
2546 *) Add support for certificate stores in CERT structure. This makes it
2547 possible to have different stores per SSL structure or one store in
2548 the parent SSL_CTX. Include distinct stores for certificate chain
2549 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2550 to build and store a certificate chain in CERT structure: returning
2551 an error if the chain cannot be built: this will allow applications
2552 to test if a chain is correctly configured.
2554 Note: if the CERT based stores are not set then the parent SSL_CTX
2555 store is used to retain compatibility with existing behaviour.
2559 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2560 mask based on the current session, check mask when sending client
2561 hello and checking the requested ciphersuite.
2564 *) New ctrls to retrieve and set certificate types in a certificate
2565 request message. Print out received values in s_client. If certificate
2566 types is not set with custom values set sensible values based on
2567 supported signature algorithms.
2570 *) Support for distinct client and server supported signature algorithms.
2573 *) Add certificate callback. If set this is called whenever a certificate
2574 is required by client or server. An application can decide which
2575 certificate chain to present based on arbitrary criteria: for example
2576 supported signature algorithms. Add very simple example to s_server.
2577 This fixes many of the problems and restrictions of the existing client
2578 certificate callback: for example you can now clear an existing
2579 certificate and specify the whole chain.
2582 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2583 the certificate can be used for (if anything). Set valid_flags field
2584 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2585 to have similar checks in it.
2587 Add new "cert_flags" field to CERT structure and include a "strict mode".
2588 This enforces some TLS certificate requirements (such as only permitting
2589 certificate signature algorithms contained in the supported algorithms
2590 extension) which some implementations ignore: this option should be used
2591 with caution as it could cause interoperability issues.
2594 *) Update and tidy signature algorithm extension processing. Work out
2595 shared signature algorithms based on preferences and peer algorithms
2596 and print them out in s_client and s_server. Abort handshake if no
2597 shared signature algorithms.
2600 *) Add new functions to allow customised supported signature algorithms
2601 for SSL and SSL_CTX structures. Add options to s_client and s_server
2605 *) New function SSL_certs_clear() to delete all references to certificates
2606 from an SSL structure. Before this once a certificate had been added
2607 it couldn't be removed.
2610 *) Integrate hostname, email address and IP address checking with certificate
2611 verification. New verify options supporting checking in openssl utility.
2614 *) Fixes and wildcard matching support to hostname and email checking
2615 functions. Add manual page.
2616 [Florian Weimer (Red Hat Product Security Team)]
2618 *) New functions to check a hostname email or IP address against a
2619 certificate. Add options x509 utility to print results of checks against
2623 *) Fix OCSP checking.
2624 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2626 *) Initial experimental support for explicitly trusted non-root CAs.
2627 OpenSSL still tries to build a complete chain to a root but if an
2628 intermediate CA has a trust setting included that is used. The first
2629 setting is used: whether to trust (e.g., -addtrust option to the x509
2633 *) Add -trusted_first option which attempts to find certificates in the
2634 trusted store even if an untrusted chain is also supplied.
2637 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2638 platform support for Linux and Android.
2641 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2644 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2645 When in FIPS mode the approved implementations are used as normal,
2646 when not in FIPS mode the internal unapproved versions are used instead.
2647 This means that the FIPS capable OpenSSL isn't forced to use the
2648 (often lower performance) FIPS implementations outside FIPS mode.
2651 *) Transparently support X9.42 DH parameters when calling
2652 PEM_read_bio_DHparameters. This means existing applications can handle
2653 the new parameter format automatically.
2656 *) Initial experimental support for X9.42 DH parameter format: mainly
2657 to support use of 'q' parameter for RFC5114 parameters.
2660 *) Add DH parameters from RFC5114 including test data to dhtest.
2663 *) Support for automatic EC temporary key parameter selection. If enabled
2664 the most preferred EC parameters are automatically used instead of
2665 hardcoded fixed parameters. Now a server just has to call:
2666 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2667 support ECDH and use the most appropriate parameters.
2670 *) Enhance and tidy EC curve and point format TLS extension code. Use
2671 static structures instead of allocation if default values are used.
2672 New ctrls to set curves we wish to support and to retrieve shared curves.
2673 Print out shared curves in s_server. New options to s_server and s_client
2674 to set list of supported curves.
2677 *) New ctrls to retrieve supported signature algorithms and
2678 supported curve values as an array of NIDs. Extend openssl utility
2679 to print out received values.
2682 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2683 between NIDs and the more common NIST names such as "P-256". Enhance
2684 ecparam utility and ECC method to recognise the NIST names for curves.
2687 *) Enhance SSL/TLS certificate chain handling to support different
2688 chains for each certificate instead of one chain in the parent SSL_CTX.
2691 *) Support for fixed DH ciphersuite client authentication: where both
2692 server and client use DH certificates with common parameters.
2695 *) Support for fixed DH ciphersuites: those requiring DH server
2699 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2701 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2702 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2703 X509_CINF_get_signature were reverted post internal team review.
2705 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2707 *) Build fixes for the Windows and OpenVMS platforms
2708 [Matt Caswell and Richard Levitte]
2710 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2712 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2713 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2714 dereference. This could lead to a Denial Of Service attack. Thanks to
2715 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2719 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2720 dtls1_buffer_record function under certain conditions. In particular this
2721 could occur if an attacker sent repeated DTLS records with the same
2722 sequence number but for the next epoch. The memory leak could be exploited
2723 by an attacker in a Denial of Service attack through memory exhaustion.
2724 Thanks to Chris Mueller for reporting this issue.
2728 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2729 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2730 method would be set to NULL which could later result in a NULL pointer
2731 dereference. Thanks to Frank Schmirler for reporting this issue.
2735 *) Abort handshake if server key exchange message is omitted for ephemeral
2738 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2739 reporting this issue.
2743 *) Remove non-export ephemeral RSA code on client and server. This code
2744 violated the TLS standard by allowing the use of temporary RSA keys in
2745 non-export ciphersuites and could be used by a server to effectively
2746 downgrade the RSA key length used to a value smaller than the server
2747 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2748 INRIA or reporting this issue.
2752 *) Fixed issue where DH client certificates are accepted without verification.
2753 An OpenSSL server will accept a DH certificate for client authentication
2754 without the certificate verify message. This effectively allows a client to
2755 authenticate without the use of a private key. This only affects servers
2756 which trust a client certificate authority which issues certificates
2757 containing DH keys: these are extremely rare and hardly ever encountered.
2758 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2763 *) Ensure that the session ID context of an SSL is updated when its
2764 SSL_CTX is updated via SSL_set_SSL_CTX.
2766 The session ID context is typically set from the parent SSL_CTX,
2767 and can vary with the CTX.
2770 *) Fix various certificate fingerprint issues.
2772 By using non-DER or invalid encodings outside the signed portion of a
2773 certificate the fingerprint can be changed without breaking the signature.
2774 Although no details of the signed portion of the certificate can be changed
2775 this can cause problems with some applications: e.g. those using the
2776 certificate fingerprint for blacklists.
2778 1. Reject signatures with non zero unused bits.
2780 If the BIT STRING containing the signature has non zero unused bits reject
2781 the signature. All current signature algorithms require zero unused bits.
2783 2. Check certificate algorithm consistency.
2785 Check the AlgorithmIdentifier inside TBS matches the one in the
2786 certificate signature. NB: this will result in signature failure
2787 errors for some broken certificates.
2789 Thanks to Konrad Kraszewski from Google for reporting this issue.
2791 3. Check DSA/ECDSA signatures use DER.
2793 Re-encode DSA/ECDSA signatures and compare with the original received
2794 signature. Return an error if there is a mismatch.
2796 This will reject various cases including garbage after signature
2797 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2798 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2799 (negative or with leading zeroes).
2801 Further analysis was conducted and fixes were developed by Stephen Henson
2802 of the OpenSSL core team.
2807 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2808 results on some platforms, including x86_64. This bug occurs at random
2809 with a very low probability, and is not known to be exploitable in any
2810 way, though its exact impact is difficult to determine. Thanks to Pieter
2811 Wuille (Blockstream) who reported this issue and also suggested an initial
2812 fix. Further analysis was conducted by the OpenSSL development team and
2813 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2814 the OpenSSL core team.
2818 *) Do not resume sessions on the server if the negotiated protocol
2819 version does not match the session's version. Resuming with a different
2820 version, while not strictly forbidden by the RFC, is of questionable
2821 sanity and breaks all known clients.
2822 [David Benjamin, Emilia Käsper]
2824 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2825 early CCS messages during renegotiation. (Note that because
2826 renegotiation is encrypted, this early CCS was not exploitable.)
2829 *) Tighten client-side session ticket handling during renegotiation:
2830 ensure that the client only accepts a session ticket if the server sends
2831 the extension anew in the ServerHello. Previously, a TLS client would
2832 reuse the old extension state and thus accept a session ticket if one was
2833 announced in the initial ServerHello.
2835 Similarly, ensure that the client requires a session ticket if one
2836 was advertised in the ServerHello. Previously, a TLS client would
2837 ignore a missing NewSessionTicket message.
2840 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2842 *) SRTP Memory Leak.
2844 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2845 sends a carefully crafted handshake message, to cause OpenSSL to fail
2846 to free up to 64k of memory causing a memory leak. This could be
2847 exploited in a Denial Of Service attack. This issue affects OpenSSL
2848 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2849 whether SRTP is used or configured. Implementations of OpenSSL that
2850 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2852 The fix was developed by the OpenSSL team.
2856 *) Session Ticket Memory Leak.
2858 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2859 integrity of that ticket is first verified. In the event of a session
2860 ticket integrity check failing, OpenSSL will fail to free memory
2861 causing a memory leak. By sending a large number of invalid session
2862 tickets an attacker could exploit this issue in a Denial Of Service
2867 *) Build option no-ssl3 is incomplete.
2869 When OpenSSL is configured with "no-ssl3" as a build option, servers
2870 could accept and complete a SSL 3.0 handshake, and clients could be
2871 configured to send them.
2873 [Akamai and the OpenSSL team]
2875 *) Add support for TLS_FALLBACK_SCSV.
2876 Client applications doing fallback retries should call
2877 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2879 [Adam Langley, Bodo Moeller]
2881 *) Add additional DigestInfo checks.
2883 Re-encode DigestInto in DER and check against the original when
2884 verifying RSA signature: this will reject any improperly encoded
2885 DigestInfo structures.
2887 Note: this is a precautionary measure and no attacks are currently known.
2891 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2893 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2894 SRP code can be overrun an internal buffer. Add sanity check that
2895 g, A, B < N to SRP code.
2897 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2898 Group for discovering this issue.
2902 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2903 TLS 1.0 instead of higher protocol versions when the ClientHello message
2904 is badly fragmented. This allows a man-in-the-middle attacker to force a
2905 downgrade to TLS 1.0 even if both the server and the client support a
2906 higher protocol version, by modifying the client's TLS records.
2908 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2909 researching this issue.
2913 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2914 to a denial of service attack. A malicious server can crash the client
2915 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2916 ciphersuite and sending carefully crafted handshake messages.
2918 Thanks to Felix Gröbert (Google) for discovering and researching this
2923 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2924 to leak memory. This can be exploited through a Denial of Service attack.
2925 Thanks to Adam Langley for discovering and researching this issue.
2929 *) An attacker can force openssl to consume large amounts of memory whilst
2930 processing DTLS handshake messages. This can be exploited through a
2931 Denial of Service attack.
2932 Thanks to Adam Langley for discovering and researching this issue.
2936 *) An attacker can force an error condition which causes openssl to crash
2937 whilst processing DTLS packets due to memory being freed twice. This
2938 can be exploited through a Denial of Service attack.
2939 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2944 *) If a multithreaded client connects to a malicious server using a resumed
2945 session and the server sends an ec point format extension it could write
2946 up to 255 bytes to freed memory.
2948 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2953 *) A malicious server can crash an OpenSSL client with a null pointer
2954 dereference (read) by specifying an SRP ciphersuite even though it was not
2955 properly negotiated with the client. This can be exploited through a
2956 Denial of Service attack.
2958 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2959 discovering and researching this issue.
2963 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2964 X509_name_oneline, X509_name_print_ex et al. to leak some information
2965 from the stack. Applications may be affected if they echo pretty printing
2966 output to the attacker.
2968 Thanks to Ivan Fratric (Google) for discovering this issue.
2970 [Emilia Käsper, and Steve Henson]
2972 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2973 for corner cases. (Certain input points at infinity could lead to
2974 bogus results, with non-infinity inputs mapped to infinity too.)
2977 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2979 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2980 handshake can force the use of weak keying material in OpenSSL
2981 SSL/TLS clients and servers.
2983 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2984 researching this issue. (CVE-2014-0224)
2985 [KIKUCHI Masashi, Steve Henson]
2987 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2988 OpenSSL DTLS client the code can be made to recurse eventually crashing
2991 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2993 [Imre Rad, Steve Henson]
2995 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2996 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2997 client or server. This is potentially exploitable to run arbitrary
2998 code on a vulnerable client or server.
3000 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3001 [Jüri Aedla, Steve Henson]
3003 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3004 are subject to a denial of service attack.
3006 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3007 this issue. (CVE-2014-3470)
3008 [Felix Gröbert, Ivan Fratric, Steve Henson]
3010 *) Harmonize version and its documentation. -f flag is used to display
3012 [mancha <mancha1@zoho.com>]
3014 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3015 in i2d_ECPrivateKey.
3016 [mancha <mancha1@zoho.com>]
3018 *) Fix some double frees. These are not thought to be exploitable.
3019 [mancha <mancha1@zoho.com>]
3021 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3023 *) A missing bounds check in the handling of the TLS heartbeat extension
3024 can be used to reveal up to 64k of memory to a connected client or
3027 Thanks for Neel Mehta of Google Security for discovering this bug and to
3028 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3029 preparing the fix (CVE-2014-0160)
3030 [Adam Langley, Bodo Moeller]
3032 *) Fix for the attack described in the paper "Recovering OpenSSL
3033 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3034 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3035 http://eprint.iacr.org/2014/140
3037 Thanks to Yuval Yarom and Naomi Benger for discovering this
3038 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3039 [Yuval Yarom and Naomi Benger]
3041 *) TLS pad extension: draft-agl-tls-padding-03
3043 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3044 TLS client Hello record length value would otherwise be > 255 and
3045 less that 512 pad with a dummy extension containing zeroes so it
3046 is at least 512 bytes long.
3048 [Adam Langley, Steve Henson]
3050 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3052 *) Fix for TLS record tampering bug. A carefully crafted invalid
3053 handshake could crash OpenSSL with a NULL pointer exception.
3054 Thanks to Anton Johansson for reporting this issues.
3057 *) Keep original DTLS digest and encryption contexts in retransmission
3058 structures so we can use the previous session parameters if they need
3059 to be resent. (CVE-2013-6450)
3062 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3063 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3064 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3065 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3066 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3067 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3068 [Rob Stradling, Adam Langley]
3070 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3072 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3073 supporting platforms or when small records were transferred.
3074 [Andy Polyakov, Steve Henson]
3076 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3078 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3080 This addresses the flaw in CBC record processing discovered by
3081 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3082 at: http://www.isg.rhul.ac.uk/tls/
3084 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3085 Security Group at Royal Holloway, University of London
3086 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3087 Emilia Käsper for the initial patch.
3089 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3091 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3092 ciphersuites which can be exploited in a denial of service attack.
3093 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3094 and detecting this bug and to Wolfgang Ettlinger
3095 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3099 *) Return an error when checking OCSP signatures when key is NULL.
3100 This fixes a DoS attack. (CVE-2013-0166)
3103 *) Make openssl verify return errors.
3104 [Chris Palmer <palmer@google.com> and Ben Laurie]
3106 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3107 the right response is stapled. Also change SSL_get_certificate()
3108 so it returns the certificate actually sent.
3109 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3110 [Rob Stradling <rob.stradling@comodo.com>]
3112 *) Fix possible deadlock when decoding public keys.
3115 *) Don't use TLS 1.0 record version number in initial client hello
3119 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3121 *) Sanity check record length before skipping explicit IV in TLS
3122 1.2, 1.1 and DTLS to fix DoS attack.
3124 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3125 fuzzing as a service testing platform.
3129 *) Initialise tkeylen properly when encrypting CMS messages.
3130 Thanks to Solar Designer of Openwall for reporting this issue.
3133 *) In FIPS mode don't try to use composite ciphers as they are not
3137 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3139 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3140 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3141 mean any application compiled against OpenSSL 1.0.0 headers setting
3142 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3143 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3144 0x10000000L Any application which was previously compiled against
3145 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3146 will need to be recompiled as a result. Letting be results in
3147 inability to disable specifically TLS 1.1 and in client context,
3148 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3151 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3152 disable just protocol X, but all protocols above X *if* there are
3153 protocols *below* X still enabled. In more practical terms it means
3154 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3155 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3156 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3160 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3162 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3163 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3164 in CRYPTO_realloc_clean.
3166 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3167 issue and to Adam Langley <agl@chromium.org> for fixing it.
3169 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3171 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3174 *) Workarounds for some broken servers that "hang" if a client hello
3175 record length exceeds 255 bytes.
3177 1. Do not use record version number > TLS 1.0 in initial client
3178 hello: some (but not all) hanging servers will now work.
3179 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3180 the number of ciphers sent in the client hello. This should be
3181 set to an even number, such as 50, for example by passing:
3182 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3183 Most broken servers should now work.
3184 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3185 TLS 1.2 client support entirely.
3188 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3191 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3193 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3194 STRING form instead of a DigestInfo.
3197 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3198 and the RSA_sign/RSA_verify functions. This was made more apparent when
3199 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3200 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3201 the correct format in RSA_verify so both forms transparently work.
3204 *) Some servers which support TLS 1.0 can choke if we initially indicate
3205 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3206 encrypted premaster secret. As a workaround use the maximum permitted
3207 client version in client hello, this should keep such servers happy
3208 and still work with previous versions of OpenSSL.
3211 *) Add support for TLS/DTLS heartbeats.
3212 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3214 *) Add support for SCTP.
3215 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3217 *) Improved PRNG seeding for VOS.
3218 [Paul Green <Paul.Green@stratus.com>]
3220 *) Extensive assembler packs updates, most notably:
3222 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3223 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3224 - x86_64: bit-sliced AES implementation;
3225 - ARM: NEON support, contemporary platforms optimizations;
3226 - s390x: z196 support;
3227 - *: GHASH and GF(2^m) multiplication implementations;
3231 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3232 (removal of unnecessary code)
3233 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3235 *) Add TLS key material exporter from RFC 5705.
3238 *) Add DTLS-SRTP negotiation from RFC 5764.
3241 *) Add Next Protocol Negotiation,
3242 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3243 disabled with a no-npn flag to config or Configure. Code donated
3245 [Adam Langley <agl@google.com> and Ben Laurie]
3247 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3248 NIST-P256, NIST-P521, with constant-time single point multiplication on
3249 typical inputs. Compiler support for the nonstandard type __uint128_t is
3250 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3251 Code made available under Apache License version 2.0.
3253 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3254 line to include this in your build of OpenSSL, and run "make depend" (or
3255 "make update"). This enables the following EC_METHODs:
3257 EC_GFp_nistp224_method()
3258 EC_GFp_nistp256_method()
3259 EC_GFp_nistp521_method()
3261 EC_GROUP_new_by_curve_name() will automatically use these (while
3262 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3264 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3266 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3267 all platforms. Move ssize_t definition from e_os.h to the public
3268 header file e_os2.h as it now appears in public header file cms.h
3271 *) New -sigopt option to the ca, req and x509 utilities. Additional
3272 signature parameters can be passed using this option and in
3276 *) Add RSA PSS signing function. This will generate and set the
3277 appropriate AlgorithmIdentifiers for PSS based on those in the
3278 corresponding EVP_MD_CTX structure. No application support yet.
3281 *) Support for companion algorithm specific ASN1 signing routines.
3282 New function ASN1_item_sign_ctx() signs a pre-initialised
3283 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3284 the appropriate parameters.
3287 *) Add new algorithm specific ASN1 verification initialisation function
3288 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3289 handling will be the same no matter what EVP_PKEY_METHOD is used.
3290 Add a PSS handler to support verification of PSS signatures: checked
3291 against a number of sample certificates.
3294 *) Add signature printing for PSS. Add PSS OIDs.
3295 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3297 *) Add algorithm specific signature printing. An individual ASN1 method
3298 can now print out signatures instead of the standard hex dump.
3300 More complex signatures (e.g. PSS) can print out more meaningful
3301 information. Include DSA version that prints out the signature
3305 *) Password based recipient info support for CMS library: implementing
3309 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3310 neatly separates the code into cipher and PBE sections and is required
3311 for some algorithms that split PBES2 into separate pieces (such as
3312 password based CMS).
3315 *) Session-handling fixes:
3316 - Fix handling of connections that are resuming with a session ID,
3317 but also support Session Tickets.
3318 - Fix a bug that suppressed issuing of a new ticket if the client
3319 presented a ticket with an expired session.
3320 - Try to set the ticket lifetime hint to something reasonable.
3321 - Make tickets shorter by excluding irrelevant information.
3322 - On the client side, don't ignore renewed tickets.
3323 [Adam Langley, Bodo Moeller (Google)]
3325 *) Fix PSK session representation.
3328 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3330 This work was sponsored by Intel.
3333 *) Add GCM support to TLS library. Some custom code is needed to split
3334 the IV between the fixed (from PRF) and explicit (from TLS record)
3335 portions. This adds all GCM ciphersuites supported by RFC5288 and
3336 RFC5289. Generalise some AES* cipherstrings to include GCM and
3337 add a special AESGCM string for GCM only.
3340 *) Expand range of ctrls for AES GCM. Permit setting invocation
3341 field on decrypt and retrieval of invocation field only on encrypt.
3344 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3345 As required by RFC5289 these ciphersuites cannot be used if for
3346 versions of TLS earlier than 1.2.
3349 *) For FIPS capable OpenSSL interpret a NULL default public key method
3350 as unset and return the appropriate default but do *not* set the default.
3351 This means we can return the appropriate method in applications that
3352 switch between FIPS and non-FIPS modes.
3355 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3356 ENGINE is used then we cannot handle that in the FIPS module so we
3357 keep original code iff non-FIPS operations are allowed.
3360 *) Add -attime option to openssl utilities.
3361 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3363 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3366 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3367 FIPS EC methods unconditionally for now.
3370 *) New build option no-ec2m to disable characteristic 2 code.
3373 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3374 all cases can be covered as some introduce binary incompatibilities.
3377 *) Redirect RSA operations to FIPS module including keygen,
3378 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3381 *) Add similar low level API blocking to ciphers.
3384 *) Low level digest APIs are not approved in FIPS mode: any attempt
3385 to use these will cause a fatal error. Applications that *really* want
3386 to use them can use the private_* version instead.
3389 *) Redirect cipher operations to FIPS module for FIPS builds.
3392 *) Redirect digest operations to FIPS module for FIPS builds.
3395 *) Update build system to add "fips" flag which will link in fipscanister.o
3396 for static and shared library builds embedding a signature if needed.
3399 *) Output TLS supported curves in preference order instead of numerical
3400 order. This is currently hardcoded for the highest order curves first.
3401 This should be configurable so applications can judge speed vs strength.
3404 *) Add TLS v1.2 server support for client authentication.
3407 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3411 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3412 FIPS modules versions.
3415 *) Add TLS v1.2 client side support for client authentication. Keep cache
3416 of handshake records longer as we don't know the hash algorithm to use
3417 until after the certificate request message is received.
3420 *) Initial TLS v1.2 client support. Add a default signature algorithms
3421 extension including all the algorithms we support. Parse new signature
3422 format in client key exchange. Relax some ECC signing restrictions for
3423 TLS v1.2 as indicated in RFC5246.
3426 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3427 to new signature format when needed using client digest preference.
3428 All server ciphersuites should now work correctly in TLS v1.2. No client
3429 support yet and no support for client certificates.
3432 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3433 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3434 ciphersuites. At present only RSA key exchange ciphersuites work with
3435 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3436 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3437 and version checking.
3440 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3441 with this defined it will not be affected by any changes to ssl internal
3442 structures. Add several utility functions to allow openssl application
3443 to work with OPENSSL_NO_SSL_INTERN defined.
3446 *) A long standing patch to add support for SRP from EdelWeb (Peter
3447 Sylvester and Christophe Renou) was integrated.
3448 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3449 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3452 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3455 *) Permit abbreviated handshakes when renegotiating using the function
3456 SSL_renegotiate_abbreviated().
3457 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3459 *) Add call to ENGINE_register_all_complete() to
3460 ENGINE_load_builtin_engines(), so some implementations get used
3461 automatically instead of needing explicit application support.
3464 *) Add support for TLS key exporter as described in RFC5705.
3465 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3467 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3468 a few changes are required:
3470 Add SSL_OP_NO_TLSv1_1 flag.
3471 Add TLSv1_1 methods.
3472 Update version checking logic to handle version 1.1.
3473 Add explicit IV handling (ported from DTLS code).
3474 Add command line options to s_client/s_server.
3477 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3479 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3480 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3481 content decryption and always return the same error. Note: this attack
3482 needs on average 2^20 messages so it only affects automated senders. The
3483 old behaviour can be re-enabled in the CMS code by setting the
3484 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3485 an MMA defence is not necessary.
3486 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3487 this issue. (CVE-2012-0884)
3490 *) Fix CVE-2011-4619: make sure we really are receiving a
3491 client hello before rejecting multiple SGC restarts. Thanks to
3492 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3495 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3497 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3498 Thanks to Antonio Martin, Enterprise Secure Access Research and
3499 Development, Cisco Systems, Inc. for discovering this bug and
3500 preparing a fix. (CVE-2012-0050)
3503 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3505 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3506 of the Vaudenay padding oracle attack on CBC mode encryption
3507 which enables an efficient plaintext recovery attack against
3508 the OpenSSL implementation of DTLS. Their attack exploits timing
3509 differences arising during decryption processing. A research
3510 paper describing this attack can be found at:
3511 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3512 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3513 Security Group at Royal Holloway, University of London
3514 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3515 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3516 for preparing the fix. (CVE-2011-4108)
3517 [Robin Seggelmann, Michael Tuexen]
3519 *) Clear bytes used for block padding of SSL 3.0 records.
3521 [Adam Langley (Google)]
3523 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3524 Kadianakis <desnacked@gmail.com> for discovering this issue and
3525 Adam Langley for preparing the fix. (CVE-2011-4619)
3526 [Adam Langley (Google)]
3528 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3529 [Andrey Kulikov <amdeich@gmail.com>]
3531 *) Prevent malformed RFC3779 data triggering an assertion failure.
3532 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3533 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3534 [Rob Austein <sra@hactrn.net>]
3536 *) Improved PRNG seeding for VOS.
3537 [Paul Green <Paul.Green@stratus.com>]
3539 *) Fix ssl_ciph.c set-up race.
3540 [Adam Langley (Google)]
3542 *) Fix spurious failures in ecdsatest.c.
3543 [Emilia Käsper (Google)]
3545 *) Fix the BIO_f_buffer() implementation (which was mixing different
3546 interpretations of the '..._len' fields).
3547 [Adam Langley (Google)]
3549 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3550 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3551 threads won't reuse the same blinding coefficients.
3553 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3554 lock to call BN_BLINDING_invert_ex, and avoids one use of
3555 BN_BLINDING_update for each BN_BLINDING structure (previously,
3556 the last update always remained unused).
3557 [Emilia Käsper (Google)]
3559 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3560 [Bob Buckholz (Google)]
3562 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3564 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3565 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3566 [Kaspar Brand <ossl@velox.ch>]
3568 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3569 for multi-threaded use of ECDH. (CVE-2011-3210)
3570 [Adam Langley (Google)]
3572 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3575 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3576 signature public key algorithm by using OID xref utilities instead.
3577 Before this you could only use some ECC ciphersuites with SHA1 only.
3580 *) Add protection against ECDSA timing attacks as mentioned in the paper
3581 by Billy Bob Brumley and Nicola Tuveri, see:
3583 http://eprint.iacr.org/2011/232.pdf
3585 [Billy Bob Brumley and Nicola Tuveri]
3587 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3589 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3590 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3592 *) Fix bug in string printing code: if *any* escaping is enabled we must
3593 escape the escape character (backslash) or the resulting string is
3597 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3599 *) Disable code workaround for ancient and obsolete Netscape browsers
3600 and servers: an attacker can use it in a ciphersuite downgrade attack.
3601 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3604 *) Fixed J-PAKE implementation error, originally discovered by
3605 Sebastien Martini, further info and confirmation from Stefan
3606 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3609 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3611 *) Fix extension code to avoid race conditions which can result in a buffer
3612 overrun vulnerability: resumed sessions must not be modified as they can
3613 be shared by multiple threads. CVE-2010-3864
3616 *) Fix WIN32 build system to correctly link an ENGINE directory into
3620 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3622 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3624 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3626 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3628 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3629 context. The operation can be customised via the ctrl mechanism in
projects / openssl.git / blob