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]
12 *) Change generating and checking of primes so that the error rate of not
13 being prime depends on the intended use based on the size of the input.
14 For larger primes this will result in more rounds of Miller-Rabin.
15 The maximal error rate for primes with more than 1080 bits is lowered
17 [Kurt Roeckx, Annie Yousar]
19 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
22 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
23 moving between systems, and to avoid confusion when a Windows build is
24 done with mingw vs with MSVC. For POSIX installs, there's still a
25 symlink or copy named 'tsget' to avoid that confusion as well.
28 *) Revert blinding in ECDSA sign and instead make problematic addition
29 length-invariant. Switch even to fixed-length Montgomery multiplication.
32 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
33 step for binary curves. The new implementation is based on formulas from
34 differential addition-and-doubling in mixed Lopez-Dahab projective
35 coordinates, modified to independently blind the operands.
36 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
38 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
39 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
40 EC_METHODs to implement their own specialized "ladder step", to take
41 advantage of more favorable coordinate systems or more efficient
42 differential addition-and-doubling algorithms.
43 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
45 *) Modified the random device based seed sources to keep the relevant
46 file descriptors open rather than reopening them on each access.
47 This allows such sources to operate in a chroot() jail without
48 the associated device nodes being available. This behaviour can be
49 controlled using RAND_keep_random_devices_open().
52 *) Numerous side-channel attack mitigations have been applied. This may have
53 performance impacts for some algorithms for the benefit of improved
54 security. Specific changes are noted in this change log by their respective
58 *) AIX shared library support overhaul. Switch to AIX "natural" way of
59 handling shared libraries, which means collecting shared objects of
60 different versions and bitnesses in one common archive. This allows to
61 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
62 doesn't affect the way 3rd party applications are linked, only how
63 multi-version installation is managed.
66 *) Make ec_group_do_inverse_ord() more robust and available to other
67 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
68 mitigations are applied to the fallback BN_mod_inverse().
69 When using this function rather than BN_mod_inverse() directly, new
70 EC cryptosystem implementations are then safer-by-default.
73 *) Add coordinate blinding for EC_POINT and implement projective
74 coordinate blinding for generic prime curves as a countermeasure to
75 chosen point SCA attacks.
76 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
78 *) Add blinding to ECDSA and DSA signatures to protect against side channel
79 attacks discovered by Keegan Ryan (NCC Group).
82 *) Enforce checking in the pkeyutl command line app to ensure that the input
83 length does not exceed the maximum supported digest length when performing
84 a sign, verify or verifyrecover operation.
87 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
88 I/O in combination with something like select() or poll() will hang. This
89 can be turned off again using SSL_CTX_clear_mode().
90 Many applications do not properly handle non-application data records, and
91 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
92 around the problems in those applications, but can also break some.
93 It's recommended to read the manpages about SSL_read(), SSL_write(),
94 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
95 SSL_CTX_set_read_ahead() again.
98 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
99 now allow empty (zero character) pass phrases.
102 *) Apply blinding to binary field modular inversion and remove patent
103 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
106 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
107 binary and prime elliptic curves.
110 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
111 constant time fixed point multiplication.
114 *) Revise elliptic curve scalar multiplication with timing attack
115 defenses: ec_wNAF_mul redirects to a constant time implementation
116 when computing fixed point and variable point multiplication (which
117 in OpenSSL are mostly used with secret scalars in keygen, sign,
118 ECDH derive operations).
119 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
122 *) Updated CONTRIBUTING
125 *) Updated DRBG / RAND to request nonce and additional low entropy
126 randomness from the system.
127 [Matthias St. Pierre]
129 *) Updated 'openssl rehash' to use OpenSSL consistent default.
132 *) Moved the load of the ssl_conf module to libcrypto, which helps
133 loading engines that libssl uses before libssl is initialised.
136 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
139 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
140 [Ingo Schwarze, Rich Salz]
142 *) Added output of accepting IP address and port for 'openssl s_server'
145 *) Added a new API for TLSv1.3 ciphersuites:
146 SSL_CTX_set_ciphersuites()
147 SSL_set_ciphersuites()
150 *) Memory allocation failures consistenly add an error to the error
154 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
155 in libcrypto when run as setuid/setgid.
158 *) Load any config file by default when libssl is used.
161 *) Added new public header file <openssl/rand_drbg.h> and documentation
162 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
163 [Matthias St. Pierre]
165 *) QNX support removed (cannot find contributors to get their approval
166 for the license change).
169 *) TLSv1.3 replay protection for early data has been implemented. See the
170 SSL_read_early_data() man page for further details.
173 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
174 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
175 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
176 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
177 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
178 configuration has been separated out. See the ciphers man page or the
179 SSL_CTX_set_ciphersuites() man page for more information.
182 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
183 in responder mode now supports the new "-multi" option, which
184 spawns the specified number of child processes to handle OCSP
185 requests. The "-timeout" option now also limits the OCSP
186 responder's patience to wait to receive the full client request
187 on a newly accepted connection. Child processes are respawned
188 as needed, and the CA index file is automatically reloaded
189 when changed. This makes it possible to run the "ocsp" responder
190 as a long-running service, making the OpenSSL CA somewhat more
191 feature-complete. In this mode, most diagnostic messages logged
192 after entering the event loop are logged via syslog(3) rather than
196 *) Added support for X448 and Ed448. Heavily based on original work by
200 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
201 objects loaded. This adds the functions OSSL_STORE_expect() and
202 OSSL_STORE_find() as well as needed tools to construct searches and
203 get the search data out of them.
206 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
207 version of OpenSSL should review their configuration settings to ensure
208 that they are still appropriate for TLSv1.3. For further information see:
209 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
211 NOTE: In this pre-release of OpenSSL a draft version of the
212 TLSv1.3 standard has been implemented. Implementations of different draft
213 versions of the standard do not inter-operate, and this version will not
214 inter-operate with an implementation of the final standard when it is
215 eventually published. Different pre-release versions may implement
216 different versions of the draft. The final version of OpenSSL 1.1.1 will
217 implement the final version of the standard.
218 TODO(TLS1.3): Remove the above note before final release
221 *) Grand redesign of the OpenSSL random generator
223 The default RAND method now utilizes an AES-CTR DRBG according to
224 NIST standard SP 800-90Ar1. The new random generator is essentially
225 a port of the default random generator from the OpenSSL FIPS 2.0
226 object module. It is a hybrid deterministic random bit generator
227 using an AES-CTR bit stream and which seeds and reseeds itself
228 automatically using trusted system entropy sources.
230 Some of its new features are:
231 o Support for multiple DRBG instances with seed chaining.
232 o The default RAND method makes use of a DRBG.
233 o There is a public and private DRBG instance.
234 o The DRBG instances are fork-safe.
235 o Keep all global DRBG instances on the secure heap if it is enabled.
236 o The public and private DRBG instance are per thread for lock free
238 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
240 *) Changed Configure so it only says what it does and doesn't dump
241 so much data. Instead, ./configdata.pm should be used as a script
242 to display all sorts of configuration data.
245 *) Added processing of "make variables" to Configure.
248 *) Added SHA512/224 and SHA512/256 algorithm support.
251 *) The last traces of Netware support, first removed in 1.1.0, have
255 *) Get rid of Makefile.shared, and in the process, make the processing
256 of certain files (rc.obj, or the .def/.map/.opt files produced from
257 the ordinal files) more visible and hopefully easier to trace and
258 debug (or make silent).
261 *) Make it possible to have environment variable assignments as
262 arguments to config / Configure.
265 *) Add multi-prime RSA (RFC 8017) support.
268 *) Add SM3 implemented according to GB/T 32905-2016
269 [ Jack Lloyd <jack.lloyd@ribose.com>,
270 Ronald Tse <ronald.tse@ribose.com>,
271 Erick Borsboom <erick.borsboom@ribose.com> ]
273 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
274 as documented in RFC6066.
275 Based on a patch from Tomasz Moń
276 [Filipe Raimundo da Silva]
278 *) Add SM4 implemented according to GB/T 32907-2016.
279 [ Jack Lloyd <jack.lloyd@ribose.com>,
280 Ronald Tse <ronald.tse@ribose.com>,
281 Erick Borsboom <erick.borsboom@ribose.com> ]
283 *) Reimplement -newreq-nodes and ERR_error_string_n; the
284 original author does not agree with the license change.
287 *) Add ARIA AEAD TLS support.
290 *) Some macro definitions to support VS6 have been removed. Visual
291 Studio 6 has not worked since 1.1.0
294 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
295 without clearing the errors.
298 *) Add "atfork" functions. If building on a system that without
299 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
300 requirements. The RAND facility now uses/requires this.
306 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
307 not possible to disable entirely. However, it's still possible to
308 disable the console reading UI method, UI_OpenSSL() (use UI_null()
311 To disable, configure with 'no-ui-console'. 'no-ui' is still
312 possible to use as an alias. Check at compile time with the
313 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
314 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
317 *) Add a STORE module, which implements a uniform and URI based reader of
318 stores that can contain keys, certificates, CRLs and numerous other
319 objects. The main API is loosely based on a few stdio functions,
320 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
321 OSSL_STORE_error and OSSL_STORE_close.
322 The implementation uses backends called "loaders" to implement arbitrary
323 URI schemes. There is one built in "loader" for the 'file' scheme.
326 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
327 then adjusted to work on FreeBSD 8.4 as well.
328 Enable by configuring with 'enable-devcryptoeng'. This is done by default
329 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
332 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
333 util/mkerr.pl, which is adapted to allow those prefixes, leading to
334 error code calls like this:
336 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
338 With this change, we claim the namespaces OSSL and OPENSSL in a manner
339 that can be encoded in C. For the foreseeable future, this will only
341 [Richard Levitte and Tim Hudson]
343 *) Removed BSD cryptodev engine.
346 *) Add a build target 'build_all_generated', to build all generated files
347 and only that. This can be used to prepare everything that requires
348 things like perl for a system that lacks perl and then move everything
349 to that system and do the rest of the build there.
352 *) In the UI interface, make it possible to duplicate the user data. This
353 can be used by engines that need to retain the data for a longer time
354 than just the call where this user data is passed.
357 *) Ignore the '-named_curve auto' value for compatibility of applications
359 [Tomas Mraz <tmraz@fedoraproject.org>]
361 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
362 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
363 alerts across multiple records (some of which could be empty). In practice
364 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
365 prohibts this altogether and other libraries (BoringSSL, NSS) do not
366 support this at all. Supporting it adds significant complexity to the
367 record layer, and its removal is unlikely to cause inter-operability
371 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
372 with Z. These are meant to replace LONG and ZLONG and to be size safe.
373 The use of LONG and ZLONG is discouraged and scheduled for deprecation
377 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
378 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
379 [Richard Levitte, Andy Polyakov]
381 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
385 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
386 platform rather than 'mingw'.
389 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
390 success if they are asked to add an object which already exists
391 in the store. This change cascades to other functions which load
392 certificates and CRLs.
395 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
396 facilitate stack unwinding even from assembly subroutines.
399 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
400 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
403 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
404 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
405 which is the minimum version we support.
408 *) Certificate time validation (X509_cmp_time) enforces stricter
409 compliance with RFC 5280. Fractional seconds and timezone offsets
410 are no longer allowed.
413 *) Add support for ARIA
416 *) s_client will now send the Server Name Indication (SNI) extension by
417 default unless the new "-noservername" option is used. The server name is
418 based on the host provided to the "-connect" option unless overridden by
422 *) Add support for SipHash
425 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
426 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
427 prevent issues where no progress is being made and the peer continually
428 sends unrecognised record types, using up resources processing them.
431 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
432 using the algorithm defined in
433 https://www.akkadia.org/drepper/SHA-crypt.txt
436 *) Heartbeat support has been removed; the ABI is changed for now.
437 [Richard Levitte, Rich Salz]
439 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
442 *) The RSA "null" method, which was partially supported to avoid patent
443 issues, has been replaced to always returns NULL.
447 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
449 *) Fixed a text canonicalisation bug in CMS
451 Where a CMS detached signature is used with text content the text goes
452 through a canonicalisation process first prior to signing or verifying a
453 signature. This process strips trailing space at the end of lines, converts
454 line terminators to CRLF and removes additional trailing line terminators
455 at the end of a file. A bug in the canonicalisation process meant that
456 some characters, such as form-feed, were incorrectly treated as whitespace
457 and removed. This is contrary to the specification (RFC5485). This fix
458 could mean that detached text data signed with an earlier version of
459 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
460 signed with a fixed OpenSSL may fail to verify with an earlier version of
461 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
462 and use the "-binary" flag (for the "cms" command line application) or set
463 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
466 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
468 *) Constructed ASN.1 types with a recursive definition could exceed the stack
470 Constructed ASN.1 types with a recursive definition (such as can be found
471 in PKCS7) could eventually exceed the stack given malicious input with
472 excessive recursion. This could result in a Denial Of Service attack. There
473 are no such structures used within SSL/TLS that come from untrusted sources
474 so this is considered safe.
476 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
481 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
483 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
484 effectively reduced to only comparing the least significant bit of each
485 byte. This allows an attacker to forge messages that would be considered as
486 authenticated in an amount of tries lower than that guaranteed by the
487 security claims of the scheme. The module can only be compiled by the
488 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
490 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
495 *) Add a build target 'build_all_generated', to build all generated files
496 and only that. This can be used to prepare everything that requires
497 things like perl for a system that lacks perl and then move everything
498 to that system and do the rest of the build there.
501 *) Backport SSL_OP_NO_RENGOTIATION
503 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
504 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
505 changes this is no longer possible in 1.1.0. Therefore the new
506 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
507 1.1.0 to provide equivalent functionality.
509 Note that if an application built against 1.1.0h headers (or above) is run
510 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
511 accepted but nothing will happen, i.e. renegotiation will not be prevented.
514 *) Removed the OS390-Unix config target. It relied on a script that doesn't
518 *) rsaz_1024_mul_avx2 overflow bug on x86_64
520 There is an overflow bug in the AVX2 Montgomery multiplication procedure
521 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
522 Analysis suggests that attacks against RSA and DSA as a result of this
523 defect would be very difficult to perform and are not believed likely.
524 Attacks against DH1024 are considered just feasible, because most of the
525 work necessary to deduce information about a private key may be performed
526 offline. The amount of resources required for such an attack would be
527 significant. However, for an attack on TLS to be meaningful, the server
528 would have to share the DH1024 private key among multiple clients, which is
529 no longer an option since CVE-2016-0701.
531 This only affects processors that support the AVX2 but not ADX extensions
532 like Intel Haswell (4th generation).
534 This issue was reported to OpenSSL by David Benjamin (Google). The issue
535 was originally found via the OSS-Fuzz project.
539 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
541 *) bn_sqrx8x_internal carry bug on x86_64
543 There is a carry propagating bug in the x86_64 Montgomery squaring
544 procedure. No EC algorithms are affected. Analysis suggests that attacks
545 against RSA and DSA as a result of this defect would be very difficult to
546 perform and are not believed likely. Attacks against DH are considered just
547 feasible (although very difficult) because most of the work necessary to
548 deduce information about a private key may be performed offline. The amount
549 of resources required for such an attack would be very significant and
550 likely only accessible to a limited number of attackers. An attacker would
551 additionally need online access to an unpatched system using the target
552 private key in a scenario with persistent DH parameters and a private
553 key that is shared between multiple clients.
555 This only affects processors that support the BMI1, BMI2 and ADX extensions
556 like Intel Broadwell (5th generation) and later or AMD Ryzen.
558 This issue was reported to OpenSSL by the OSS-Fuzz project.
562 *) Malformed X.509 IPAddressFamily could cause OOB read
564 If an X.509 certificate has a malformed IPAddressFamily extension,
565 OpenSSL could do a one-byte buffer overread. The most likely result
566 would be an erroneous display of the certificate in text format.
568 This issue was reported to OpenSSL by the OSS-Fuzz project.
572 Changes between 1.1.0e and 1.1.0f [25 May 2017]
574 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
575 platform rather than 'mingw'.
578 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
579 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
580 which is the minimum version we support.
583 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
585 *) Encrypt-Then-Mac renegotiation crash
587 During a renegotiation handshake if the Encrypt-Then-Mac extension is
588 negotiated where it was not in the original handshake (or vice-versa) then
589 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
590 and servers are affected.
592 This issue was reported to OpenSSL by Joe Orton (Red Hat).
596 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
598 *) Truncated packet could crash via OOB read
600 If one side of an SSL/TLS path is running on a 32-bit host and a specific
601 cipher is being used, then a truncated packet can cause that host to
602 perform an out-of-bounds read, usually resulting in a crash.
604 This issue was reported to OpenSSL by Robert Święcki of Google.
608 *) Bad (EC)DHE parameters cause a client crash
610 If a malicious server supplies bad parameters for a DHE or ECDHE key
611 exchange then this can result in the client attempting to dereference a
612 NULL pointer leading to a client crash. This could be exploited in a Denial
615 This issue was reported to OpenSSL by Guido Vranken.
619 *) BN_mod_exp may produce incorrect results on x86_64
621 There is a carry propagating bug in the x86_64 Montgomery squaring
622 procedure. No EC algorithms are affected. Analysis suggests that attacks
623 against RSA and DSA as a result of this defect would be very difficult to
624 perform and are not believed likely. Attacks against DH are considered just
625 feasible (although very difficult) because most of the work necessary to
626 deduce information about a private key may be performed offline. The amount
627 of resources required for such an attack would be very significant and
628 likely only accessible to a limited number of attackers. An attacker would
629 additionally need online access to an unpatched system using the target
630 private key in a scenario with persistent DH parameters and a private
631 key that is shared between multiple clients. For example this can occur by
632 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
633 similar to CVE-2015-3193 but must be treated as a separate problem.
635 This issue was reported to OpenSSL by the OSS-Fuzz project.
639 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
641 *) ChaCha20/Poly1305 heap-buffer-overflow
643 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
644 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
645 crash. This issue is not considered to be exploitable beyond a DoS.
647 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
651 *) CMS Null dereference
653 Applications parsing invalid CMS structures can crash with a NULL pointer
654 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
655 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
656 structure callback if an attempt is made to free certain invalid encodings.
657 Only CHOICE structures using a callback which do not handle NULL value are
660 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
664 *) Montgomery multiplication may produce incorrect results
666 There is a carry propagating bug in the Broadwell-specific Montgomery
667 multiplication procedure that handles input lengths divisible by, but
668 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
669 and DH private keys are impossible. This is because the subroutine in
670 question is not used in operations with the private key itself and an input
671 of the attacker's direct choice. Otherwise the bug can manifest itself as
672 transient authentication and key negotiation failures or reproducible
673 erroneous outcome of public-key operations with specially crafted input.
674 Among EC algorithms only Brainpool P-512 curves are affected and one
675 presumably can attack ECDH key negotiation. Impact was not analyzed in
676 detail, because pre-requisites for attack are considered unlikely. Namely
677 multiple clients have to choose the curve in question and the server has to
678 share the private key among them, neither of which is default behaviour.
679 Even then only clients that chose the curve will be affected.
681 This issue was publicly reported as transient failures and was not
682 initially recognized as a security issue. Thanks to Richard Morgan for
683 providing reproducible case.
687 *) Removed automatic addition of RPATH in shared libraries and executables,
688 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
691 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
693 *) Fix Use After Free for large message sizes
695 The patch applied to address CVE-2016-6307 resulted in an issue where if a
696 message larger than approx 16k is received then the underlying buffer to
697 store the incoming message is reallocated and moved. Unfortunately a
698 dangling pointer to the old location is left which results in an attempt to
699 write to the previously freed location. This is likely to result in a
700 crash, however it could potentially lead to execution of arbitrary code.
702 This issue only affects OpenSSL 1.1.0a.
704 This issue was reported to OpenSSL by Robert Święcki.
708 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
710 *) OCSP Status Request extension unbounded memory growth
712 A malicious client can send an excessively large OCSP Status Request
713 extension. If that client continually requests renegotiation, sending a
714 large OCSP Status Request extension each time, then there will be unbounded
715 memory growth on the server. This will eventually lead to a Denial Of
716 Service attack through memory exhaustion. Servers with a default
717 configuration are vulnerable even if they do not support OCSP. Builds using
718 the "no-ocsp" build time option are not affected.
720 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
724 *) SSL_peek() hang on empty record
726 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
727 sends an empty record. This could be exploited by a malicious peer in a
728 Denial Of Service attack.
730 This issue was reported to OpenSSL by Alex Gaynor.
734 *) Excessive allocation of memory in tls_get_message_header() and
735 dtls1_preprocess_fragment()
737 A (D)TLS message includes 3 bytes for its length in the header for the
738 message. This would allow for messages up to 16Mb in length. Messages of
739 this length are excessive and OpenSSL includes a check to ensure that a
740 peer is sending reasonably sized messages in order to avoid too much memory
741 being consumed to service a connection. A flaw in the logic of version
742 1.1.0 means that memory for the message is allocated too early, prior to
743 the excessive message length check. Due to way memory is allocated in
744 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
745 to service a connection. This could lead to a Denial of Service through
746 memory exhaustion. However, the excessive message length check still takes
747 place, and this would cause the connection to immediately fail. Assuming
748 that the application calls SSL_free() on the failed connection in a timely
749 manner then the 21Mb of allocated memory will then be immediately freed
750 again. Therefore the excessive memory allocation will be transitory in
751 nature. This then means that there is only a security impact if:
753 1) The application does not call SSL_free() in a timely manner in the event
754 that the connection fails
756 2) The application is working in a constrained environment where there is
757 very little free memory
759 3) The attacker initiates multiple connection attempts such that there are
760 multiple connections in a state where memory has been allocated for the
761 connection; SSL_free() has not yet been called; and there is insufficient
762 memory to service the multiple requests.
764 Except in the instance of (1) above any Denial Of Service is likely to be
765 transitory because as soon as the connection fails the memory is
766 subsequently freed again in the SSL_free() call. However there is an
767 increased risk during this period of application crashes due to the lack of
768 memory - which would then mean a more serious Denial of Service.
770 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
771 (CVE-2016-6307 and CVE-2016-6308)
774 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
775 had to be removed. Primary reason is that vendor assembler can't
776 assemble our modules with -KPIC flag. As result it, assembly
777 support, was not even available as option. But its lack means
778 lack of side-channel resistant code, which is incompatible with
779 security by todays standards. Fortunately gcc is readily available
780 prepackaged option, which we firmly point at...
783 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
785 *) Windows command-line tool supports UTF-8 opt-in option for arguments
786 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
787 (to any value) allows Windows user to access PKCS#12 file generated
788 with Windows CryptoAPI and protected with non-ASCII password, as well
789 as files generated under UTF-8 locale on Linux also protected with
793 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
794 have been disabled by default and removed from DEFAULT, just like RC4.
795 See the RC4 item below to re-enable both.
798 *) The method for finding the storage location for the Windows RAND seed file
799 has changed. First we check %RANDFILE%. If that is not set then we check
800 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
801 all else fails we fall back to C:\.
804 *) The EVP_EncryptUpdate() function has had its return type changed from void
805 to int. A return of 0 indicates and error while a return of 1 indicates
809 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
810 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
811 off the constant time implementation for RSA, DSA and DH have been made
812 no-ops and deprecated.
815 *) Windows RAND implementation was simplified to only get entropy by
816 calling CryptGenRandom(). Various other RAND-related tickets
818 [Joseph Wylie Yandle, Rich Salz]
820 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
821 and OPENSSL_LH_, respectively. The old names are available
822 with API compatibility. They new names are now completely documented.
825 *) Unify TYPE_up_ref(obj) methods signature.
826 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
827 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
828 int (instead of void) like all others TYPE_up_ref() methods.
829 So now these methods also check the return value of CRYPTO_atomic_add(),
830 and the validity of object reference counter.
831 [fdasilvayy@gmail.com]
833 *) With Windows Visual Studio builds, the .pdb files are installed
834 alongside the installed libraries and executables. For a static
835 library installation, ossl_static.pdb is the associate compiler
836 generated .pdb file to be used when linking programs.
839 *) Remove openssl.spec. Packaging files belong with the packagers.
842 *) Automatic Darwin/OSX configuration has had a refresh, it will now
843 recognise x86_64 architectures automatically. You can still decide
844 to build for a different bitness with the environment variable
845 KERNEL_BITS (can be 32 or 64), for example:
847 KERNEL_BITS=32 ./config
851 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
852 256 bit AES and HMAC with SHA256.
855 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
858 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
861 *) To enable users to have their own config files and build file templates,
862 Configure looks in the directory indicated by the environment variable
863 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
864 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
865 name and is used as is.
868 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
869 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
870 X509_CERT_FILE_CTX was removed.
873 *) "shared" builds are now the default. To create only static libraries use
874 the "no-shared" Configure option.
877 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
878 All of these option have not worked for some while and are fundamental
882 *) Make various cleanup routines no-ops and mark them as deprecated. Most
883 global cleanup functions are no longer required because they are handled
884 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
885 Explicitly de-initing can cause problems (e.g. where a library that uses
886 OpenSSL de-inits, but an application is still using it). The affected
887 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
888 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
889 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
893 *) --strict-warnings no longer enables runtime debugging options
894 such as REF_DEBUG. Instead, debug options are automatically
895 enabled with '--debug' builds.
896 [Andy Polyakov, Emilia Käsper]
898 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
899 have been moved out of the public header files. New functions for managing
900 these have been added.
903 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
904 objects have been moved out of the public header files. New
905 functions for managing these have been added.
908 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
909 have been moved out of the public header files. New functions for managing
910 these have been added.
913 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
914 moved out of the public header files. New functions for managing these
918 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
921 *) Removed the mk1mf build scripts.
924 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
925 it is always safe to #include a header now.
928 *) Removed the aged BC-32 config and all its supporting scripts
931 *) Removed support for Ultrix, Netware, and OS/2.
934 *) Add support for HKDF.
937 *) Add support for blake2b and blake2s
940 *) Added support for "pipelining". Ciphers that have the
941 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
942 encryptions/decryptions simultaneously. There are currently no built-in
943 ciphers with this property but the expectation is that engines will be able
944 to offer it to significantly improve throughput. Support has been extended
945 into libssl so that multiple records for a single connection can be
946 processed in one go (for >=TLS 1.1).
949 *) Added the AFALG engine. This is an async capable engine which is able to
950 offload work to the Linux kernel. In this initial version it only supports
951 AES128-CBC. The kernel must be version 4.1.0 or greater.
954 *) OpenSSL now uses a new threading API. It is no longer necessary to
955 set locking callbacks to use OpenSSL in a multi-threaded environment. There
956 are two supported threading models: pthreads and windows threads. It is
957 also possible to configure OpenSSL at compile time for "no-threads". The
958 old threading API should no longer be used. The functions have been
959 replaced with "no-op" compatibility macros.
960 [Alessandro Ghedini, Matt Caswell]
962 *) Modify behavior of ALPN to invoke callback after SNI/servername
963 callback, such that updates to the SSL_CTX affect ALPN.
966 *) Add SSL_CIPHER queries for authentication and key-exchange.
969 *) Changes to the DEFAULT cipherlist:
970 - Prefer (EC)DHE handshakes over plain RSA.
971 - Prefer AEAD ciphers over legacy ciphers.
972 - Prefer ECDSA over RSA when both certificates are available.
973 - Prefer TLSv1.2 ciphers/PRF.
974 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
978 *) Change the ECC default curve list to be this, in order: x25519,
979 secp256r1, secp521r1, secp384r1.
982 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
983 disabled by default. They can be re-enabled using the
984 enable-weak-ssl-ciphers option to Configure.
987 *) If the server has ALPN configured, but supports no protocols that the
988 client advertises, send a fatal "no_application_protocol" alert.
989 This behaviour is SHALL in RFC 7301, though it isn't universally
990 implemented by other servers.
993 *) Add X25519 support.
994 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
995 for public and private key encoding using the format documented in
996 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
997 key generation and key derivation.
999 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1003 *) Deprecate SRP_VBASE_get_by_user.
1004 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1005 In order to fix an unavoidable memory leak (CVE-2016-0798),
1006 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1007 seed, even if the seed is configured.
1009 Users should use SRP_VBASE_get1_by_user instead. Note that in
1010 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1011 also that even though configuring the SRP seed attempts to hide
1012 invalid usernames by continuing the handshake with fake
1013 credentials, this behaviour is not constant time and no strong
1014 guarantees are made that the handshake is indistinguishable from
1015 that of a valid user.
1018 *) Configuration change; it's now possible to build dynamic engines
1019 without having to build shared libraries and vice versa. This
1020 only applies to the engines in engines/, those in crypto/engine/
1021 will always be built into libcrypto (i.e. "static").
1023 Building dynamic engines is enabled by default; to disable, use
1024 the configuration option "disable-dynamic-engine".
1026 The only requirements for building dynamic engines are the
1027 presence of the DSO module and building with position independent
1028 code, so they will also automatically be disabled if configuring
1029 with "disable-dso" or "disable-pic".
1031 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1032 are also taken away from openssl/opensslconf.h, as they are
1036 *) Configuration change; if there is a known flag to compile
1037 position independent code, it will always be applied on the
1038 libcrypto and libssl object files, and never on the application
1039 object files. This means other libraries that use routines from
1040 libcrypto / libssl can be made into shared libraries regardless
1041 of how OpenSSL was configured.
1043 If this isn't desirable, the configuration options "disable-pic"
1044 or "no-pic" can be used to disable the use of PIC. This will
1045 also disable building shared libraries and dynamic engines.
1048 *) Removed JPAKE code. It was experimental and has no wide use.
1051 *) The INSTALL_PREFIX Makefile variable has been renamed to
1052 DESTDIR. That makes for less confusion on what this variable
1053 is for. Also, the configuration option --install_prefix is
1057 *) Heartbeat for TLS has been removed and is disabled by default
1058 for DTLS; configure with enable-heartbeats. Code that uses the
1059 old #define's might need to be updated.
1060 [Emilia Käsper, Rich Salz]
1062 *) Rename REF_CHECK to REF_DEBUG.
1065 *) New "unified" build system
1067 The "unified" build system is aimed to be a common system for all
1068 platforms we support. With it comes new support for VMS.
1070 This system builds supports building in a different directory tree
1071 than the source tree. It produces one Makefile (for unix family
1072 or lookalikes), or one descrip.mms (for VMS).
1074 The source of information to make the Makefile / descrip.mms is
1075 small files called 'build.info', holding the necessary
1076 information for each directory with source to compile, and a
1077 template in Configurations, like unix-Makefile.tmpl or
1080 With this change, the library names were also renamed on Windows
1081 and on VMS. They now have names that are closer to the standard
1082 on Unix, and include the major version number, and in certain
1083 cases, the architecture they are built for. See "Notes on shared
1084 libraries" in INSTALL.
1086 We rely heavily on the perl module Text::Template.
1089 *) Added support for auto-initialisation and de-initialisation of the library.
1090 OpenSSL no longer requires explicit init or deinit routines to be called,
1091 except in certain circumstances. See the OPENSSL_init_crypto() and
1092 OPENSSL_init_ssl() man pages for further information.
1095 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1096 "peer" argument is now expected to be a BIO_ADDR object.
1098 *) Rewrite of BIO networking library. The BIO library lacked consistent
1099 support of IPv6, and adding it required some more extensive
1100 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1101 which hold all types of addresses and chains of address information.
1102 It also introduces a new API, with functions like BIO_socket,
1103 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1104 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1105 have been adapted accordingly.
1108 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1112 *) CRIME protection: disable compression by default, even if OpenSSL is
1113 compiled with zlib enabled. Applications can still enable compression
1114 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1115 using the SSL_CONF library to configure compression.
1118 *) The signature of the session callback configured with
1119 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1120 was explicitly marked as 'const unsigned char*' instead of
1124 *) Always DPURIFY. Remove the use of uninitialized memory in the
1125 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1128 *) Removed many obsolete configuration items, including
1129 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1130 MD2_CHAR, MD2_INT, MD2_LONG
1132 IDEA_SHORT, IDEA_LONG
1133 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1134 [Rich Salz, with advice from Andy Polyakov]
1136 *) Many BN internals have been moved to an internal header file.
1137 [Rich Salz with help from Andy Polyakov]
1139 *) Configuration and writing out the results from it has changed.
1140 Files such as Makefile include/openssl/opensslconf.h and are now
1141 produced through general templates, such as Makefile.in and
1142 crypto/opensslconf.h.in and some help from the perl module
1145 Also, the center of configuration information is no longer
1146 Makefile. Instead, Configure produces a perl module in
1147 configdata.pm which holds most of the config data (in the hash
1148 table %config), the target data that comes from the target
1149 configuration in one of the Configurations/*.conf files (in
1153 *) To clarify their intended purposes, the Configure options
1154 --prefix and --openssldir change their semantics, and become more
1155 straightforward and less interdependent.
1157 --prefix shall be used exclusively to give the location INSTALLTOP
1158 where programs, scripts, libraries, include files and manuals are
1159 going to be installed. The default is now /usr/local.
1161 --openssldir shall be used exclusively to give the default
1162 location OPENSSLDIR where certificates, private keys, CRLs are
1163 managed. This is also where the default openssl.cnf gets
1165 If the directory given with this option is a relative path, the
1166 values of both the --prefix value and the --openssldir value will
1167 be combined to become OPENSSLDIR.
1168 The default for --openssldir is INSTALLTOP/ssl.
1170 Anyone who uses --openssldir to specify where OpenSSL is to be
1171 installed MUST change to use --prefix instead.
1174 *) The GOST engine was out of date and therefore it has been removed. An up
1175 to date GOST engine is now being maintained in an external repository.
1176 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1177 support for GOST ciphersuites (these are only activated if a GOST engine
1181 *) EGD is no longer supported by default; use enable-egd when
1183 [Ben Kaduk and Rich Salz]
1185 *) The distribution now has Makefile.in files, which are used to
1186 create Makefile's when Configure is run. *Configure must be run
1187 before trying to build now.*
1190 *) The return value for SSL_CIPHER_description() for error conditions
1194 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1196 Obtaining and performing DNSSEC validation of TLSA records is
1197 the application's responsibility. The application provides
1198 the TLSA records of its choice to OpenSSL, and these are then
1199 used to authenticate the peer.
1201 The TLSA records need not even come from DNS. They can, for
1202 example, be used to implement local end-entity certificate or
1203 trust-anchor "pinning", where the "pin" data takes the form
1204 of TLSA records, which can augment or replace verification
1205 based on the usual WebPKI public certification authorities.
1208 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1209 continues to support deprecated interfaces in default builds.
1210 However, applications are strongly advised to compile their
1211 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1212 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1213 or the 1.1.0 releases.
1215 In environments in which all applications have been ported to
1216 not use any deprecated interfaces OpenSSL's Configure script
1217 should be used with the --api=1.1.0 option to entirely remove
1218 support for the deprecated features from the library and
1219 unconditionally disable them in the installed headers.
1220 Essentially the same effect can be achieved with the "no-deprecated"
1221 argument to Configure, except that this will always restrict
1222 the build to just the latest API, rather than a fixed API
1225 As applications are ported to future revisions of the API,
1226 they should update their compile-time OPENSSL_API_COMPAT define
1227 accordingly, but in most cases should be able to continue to
1228 compile with later releases.
1230 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1231 0x10000000L and 0x00908000L, respectively. However those
1232 versions did not support the OPENSSL_API_COMPAT feature, and
1233 so applications are not typically tested for explicit support
1234 of just the undeprecated features of either release.
1237 *) Add support for setting the minimum and maximum supported protocol.
1238 It can bet set via the SSL_set_min_proto_version() and
1239 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1240 MaxProtocol. It's recommended to use the new APIs to disable
1241 protocols instead of disabling individual protocols using
1242 SSL_set_options() or SSL_CONF's Protocol. This change also
1243 removes support for disabling TLS 1.2 in the OpenSSL TLS
1244 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1247 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1250 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1251 and integrates ECDSA and ECDH functionality into EC. Implementations can
1252 now redirect key generation and no longer need to convert to or from
1255 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1256 include the ec.h header file instead.
1259 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1260 ciphers who are no longer supported and drops support the ephemeral RSA key
1261 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1264 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1265 opaque. For HMAC_CTX, the following constructors and destructors
1268 HMAC_CTX *HMAC_CTX_new(void);
1269 void HMAC_CTX_free(HMAC_CTX *ctx);
1271 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1272 destroy such methods has been added. See EVP_MD_meth_new(3) and
1273 EVP_CIPHER_meth_new(3) for documentation.
1276 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1277 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1278 EVP_MD_CTX_reset() should be called instead to reinitialise
1279 an already created structure.
1280 2) For consistency with the majority of our object creators and
1281 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1282 EVP_MD_CTX_(new|free). The old names are retained as macros
1283 for deprecated builds.
1286 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1287 cryptographic operations to be performed asynchronously as long as an
1288 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1289 further details. Libssl has also had this capability integrated with the
1290 introduction of the new mode SSL_MODE_ASYNC and associated error
1291 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1292 pages. This work was developed in partnership with Intel Corp.
1295 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1296 always enabled now. If you want to disable the support you should
1297 exclude it using the list of supported ciphers. This also means that the
1298 "-no_ecdhe" option has been removed from s_server.
1301 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1302 SSL_{CTX_}set1_curves() which can set a list.
1305 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1306 curve you want to support using SSL_{CTX_}set1_curves().
1309 *) State machine rewrite. The state machine code has been significantly
1310 refactored in order to remove much duplication of code and solve issues
1311 with the old code (see ssl/statem/README for further details). This change
1312 does have some associated API changes. Notably the SSL_state() function
1313 has been removed and replaced by SSL_get_state which now returns an
1314 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1315 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1319 *) All instances of the string "ssleay" in the public API were replaced
1320 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1321 Some error codes related to internal RSA_eay API's were renamed.
1324 *) The demo files in crypto/threads were moved to demo/threads.
1327 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1329 [Matt Caswell, Rich Salz]
1331 *) New ASN.1 embed macro.
1333 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1334 structure is not allocated: it is part of the parent. That is instead of
1342 This reduces memory fragmentation and make it impossible to accidentally
1343 set a mandatory field to NULL.
1345 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1346 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1347 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1351 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1354 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1355 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1356 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1357 DES and RC4 ciphersuites.
1360 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1361 This changes the decoding behaviour for some invalid messages,
1362 though the change is mostly in the more lenient direction, and
1363 legacy behaviour is preserved as much as possible.
1366 *) Fix no-stdio build.
1367 [ David Woodhouse <David.Woodhouse@intel.com> and also
1368 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1370 *) New testing framework
1371 The testing framework has been largely rewritten and is now using
1372 perl and the perl modules Test::Harness and an extended variant of
1373 Test::More called OpenSSL::Test to do its work. All test scripts in
1374 test/ have been rewritten into test recipes, and all direct calls to
1375 executables in test/Makefile have become individual recipes using the
1376 simplified testing OpenSSL::Test::Simple.
1378 For documentation on our testing modules, do:
1380 perldoc test/testlib/OpenSSL/Test/Simple.pm
1381 perldoc test/testlib/OpenSSL/Test.pm
1385 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1386 are used; the latter aborts on memory leaks (usually checked on exit).
1387 Some undocumented "set malloc, etc., hooks" functions were removed
1388 and others were changed. All are now documented.
1391 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1393 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1395 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1396 from RFC4279, RFC4785, RFC5487, RFC5489.
1398 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1399 original RSA_PSK patch.
1402 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1403 era flag was never set throughout the codebase (only read). Also removed
1404 SSL3_FLAGS_POP_BUFFER which was only used if
1405 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1408 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1409 to be "oneline" instead of "compat".
1412 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1413 not aware of clients that still exhibit this bug, and the workaround
1414 hasn't been working properly for a while.
1417 *) The return type of BIO_number_read() and BIO_number_written() as well as
1418 the corresponding num_read and num_write members in the BIO structure has
1419 changed from unsigned long to uint64_t. On platforms where an unsigned
1420 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1424 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1425 OpenSSL without support for them. It also means that maintaining
1426 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1427 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1430 *) Removed support for the two export grade static DH ciphersuites
1431 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1432 were newly added (along with a number of other static DH ciphersuites) to
1433 1.0.2. However the two export ones have *never* worked since they were
1434 introduced. It seems strange in any case to be adding new export
1435 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1438 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1439 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1440 and turned into macros which simply call the new preferred function names
1441 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1442 should use the new names instead. Also as part of this change the ssl23.h
1443 header file has been removed.
1446 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1447 code and the associated standard is no longer considered fit-for-purpose.
1450 *) RT2547 was closed. When generating a private key, try to make the
1451 output file readable only by the owner. This behavior change might
1452 be noticeable when interacting with other software.
1454 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1458 *) Added HTTP GET support to the ocsp command.
1461 *) Changed default digest for the dgst and enc commands from MD5 to
1465 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1468 *) Added support for TLS extended master secret from
1469 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1470 initial patch which was a great help during development.
1473 *) All libssl internal structures have been removed from the public header
1474 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1475 now redundant). Users should not attempt to access internal structures
1476 directly. Instead they should use the provided API functions.
1479 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1480 Access to deprecated functions can be re-enabled by running config with
1481 "enable-deprecated". In addition applications wishing to use deprecated
1482 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1483 will, by default, disable some transitive includes that previously existed
1484 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1487 *) Added support for OCB mode. OpenSSL has been granted a patent license
1488 compatible with the OpenSSL license for use of OCB. Details are available
1489 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1490 for OCB can be removed by calling config with no-ocb.
1493 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1494 compatible client hello.
1497 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1498 done while fixing the error code for the key-too-small case.
1499 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1501 *) CA.sh has been removed; use CA.pl instead.
1504 *) Removed old DES API.
1507 *) Remove various unsupported platforms:
1513 Sinix/ReliantUNIX RM400
1518 16-bit platforms such as WIN16
1521 *) Clean up OPENSSL_NO_xxx #define's
1522 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1523 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1524 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1525 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1526 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1527 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1528 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1529 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1530 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1531 Remove MS_STATIC; it's a relic from platforms <32 bits.
1534 *) Cleaned up dead code
1535 Remove all but one '#ifdef undef' which is to be looked at.
1538 *) Clean up calling of xxx_free routines.
1539 Just like free(), fix most of the xxx_free routines to accept
1540 NULL. Remove the non-null checks from callers. Save much code.
1543 *) Add secure heap for storage of private keys (when possible).
1544 Add BIO_s_secmem(), CBIGNUM, etc.
1545 Contributed by Akamai Technologies under our Corporate CLA.
1548 *) Experimental support for a new, fast, unbiased prime candidate generator,
1549 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1550 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1552 *) New output format NSS in the sess_id command line tool. This allows
1553 exporting the session id and the master key in NSS keylog format.
1554 [Martin Kaiser <martin@kaiser.cx>]
1556 *) Harmonize version and its documentation. -f flag is used to display
1558 [mancha <mancha1@zoho.com>]
1560 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1561 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1562 [mancha <mancha1@zoho.com>]
1564 *) Fix some double frees. These are not thought to be exploitable.
1565 [mancha <mancha1@zoho.com>]
1567 *) A missing bounds check in the handling of the TLS heartbeat extension
1568 can be used to reveal up to 64k of memory to a connected client or
1571 Thanks for Neel Mehta of Google Security for discovering this bug and to
1572 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1573 preparing the fix (CVE-2014-0160)
1574 [Adam Langley, Bodo Moeller]
1576 *) Fix for the attack described in the paper "Recovering OpenSSL
1577 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1578 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1579 http://eprint.iacr.org/2014/140
1581 Thanks to Yuval Yarom and Naomi Benger for discovering this
1582 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1583 [Yuval Yarom and Naomi Benger]
1585 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1586 this fixes a limitation in previous versions of OpenSSL.
1589 *) Experimental encrypt-then-mac support.
1591 Experimental support for encrypt then mac from
1592 draft-gutmann-tls-encrypt-then-mac-02.txt
1594 To enable it set the appropriate extension number (0x42 for the test
1595 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1597 For non-compliant peers (i.e. just about everything) this should have no
1600 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1604 *) Add EVP support for key wrapping algorithms, to avoid problems with
1605 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1606 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1607 algorithms and include tests cases.
1610 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1614 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1615 MGF1 digest and OAEP label.
1618 *) Make openssl verify return errors.
1619 [Chris Palmer <palmer@google.com> and Ben Laurie]
1621 *) New function ASN1_TIME_diff to calculate the difference between two
1622 ASN1_TIME structures or one structure and the current time.
1625 *) Update fips_test_suite to support multiple command line options. New
1626 test to induce all self test errors in sequence and check expected
1630 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1631 sign or verify all in one operation.
1634 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1635 test programs and fips_test_suite. Includes functionality to parse
1636 the minimal script output of fipsalgest.pl directly.
1639 *) Add authorisation parameter to FIPS_module_mode_set().
1642 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1645 *) Use separate DRBG fields for internal and external flags. New function
1646 FIPS_drbg_health_check() to perform on demand health checking. Add
1647 generation tests to fips_test_suite with reduced health check interval to
1648 demonstrate periodic health checking. Add "nodh" option to
1649 fips_test_suite to skip very slow DH test.
1652 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1656 *) More extensive health check for DRBG checking many more failure modes.
1657 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1658 combination: call this in fips_test_suite.
1661 *) Add support for canonical generation of DSA parameter 'g'. See
1664 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1665 POST to handle HMAC cases.
1668 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1669 to return numerical and string versions of the FIPS module number.
1672 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1673 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1674 outside the validated module in the FIPS capable OpenSSL.
1677 *) Minor change to DRBG entropy callback semantics. In some cases
1678 there is no multiple of the block length between min_len and
1679 max_len. Allow the callback to return more than max_len bytes
1680 of entropy but discard any extra: it is the callback's responsibility
1681 to ensure that the extra data discarded does not impact the
1682 requested amount of entropy.
1685 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1686 information in FIPS186-3, SP800-57 and SP800-131A.
1689 *) CCM support via EVP. Interface is very similar to GCM case except we
1690 must supply all data in one chunk (i.e. no update, final) and the
1691 message length must be supplied if AAD is used. Add algorithm test
1695 *) Initial version of POST overhaul. Add POST callback to allow the status
1696 of POST to be monitored and/or failures induced. Modify fips_test_suite
1697 to use callback. Always run all selftests even if one fails.
1700 *) XTS support including algorithm test driver in the fips_gcmtest program.
1701 Note: this does increase the maximum key length from 32 to 64 bytes but
1702 there should be no binary compatibility issues as existing applications
1703 will never use XTS mode.
1706 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1707 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1708 performs algorithm blocking for unapproved PRNG types. Also do not
1709 set PRNG type in FIPS_mode_set(): leave this to the application.
1710 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1711 the standard OpenSSL PRNG: set additional data to a date time vector.
1714 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1715 This shouldn't present any incompatibility problems because applications
1716 shouldn't be using these directly and any that are will need to rethink
1717 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1720 *) Extensive self tests and health checking required by SP800-90 DRBG.
1721 Remove strength parameter from FIPS_drbg_instantiate and always
1722 instantiate at maximum supported strength.
1725 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1728 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1731 *) New function DH_compute_key_padded() to compute a DH key and pad with
1732 leading zeroes if needed: this complies with SP800-56A et al.
1735 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1736 anything, incomplete, subject to change and largely untested at present.
1739 *) Modify fipscanisteronly build option to only build the necessary object
1740 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1743 *) Add experimental option FIPSSYMS to give all symbols in
1744 fipscanister.o and FIPS or fips prefix. This will avoid
1745 conflicts with future versions of OpenSSL. Add perl script
1746 util/fipsas.pl to preprocess assembly language source files
1747 and rename any affected symbols.
1750 *) Add selftest checks and algorithm block of non-fips algorithms in
1751 FIPS mode. Remove DES2 from selftests.
1754 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1755 return internal method without any ENGINE dependencies. Add new
1756 tiny fips sign and verify functions.
1759 *) New build option no-ec2m to disable characteristic 2 code.
1762 *) New build option "fipscanisteronly". This only builds fipscanister.o
1763 and (currently) associated fips utilities. Uses the file Makefile.fips
1764 instead of Makefile.org as the prototype.
1767 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1768 Update fips_gcmtest to use IV generator.
1771 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1772 setting output buffer to NULL. The *Final function must be
1773 called although it will not retrieve any additional data. The tag
1774 can be set or retrieved with a ctrl. The IV length is by default 12
1775 bytes (96 bits) but can be set to an alternative value. If the IV
1776 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1780 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1781 underlying do_cipher function handles all cipher semantics itself
1782 including padding and finalisation. This is useful if (for example)
1783 an ENGINE cipher handles block padding itself. The behaviour of
1784 do_cipher is subtly changed if this flag is set: the return value
1785 is the number of characters written to the output buffer (zero is
1786 no longer an error code) or a negative error code. Also if the
1787 input buffer is NULL and length 0 finalisation should be performed.
1790 *) If a candidate issuer certificate is already part of the constructed
1791 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1794 *) Improve forward-security support: add functions
1796 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1797 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1799 for use by SSL/TLS servers; the callback function will be called whenever a
1800 new session is created, and gets to decide whether the session may be
1801 cached to make it resumable (return 0) or not (return 1). (As by the
1802 SSL/TLS protocol specifications, the session_id sent by the server will be
1803 empty to indicate that the session is not resumable; also, the server will
1804 not generate RFC 4507 (RFC 5077) session tickets.)
1806 A simple reasonable callback implementation is to return is_forward_secure.
1807 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1808 by the SSL/TLS server library, indicating whether it can provide forward
1810 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1812 *) New -verify_name option in command line utilities to set verification
1816 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1817 Add CMAC pkey methods.
1820 *) Experimental renegotiation in s_server -www mode. If the client
1821 browses /reneg connection is renegotiated. If /renegcert it is
1822 renegotiated requesting a certificate.
1825 *) Add an "external" session cache for debugging purposes to s_server. This
1826 should help trace issues which normally are only apparent in deployed
1827 multi-process servers.
1830 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1831 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1832 BIO_set_cipher() and some obscure PEM functions were changed so they
1833 can now return an error. The RAND changes required a change to the
1834 RAND_METHOD structure.
1837 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1838 a gcc attribute to warn if the result of a function is ignored. This
1839 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1840 whose return value is often ignored.
1843 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1844 These allow SCTs (signed certificate timestamps) to be requested and
1845 validated when establishing a connection.
1846 [Rob Percival <robpercival@google.com>]
1848 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1850 *) Prevent padding oracle in AES-NI CBC MAC check
1852 A MITM attacker can use a padding oracle attack to decrypt traffic
1853 when the connection uses an AES CBC cipher and the server support
1856 This issue was introduced as part of the fix for Lucky 13 padding
1857 attack (CVE-2013-0169). The padding check was rewritten to be in
1858 constant time by making sure that always the same bytes are read and
1859 compared against either the MAC or padding bytes. But it no longer
1860 checked that there was enough data to have both the MAC and padding
1863 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1867 *) Fix EVP_EncodeUpdate overflow
1869 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1870 Base64 encoding of binary data. If an attacker is able to supply very large
1871 amounts of input data then a length check can overflow resulting in a heap
1874 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1875 the PEM_write_bio* family of functions. These are mainly used within the
1876 OpenSSL command line applications, so any application which processes data
1877 from an untrusted source and outputs it as a PEM file should be considered
1878 vulnerable to this issue. User applications that call these APIs directly
1879 with large amounts of untrusted data may also be vulnerable.
1881 This issue was reported by Guido Vranken.
1885 *) Fix EVP_EncryptUpdate overflow
1887 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1888 is able to supply very large amounts of input data after a previous call to
1889 EVP_EncryptUpdate() with a partial block then a length check can overflow
1890 resulting in a heap corruption. Following an analysis of all OpenSSL
1891 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1892 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1893 the first called function after an EVP_EncryptInit(), and therefore that
1894 specific call must be safe. The second form is where the length passed to
1895 EVP_EncryptUpdate() can be seen from the code to be some small value and
1896 therefore there is no possibility of an overflow. Since all instances are
1897 one of these two forms, it is believed that there can be no overflows in
1898 internal code due to this problem. It should be noted that
1899 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1900 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1901 of these calls have also been analysed too and it is believed there are no
1902 instances in internal usage where an overflow could occur.
1904 This issue was reported by Guido Vranken.
1908 *) Prevent ASN.1 BIO excessive memory allocation
1910 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1911 a short invalid encoding can cause allocation of large amounts of memory
1912 potentially consuming excessive resources or exhausting memory.
1914 Any application parsing untrusted data through d2i BIO functions is
1915 affected. The memory based functions such as d2i_X509() are *not* affected.
1916 Since the memory based functions are used by the TLS library, TLS
1917 applications are not affected.
1919 This issue was reported by Brian Carpenter.
1925 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1926 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1927 in arbitrary stack data being returned in the buffer.
1929 This issue was reported by Guido Vranken.
1933 *) Modify behavior of ALPN to invoke callback after SNI/servername
1934 callback, such that updates to the SSL_CTX affect ALPN.
1937 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1941 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1942 methods are enabled and ssl2 is disabled the methods return NULL.
1945 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1947 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1948 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1949 provide any "EXPORT" or "LOW" strength ciphers.
1952 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1953 is by default disabled at build-time. Builds that are not configured with
1954 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1955 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1956 will need to explicitly call either of:
1958 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1960 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1962 as appropriate. Even if either of those is used, or the application
1963 explicitly uses the version-specific SSLv2_method() or its client and
1964 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1965 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1966 ciphers, and SSLv2 56-bit DES are no longer available.
1970 *) Fix a double-free in DSA code
1972 A double free bug was discovered when OpenSSL parses malformed DSA private
1973 keys and could lead to a DoS attack or memory corruption for applications
1974 that receive DSA private keys from untrusted sources. This scenario is
1977 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1982 *) Disable SRP fake user seed to address a server memory leak.
1984 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1986 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1987 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1988 was changed to ignore the "fake user" SRP seed, even if the seed
1991 Users should use SRP_VBASE_get1_by_user instead. Note that in
1992 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1993 also that even though configuring the SRP seed attempts to hide
1994 invalid usernames by continuing the handshake with fake
1995 credentials, this behaviour is not constant time and no strong
1996 guarantees are made that the handshake is indistinguishable from
1997 that of a valid user.
2001 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2003 In the BN_hex2bn function the number of hex digits is calculated using an
2004 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2005 large values of |i| this can result in |bn_expand| not allocating any
2006 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2007 field as NULL leading to a subsequent NULL ptr deref. For very large values
2008 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2009 In this case memory is allocated to the internal BIGNUM data field, but it
2010 is insufficiently sized leading to heap corruption. A similar issue exists
2011 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2012 is ever called by user applications with very large untrusted hex/dec data.
2013 This is anticipated to be a rare occurrence.
2015 All OpenSSL internal usage of these functions use data that is not expected
2016 to be untrusted, e.g. config file data or application command line
2017 arguments. If user developed applications generate config file data based
2018 on untrusted data then it is possible that this could also lead to security
2019 consequences. This is also anticipated to be rare.
2021 This issue was reported to OpenSSL by Guido Vranken.
2025 *) Fix memory issues in BIO_*printf functions
2027 The internal |fmtstr| function used in processing a "%s" format string in
2028 the BIO_*printf functions could overflow while calculating the length of a
2029 string and cause an OOB read when printing very long strings.
2031 Additionally the internal |doapr_outch| function can attempt to write to an
2032 OOB memory location (at an offset from the NULL pointer) in the event of a
2033 memory allocation failure. In 1.0.2 and below this could be caused where
2034 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2035 could be in processing a very long "%s" format string. Memory leaks can
2038 The first issue may mask the second issue dependent on compiler behaviour.
2039 These problems could enable attacks where large amounts of untrusted data
2040 is passed to the BIO_*printf functions. If applications use these functions
2041 in this way then they could be vulnerable. OpenSSL itself uses these
2042 functions when printing out human-readable dumps of ASN.1 data. Therefore
2043 applications that print this data could be vulnerable if the data is from
2044 untrusted sources. OpenSSL command line applications could also be
2045 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2046 as command line arguments.
2048 Libssl is not considered directly vulnerable. Additionally certificates etc
2049 received via remote connections via libssl are also unlikely to be able to
2050 trigger these issues because of message size limits enforced within libssl.
2052 This issue was reported to OpenSSL Guido Vranken.
2056 *) Side channel attack on modular exponentiation
2058 A side-channel attack was found which makes use of cache-bank conflicts on
2059 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2060 of RSA keys. The ability to exploit this issue is limited as it relies on
2061 an attacker who has control of code in a thread running on the same
2062 hyper-threaded core as the victim thread which is performing decryptions.
2064 This issue was reported to OpenSSL by Yuval Yarom, The University of
2065 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2066 Nadia Heninger, University of Pennsylvania with more information at
2067 http://cachebleed.info.
2071 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2072 if no keysize is specified with default_bits. This fixes an
2073 omission in an earlier change that changed all RSA/DSA key generation
2074 apps to use 2048 bits by default.
2077 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2078 *) DH small subgroups
2080 Historically OpenSSL only ever generated DH parameters based on "safe"
2081 primes. More recently (in version 1.0.2) support was provided for
2082 generating X9.42 style parameter files such as those required for RFC 5114
2083 support. The primes used in such files may not be "safe". Where an
2084 application is using DH configured with parameters based on primes that are
2085 not "safe" then an attacker could use this fact to find a peer's private
2086 DH exponent. This attack requires that the attacker complete multiple
2087 handshakes in which the peer uses the same private DH exponent. For example
2088 this could be used to discover a TLS server's private DH exponent if it's
2089 reusing the private DH exponent or it's using a static DH ciphersuite.
2091 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2092 TLS. It is not on by default. If the option is not set then the server
2093 reuses the same private DH exponent for the life of the server process and
2094 would be vulnerable to this attack. It is believed that many popular
2095 applications do set this option and would therefore not be at risk.
2097 The fix for this issue adds an additional check where a "q" parameter is
2098 available (as is the case in X9.42 based parameters). This detects the
2099 only known attack, and is the only possible defense for static DH
2100 ciphersuites. This could have some performance impact.
2102 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2103 default and cannot be disabled. This could have some performance impact.
2105 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2109 *) SSLv2 doesn't block disabled ciphers
2111 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2112 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2113 been disabled, provided that the SSLv2 protocol was not also disabled via
2116 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2117 and Sebastian Schinzel.
2121 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2123 *) BN_mod_exp may produce incorrect results on x86_64
2125 There is a carry propagating bug in the x86_64 Montgomery squaring
2126 procedure. No EC algorithms are affected. Analysis suggests that attacks
2127 against RSA and DSA as a result of this defect would be very difficult to
2128 perform and are not believed likely. Attacks against DH are considered just
2129 feasible (although very difficult) because most of the work necessary to
2130 deduce information about a private key may be performed offline. The amount
2131 of resources required for such an attack would be very significant and
2132 likely only accessible to a limited number of attackers. An attacker would
2133 additionally need online access to an unpatched system using the target
2134 private key in a scenario with persistent DH parameters and a private
2135 key that is shared between multiple clients. For example this can occur by
2136 default in OpenSSL DHE based SSL/TLS ciphersuites.
2138 This issue was reported to OpenSSL by Hanno Böck.
2142 *) Certificate verify crash with missing PSS parameter
2144 The signature verification routines will crash with a NULL pointer
2145 dereference if presented with an ASN.1 signature using the RSA PSS
2146 algorithm and absent mask generation function parameter. Since these
2147 routines are used to verify certificate signature algorithms this can be
2148 used to crash any certificate verification operation and exploited in a
2149 DoS attack. Any application which performs certificate verification is
2150 vulnerable including OpenSSL clients and servers which enable client
2153 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2157 *) X509_ATTRIBUTE memory leak
2159 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2160 memory. This structure is used by the PKCS#7 and CMS routines so any
2161 application which reads PKCS#7 or CMS data from untrusted sources is
2162 affected. SSL/TLS is not affected.
2164 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2169 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2170 This changes the decoding behaviour for some invalid messages,
2171 though the change is mostly in the more lenient direction, and
2172 legacy behaviour is preserved as much as possible.
2175 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2177 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2179 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2181 *) Alternate chains certificate forgery
2183 During certificate verification, OpenSSL will attempt to find an
2184 alternative certificate chain if the first attempt to build such a chain
2185 fails. An error in the implementation of this logic can mean that an
2186 attacker could cause certain checks on untrusted certificates to be
2187 bypassed, such as the CA flag, enabling them to use a valid leaf
2188 certificate to act as a CA and "issue" an invalid certificate.
2190 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2194 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2196 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2197 incompatibility in the handling of HMAC. The previous ABI has now been
2201 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2203 *) Malformed ECParameters causes infinite loop
2205 When processing an ECParameters structure OpenSSL enters an infinite loop
2206 if the curve specified is over a specially malformed binary polynomial
2209 This can be used to perform denial of service against any
2210 system which processes public keys, certificate requests or
2211 certificates. This includes TLS clients and TLS servers with
2212 client authentication enabled.
2214 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2218 *) Exploitable out-of-bounds read in X509_cmp_time
2220 X509_cmp_time does not properly check the length of the ASN1_TIME
2221 string and can read a few bytes out of bounds. In addition,
2222 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2225 An attacker can use this to craft malformed certificates and CRLs of
2226 various sizes and potentially cause a segmentation fault, resulting in
2227 a DoS on applications that verify certificates or CRLs. TLS clients
2228 that verify CRLs are affected. TLS clients and servers with client
2229 authentication enabled may be affected if they use custom verification
2232 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2233 independently by Hanno Böck.
2237 *) PKCS7 crash with missing EnvelopedContent
2239 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2240 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2241 with missing content and trigger a NULL pointer dereference on parsing.
2243 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2244 structures from untrusted sources are affected. OpenSSL clients and
2245 servers are not affected.
2247 This issue was reported to OpenSSL by Michal Zalewski (Google).
2251 *) CMS verify infinite loop with unknown hash function
2253 When verifying a signedData message the CMS code can enter an infinite loop
2254 if presented with an unknown hash function OID. This can be used to perform
2255 denial of service against any system which verifies signedData messages using
2257 This issue was reported to OpenSSL by Johannes Bauer.
2261 *) Race condition handling NewSessionTicket
2263 If a NewSessionTicket is received by a multi-threaded client when attempting to
2264 reuse a previous ticket then a race condition can occur potentially leading to
2265 a double free of the ticket data.
2269 *) Only support 256-bit or stronger elliptic curves with the
2270 'ecdh_auto' setting (server) or by default (client). Of supported
2271 curves, prefer P-256 (both).
2274 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2276 *) ClientHello sigalgs DoS fix
2278 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2279 invalid signature algorithms extension a NULL pointer dereference will
2280 occur. This can be exploited in a DoS attack against the server.
2282 This issue was was reported to OpenSSL by David Ramos of Stanford
2285 [Stephen Henson and Matt Caswell]
2287 *) Multiblock corrupted pointer fix
2289 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2290 feature only applies on 64 bit x86 architecture platforms that support AES
2291 NI instructions. A defect in the implementation of "multiblock" can cause
2292 OpenSSL's internal write buffer to become incorrectly set to NULL when
2293 using non-blocking IO. Typically, when the user application is using a
2294 socket BIO for writing, this will only result in a failed connection.
2295 However if some other BIO is used then it is likely that a segmentation
2296 fault will be triggered, thus enabling a potential DoS attack.
2298 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2302 *) Segmentation fault in DTLSv1_listen fix
2304 The DTLSv1_listen function is intended to be stateless and processes the
2305 initial ClientHello from many peers. It is common for user code to loop
2306 over the call to DTLSv1_listen until a valid ClientHello is received with
2307 an associated cookie. A defect in the implementation of DTLSv1_listen means
2308 that state is preserved in the SSL object from one invocation to the next
2309 that can lead to a segmentation fault. Errors processing the initial
2310 ClientHello can trigger this scenario. An example of such an error could be
2311 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2314 This issue was reported to OpenSSL by Per Allansson.
2318 *) Segmentation fault in ASN1_TYPE_cmp fix
2320 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2321 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2322 certificate signature algorithm consistency this can be used to crash any
2323 certificate verification operation and exploited in a DoS attack. Any
2324 application which performs certificate verification is vulnerable including
2325 OpenSSL clients and servers which enable client authentication.
2329 *) Segmentation fault for invalid PSS parameters fix
2331 The signature verification routines will crash with a NULL pointer
2332 dereference if presented with an ASN.1 signature using the RSA PSS
2333 algorithm and invalid parameters. Since these routines are used to verify
2334 certificate signature algorithms this can be used to crash any
2335 certificate verification operation and exploited in a DoS attack. Any
2336 application which performs certificate verification is vulnerable including
2337 OpenSSL clients and servers which enable client authentication.
2339 This issue was was reported to OpenSSL by Brian Carpenter.
2343 *) ASN.1 structure reuse memory corruption fix
2345 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2346 memory corruption via an invalid write. Such reuse is and has been
2347 strongly discouraged and is believed to be rare.
2349 Applications that parse structures containing CHOICE or ANY DEFINED BY
2350 components may be affected. Certificate parsing (d2i_X509 and related
2351 functions) are however not affected. OpenSSL clients and servers are
2356 *) PKCS7 NULL pointer dereferences fix
2358 The PKCS#7 parsing code does not handle missing outer ContentInfo
2359 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2360 missing content and trigger a NULL pointer dereference on parsing.
2362 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2363 otherwise parse PKCS#7 structures from untrusted sources are
2364 affected. OpenSSL clients and servers are not affected.
2366 This issue was reported to OpenSSL by Michal Zalewski (Google).
2370 *) DoS via reachable assert in SSLv2 servers fix
2372 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2373 servers that both support SSLv2 and enable export cipher suites by sending
2374 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2376 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2377 (OpenSSL development team).
2381 *) Empty CKE with client auth and DHE fix
2383 If client auth is used then a server can seg fault in the event of a DHE
2384 ciphersuite being selected and a zero length ClientKeyExchange message
2385 being sent by the client. This could be exploited in a DoS attack.
2389 *) Handshake with unseeded PRNG fix
2391 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2392 with an unseeded PRNG. The conditions are:
2393 - The client is on a platform where the PRNG has not been seeded
2394 automatically, and the user has not seeded manually
2395 - A protocol specific client method version has been used (i.e. not
2396 SSL_client_methodv23)
2397 - A ciphersuite is used that does not require additional random data from
2398 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2400 If the handshake succeeds then the client random that has been used will
2401 have been generated from a PRNG with insufficient entropy and therefore the
2402 output may be predictable.
2404 For example using the following command with an unseeded openssl will
2405 succeed on an unpatched platform:
2407 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2411 *) Use After Free following d2i_ECPrivatekey error fix
2413 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2414 could cause a use after free condition. This, in turn, could cause a double
2415 free in several private key parsing functions (such as d2i_PrivateKey
2416 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2417 for applications that receive EC private keys from untrusted
2418 sources. This scenario is considered rare.
2420 This issue was discovered by the BoringSSL project and fixed in their
2425 *) X509_to_X509_REQ NULL pointer deref fix
2427 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2428 the certificate key is invalid. This function is rarely used in practice.
2430 This issue was discovered by Brian Carpenter.
2434 *) Removed the export ciphers from the DEFAULT ciphers
2437 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2439 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2440 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2441 So far those who have to target multiple platforms would compromise
2442 and argue that binary targeting say ARMv5 would still execute on
2443 ARMv8. "Universal" build resolves this compromise by providing
2444 near-optimal performance even on newer platforms.
2447 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2448 (other platforms pending).
2449 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2451 *) Add support for the SignedCertificateTimestampList certificate and
2452 OCSP response extensions from RFC6962.
2455 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2456 for corner cases. (Certain input points at infinity could lead to
2457 bogus results, with non-infinity inputs mapped to infinity too.)
2460 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2461 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2462 common cases are optimized and there still is room for further
2463 improvements. Vector Permutation AES for Altivec is also added.
2466 *) Add support for little-endian ppc64 Linux target.
2467 [Marcelo Cerri (IBM)]
2469 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2470 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2471 are optimized and there still is room for further improvements.
2472 Both 32- and 64-bit modes are supported.
2473 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2475 *) Improved ARMv7 NEON support.
2478 *) Support for SPARC Architecture 2011 crypto extensions, first
2479 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2480 SHA256/512, MD5, GHASH and modular exponentiation.
2481 [Andy Polyakov, David Miller]
2483 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2485 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2487 *) Support for new and upcoming Intel processors, including AVX2,
2488 BMI and SHA ISA extensions. This includes additional "stitched"
2489 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2492 This work was sponsored by Intel Corp.
2495 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2496 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2497 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2500 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2501 this fixes a limitation in previous versions of OpenSSL.
2504 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2505 MGF1 digest and OAEP label.
2508 *) Add EVP support for key wrapping algorithms, to avoid problems with
2509 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2510 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2511 algorithms and include tests cases.
2514 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2516 [Douglas E. Engert, Steve Henson]
2518 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2519 difference in days and seconds between two tm or ASN1_TIME structures.
2522 *) Add -rev test option to s_server to just reverse order of characters
2523 received by client and send back to server. Also prints an abbreviated
2524 summary of the connection parameters.
2527 *) New option -brief for s_client and s_server to print out a brief summary
2528 of connection parameters.
2531 *) Add callbacks for arbitrary TLS extensions.
2532 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2534 *) New option -crl_download in several openssl utilities to download CRLs
2535 from CRLDP extension in certificates.
2538 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2541 *) New function X509_CRL_diff to generate a delta CRL from the difference
2542 of two full CRLs. Add support to "crl" utility.
2545 *) New functions to set lookup_crls function and to retrieve
2546 X509_STORE from X509_STORE_CTX.
2549 *) Print out deprecated issuer and subject unique ID fields in
2553 *) Extend OCSP I/O functions so they can be used for simple general purpose
2554 HTTP as well as OCSP. New wrapper function which can be used to download
2555 CRLs using the OCSP API.
2558 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2561 *) SSL_CONF* functions. These provide a common framework for application
2562 configuration using configuration files or command lines.
2565 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2566 message callback and prints the results. Needs compile time option
2567 "enable-ssl-trace". New options to s_client and s_server to enable
2571 *) New ctrl and macro to retrieve supported points extensions.
2572 Print out extension in s_server and s_client.
2575 *) New functions to retrieve certificate signature and signature
2579 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2583 *) New Suite B modes for TLS code. These use and enforce the requirements
2584 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2585 only use Suite B curves. The Suite B modes can be set by using the
2586 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2589 *) New chain verification flags for Suite B levels of security. Check
2590 algorithms are acceptable when flags are set in X509_verify_cert.
2593 *) Make tls1_check_chain return a set of flags indicating checks passed
2594 by a certificate chain. Add additional tests to handle client
2595 certificates: checks for matching certificate type and issuer name
2599 *) If an attempt is made to use a signature algorithm not in the peer
2600 preference list abort the handshake. If client has no suitable
2601 signature algorithms in response to a certificate request do not
2602 use the certificate.
2605 *) If server EC tmp key is not in client preference list abort handshake.
2608 *) Add support for certificate stores in CERT structure. This makes it
2609 possible to have different stores per SSL structure or one store in
2610 the parent SSL_CTX. Include distinct stores for certificate chain
2611 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2612 to build and store a certificate chain in CERT structure: returning
2613 an error if the chain cannot be built: this will allow applications
2614 to test if a chain is correctly configured.
2616 Note: if the CERT based stores are not set then the parent SSL_CTX
2617 store is used to retain compatibility with existing behaviour.
2621 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2622 mask based on the current session, check mask when sending client
2623 hello and checking the requested ciphersuite.
2626 *) New ctrls to retrieve and set certificate types in a certificate
2627 request message. Print out received values in s_client. If certificate
2628 types is not set with custom values set sensible values based on
2629 supported signature algorithms.
2632 *) Support for distinct client and server supported signature algorithms.
2635 *) Add certificate callback. If set this is called whenever a certificate
2636 is required by client or server. An application can decide which
2637 certificate chain to present based on arbitrary criteria: for example
2638 supported signature algorithms. Add very simple example to s_server.
2639 This fixes many of the problems and restrictions of the existing client
2640 certificate callback: for example you can now clear an existing
2641 certificate and specify the whole chain.
2644 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2645 the certificate can be used for (if anything). Set valid_flags field
2646 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2647 to have similar checks in it.
2649 Add new "cert_flags" field to CERT structure and include a "strict mode".
2650 This enforces some TLS certificate requirements (such as only permitting
2651 certificate signature algorithms contained in the supported algorithms
2652 extension) which some implementations ignore: this option should be used
2653 with caution as it could cause interoperability issues.
2656 *) Update and tidy signature algorithm extension processing. Work out
2657 shared signature algorithms based on preferences and peer algorithms
2658 and print them out in s_client and s_server. Abort handshake if no
2659 shared signature algorithms.
2662 *) Add new functions to allow customised supported signature algorithms
2663 for SSL and SSL_CTX structures. Add options to s_client and s_server
2667 *) New function SSL_certs_clear() to delete all references to certificates
2668 from an SSL structure. Before this once a certificate had been added
2669 it couldn't be removed.
2672 *) Integrate hostname, email address and IP address checking with certificate
2673 verification. New verify options supporting checking in openssl utility.
2676 *) Fixes and wildcard matching support to hostname and email checking
2677 functions. Add manual page.
2678 [Florian Weimer (Red Hat Product Security Team)]
2680 *) New functions to check a hostname email or IP address against a
2681 certificate. Add options x509 utility to print results of checks against
2685 *) Fix OCSP checking.
2686 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2688 *) Initial experimental support for explicitly trusted non-root CAs.
2689 OpenSSL still tries to build a complete chain to a root but if an
2690 intermediate CA has a trust setting included that is used. The first
2691 setting is used: whether to trust (e.g., -addtrust option to the x509
2695 *) Add -trusted_first option which attempts to find certificates in the
2696 trusted store even if an untrusted chain is also supplied.
2699 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2700 platform support for Linux and Android.
2703 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2706 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2707 When in FIPS mode the approved implementations are used as normal,
2708 when not in FIPS mode the internal unapproved versions are used instead.
2709 This means that the FIPS capable OpenSSL isn't forced to use the
2710 (often lower performance) FIPS implementations outside FIPS mode.
2713 *) Transparently support X9.42 DH parameters when calling
2714 PEM_read_bio_DHparameters. This means existing applications can handle
2715 the new parameter format automatically.
2718 *) Initial experimental support for X9.42 DH parameter format: mainly
2719 to support use of 'q' parameter for RFC5114 parameters.
2722 *) Add DH parameters from RFC5114 including test data to dhtest.
2725 *) Support for automatic EC temporary key parameter selection. If enabled
2726 the most preferred EC parameters are automatically used instead of
2727 hardcoded fixed parameters. Now a server just has to call:
2728 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2729 support ECDH and use the most appropriate parameters.
2732 *) Enhance and tidy EC curve and point format TLS extension code. Use
2733 static structures instead of allocation if default values are used.
2734 New ctrls to set curves we wish to support and to retrieve shared curves.
2735 Print out shared curves in s_server. New options to s_server and s_client
2736 to set list of supported curves.
2739 *) New ctrls to retrieve supported signature algorithms and
2740 supported curve values as an array of NIDs. Extend openssl utility
2741 to print out received values.
2744 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2745 between NIDs and the more common NIST names such as "P-256". Enhance
2746 ecparam utility and ECC method to recognise the NIST names for curves.
2749 *) Enhance SSL/TLS certificate chain handling to support different
2750 chains for each certificate instead of one chain in the parent SSL_CTX.
2753 *) Support for fixed DH ciphersuite client authentication: where both
2754 server and client use DH certificates with common parameters.
2757 *) Support for fixed DH ciphersuites: those requiring DH server
2761 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2763 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2764 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2765 X509_CINF_get_signature were reverted post internal team review.
2767 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2769 *) Build fixes for the Windows and OpenVMS platforms
2770 [Matt Caswell and Richard Levitte]
2772 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2774 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2775 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2776 dereference. This could lead to a Denial Of Service attack. Thanks to
2777 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2781 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2782 dtls1_buffer_record function under certain conditions. In particular this
2783 could occur if an attacker sent repeated DTLS records with the same
2784 sequence number but for the next epoch. The memory leak could be exploited
2785 by an attacker in a Denial of Service attack through memory exhaustion.
2786 Thanks to Chris Mueller for reporting this issue.
2790 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2791 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2792 method would be set to NULL which could later result in a NULL pointer
2793 dereference. Thanks to Frank Schmirler for reporting this issue.
2797 *) Abort handshake if server key exchange message is omitted for ephemeral
2800 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2801 reporting this issue.
2805 *) Remove non-export ephemeral RSA code on client and server. This code
2806 violated the TLS standard by allowing the use of temporary RSA keys in
2807 non-export ciphersuites and could be used by a server to effectively
2808 downgrade the RSA key length used to a value smaller than the server
2809 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2810 INRIA or reporting this issue.
2814 *) Fixed issue where DH client certificates are accepted without verification.
2815 An OpenSSL server will accept a DH certificate for client authentication
2816 without the certificate verify message. This effectively allows a client to
2817 authenticate without the use of a private key. This only affects servers
2818 which trust a client certificate authority which issues certificates
2819 containing DH keys: these are extremely rare and hardly ever encountered.
2820 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2825 *) Ensure that the session ID context of an SSL is updated when its
2826 SSL_CTX is updated via SSL_set_SSL_CTX.
2828 The session ID context is typically set from the parent SSL_CTX,
2829 and can vary with the CTX.
2832 *) Fix various certificate fingerprint issues.
2834 By using non-DER or invalid encodings outside the signed portion of a
2835 certificate the fingerprint can be changed without breaking the signature.
2836 Although no details of the signed portion of the certificate can be changed
2837 this can cause problems with some applications: e.g. those using the
2838 certificate fingerprint for blacklists.
2840 1. Reject signatures with non zero unused bits.
2842 If the BIT STRING containing the signature has non zero unused bits reject
2843 the signature. All current signature algorithms require zero unused bits.
2845 2. Check certificate algorithm consistency.
2847 Check the AlgorithmIdentifier inside TBS matches the one in the
2848 certificate signature. NB: this will result in signature failure
2849 errors for some broken certificates.
2851 Thanks to Konrad Kraszewski from Google for reporting this issue.
2853 3. Check DSA/ECDSA signatures use DER.
2855 Re-encode DSA/ECDSA signatures and compare with the original received
2856 signature. Return an error if there is a mismatch.
2858 This will reject various cases including garbage after signature
2859 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2860 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2861 (negative or with leading zeroes).
2863 Further analysis was conducted and fixes were developed by Stephen Henson
2864 of the OpenSSL core team.
2869 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2870 results on some platforms, including x86_64. This bug occurs at random
2871 with a very low probability, and is not known to be exploitable in any
2872 way, though its exact impact is difficult to determine. Thanks to Pieter
2873 Wuille (Blockstream) who reported this issue and also suggested an initial
2874 fix. Further analysis was conducted by the OpenSSL development team and
2875 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2876 the OpenSSL core team.
2880 *) Do not resume sessions on the server if the negotiated protocol
2881 version does not match the session's version. Resuming with a different
2882 version, while not strictly forbidden by the RFC, is of questionable
2883 sanity and breaks all known clients.
2884 [David Benjamin, Emilia Käsper]
2886 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2887 early CCS messages during renegotiation. (Note that because
2888 renegotiation is encrypted, this early CCS was not exploitable.)
2891 *) Tighten client-side session ticket handling during renegotiation:
2892 ensure that the client only accepts a session ticket if the server sends
2893 the extension anew in the ServerHello. Previously, a TLS client would
2894 reuse the old extension state and thus accept a session ticket if one was
2895 announced in the initial ServerHello.
2897 Similarly, ensure that the client requires a session ticket if one
2898 was advertised in the ServerHello. Previously, a TLS client would
2899 ignore a missing NewSessionTicket message.
2902 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2904 *) SRTP Memory Leak.
2906 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2907 sends a carefully crafted handshake message, to cause OpenSSL to fail
2908 to free up to 64k of memory causing a memory leak. This could be
2909 exploited in a Denial Of Service attack. This issue affects OpenSSL
2910 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2911 whether SRTP is used or configured. Implementations of OpenSSL that
2912 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2914 The fix was developed by the OpenSSL team.
2918 *) Session Ticket Memory Leak.
2920 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2921 integrity of that ticket is first verified. In the event of a session
2922 ticket integrity check failing, OpenSSL will fail to free memory
2923 causing a memory leak. By sending a large number of invalid session
2924 tickets an attacker could exploit this issue in a Denial Of Service
2929 *) Build option no-ssl3 is incomplete.
2931 When OpenSSL is configured with "no-ssl3" as a build option, servers
2932 could accept and complete a SSL 3.0 handshake, and clients could be
2933 configured to send them.
2935 [Akamai and the OpenSSL team]
2937 *) Add support for TLS_FALLBACK_SCSV.
2938 Client applications doing fallback retries should call
2939 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2941 [Adam Langley, Bodo Moeller]
2943 *) Add additional DigestInfo checks.
2945 Re-encode DigestInto in DER and check against the original when
2946 verifying RSA signature: this will reject any improperly encoded
2947 DigestInfo structures.
2949 Note: this is a precautionary measure and no attacks are currently known.
2953 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2955 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2956 SRP code can be overrun an internal buffer. Add sanity check that
2957 g, A, B < N to SRP code.
2959 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2960 Group for discovering this issue.
2964 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2965 TLS 1.0 instead of higher protocol versions when the ClientHello message
2966 is badly fragmented. This allows a man-in-the-middle attacker to force a
2967 downgrade to TLS 1.0 even if both the server and the client support a
2968 higher protocol version, by modifying the client's TLS records.
2970 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2971 researching this issue.
2975 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2976 to a denial of service attack. A malicious server can crash the client
2977 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2978 ciphersuite and sending carefully crafted handshake messages.
2980 Thanks to Felix Gröbert (Google) for discovering and researching this
2985 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2986 to leak memory. This can be exploited through a Denial of Service attack.
2987 Thanks to Adam Langley for discovering and researching this issue.
2991 *) An attacker can force openssl to consume large amounts of memory whilst
2992 processing DTLS handshake messages. This can be exploited through a
2993 Denial of Service attack.
2994 Thanks to Adam Langley for discovering and researching this issue.
2998 *) An attacker can force an error condition which causes openssl to crash
2999 whilst processing DTLS packets due to memory being freed twice. This
3000 can be exploited through a Denial of Service attack.
3001 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3006 *) If a multithreaded client connects to a malicious server using a resumed
3007 session and the server sends an ec point format extension it could write
3008 up to 255 bytes to freed memory.
3010 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3015 *) A malicious server can crash an OpenSSL client with a null pointer
3016 dereference (read) by specifying an SRP ciphersuite even though it was not
3017 properly negotiated with the client. This can be exploited through a
3018 Denial of Service attack.
3020 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3021 discovering and researching this issue.
3025 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3026 X509_name_oneline, X509_name_print_ex et al. to leak some information
3027 from the stack. Applications may be affected if they echo pretty printing
3028 output to the attacker.
3030 Thanks to Ivan Fratric (Google) for discovering this issue.
3032 [Emilia Käsper, and Steve Henson]
3034 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3035 for corner cases. (Certain input points at infinity could lead to
3036 bogus results, with non-infinity inputs mapped to infinity too.)
3039 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3041 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3042 handshake can force the use of weak keying material in OpenSSL
3043 SSL/TLS clients and servers.
3045 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3046 researching this issue. (CVE-2014-0224)
3047 [KIKUCHI Masashi, Steve Henson]
3049 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3050 OpenSSL DTLS client the code can be made to recurse eventually crashing
3053 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3055 [Imre Rad, Steve Henson]
3057 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3058 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3059 client or server. This is potentially exploitable to run arbitrary
3060 code on a vulnerable client or server.
3062 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3063 [Jüri Aedla, Steve Henson]
3065 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3066 are subject to a denial of service attack.
3068 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3069 this issue. (CVE-2014-3470)
3070 [Felix Gröbert, Ivan Fratric, Steve Henson]
3072 *) Harmonize version and its documentation. -f flag is used to display
3074 [mancha <mancha1@zoho.com>]
3076 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3077 in i2d_ECPrivateKey.
3078 [mancha <mancha1@zoho.com>]
3080 *) Fix some double frees. These are not thought to be exploitable.
3081 [mancha <mancha1@zoho.com>]
3083 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3085 *) A missing bounds check in the handling of the TLS heartbeat extension
3086 can be used to reveal up to 64k of memory to a connected client or
3089 Thanks for Neel Mehta of Google Security for discovering this bug and to
3090 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3091 preparing the fix (CVE-2014-0160)
3092 [Adam Langley, Bodo Moeller]
3094 *) Fix for the attack described in the paper "Recovering OpenSSL
3095 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3096 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3097 http://eprint.iacr.org/2014/140
3099 Thanks to Yuval Yarom and Naomi Benger for discovering this
3100 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3101 [Yuval Yarom and Naomi Benger]
3103 *) TLS pad extension: draft-agl-tls-padding-03
3105 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3106 TLS client Hello record length value would otherwise be > 255 and
3107 less that 512 pad with a dummy extension containing zeroes so it
3108 is at least 512 bytes long.
3110 [Adam Langley, Steve Henson]
3112 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3114 *) Fix for TLS record tampering bug. A carefully crafted invalid
3115 handshake could crash OpenSSL with a NULL pointer exception.
3116 Thanks to Anton Johansson for reporting this issues.
3119 *) Keep original DTLS digest and encryption contexts in retransmission
3120 structures so we can use the previous session parameters if they need
3121 to be resent. (CVE-2013-6450)
3124 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3125 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3126 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3127 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3128 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3129 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3130 [Rob Stradling, Adam Langley]
3132 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3134 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3135 supporting platforms or when small records were transferred.
3136 [Andy Polyakov, Steve Henson]
3138 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3140 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3142 This addresses the flaw in CBC record processing discovered by
3143 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3144 at: http://www.isg.rhul.ac.uk/tls/
3146 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3147 Security Group at Royal Holloway, University of London
3148 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3149 Emilia Käsper for the initial patch.
3151 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3153 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3154 ciphersuites which can be exploited in a denial of service attack.
3155 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3156 and detecting this bug and to Wolfgang Ettlinger
3157 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3161 *) Return an error when checking OCSP signatures when key is NULL.
3162 This fixes a DoS attack. (CVE-2013-0166)
3165 *) Make openssl verify return errors.
3166 [Chris Palmer <palmer@google.com> and Ben Laurie]
3168 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3169 the right response is stapled. Also change SSL_get_certificate()
3170 so it returns the certificate actually sent.
3171 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3172 [Rob Stradling <rob.stradling@comodo.com>]
3174 *) Fix possible deadlock when decoding public keys.
3177 *) Don't use TLS 1.0 record version number in initial client hello
3181 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3183 *) Sanity check record length before skipping explicit IV in TLS
3184 1.2, 1.1 and DTLS to fix DoS attack.
3186 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3187 fuzzing as a service testing platform.
3191 *) Initialise tkeylen properly when encrypting CMS messages.
3192 Thanks to Solar Designer of Openwall for reporting this issue.
3195 *) In FIPS mode don't try to use composite ciphers as they are not
3199 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3201 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3202 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3203 mean any application compiled against OpenSSL 1.0.0 headers setting
3204 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3205 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3206 0x10000000L Any application which was previously compiled against
3207 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3208 will need to be recompiled as a result. Letting be results in
3209 inability to disable specifically TLS 1.1 and in client context,
3210 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3213 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3214 disable just protocol X, but all protocols above X *if* there are
3215 protocols *below* X still enabled. In more practical terms it means
3216 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3217 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3218 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3222 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3224 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3225 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3226 in CRYPTO_realloc_clean.
3228 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3229 issue and to Adam Langley <agl@chromium.org> for fixing it.
3231 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3233 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3236 *) Workarounds for some broken servers that "hang" if a client hello
3237 record length exceeds 255 bytes.
3239 1. Do not use record version number > TLS 1.0 in initial client
3240 hello: some (but not all) hanging servers will now work.
3241 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3242 the number of ciphers sent in the client hello. This should be
3243 set to an even number, such as 50, for example by passing:
3244 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3245 Most broken servers should now work.
3246 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3247 TLS 1.2 client support entirely.
3250 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3253 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3255 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3256 STRING form instead of a DigestInfo.
3259 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3260 and the RSA_sign/RSA_verify functions. This was made more apparent when
3261 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3262 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3263 the correct format in RSA_verify so both forms transparently work.
3266 *) Some servers which support TLS 1.0 can choke if we initially indicate
3267 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3268 encrypted premaster secret. As a workaround use the maximum permitted
3269 client version in client hello, this should keep such servers happy
3270 and still work with previous versions of OpenSSL.
3273 *) Add support for TLS/DTLS heartbeats.
3274 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3276 *) Add support for SCTP.
3277 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3279 *) Improved PRNG seeding for VOS.
3280 [Paul Green <Paul.Green@stratus.com>]
3282 *) Extensive assembler packs updates, most notably:
3284 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3285 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3286 - x86_64: bit-sliced AES implementation;
3287 - ARM: NEON support, contemporary platforms optimizations;
3288 - s390x: z196 support;
3289 - *: GHASH and GF(2^m) multiplication implementations;
3293 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3294 (removal of unnecessary code)
3295 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3297 *) Add TLS key material exporter from RFC 5705.
3300 *) Add DTLS-SRTP negotiation from RFC 5764.
3303 *) Add Next Protocol Negotiation,
3304 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3305 disabled with a no-npn flag to config or Configure. Code donated
3307 [Adam Langley <agl@google.com> and Ben Laurie]
3309 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3310 NIST-P256, NIST-P521, with constant-time single point multiplication on
3311 typical inputs. Compiler support for the nonstandard type __uint128_t is
3312 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3313 Code made available under Apache License version 2.0.
3315 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3316 line to include this in your build of OpenSSL, and run "make depend" (or
3317 "make update"). This enables the following EC_METHODs:
3319 EC_GFp_nistp224_method()
3320 EC_GFp_nistp256_method()
3321 EC_GFp_nistp521_method()
3323 EC_GROUP_new_by_curve_name() will automatically use these (while
3324 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3326 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3328 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3329 all platforms. Move ssize_t definition from e_os.h to the public
3330 header file e_os2.h as it now appears in public header file cms.h
3333 *) New -sigopt option to the ca, req and x509 utilities. Additional
3334 signature parameters can be passed using this option and in
3338 *) Add RSA PSS signing function. This will generate and set the
3339 appropriate AlgorithmIdentifiers for PSS based on those in the
3340 corresponding EVP_MD_CTX structure. No application support yet.
3343 *) Support for companion algorithm specific ASN1 signing routines.
3344 New function ASN1_item_sign_ctx() signs a pre-initialised
3345 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3346 the appropriate parameters.
3349 *) Add new algorithm specific ASN1 verification initialisation function
3350 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3351 handling will be the same no matter what EVP_PKEY_METHOD is used.
3352 Add a PSS handler to support verification of PSS signatures: checked
3353 against a number of sample certificates.
3356 *) Add signature printing for PSS. Add PSS OIDs.
3357 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3359 *) Add algorithm specific signature printing. An individual ASN1 method
3360 can now print out signatures instead of the standard hex dump.
3362 More complex signatures (e.g. PSS) can print out more meaningful
3363 information. Include DSA version that prints out the signature
3367 *) Password based recipient info support for CMS library: implementing
3371 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3372 neatly separates the code into cipher and PBE sections and is required
3373 for some algorithms that split PBES2 into separate pieces (such as
3374 password based CMS).
3377 *) Session-handling fixes:
3378 - Fix handling of connections that are resuming with a session ID,
3379 but also support Session Tickets.
3380 - Fix a bug that suppressed issuing of a new ticket if the client
3381 presented a ticket with an expired session.
3382 - Try to set the ticket lifetime hint to something reasonable.
3383 - Make tickets shorter by excluding irrelevant information.
3384 - On the client side, don't ignore renewed tickets.
3385 [Adam Langley, Bodo Moeller (Google)]
3387 *) Fix PSK session representation.
3390 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3392 This work was sponsored by Intel.
3395 *) Add GCM support to TLS library. Some custom code is needed to split
3396 the IV between the fixed (from PRF) and explicit (from TLS record)
3397 portions. This adds all GCM ciphersuites supported by RFC5288 and
3398 RFC5289. Generalise some AES* cipherstrings to include GCM and
3399 add a special AESGCM string for GCM only.
3402 *) Expand range of ctrls for AES GCM. Permit setting invocation
3403 field on decrypt and retrieval of invocation field only on encrypt.
3406 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3407 As required by RFC5289 these ciphersuites cannot be used if for
3408 versions of TLS earlier than 1.2.
3411 *) For FIPS capable OpenSSL interpret a NULL default public key method
3412 as unset and return the appropriate default but do *not* set the default.
3413 This means we can return the appropriate method in applications that
3414 switch between FIPS and non-FIPS modes.
3417 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3418 ENGINE is used then we cannot handle that in the FIPS module so we
3419 keep original code iff non-FIPS operations are allowed.
3422 *) Add -attime option to openssl utilities.
3423 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3425 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3428 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3429 FIPS EC methods unconditionally for now.
3432 *) New build option no-ec2m to disable characteristic 2 code.
3435 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3436 all cases can be covered as some introduce binary incompatibilities.
3439 *) Redirect RSA operations to FIPS module including keygen,
3440 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3443 *) Add similar low level API blocking to ciphers.
3446 *) Low level digest APIs are not approved in FIPS mode: any attempt
3447 to use these will cause a fatal error. Applications that *really* want
3448 to use them can use the private_* version instead.
3451 *) Redirect cipher operations to FIPS module for FIPS builds.
3454 *) Redirect digest operations to FIPS module for FIPS builds.
3457 *) Update build system to add "fips" flag which will link in fipscanister.o
3458 for static and shared library builds embedding a signature if needed.
3461 *) Output TLS supported curves in preference order instead of numerical
3462 order. This is currently hardcoded for the highest order curves first.
3463 This should be configurable so applications can judge speed vs strength.
3466 *) Add TLS v1.2 server support for client authentication.
3469 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3473 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3474 FIPS modules versions.
3477 *) Add TLS v1.2 client side support for client authentication. Keep cache
3478 of handshake records longer as we don't know the hash algorithm to use
3479 until after the certificate request message is received.
3482 *) Initial TLS v1.2 client support. Add a default signature algorithms
3483 extension including all the algorithms we support. Parse new signature
3484 format in client key exchange. Relax some ECC signing restrictions for
3485 TLS v1.2 as indicated in RFC5246.
3488 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3489 to new signature format when needed using client digest preference.
3490 All server ciphersuites should now work correctly in TLS v1.2. No client
3491 support yet and no support for client certificates.
3494 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3495 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3496 ciphersuites. At present only RSA key exchange ciphersuites work with
3497 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3498 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3499 and version checking.
3502 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3503 with this defined it will not be affected by any changes to ssl internal
3504 structures. Add several utility functions to allow openssl application
3505 to work with OPENSSL_NO_SSL_INTERN defined.
3508 *) A long standing patch to add support for SRP from EdelWeb (Peter
3509 Sylvester and Christophe Renou) was integrated.
3510 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3511 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3514 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3517 *) Permit abbreviated handshakes when renegotiating using the function
3518 SSL_renegotiate_abbreviated().
3519 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3521 *) Add call to ENGINE_register_all_complete() to
3522 ENGINE_load_builtin_engines(), so some implementations get used
3523 automatically instead of needing explicit application support.
3526 *) Add support for TLS key exporter as described in RFC5705.
3527 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3529 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3530 a few changes are required:
3532 Add SSL_OP_NO_TLSv1_1 flag.
3533 Add TLSv1_1 methods.
3534 Update version checking logic to handle version 1.1.
3535 Add explicit IV handling (ported from DTLS code).
3536 Add command line options to s_client/s_server.
3539 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3541 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3542 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3543 content decryption and always return the same error. Note: this attack
3544 needs on average 2^20 messages so it only affects automated senders. The
3545 old behaviour can be re-enabled in the CMS code by setting the
3546 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3547 an MMA defence is not necessary.
3548 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3549 this issue. (CVE-2012-0884)
3552 *) Fix CVE-2011-4619: make sure we really are receiving a
3553 client hello before rejecting multiple SGC restarts. Thanks to
3554 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3557 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3559 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3560 Thanks to Antonio Martin, Enterprise Secure Access Research and
3561 Development, Cisco Systems, Inc. for discovering this bug and
3562 preparing a fix. (CVE-2012-0050)
3565 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3567 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3568 of the Vaudenay padding oracle attack on CBC mode encryption
3569 which enables an efficient plaintext recovery attack against
3570 the OpenSSL implementation of DTLS. Their attack exploits timing
3571 differences arising during decryption processing. A research
3572 paper describing this attack can be found at:
3573 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3574 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3575 Security Group at Royal Holloway, University of London
3576 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3577 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3578 for preparing the fix. (CVE-2011-4108)
3579 [Robin Seggelmann, Michael Tuexen]
3581 *) Clear bytes used for block padding of SSL 3.0 records.
3583 [Adam Langley (Google)]
3585 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3586 Kadianakis <desnacked@gmail.com> for discovering this issue and
3587 Adam Langley for preparing the fix. (CVE-2011-4619)
3588 [Adam Langley (Google)]
3590 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3591 [Andrey Kulikov <amdeich@gmail.com>]
3593 *) Prevent malformed RFC3779 data triggering an assertion failure.
3594 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3595 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3596 [Rob Austein <sra@hactrn.net>]
3598 *) Improved PRNG seeding for VOS.
3599 [Paul Green <Paul.Green@stratus.com>]
3601 *) Fix ssl_ciph.c set-up race.
3602 [Adam Langley (Google)]
3604 *) Fix spurious failures in ecdsatest.c.
3605 [Emilia Käsper (Google)]
3607 *) Fix the BIO_f_buffer() implementation (which was mixing different
3608 interpretations of the '..._len' fields).
3609 [Adam Langley (Google)]
3611 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3612 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3613 threads won't reuse the same blinding coefficients.
3615 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3616 lock to call BN_BLINDING_invert_ex, and avoids one use of
3617 BN_BLINDING_update for each BN_BLINDING structure (previously,
3618 the last update always remained unused).
3619 [Emilia Käsper (Google)]
3621 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3622 [Bob Buckholz (Google)]
3624 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
projects / openssl.git / blob