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 *) Modified the random device based seed sources to keep the relevant
13 file descriptors open rather than reopening them on each access.
14 This allows such sources to operate in a chroot() jail without
15 the associated device nodes being available. This behaviour can be
16 controlled using RAND_keep_random_devices_open().
19 *) Numerous side-channel attack mitigations have been applied. This may have
20 performance impacts for some algorithms for the benefit of improved
21 security. Specific changes are noted in this change log by their respective
25 *) AIX shared library support overhaul. Switch to AIX "natural" way of
26 handling shared libraries, which means collecting shared objects of
27 different versions and bitnesses in one common archive. This allows to
28 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
29 doesn't affect the way 3rd party applications are linked, only how
30 multi-version installation is managed.
33 *) Make ec_group_do_inverse_ord() more robust and available to other
34 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
35 mitigations are applied to the fallback BN_mod_inverse().
36 When using this function rather than BN_mod_inverse() directly, new
37 EC cryptosystem implementations are then safer-by-default.
40 *) Add coordinate blinding for EC_POINT and implement projective
41 coordinate blinding for generic prime curves as a countermeasure to
42 chosen point SCA attacks.
43 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
45 *) Add blinding to ECDSA and DSA signatures to protect against side channel
46 attacks discovered by Keegan Ryan (NCC Group).
49 *) Enforce checking in the pkeyutl command line app to ensure that the input
50 length does not exceed the maximum supported digest length when performing
51 a sign, verify or verifyrecover operation.
54 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
55 I/O in combination with something like select() or poll() will hang. This
56 can be turned off again using SSL_CTX_clear_mode().
57 Many applications do not properly handle non-application data records, and
58 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
59 around the problems in those applications, but can also break some.
60 It's recommended to read the manpages about SSL_read(), SSL_write(),
61 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
62 SSL_CTX_set_read_ahead() again.
65 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
66 now allow empty (zero character) pass phrases.
69 *) Apply blinding to binary field modular inversion and remove patent
70 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
73 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
74 binary and prime elliptic curves.
77 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
78 constant time fixed point multiplication.
81 *) Revise elliptic curve scalar multiplication with timing attack
82 defenses: ec_wNAF_mul redirects to a constant time implementation
83 when computing fixed point and variable point multiplication (which
84 in OpenSSL are mostly used with secret scalars in keygen, sign,
85 ECDH derive operations).
86 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
89 *) Updated CONTRIBUTING
92 *) Updated DRBG / RAND to request nonce and additional low entropy
93 randomness from the system.
96 *) Updated 'openssl rehash' to use OpenSSL consistent default.
99 *) Moved the load of the ssl_conf module to libcrypto, which helps
100 loading engines that libssl uses before libssl is initialised.
103 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
106 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
107 [Ingo Schwarze, Rich Salz]
109 *) Added output of accepting IP address and port for 'openssl s_server'
112 *) Added a new API for TLSv1.3 ciphersuites:
113 SSL_CTX_set_ciphersuites()
114 SSL_set_ciphersuites()
117 *) Memory allocation failures consistenly add an error to the error
121 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
122 in libcrypto when run as setuid/setgid.
125 *) Load any config file by default when libssl is used.
128 *) Added new public header file <openssl/rand_drbg.h> and documentation
129 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
130 [Matthias St. Pierre]
132 *) QNX support removed (cannot find contributors to get their approval
133 for the license change).
136 *) TLSv1.3 replay protection for early data has been implemented. See the
137 SSL_read_early_data() man page for further details.
140 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
141 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
142 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
143 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
144 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
145 configuration has been separated out. See the ciphers man page or the
146 SSL_CTX_set_ciphersuites() man page for more information.
149 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
150 in responder mode now supports the new "-multi" option, which
151 spawns the specified number of child processes to handle OCSP
152 requests. The "-timeout" option now also limits the OCSP
153 responder's patience to wait to receive the full client request
154 on a newly accepted connection. Child processes are respawned
155 as needed, and the CA index file is automatically reloaded
156 when changed. This makes it possible to run the "ocsp" responder
157 as a long-running service, making the OpenSSL CA somewhat more
158 feature-complete. In this mode, most diagnostic messages logged
159 after entering the event loop are logged via syslog(3) rather than
163 *) Added support for X448 and Ed448. Heavily based on original work by
167 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
168 objects loaded. This adds the functions OSSL_STORE_expect() and
169 OSSL_STORE_find() as well as needed tools to construct searches and
170 get the search data out of them.
173 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
174 version of OpenSSL should review their configuration settings to ensure
175 that they are still appropriate for TLSv1.3. For further information see:
176 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
178 NOTE: In this pre-release of OpenSSL a draft version of the
179 TLSv1.3 standard has been implemented. Implementations of different draft
180 versions of the standard do not inter-operate, and this version will not
181 inter-operate with an implementation of the final standard when it is
182 eventually published. Different pre-release versions may implement
183 different versions of the draft. The final version of OpenSSL 1.1.1 will
184 implement the final version of the standard.
185 TODO(TLS1.3): Remove the above note before final release
188 *) Grand redesign of the OpenSSL random generator
190 The default RAND method now utilizes an AES-CTR DRBG according to
191 NIST standard SP 800-90Ar1. The new random generator is essentially
192 a port of the default random generator from the OpenSSL FIPS 2.0
193 object module. It is a hybrid deterministic random bit generator
194 using an AES-CTR bit stream and which seeds and reseeds itself
195 automatically using trusted system entropy sources.
197 Some of its new features are:
198 o Support for multiple DRBG instances with seed chaining.
199 o Add a public DRBG instance for the default RAND method.
200 o Add a dedicated DRBG instance for generating long term private keys.
201 o Make the DRBG instances fork-safe.
202 o Keep all global DRBG instances on the secure heap if it is enabled.
203 o Add a DRBG instance to every SSL instance for lock free operation
204 and to increase unpredictability.
205 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
207 *) Changed Configure so it only says what it does and doesn't dump
208 so much data. Instead, ./configdata.pm should be used as a script
209 to display all sorts of configuration data.
212 *) Added processing of "make variables" to Configure.
215 *) Added SHA512/224 and SHA512/256 algorithm support.
218 *) The last traces of Netware support, first removed in 1.1.0, have
222 *) Get rid of Makefile.shared, and in the process, make the processing
223 of certain files (rc.obj, or the .def/.map/.opt files produced from
224 the ordinal files) more visible and hopefully easier to trace and
225 debug (or make silent).
228 *) Make it possible to have environment variable assignments as
229 arguments to config / Configure.
232 *) Add multi-prime RSA (RFC 8017) support.
235 *) Add SM3 implemented according to GB/T 32905-2016
236 [ Jack Lloyd <jack.lloyd@ribose.com>,
237 Ronald Tse <ronald.tse@ribose.com>,
238 Erick Borsboom <erick.borsboom@ribose.com> ]
240 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
241 as documented in RFC6066.
242 Based on a patch from Tomasz Moń
243 [Filipe Raimundo da Silva]
245 *) Add SM4 implemented according to GB/T 32907-2016.
246 [ Jack Lloyd <jack.lloyd@ribose.com>,
247 Ronald Tse <ronald.tse@ribose.com>,
248 Erick Borsboom <erick.borsboom@ribose.com> ]
250 *) Reimplement -newreq-nodes and ERR_error_string_n; the
251 original author does not agree with the license change.
254 *) Add ARIA AEAD TLS support.
257 *) Some macro definitions to support VS6 have been removed. Visual
258 Studio 6 has not worked since 1.1.0
261 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
262 without clearing the errors.
265 *) Add "atfork" functions. If building on a system that without
266 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
267 requirements. The RAND facility now uses/requires this.
273 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
274 not possible to disable entirely. However, it's still possible to
275 disable the console reading UI method, UI_OpenSSL() (use UI_null()
278 To disable, configure with 'no-ui-console'. 'no-ui' is still
279 possible to use as an alias. Check at compile time with the
280 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
281 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
284 *) Add a STORE module, which implements a uniform and URI based reader of
285 stores that can contain keys, certificates, CRLs and numerous other
286 objects. The main API is loosely based on a few stdio functions,
287 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
288 OSSL_STORE_error and OSSL_STORE_close.
289 The implementation uses backends called "loaders" to implement arbitrary
290 URI schemes. There is one built in "loader" for the 'file' scheme.
293 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
294 then adjusted to work on FreeBSD 8.4 as well.
295 Enable by configuring with 'enable-devcryptoeng'. This is done by default
296 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
299 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
300 util/mkerr.pl, which is adapted to allow those prefixes, leading to
301 error code calls like this:
303 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
305 With this change, we claim the namespaces OSSL and OPENSSL in a manner
306 that can be encoded in C. For the foreseeable future, this will only
308 [Richard Levitte and Tim Hudson]
310 *) Removed BSD cryptodev engine.
313 *) Add a build target 'build_all_generated', to build all generated files
314 and only that. This can be used to prepare everything that requires
315 things like perl for a system that lacks perl and then move everything
316 to that system and do the rest of the build there.
319 *) In the UI interface, make it possible to duplicate the user data. This
320 can be used by engines that need to retain the data for a longer time
321 than just the call where this user data is passed.
324 *) Ignore the '-named_curve auto' value for compatibility of applications
326 [Tomas Mraz <tmraz@fedoraproject.org>]
328 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
329 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
330 alerts across multiple records (some of which could be empty). In practice
331 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
332 prohibts this altogether and other libraries (BoringSSL, NSS) do not
333 support this at all. Supporting it adds significant complexity to the
334 record layer, and its removal is unlikely to cause inter-operability
338 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
339 with Z. These are meant to replace LONG and ZLONG and to be size safe.
340 The use of LONG and ZLONG is discouraged and scheduled for deprecation
344 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
345 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
346 [Richard Levitte, Andy Polyakov]
348 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
352 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
353 platform rather than 'mingw'.
356 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
357 success if they are asked to add an object which already exists
358 in the store. This change cascades to other functions which load
359 certificates and CRLs.
362 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
363 facilitate stack unwinding even from assembly subroutines.
366 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
367 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
370 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
371 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
372 which is the minimum version we support.
375 *) Certificate time validation (X509_cmp_time) enforces stricter
376 compliance with RFC 5280. Fractional seconds and timezone offsets
377 are no longer allowed.
380 *) Add support for ARIA
383 *) s_client will now send the Server Name Indication (SNI) extension by
384 default unless the new "-noservername" option is used. The server name is
385 based on the host provided to the "-connect" option unless overridden by
389 *) Add support for SipHash
392 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
393 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
394 prevent issues where no progress is being made and the peer continually
395 sends unrecognised record types, using up resources processing them.
398 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
399 using the algorithm defined in
400 https://www.akkadia.org/drepper/SHA-crypt.txt
403 *) Heartbeat support has been removed; the ABI is changed for now.
404 [Richard Levitte, Rich Salz]
406 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
409 *) The RSA "null" method, which was partially supported to avoid patent
410 issues, has been replaced to always returns NULL.
414 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
416 *) Fixed a text canonicalisation bug in CMS
418 Where a CMS detached signature is used with text content the text goes
419 through a canonicalisation process first prior to signing or verifying a
420 signature. This process strips trailing space at the end of lines, converts
421 line terminators to CRLF and removes additional trailing line terminators
422 at the end of a file. A bug in the canonicalisation process meant that
423 some characters, such as form-feed, were incorrectly treated as whitespace
424 and removed. This is contrary to the specification (RFC5485). This fix
425 could mean that detached text data signed with an earlier version of
426 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
427 signed with a fixed OpenSSL may fail to verify with an earlier version of
428 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
429 and use the "-binary" flag (for the "cms" command line application) or set
430 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
433 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
435 *) Constructed ASN.1 types with a recursive definition could exceed the stack
437 Constructed ASN.1 types with a recursive definition (such as can be found
438 in PKCS7) could eventually exceed the stack given malicious input with
439 excessive recursion. This could result in a Denial Of Service attack. There
440 are no such structures used within SSL/TLS that come from untrusted sources
441 so this is considered safe.
443 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
448 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
450 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
451 effectively reduced to only comparing the least significant bit of each
452 byte. This allows an attacker to forge messages that would be considered as
453 authenticated in an amount of tries lower than that guaranteed by the
454 security claims of the scheme. The module can only be compiled by the
455 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
457 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
462 *) Add a build target 'build_all_generated', to build all generated files
463 and only that. This can be used to prepare everything that requires
464 things like perl for a system that lacks perl and then move everything
465 to that system and do the rest of the build there.
468 *) Backport SSL_OP_NO_RENGOTIATION
470 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
471 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
472 changes this is no longer possible in 1.1.0. Therefore the new
473 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
474 1.1.0 to provide equivalent functionality.
476 Note that if an application built against 1.1.0h headers (or above) is run
477 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
478 accepted but nothing will happen, i.e. renegotiation will not be prevented.
481 *) Removed the OS390-Unix config target. It relied on a script that doesn't
485 *) rsaz_1024_mul_avx2 overflow bug on x86_64
487 There is an overflow bug in the AVX2 Montgomery multiplication procedure
488 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
489 Analysis suggests that attacks against RSA and DSA as a result of this
490 defect would be very difficult to perform and are not believed likely.
491 Attacks against DH1024 are considered just feasible, because most of the
492 work necessary to deduce information about a private key may be performed
493 offline. The amount of resources required for such an attack would be
494 significant. However, for an attack on TLS to be meaningful, the server
495 would have to share the DH1024 private key among multiple clients, which is
496 no longer an option since CVE-2016-0701.
498 This only affects processors that support the AVX2 but not ADX extensions
499 like Intel Haswell (4th generation).
501 This issue was reported to OpenSSL by David Benjamin (Google). The issue
502 was originally found via the OSS-Fuzz project.
506 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
508 *) bn_sqrx8x_internal carry bug on x86_64
510 There is a carry propagating bug in the x86_64 Montgomery squaring
511 procedure. No EC algorithms are affected. Analysis suggests that attacks
512 against RSA and DSA as a result of this defect would be very difficult to
513 perform and are not believed likely. Attacks against DH are considered just
514 feasible (although very difficult) because most of the work necessary to
515 deduce information about a private key may be performed offline. The amount
516 of resources required for such an attack would be very significant and
517 likely only accessible to a limited number of attackers. An attacker would
518 additionally need online access to an unpatched system using the target
519 private key in a scenario with persistent DH parameters and a private
520 key that is shared between multiple clients.
522 This only affects processors that support the BMI1, BMI2 and ADX extensions
523 like Intel Broadwell (5th generation) and later or AMD Ryzen.
525 This issue was reported to OpenSSL by the OSS-Fuzz project.
529 *) Malformed X.509 IPAddressFamily could cause OOB read
531 If an X.509 certificate has a malformed IPAddressFamily extension,
532 OpenSSL could do a one-byte buffer overread. The most likely result
533 would be an erroneous display of the certificate in text format.
535 This issue was reported to OpenSSL by the OSS-Fuzz project.
539 Changes between 1.1.0e and 1.1.0f [25 May 2017]
541 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
542 platform rather than 'mingw'.
545 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
546 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
547 which is the minimum version we support.
550 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
552 *) Encrypt-Then-Mac renegotiation crash
554 During a renegotiation handshake if the Encrypt-Then-Mac extension is
555 negotiated where it was not in the original handshake (or vice-versa) then
556 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
557 and servers are affected.
559 This issue was reported to OpenSSL by Joe Orton (Red Hat).
563 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
565 *) Truncated packet could crash via OOB read
567 If one side of an SSL/TLS path is running on a 32-bit host and a specific
568 cipher is being used, then a truncated packet can cause that host to
569 perform an out-of-bounds read, usually resulting in a crash.
571 This issue was reported to OpenSSL by Robert Święcki of Google.
575 *) Bad (EC)DHE parameters cause a client crash
577 If a malicious server supplies bad parameters for a DHE or ECDHE key
578 exchange then this can result in the client attempting to dereference a
579 NULL pointer leading to a client crash. This could be exploited in a Denial
582 This issue was reported to OpenSSL by Guido Vranken.
586 *) BN_mod_exp may produce incorrect results on x86_64
588 There is a carry propagating bug in the x86_64 Montgomery squaring
589 procedure. No EC algorithms are affected. Analysis suggests that attacks
590 against RSA and DSA as a result of this defect would be very difficult to
591 perform and are not believed likely. Attacks against DH are considered just
592 feasible (although very difficult) because most of the work necessary to
593 deduce information about a private key may be performed offline. The amount
594 of resources required for such an attack would be very significant and
595 likely only accessible to a limited number of attackers. An attacker would
596 additionally need online access to an unpatched system using the target
597 private key in a scenario with persistent DH parameters and a private
598 key that is shared between multiple clients. For example this can occur by
599 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
600 similar to CVE-2015-3193 but must be treated as a separate problem.
602 This issue was reported to OpenSSL by the OSS-Fuzz project.
606 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
608 *) ChaCha20/Poly1305 heap-buffer-overflow
610 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
611 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
612 crash. This issue is not considered to be exploitable beyond a DoS.
614 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
618 *) CMS Null dereference
620 Applications parsing invalid CMS structures can crash with a NULL pointer
621 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
622 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
623 structure callback if an attempt is made to free certain invalid encodings.
624 Only CHOICE structures using a callback which do not handle NULL value are
627 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
631 *) Montgomery multiplication may produce incorrect results
633 There is a carry propagating bug in the Broadwell-specific Montgomery
634 multiplication procedure that handles input lengths divisible by, but
635 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
636 and DH private keys are impossible. This is because the subroutine in
637 question is not used in operations with the private key itself and an input
638 of the attacker's direct choice. Otherwise the bug can manifest itself as
639 transient authentication and key negotiation failures or reproducible
640 erroneous outcome of public-key operations with specially crafted input.
641 Among EC algorithms only Brainpool P-512 curves are affected and one
642 presumably can attack ECDH key negotiation. Impact was not analyzed in
643 detail, because pre-requisites for attack are considered unlikely. Namely
644 multiple clients have to choose the curve in question and the server has to
645 share the private key among them, neither of which is default behaviour.
646 Even then only clients that chose the curve will be affected.
648 This issue was publicly reported as transient failures and was not
649 initially recognized as a security issue. Thanks to Richard Morgan for
650 providing reproducible case.
654 *) Removed automatic addition of RPATH in shared libraries and executables,
655 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
658 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
660 *) Fix Use After Free for large message sizes
662 The patch applied to address CVE-2016-6307 resulted in an issue where if a
663 message larger than approx 16k is received then the underlying buffer to
664 store the incoming message is reallocated and moved. Unfortunately a
665 dangling pointer to the old location is left which results in an attempt to
666 write to the previously freed location. This is likely to result in a
667 crash, however it could potentially lead to execution of arbitrary code.
669 This issue only affects OpenSSL 1.1.0a.
671 This issue was reported to OpenSSL by Robert Święcki.
675 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
677 *) OCSP Status Request extension unbounded memory growth
679 A malicious client can send an excessively large OCSP Status Request
680 extension. If that client continually requests renegotiation, sending a
681 large OCSP Status Request extension each time, then there will be unbounded
682 memory growth on the server. This will eventually lead to a Denial Of
683 Service attack through memory exhaustion. Servers with a default
684 configuration are vulnerable even if they do not support OCSP. Builds using
685 the "no-ocsp" build time option are not affected.
687 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
691 *) SSL_peek() hang on empty record
693 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
694 sends an empty record. This could be exploited by a malicious peer in a
695 Denial Of Service attack.
697 This issue was reported to OpenSSL by Alex Gaynor.
701 *) Excessive allocation of memory in tls_get_message_header() and
702 dtls1_preprocess_fragment()
704 A (D)TLS message includes 3 bytes for its length in the header for the
705 message. This would allow for messages up to 16Mb in length. Messages of
706 this length are excessive and OpenSSL includes a check to ensure that a
707 peer is sending reasonably sized messages in order to avoid too much memory
708 being consumed to service a connection. A flaw in the logic of version
709 1.1.0 means that memory for the message is allocated too early, prior to
710 the excessive message length check. Due to way memory is allocated in
711 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
712 to service a connection. This could lead to a Denial of Service through
713 memory exhaustion. However, the excessive message length check still takes
714 place, and this would cause the connection to immediately fail. Assuming
715 that the application calls SSL_free() on the failed connection in a timely
716 manner then the 21Mb of allocated memory will then be immediately freed
717 again. Therefore the excessive memory allocation will be transitory in
718 nature. This then means that there is only a security impact if:
720 1) The application does not call SSL_free() in a timely manner in the event
721 that the connection fails
723 2) The application is working in a constrained environment where there is
724 very little free memory
726 3) The attacker initiates multiple connection attempts such that there are
727 multiple connections in a state where memory has been allocated for the
728 connection; SSL_free() has not yet been called; and there is insufficient
729 memory to service the multiple requests.
731 Except in the instance of (1) above any Denial Of Service is likely to be
732 transitory because as soon as the connection fails the memory is
733 subsequently freed again in the SSL_free() call. However there is an
734 increased risk during this period of application crashes due to the lack of
735 memory - which would then mean a more serious Denial of Service.
737 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
738 (CVE-2016-6307 and CVE-2016-6308)
741 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
742 had to be removed. Primary reason is that vendor assembler can't
743 assemble our modules with -KPIC flag. As result it, assembly
744 support, was not even available as option. But its lack means
745 lack of side-channel resistant code, which is incompatible with
746 security by todays standards. Fortunately gcc is readily available
747 prepackaged option, which we firmly point at...
750 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
752 *) Windows command-line tool supports UTF-8 opt-in option for arguments
753 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
754 (to any value) allows Windows user to access PKCS#12 file generated
755 with Windows CryptoAPI and protected with non-ASCII password, as well
756 as files generated under UTF-8 locale on Linux also protected with
760 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
761 have been disabled by default and removed from DEFAULT, just like RC4.
762 See the RC4 item below to re-enable both.
765 *) The method for finding the storage location for the Windows RAND seed file
766 has changed. First we check %RANDFILE%. If that is not set then we check
767 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
768 all else fails we fall back to C:\.
771 *) The EVP_EncryptUpdate() function has had its return type changed from void
772 to int. A return of 0 indicates and error while a return of 1 indicates
776 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
777 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
778 off the constant time implementation for RSA, DSA and DH have been made
779 no-ops and deprecated.
782 *) Windows RAND implementation was simplified to only get entropy by
783 calling CryptGenRandom(). Various other RAND-related tickets
785 [Joseph Wylie Yandle, Rich Salz]
787 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
788 and OPENSSL_LH_, respectively. The old names are available
789 with API compatibility. They new names are now completely documented.
792 *) Unify TYPE_up_ref(obj) methods signature.
793 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
794 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
795 int (instead of void) like all others TYPE_up_ref() methods.
796 So now these methods also check the return value of CRYPTO_atomic_add(),
797 and the validity of object reference counter.
798 [fdasilvayy@gmail.com]
800 *) With Windows Visual Studio builds, the .pdb files are installed
801 alongside the installed libraries and executables. For a static
802 library installation, ossl_static.pdb is the associate compiler
803 generated .pdb file to be used when linking programs.
806 *) Remove openssl.spec. Packaging files belong with the packagers.
809 *) Automatic Darwin/OSX configuration has had a refresh, it will now
810 recognise x86_64 architectures automatically. You can still decide
811 to build for a different bitness with the environment variable
812 KERNEL_BITS (can be 32 or 64), for example:
814 KERNEL_BITS=32 ./config
818 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
819 256 bit AES and HMAC with SHA256.
822 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
825 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
828 *) To enable users to have their own config files and build file templates,
829 Configure looks in the directory indicated by the environment variable
830 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
831 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
832 name and is used as is.
835 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
836 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
837 X509_CERT_FILE_CTX was removed.
840 *) "shared" builds are now the default. To create only static libraries use
841 the "no-shared" Configure option.
844 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
845 All of these option have not worked for some while and are fundamental
849 *) Make various cleanup routines no-ops and mark them as deprecated. Most
850 global cleanup functions are no longer required because they are handled
851 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
852 Explicitly de-initing can cause problems (e.g. where a library that uses
853 OpenSSL de-inits, but an application is still using it). The affected
854 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
855 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
856 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
860 *) --strict-warnings no longer enables runtime debugging options
861 such as REF_DEBUG. Instead, debug options are automatically
862 enabled with '--debug' builds.
863 [Andy Polyakov, Emilia Käsper]
865 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
866 have been moved out of the public header files. New functions for managing
867 these have been added.
870 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
871 objects have been moved out of the public header files. New
872 functions for managing these have been added.
875 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
876 have been moved out of the public header files. New functions for managing
877 these have been added.
880 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
881 moved out of the public header files. New functions for managing these
885 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
888 *) Removed the mk1mf build scripts.
891 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
892 it is always safe to #include a header now.
895 *) Removed the aged BC-32 config and all its supporting scripts
898 *) Removed support for Ultrix, Netware, and OS/2.
901 *) Add support for HKDF.
904 *) Add support for blake2b and blake2s
907 *) Added support for "pipelining". Ciphers that have the
908 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
909 encryptions/decryptions simultaneously. There are currently no built-in
910 ciphers with this property but the expectation is that engines will be able
911 to offer it to significantly improve throughput. Support has been extended
912 into libssl so that multiple records for a single connection can be
913 processed in one go (for >=TLS 1.1).
916 *) Added the AFALG engine. This is an async capable engine which is able to
917 offload work to the Linux kernel. In this initial version it only supports
918 AES128-CBC. The kernel must be version 4.1.0 or greater.
921 *) OpenSSL now uses a new threading API. It is no longer necessary to
922 set locking callbacks to use OpenSSL in a multi-threaded environment. There
923 are two supported threading models: pthreads and windows threads. It is
924 also possible to configure OpenSSL at compile time for "no-threads". The
925 old threading API should no longer be used. The functions have been
926 replaced with "no-op" compatibility macros.
927 [Alessandro Ghedini, Matt Caswell]
929 *) Modify behavior of ALPN to invoke callback after SNI/servername
930 callback, such that updates to the SSL_CTX affect ALPN.
933 *) Add SSL_CIPHER queries for authentication and key-exchange.
936 *) Changes to the DEFAULT cipherlist:
937 - Prefer (EC)DHE handshakes over plain RSA.
938 - Prefer AEAD ciphers over legacy ciphers.
939 - Prefer ECDSA over RSA when both certificates are available.
940 - Prefer TLSv1.2 ciphers/PRF.
941 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
945 *) Change the ECC default curve list to be this, in order: x25519,
946 secp256r1, secp521r1, secp384r1.
949 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
950 disabled by default. They can be re-enabled using the
951 enable-weak-ssl-ciphers option to Configure.
954 *) If the server has ALPN configured, but supports no protocols that the
955 client advertises, send a fatal "no_application_protocol" alert.
956 This behaviour is SHALL in RFC 7301, though it isn't universally
957 implemented by other servers.
960 *) Add X25519 support.
961 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
962 for public and private key encoding using the format documented in
963 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
964 key generation and key derivation.
966 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
970 *) Deprecate SRP_VBASE_get_by_user.
971 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
972 In order to fix an unavoidable memory leak (CVE-2016-0798),
973 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
974 seed, even if the seed is configured.
976 Users should use SRP_VBASE_get1_by_user instead. Note that in
977 SRP_VBASE_get1_by_user, caller must free the returned value. Note
978 also that even though configuring the SRP seed attempts to hide
979 invalid usernames by continuing the handshake with fake
980 credentials, this behaviour is not constant time and no strong
981 guarantees are made that the handshake is indistinguishable from
982 that of a valid user.
985 *) Configuration change; it's now possible to build dynamic engines
986 without having to build shared libraries and vice versa. This
987 only applies to the engines in engines/, those in crypto/engine/
988 will always be built into libcrypto (i.e. "static").
990 Building dynamic engines is enabled by default; to disable, use
991 the configuration option "disable-dynamic-engine".
993 The only requirements for building dynamic engines are the
994 presence of the DSO module and building with position independent
995 code, so they will also automatically be disabled if configuring
996 with "disable-dso" or "disable-pic".
998 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
999 are also taken away from openssl/opensslconf.h, as they are
1003 *) Configuration change; if there is a known flag to compile
1004 position independent code, it will always be applied on the
1005 libcrypto and libssl object files, and never on the application
1006 object files. This means other libraries that use routines from
1007 libcrypto / libssl can be made into shared libraries regardless
1008 of how OpenSSL was configured.
1010 If this isn't desirable, the configuration options "disable-pic"
1011 or "no-pic" can be used to disable the use of PIC. This will
1012 also disable building shared libraries and dynamic engines.
1015 *) Removed JPAKE code. It was experimental and has no wide use.
1018 *) The INSTALL_PREFIX Makefile variable has been renamed to
1019 DESTDIR. That makes for less confusion on what this variable
1020 is for. Also, the configuration option --install_prefix is
1024 *) Heartbeat for TLS has been removed and is disabled by default
1025 for DTLS; configure with enable-heartbeats. Code that uses the
1026 old #define's might need to be updated.
1027 [Emilia Käsper, Rich Salz]
1029 *) Rename REF_CHECK to REF_DEBUG.
1032 *) New "unified" build system
1034 The "unified" build system is aimed to be a common system for all
1035 platforms we support. With it comes new support for VMS.
1037 This system builds supports building in a different directory tree
1038 than the source tree. It produces one Makefile (for unix family
1039 or lookalikes), or one descrip.mms (for VMS).
1041 The source of information to make the Makefile / descrip.mms is
1042 small files called 'build.info', holding the necessary
1043 information for each directory with source to compile, and a
1044 template in Configurations, like unix-Makefile.tmpl or
1047 With this change, the library names were also renamed on Windows
1048 and on VMS. They now have names that are closer to the standard
1049 on Unix, and include the major version number, and in certain
1050 cases, the architecture they are built for. See "Notes on shared
1051 libraries" in INSTALL.
1053 We rely heavily on the perl module Text::Template.
1056 *) Added support for auto-initialisation and de-initialisation of the library.
1057 OpenSSL no longer requires explicit init or deinit routines to be called,
1058 except in certain circumstances. See the OPENSSL_init_crypto() and
1059 OPENSSL_init_ssl() man pages for further information.
1062 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1063 "peer" argument is now expected to be a BIO_ADDR object.
1065 *) Rewrite of BIO networking library. The BIO library lacked consistent
1066 support of IPv6, and adding it required some more extensive
1067 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1068 which hold all types of addresses and chains of address information.
1069 It also introduces a new API, with functions like BIO_socket,
1070 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1071 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1072 have been adapted accordingly.
1075 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1079 *) CRIME protection: disable compression by default, even if OpenSSL is
1080 compiled with zlib enabled. Applications can still enable compression
1081 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1082 using the SSL_CONF library to configure compression.
1085 *) The signature of the session callback configured with
1086 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1087 was explicitly marked as 'const unsigned char*' instead of
1091 *) Always DPURIFY. Remove the use of uninitialized memory in the
1092 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1095 *) Removed many obsolete configuration items, including
1096 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1097 MD2_CHAR, MD2_INT, MD2_LONG
1099 IDEA_SHORT, IDEA_LONG
1100 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1101 [Rich Salz, with advice from Andy Polyakov]
1103 *) Many BN internals have been moved to an internal header file.
1104 [Rich Salz with help from Andy Polyakov]
1106 *) Configuration and writing out the results from it has changed.
1107 Files such as Makefile include/openssl/opensslconf.h and are now
1108 produced through general templates, such as Makefile.in and
1109 crypto/opensslconf.h.in and some help from the perl module
1112 Also, the center of configuration information is no longer
1113 Makefile. Instead, Configure produces a perl module in
1114 configdata.pm which holds most of the config data (in the hash
1115 table %config), the target data that comes from the target
1116 configuration in one of the Configurations/*.conf files (in
1120 *) To clarify their intended purposes, the Configure options
1121 --prefix and --openssldir change their semantics, and become more
1122 straightforward and less interdependent.
1124 --prefix shall be used exclusively to give the location INSTALLTOP
1125 where programs, scripts, libraries, include files and manuals are
1126 going to be installed. The default is now /usr/local.
1128 --openssldir shall be used exclusively to give the default
1129 location OPENSSLDIR where certificates, private keys, CRLs are
1130 managed. This is also where the default openssl.cnf gets
1132 If the directory given with this option is a relative path, the
1133 values of both the --prefix value and the --openssldir value will
1134 be combined to become OPENSSLDIR.
1135 The default for --openssldir is INSTALLTOP/ssl.
1137 Anyone who uses --openssldir to specify where OpenSSL is to be
1138 installed MUST change to use --prefix instead.
1141 *) The GOST engine was out of date and therefore it has been removed. An up
1142 to date GOST engine is now being maintained in an external repository.
1143 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1144 support for GOST ciphersuites (these are only activated if a GOST engine
1148 *) EGD is no longer supported by default; use enable-egd when
1150 [Ben Kaduk and Rich Salz]
1152 *) The distribution now has Makefile.in files, which are used to
1153 create Makefile's when Configure is run. *Configure must be run
1154 before trying to build now.*
1157 *) The return value for SSL_CIPHER_description() for error conditions
1161 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1163 Obtaining and performing DNSSEC validation of TLSA records is
1164 the application's responsibility. The application provides
1165 the TLSA records of its choice to OpenSSL, and these are then
1166 used to authenticate the peer.
1168 The TLSA records need not even come from DNS. They can, for
1169 example, be used to implement local end-entity certificate or
1170 trust-anchor "pinning", where the "pin" data takes the form
1171 of TLSA records, which can augment or replace verification
1172 based on the usual WebPKI public certification authorities.
1175 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1176 continues to support deprecated interfaces in default builds.
1177 However, applications are strongly advised to compile their
1178 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1179 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1180 or the 1.1.0 releases.
1182 In environments in which all applications have been ported to
1183 not use any deprecated interfaces OpenSSL's Configure script
1184 should be used with the --api=1.1.0 option to entirely remove
1185 support for the deprecated features from the library and
1186 unconditionally disable them in the installed headers.
1187 Essentially the same effect can be achieved with the "no-deprecated"
1188 argument to Configure, except that this will always restrict
1189 the build to just the latest API, rather than a fixed API
1192 As applications are ported to future revisions of the API,
1193 they should update their compile-time OPENSSL_API_COMPAT define
1194 accordingly, but in most cases should be able to continue to
1195 compile with later releases.
1197 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1198 0x10000000L and 0x00908000L, respectively. However those
1199 versions did not support the OPENSSL_API_COMPAT feature, and
1200 so applications are not typically tested for explicit support
1201 of just the undeprecated features of either release.
1204 *) Add support for setting the minimum and maximum supported protocol.
1205 It can bet set via the SSL_set_min_proto_version() and
1206 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1207 MaxProtocol. It's recommended to use the new APIs to disable
1208 protocols instead of disabling individual protocols using
1209 SSL_set_options() or SSL_CONF's Protocol. This change also
1210 removes support for disabling TLS 1.2 in the OpenSSL TLS
1211 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1214 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1217 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1218 and integrates ECDSA and ECDH functionality into EC. Implementations can
1219 now redirect key generation and no longer need to convert to or from
1222 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1223 include the ec.h header file instead.
1226 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1227 ciphers who are no longer supported and drops support the ephemeral RSA key
1228 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1231 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1232 opaque. For HMAC_CTX, the following constructors and destructors
1235 HMAC_CTX *HMAC_CTX_new(void);
1236 void HMAC_CTX_free(HMAC_CTX *ctx);
1238 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1239 destroy such methods has been added. See EVP_MD_meth_new(3) and
1240 EVP_CIPHER_meth_new(3) for documentation.
1243 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1244 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1245 EVP_MD_CTX_reset() should be called instead to reinitialise
1246 an already created structure.
1247 2) For consistency with the majority of our object creators and
1248 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1249 EVP_MD_CTX_(new|free). The old names are retained as macros
1250 for deprecated builds.
1253 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1254 cryptographic operations to be performed asynchronously as long as an
1255 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1256 further details. Libssl has also had this capability integrated with the
1257 introduction of the new mode SSL_MODE_ASYNC and associated error
1258 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1259 pages. This work was developed in partnership with Intel Corp.
1262 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1263 always enabled now. If you want to disable the support you should
1264 exclude it using the list of supported ciphers. This also means that the
1265 "-no_ecdhe" option has been removed from s_server.
1268 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1269 SSL_{CTX_}set1_curves() which can set a list.
1272 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1273 curve you want to support using SSL_{CTX_}set1_curves().
1276 *) State machine rewrite. The state machine code has been significantly
1277 refactored in order to remove much duplication of code and solve issues
1278 with the old code (see ssl/statem/README for further details). This change
1279 does have some associated API changes. Notably the SSL_state() function
1280 has been removed and replaced by SSL_get_state which now returns an
1281 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1282 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1286 *) All instances of the string "ssleay" in the public API were replaced
1287 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1288 Some error codes related to internal RSA_eay API's were renamed.
1291 *) The demo files in crypto/threads were moved to demo/threads.
1294 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1296 [Matt Caswell, Rich Salz]
1298 *) New ASN.1 embed macro.
1300 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1301 structure is not allocated: it is part of the parent. That is instead of
1309 This reduces memory fragmentation and make it impossible to accidentally
1310 set a mandatory field to NULL.
1312 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1313 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1314 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1318 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1321 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1322 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1323 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1324 DES and RC4 ciphersuites.
1327 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1328 This changes the decoding behaviour for some invalid messages,
1329 though the change is mostly in the more lenient direction, and
1330 legacy behaviour is preserved as much as possible.
1333 *) Fix no-stdio build.
1334 [ David Woodhouse <David.Woodhouse@intel.com> and also
1335 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1337 *) New testing framework
1338 The testing framework has been largely rewritten and is now using
1339 perl and the perl modules Test::Harness and an extended variant of
1340 Test::More called OpenSSL::Test to do its work. All test scripts in
1341 test/ have been rewritten into test recipes, and all direct calls to
1342 executables in test/Makefile have become individual recipes using the
1343 simplified testing OpenSSL::Test::Simple.
1345 For documentation on our testing modules, do:
1347 perldoc test/testlib/OpenSSL/Test/Simple.pm
1348 perldoc test/testlib/OpenSSL/Test.pm
1352 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1353 are used; the latter aborts on memory leaks (usually checked on exit).
1354 Some undocumented "set malloc, etc., hooks" functions were removed
1355 and others were changed. All are now documented.
1358 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1360 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1362 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1363 from RFC4279, RFC4785, RFC5487, RFC5489.
1365 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1366 original RSA_PSK patch.
1369 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1370 era flag was never set throughout the codebase (only read). Also removed
1371 SSL3_FLAGS_POP_BUFFER which was only used if
1372 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1375 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1376 to be "oneline" instead of "compat".
1379 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1380 not aware of clients that still exhibit this bug, and the workaround
1381 hasn't been working properly for a while.
1384 *) The return type of BIO_number_read() and BIO_number_written() as well as
1385 the corresponding num_read and num_write members in the BIO structure has
1386 changed from unsigned long to uint64_t. On platforms where an unsigned
1387 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1391 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1392 OpenSSL without support for them. It also means that maintaining
1393 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1394 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1397 *) Removed support for the two export grade static DH ciphersuites
1398 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1399 were newly added (along with a number of other static DH ciphersuites) to
1400 1.0.2. However the two export ones have *never* worked since they were
1401 introduced. It seems strange in any case to be adding new export
1402 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1405 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1406 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1407 and turned into macros which simply call the new preferred function names
1408 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1409 should use the new names instead. Also as part of this change the ssl23.h
1410 header file has been removed.
1413 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1414 code and the associated standard is no longer considered fit-for-purpose.
1417 *) RT2547 was closed. When generating a private key, try to make the
1418 output file readable only by the owner. This behavior change might
1419 be noticeable when interacting with other software.
1421 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1425 *) Added HTTP GET support to the ocsp command.
1428 *) Changed default digest for the dgst and enc commands from MD5 to
1432 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1435 *) Added support for TLS extended master secret from
1436 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1437 initial patch which was a great help during development.
1440 *) All libssl internal structures have been removed from the public header
1441 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1442 now redundant). Users should not attempt to access internal structures
1443 directly. Instead they should use the provided API functions.
1446 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1447 Access to deprecated functions can be re-enabled by running config with
1448 "enable-deprecated". In addition applications wishing to use deprecated
1449 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1450 will, by default, disable some transitive includes that previously existed
1451 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1454 *) Added support for OCB mode. OpenSSL has been granted a patent license
1455 compatible with the OpenSSL license for use of OCB. Details are available
1456 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1457 for OCB can be removed by calling config with no-ocb.
1460 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1461 compatible client hello.
1464 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1465 done while fixing the error code for the key-too-small case.
1466 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1468 *) CA.sh has been removed; use CA.pl instead.
1471 *) Removed old DES API.
1474 *) Remove various unsupported platforms:
1480 Sinix/ReliantUNIX RM400
1485 16-bit platforms such as WIN16
1488 *) Clean up OPENSSL_NO_xxx #define's
1489 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1490 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1491 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1492 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1493 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1494 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1495 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1496 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1497 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1498 Remove MS_STATIC; it's a relic from platforms <32 bits.
1501 *) Cleaned up dead code
1502 Remove all but one '#ifdef undef' which is to be looked at.
1505 *) Clean up calling of xxx_free routines.
1506 Just like free(), fix most of the xxx_free routines to accept
1507 NULL. Remove the non-null checks from callers. Save much code.
1510 *) Add secure heap for storage of private keys (when possible).
1511 Add BIO_s_secmem(), CBIGNUM, etc.
1512 Contributed by Akamai Technologies under our Corporate CLA.
1515 *) Experimental support for a new, fast, unbiased prime candidate generator,
1516 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1517 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1519 *) New output format NSS in the sess_id command line tool. This allows
1520 exporting the session id and the master key in NSS keylog format.
1521 [Martin Kaiser <martin@kaiser.cx>]
1523 *) Harmonize version and its documentation. -f flag is used to display
1525 [mancha <mancha1@zoho.com>]
1527 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1528 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1529 [mancha <mancha1@zoho.com>]
1531 *) Fix some double frees. These are not thought to be exploitable.
1532 [mancha <mancha1@zoho.com>]
1534 *) A missing bounds check in the handling of the TLS heartbeat extension
1535 can be used to reveal up to 64k of memory to a connected client or
1538 Thanks for Neel Mehta of Google Security for discovering this bug and to
1539 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1540 preparing the fix (CVE-2014-0160)
1541 [Adam Langley, Bodo Moeller]
1543 *) Fix for the attack described in the paper "Recovering OpenSSL
1544 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1545 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1546 http://eprint.iacr.org/2014/140
1548 Thanks to Yuval Yarom and Naomi Benger for discovering this
1549 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1550 [Yuval Yarom and Naomi Benger]
1552 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1553 this fixes a limitation in previous versions of OpenSSL.
1556 *) Experimental encrypt-then-mac support.
1558 Experimental support for encrypt then mac from
1559 draft-gutmann-tls-encrypt-then-mac-02.txt
1561 To enable it set the appropriate extension number (0x42 for the test
1562 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1564 For non-compliant peers (i.e. just about everything) this should have no
1567 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1571 *) Add EVP support for key wrapping algorithms, to avoid problems with
1572 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1573 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1574 algorithms and include tests cases.
1577 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1581 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1582 MGF1 digest and OAEP label.
1585 *) Make openssl verify return errors.
1586 [Chris Palmer <palmer@google.com> and Ben Laurie]
1588 *) New function ASN1_TIME_diff to calculate the difference between two
1589 ASN1_TIME structures or one structure and the current time.
1592 *) Update fips_test_suite to support multiple command line options. New
1593 test to induce all self test errors in sequence and check expected
1597 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1598 sign or verify all in one operation.
1601 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1602 test programs and fips_test_suite. Includes functionality to parse
1603 the minimal script output of fipsalgest.pl directly.
1606 *) Add authorisation parameter to FIPS_module_mode_set().
1609 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1612 *) Use separate DRBG fields for internal and external flags. New function
1613 FIPS_drbg_health_check() to perform on demand health checking. Add
1614 generation tests to fips_test_suite with reduced health check interval to
1615 demonstrate periodic health checking. Add "nodh" option to
1616 fips_test_suite to skip very slow DH test.
1619 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1623 *) More extensive health check for DRBG checking many more failure modes.
1624 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1625 combination: call this in fips_test_suite.
1628 *) Add support for canonical generation of DSA parameter 'g'. See
1631 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1632 POST to handle HMAC cases.
1635 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1636 to return numerical and string versions of the FIPS module number.
1639 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1640 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1641 outside the validated module in the FIPS capable OpenSSL.
1644 *) Minor change to DRBG entropy callback semantics. In some cases
1645 there is no multiple of the block length between min_len and
1646 max_len. Allow the callback to return more than max_len bytes
1647 of entropy but discard any extra: it is the callback's responsibility
1648 to ensure that the extra data discarded does not impact the
1649 requested amount of entropy.
1652 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1653 information in FIPS186-3, SP800-57 and SP800-131A.
1656 *) CCM support via EVP. Interface is very similar to GCM case except we
1657 must supply all data in one chunk (i.e. no update, final) and the
1658 message length must be supplied if AAD is used. Add algorithm test
1662 *) Initial version of POST overhaul. Add POST callback to allow the status
1663 of POST to be monitored and/or failures induced. Modify fips_test_suite
1664 to use callback. Always run all selftests even if one fails.
1667 *) XTS support including algorithm test driver in the fips_gcmtest program.
1668 Note: this does increase the maximum key length from 32 to 64 bytes but
1669 there should be no binary compatibility issues as existing applications
1670 will never use XTS mode.
1673 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1674 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1675 performs algorithm blocking for unapproved PRNG types. Also do not
1676 set PRNG type in FIPS_mode_set(): leave this to the application.
1677 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1678 the standard OpenSSL PRNG: set additional data to a date time vector.
1681 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1682 This shouldn't present any incompatibility problems because applications
1683 shouldn't be using these directly and any that are will need to rethink
1684 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1687 *) Extensive self tests and health checking required by SP800-90 DRBG.
1688 Remove strength parameter from FIPS_drbg_instantiate and always
1689 instantiate at maximum supported strength.
1692 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1695 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1698 *) New function DH_compute_key_padded() to compute a DH key and pad with
1699 leading zeroes if needed: this complies with SP800-56A et al.
1702 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1703 anything, incomplete, subject to change and largely untested at present.
1706 *) Modify fipscanisteronly build option to only build the necessary object
1707 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1710 *) Add experimental option FIPSSYMS to give all symbols in
1711 fipscanister.o and FIPS or fips prefix. This will avoid
1712 conflicts with future versions of OpenSSL. Add perl script
1713 util/fipsas.pl to preprocess assembly language source files
1714 and rename any affected symbols.
1717 *) Add selftest checks and algorithm block of non-fips algorithms in
1718 FIPS mode. Remove DES2 from selftests.
1721 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1722 return internal method without any ENGINE dependencies. Add new
1723 tiny fips sign and verify functions.
1726 *) New build option no-ec2m to disable characteristic 2 code.
1729 *) New build option "fipscanisteronly". This only builds fipscanister.o
1730 and (currently) associated fips utilities. Uses the file Makefile.fips
1731 instead of Makefile.org as the prototype.
1734 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1735 Update fips_gcmtest to use IV generator.
1738 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1739 setting output buffer to NULL. The *Final function must be
1740 called although it will not retrieve any additional data. The tag
1741 can be set or retrieved with a ctrl. The IV length is by default 12
1742 bytes (96 bits) but can be set to an alternative value. If the IV
1743 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1747 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1748 underlying do_cipher function handles all cipher semantics itself
1749 including padding and finalisation. This is useful if (for example)
1750 an ENGINE cipher handles block padding itself. The behaviour of
1751 do_cipher is subtly changed if this flag is set: the return value
1752 is the number of characters written to the output buffer (zero is
1753 no longer an error code) or a negative error code. Also if the
1754 input buffer is NULL and length 0 finalisation should be performed.
1757 *) If a candidate issuer certificate is already part of the constructed
1758 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1761 *) Improve forward-security support: add functions
1763 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1764 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1766 for use by SSL/TLS servers; the callback function will be called whenever a
1767 new session is created, and gets to decide whether the session may be
1768 cached to make it resumable (return 0) or not (return 1). (As by the
1769 SSL/TLS protocol specifications, the session_id sent by the server will be
1770 empty to indicate that the session is not resumable; also, the server will
1771 not generate RFC 4507 (RFC 5077) session tickets.)
1773 A simple reasonable callback implementation is to return is_forward_secure.
1774 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1775 by the SSL/TLS server library, indicating whether it can provide forward
1777 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1779 *) New -verify_name option in command line utilities to set verification
1783 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1784 Add CMAC pkey methods.
1787 *) Experimental renegotiation in s_server -www mode. If the client
1788 browses /reneg connection is renegotiated. If /renegcert it is
1789 renegotiated requesting a certificate.
1792 *) Add an "external" session cache for debugging purposes to s_server. This
1793 should help trace issues which normally are only apparent in deployed
1794 multi-process servers.
1797 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1798 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1799 BIO_set_cipher() and some obscure PEM functions were changed so they
1800 can now return an error. The RAND changes required a change to the
1801 RAND_METHOD structure.
1804 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1805 a gcc attribute to warn if the result of a function is ignored. This
1806 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1807 whose return value is often ignored.
1810 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1811 These allow SCTs (signed certificate timestamps) to be requested and
1812 validated when establishing a connection.
1813 [Rob Percival <robpercival@google.com>]
1815 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1817 *) Prevent padding oracle in AES-NI CBC MAC check
1819 A MITM attacker can use a padding oracle attack to decrypt traffic
1820 when the connection uses an AES CBC cipher and the server support
1823 This issue was introduced as part of the fix for Lucky 13 padding
1824 attack (CVE-2013-0169). The padding check was rewritten to be in
1825 constant time by making sure that always the same bytes are read and
1826 compared against either the MAC or padding bytes. But it no longer
1827 checked that there was enough data to have both the MAC and padding
1830 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1834 *) Fix EVP_EncodeUpdate overflow
1836 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1837 Base64 encoding of binary data. If an attacker is able to supply very large
1838 amounts of input data then a length check can overflow resulting in a heap
1841 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1842 the PEM_write_bio* family of functions. These are mainly used within the
1843 OpenSSL command line applications, so any application which processes data
1844 from an untrusted source and outputs it as a PEM file should be considered
1845 vulnerable to this issue. User applications that call these APIs directly
1846 with large amounts of untrusted data may also be vulnerable.
1848 This issue was reported by Guido Vranken.
1852 *) Fix EVP_EncryptUpdate overflow
1854 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1855 is able to supply very large amounts of input data after a previous call to
1856 EVP_EncryptUpdate() with a partial block then a length check can overflow
1857 resulting in a heap corruption. Following an analysis of all OpenSSL
1858 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1859 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1860 the first called function after an EVP_EncryptInit(), and therefore that
1861 specific call must be safe. The second form is where the length passed to
1862 EVP_EncryptUpdate() can be seen from the code to be some small value and
1863 therefore there is no possibility of an overflow. Since all instances are
1864 one of these two forms, it is believed that there can be no overflows in
1865 internal code due to this problem. It should be noted that
1866 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1867 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1868 of these calls have also been analysed too and it is believed there are no
1869 instances in internal usage where an overflow could occur.
1871 This issue was reported by Guido Vranken.
1875 *) Prevent ASN.1 BIO excessive memory allocation
1877 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1878 a short invalid encoding can cause allocation of large amounts of memory
1879 potentially consuming excessive resources or exhausting memory.
1881 Any application parsing untrusted data through d2i BIO functions is
1882 affected. The memory based functions such as d2i_X509() are *not* affected.
1883 Since the memory based functions are used by the TLS library, TLS
1884 applications are not affected.
1886 This issue was reported by Brian Carpenter.
1892 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1893 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1894 in arbitrary stack data being returned in the buffer.
1896 This issue was reported by Guido Vranken.
1900 *) Modify behavior of ALPN to invoke callback after SNI/servername
1901 callback, such that updates to the SSL_CTX affect ALPN.
1904 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1908 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1909 methods are enabled and ssl2 is disabled the methods return NULL.
1912 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1914 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1915 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1916 provide any "EXPORT" or "LOW" strength ciphers.
1919 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1920 is by default disabled at build-time. Builds that are not configured with
1921 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1922 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1923 will need to explicitly call either of:
1925 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1927 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1929 as appropriate. Even if either of those is used, or the application
1930 explicitly uses the version-specific SSLv2_method() or its client and
1931 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1932 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1933 ciphers, and SSLv2 56-bit DES are no longer available.
1937 *) Fix a double-free in DSA code
1939 A double free bug was discovered when OpenSSL parses malformed DSA private
1940 keys and could lead to a DoS attack or memory corruption for applications
1941 that receive DSA private keys from untrusted sources. This scenario is
1944 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1949 *) Disable SRP fake user seed to address a server memory leak.
1951 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1953 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1954 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1955 was changed to ignore the "fake user" SRP seed, even if the seed
1958 Users should use SRP_VBASE_get1_by_user instead. Note that in
1959 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1960 also that even though configuring the SRP seed attempts to hide
1961 invalid usernames by continuing the handshake with fake
1962 credentials, this behaviour is not constant time and no strong
1963 guarantees are made that the handshake is indistinguishable from
1964 that of a valid user.
1968 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1970 In the BN_hex2bn function the number of hex digits is calculated using an
1971 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1972 large values of |i| this can result in |bn_expand| not allocating any
1973 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1974 field as NULL leading to a subsequent NULL ptr deref. For very large values
1975 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1976 In this case memory is allocated to the internal BIGNUM data field, but it
1977 is insufficiently sized leading to heap corruption. A similar issue exists
1978 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1979 is ever called by user applications with very large untrusted hex/dec data.
1980 This is anticipated to be a rare occurrence.
1982 All OpenSSL internal usage of these functions use data that is not expected
1983 to be untrusted, e.g. config file data or application command line
1984 arguments. If user developed applications generate config file data based
1985 on untrusted data then it is possible that this could also lead to security
1986 consequences. This is also anticipated to be rare.
1988 This issue was reported to OpenSSL by Guido Vranken.
1992 *) Fix memory issues in BIO_*printf functions
1994 The internal |fmtstr| function used in processing a "%s" format string in
1995 the BIO_*printf functions could overflow while calculating the length of a
1996 string and cause an OOB read when printing very long strings.
1998 Additionally the internal |doapr_outch| function can attempt to write to an
1999 OOB memory location (at an offset from the NULL pointer) in the event of a
2000 memory allocation failure. In 1.0.2 and below this could be caused where
2001 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2002 could be in processing a very long "%s" format string. Memory leaks can
2005 The first issue may mask the second issue dependent on compiler behaviour.
2006 These problems could enable attacks where large amounts of untrusted data
2007 is passed to the BIO_*printf functions. If applications use these functions
2008 in this way then they could be vulnerable. OpenSSL itself uses these
2009 functions when printing out human-readable dumps of ASN.1 data. Therefore
2010 applications that print this data could be vulnerable if the data is from
2011 untrusted sources. OpenSSL command line applications could also be
2012 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2013 as command line arguments.
2015 Libssl is not considered directly vulnerable. Additionally certificates etc
2016 received via remote connections via libssl are also unlikely to be able to
2017 trigger these issues because of message size limits enforced within libssl.
2019 This issue was reported to OpenSSL Guido Vranken.
2023 *) Side channel attack on modular exponentiation
2025 A side-channel attack was found which makes use of cache-bank conflicts on
2026 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2027 of RSA keys. The ability to exploit this issue is limited as it relies on
2028 an attacker who has control of code in a thread running on the same
2029 hyper-threaded core as the victim thread which is performing decryptions.
2031 This issue was reported to OpenSSL by Yuval Yarom, The University of
2032 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2033 Nadia Heninger, University of Pennsylvania with more information at
2034 http://cachebleed.info.
2038 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2039 if no keysize is specified with default_bits. This fixes an
2040 omission in an earlier change that changed all RSA/DSA key generation
2041 apps to use 2048 bits by default.
2044 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2045 *) DH small subgroups
2047 Historically OpenSSL only ever generated DH parameters based on "safe"
2048 primes. More recently (in version 1.0.2) support was provided for
2049 generating X9.42 style parameter files such as those required for RFC 5114
2050 support. The primes used in such files may not be "safe". Where an
2051 application is using DH configured with parameters based on primes that are
2052 not "safe" then an attacker could use this fact to find a peer's private
2053 DH exponent. This attack requires that the attacker complete multiple
2054 handshakes in which the peer uses the same private DH exponent. For example
2055 this could be used to discover a TLS server's private DH exponent if it's
2056 reusing the private DH exponent or it's using a static DH ciphersuite.
2058 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2059 TLS. It is not on by default. If the option is not set then the server
2060 reuses the same private DH exponent for the life of the server process and
2061 would be vulnerable to this attack. It is believed that many popular
2062 applications do set this option and would therefore not be at risk.
2064 The fix for this issue adds an additional check where a "q" parameter is
2065 available (as is the case in X9.42 based parameters). This detects the
2066 only known attack, and is the only possible defense for static DH
2067 ciphersuites. This could have some performance impact.
2069 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2070 default and cannot be disabled. This could have some performance impact.
2072 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2076 *) SSLv2 doesn't block disabled ciphers
2078 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2079 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2080 been disabled, provided that the SSLv2 protocol was not also disabled via
2083 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2084 and Sebastian Schinzel.
2088 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2090 *) BN_mod_exp may produce incorrect results on x86_64
2092 There is a carry propagating bug in the x86_64 Montgomery squaring
2093 procedure. No EC algorithms are affected. Analysis suggests that attacks
2094 against RSA and DSA as a result of this defect would be very difficult to
2095 perform and are not believed likely. Attacks against DH are considered just
2096 feasible (although very difficult) because most of the work necessary to
2097 deduce information about a private key may be performed offline. The amount
2098 of resources required for such an attack would be very significant and
2099 likely only accessible to a limited number of attackers. An attacker would
2100 additionally need online access to an unpatched system using the target
2101 private key in a scenario with persistent DH parameters and a private
2102 key that is shared between multiple clients. For example this can occur by
2103 default in OpenSSL DHE based SSL/TLS ciphersuites.
2105 This issue was reported to OpenSSL by Hanno Böck.
2109 *) Certificate verify crash with missing PSS parameter
2111 The signature verification routines will crash with a NULL pointer
2112 dereference if presented with an ASN.1 signature using the RSA PSS
2113 algorithm and absent mask generation function parameter. Since these
2114 routines are used to verify certificate signature algorithms this can be
2115 used to crash any certificate verification operation and exploited in a
2116 DoS attack. Any application which performs certificate verification is
2117 vulnerable including OpenSSL clients and servers which enable client
2120 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2124 *) X509_ATTRIBUTE memory leak
2126 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2127 memory. This structure is used by the PKCS#7 and CMS routines so any
2128 application which reads PKCS#7 or CMS data from untrusted sources is
2129 affected. SSL/TLS is not affected.
2131 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2136 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2137 This changes the decoding behaviour for some invalid messages,
2138 though the change is mostly in the more lenient direction, and
2139 legacy behaviour is preserved as much as possible.
2142 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2144 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2146 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2148 *) Alternate chains certificate forgery
2150 During certificate verification, OpenSSL will attempt to find an
2151 alternative certificate chain if the first attempt to build such a chain
2152 fails. An error in the implementation of this logic can mean that an
2153 attacker could cause certain checks on untrusted certificates to be
2154 bypassed, such as the CA flag, enabling them to use a valid leaf
2155 certificate to act as a CA and "issue" an invalid certificate.
2157 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2161 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2163 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2164 incompatibility in the handling of HMAC. The previous ABI has now been
2168 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2170 *) Malformed ECParameters causes infinite loop
2172 When processing an ECParameters structure OpenSSL enters an infinite loop
2173 if the curve specified is over a specially malformed binary polynomial
2176 This can be used to perform denial of service against any
2177 system which processes public keys, certificate requests or
2178 certificates. This includes TLS clients and TLS servers with
2179 client authentication enabled.
2181 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2185 *) Exploitable out-of-bounds read in X509_cmp_time
2187 X509_cmp_time does not properly check the length of the ASN1_TIME
2188 string and can read a few bytes out of bounds. In addition,
2189 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2192 An attacker can use this to craft malformed certificates and CRLs of
2193 various sizes and potentially cause a segmentation fault, resulting in
2194 a DoS on applications that verify certificates or CRLs. TLS clients
2195 that verify CRLs are affected. TLS clients and servers with client
2196 authentication enabled may be affected if they use custom verification
2199 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2200 independently by Hanno Böck.
2204 *) PKCS7 crash with missing EnvelopedContent
2206 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2207 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2208 with missing content and trigger a NULL pointer dereference on parsing.
2210 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2211 structures from untrusted sources are affected. OpenSSL clients and
2212 servers are not affected.
2214 This issue was reported to OpenSSL by Michal Zalewski (Google).
2218 *) CMS verify infinite loop with unknown hash function
2220 When verifying a signedData message the CMS code can enter an infinite loop
2221 if presented with an unknown hash function OID. This can be used to perform
2222 denial of service against any system which verifies signedData messages using
2224 This issue was reported to OpenSSL by Johannes Bauer.
2228 *) Race condition handling NewSessionTicket
2230 If a NewSessionTicket is received by a multi-threaded client when attempting to
2231 reuse a previous ticket then a race condition can occur potentially leading to
2232 a double free of the ticket data.
2236 *) Only support 256-bit or stronger elliptic curves with the
2237 'ecdh_auto' setting (server) or by default (client). Of supported
2238 curves, prefer P-256 (both).
2241 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2243 *) ClientHello sigalgs DoS fix
2245 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2246 invalid signature algorithms extension a NULL pointer dereference will
2247 occur. This can be exploited in a DoS attack against the server.
2249 This issue was was reported to OpenSSL by David Ramos of Stanford
2252 [Stephen Henson and Matt Caswell]
2254 *) Multiblock corrupted pointer fix
2256 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2257 feature only applies on 64 bit x86 architecture platforms that support AES
2258 NI instructions. A defect in the implementation of "multiblock" can cause
2259 OpenSSL's internal write buffer to become incorrectly set to NULL when
2260 using non-blocking IO. Typically, when the user application is using a
2261 socket BIO for writing, this will only result in a failed connection.
2262 However if some other BIO is used then it is likely that a segmentation
2263 fault will be triggered, thus enabling a potential DoS attack.
2265 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2269 *) Segmentation fault in DTLSv1_listen fix
2271 The DTLSv1_listen function is intended to be stateless and processes the
2272 initial ClientHello from many peers. It is common for user code to loop
2273 over the call to DTLSv1_listen until a valid ClientHello is received with
2274 an associated cookie. A defect in the implementation of DTLSv1_listen means
2275 that state is preserved in the SSL object from one invocation to the next
2276 that can lead to a segmentation fault. Errors processing the initial
2277 ClientHello can trigger this scenario. An example of such an error could be
2278 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2281 This issue was reported to OpenSSL by Per Allansson.
2285 *) Segmentation fault in ASN1_TYPE_cmp fix
2287 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2288 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2289 certificate signature algorithm consistency this can be used to crash any
2290 certificate verification operation and exploited in a DoS attack. Any
2291 application which performs certificate verification is vulnerable including
2292 OpenSSL clients and servers which enable client authentication.
2296 *) Segmentation fault for invalid PSS parameters fix
2298 The signature verification routines will crash with a NULL pointer
2299 dereference if presented with an ASN.1 signature using the RSA PSS
2300 algorithm and invalid parameters. Since these routines are used to verify
2301 certificate signature algorithms this can be used to crash any
2302 certificate verification operation and exploited in a DoS attack. Any
2303 application which performs certificate verification is vulnerable including
2304 OpenSSL clients and servers which enable client authentication.
2306 This issue was was reported to OpenSSL by Brian Carpenter.
2310 *) ASN.1 structure reuse memory corruption fix
2312 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2313 memory corruption via an invalid write. Such reuse is and has been
2314 strongly discouraged and is believed to be rare.
2316 Applications that parse structures containing CHOICE or ANY DEFINED BY
2317 components may be affected. Certificate parsing (d2i_X509 and related
2318 functions) are however not affected. OpenSSL clients and servers are
2323 *) PKCS7 NULL pointer dereferences fix
2325 The PKCS#7 parsing code does not handle missing outer ContentInfo
2326 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2327 missing content and trigger a NULL pointer dereference on parsing.
2329 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2330 otherwise parse PKCS#7 structures from untrusted sources are
2331 affected. OpenSSL clients and servers are not affected.
2333 This issue was reported to OpenSSL by Michal Zalewski (Google).
2337 *) DoS via reachable assert in SSLv2 servers fix
2339 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2340 servers that both support SSLv2 and enable export cipher suites by sending
2341 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2343 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2344 (OpenSSL development team).
2348 *) Empty CKE with client auth and DHE fix
2350 If client auth is used then a server can seg fault in the event of a DHE
2351 ciphersuite being selected and a zero length ClientKeyExchange message
2352 being sent by the client. This could be exploited in a DoS attack.
2356 *) Handshake with unseeded PRNG fix
2358 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2359 with an unseeded PRNG. The conditions are:
2360 - The client is on a platform where the PRNG has not been seeded
2361 automatically, and the user has not seeded manually
2362 - A protocol specific client method version has been used (i.e. not
2363 SSL_client_methodv23)
2364 - A ciphersuite is used that does not require additional random data from
2365 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2367 If the handshake succeeds then the client random that has been used will
2368 have been generated from a PRNG with insufficient entropy and therefore the
2369 output may be predictable.
2371 For example using the following command with an unseeded openssl will
2372 succeed on an unpatched platform:
2374 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2378 *) Use After Free following d2i_ECPrivatekey error fix
2380 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2381 could cause a use after free condition. This, in turn, could cause a double
2382 free in several private key parsing functions (such as d2i_PrivateKey
2383 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2384 for applications that receive EC private keys from untrusted
2385 sources. This scenario is considered rare.
2387 This issue was discovered by the BoringSSL project and fixed in their
2392 *) X509_to_X509_REQ NULL pointer deref fix
2394 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2395 the certificate key is invalid. This function is rarely used in practice.
2397 This issue was discovered by Brian Carpenter.
2401 *) Removed the export ciphers from the DEFAULT ciphers
2404 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2406 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2407 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2408 So far those who have to target multiple platforms would compromise
2409 and argue that binary targeting say ARMv5 would still execute on
2410 ARMv8. "Universal" build resolves this compromise by providing
2411 near-optimal performance even on newer platforms.
2414 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2415 (other platforms pending).
2416 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2418 *) Add support for the SignedCertificateTimestampList certificate and
2419 OCSP response extensions from RFC6962.
2422 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2423 for corner cases. (Certain input points at infinity could lead to
2424 bogus results, with non-infinity inputs mapped to infinity too.)
2427 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2428 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2429 common cases are optimized and there still is room for further
2430 improvements. Vector Permutation AES for Altivec is also added.
2433 *) Add support for little-endian ppc64 Linux target.
2434 [Marcelo Cerri (IBM)]
2436 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2437 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2438 are optimized and there still is room for further improvements.
2439 Both 32- and 64-bit modes are supported.
2440 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2442 *) Improved ARMv7 NEON support.
2445 *) Support for SPARC Architecture 2011 crypto extensions, first
2446 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2447 SHA256/512, MD5, GHASH and modular exponentiation.
2448 [Andy Polyakov, David Miller]
2450 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2452 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2454 *) Support for new and upcoming Intel processors, including AVX2,
2455 BMI and SHA ISA extensions. This includes additional "stitched"
2456 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2459 This work was sponsored by Intel Corp.
2462 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2463 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2464 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2467 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2468 this fixes a limitation in previous versions of OpenSSL.
2471 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2472 MGF1 digest and OAEP label.
2475 *) Add EVP support for key wrapping algorithms, to avoid problems with
2476 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2477 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2478 algorithms and include tests cases.
2481 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2483 [Douglas E. Engert, Steve Henson]
2485 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2486 difference in days and seconds between two tm or ASN1_TIME structures.
2489 *) Add -rev test option to s_server to just reverse order of characters
2490 received by client and send back to server. Also prints an abbreviated
2491 summary of the connection parameters.
2494 *) New option -brief for s_client and s_server to print out a brief summary
2495 of connection parameters.
2498 *) Add callbacks for arbitrary TLS extensions.
2499 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2501 *) New option -crl_download in several openssl utilities to download CRLs
2502 from CRLDP extension in certificates.
2505 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2508 *) New function X509_CRL_diff to generate a delta CRL from the difference
2509 of two full CRLs. Add support to "crl" utility.
2512 *) New functions to set lookup_crls function and to retrieve
2513 X509_STORE from X509_STORE_CTX.
2516 *) Print out deprecated issuer and subject unique ID fields in
2520 *) Extend OCSP I/O functions so they can be used for simple general purpose
2521 HTTP as well as OCSP. New wrapper function which can be used to download
2522 CRLs using the OCSP API.
2525 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2528 *) SSL_CONF* functions. These provide a common framework for application
2529 configuration using configuration files or command lines.
2532 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2533 message callback and prints the results. Needs compile time option
2534 "enable-ssl-trace". New options to s_client and s_server to enable
2538 *) New ctrl and macro to retrieve supported points extensions.
2539 Print out extension in s_server and s_client.
2542 *) New functions to retrieve certificate signature and signature
2546 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2550 *) New Suite B modes for TLS code. These use and enforce the requirements
2551 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2552 only use Suite B curves. The Suite B modes can be set by using the
2553 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2556 *) New chain verification flags for Suite B levels of security. Check
2557 algorithms are acceptable when flags are set in X509_verify_cert.
2560 *) Make tls1_check_chain return a set of flags indicating checks passed
2561 by a certificate chain. Add additional tests to handle client
2562 certificates: checks for matching certificate type and issuer name
2566 *) If an attempt is made to use a signature algorithm not in the peer
2567 preference list abort the handshake. If client has no suitable
2568 signature algorithms in response to a certificate request do not
2569 use the certificate.
2572 *) If server EC tmp key is not in client preference list abort handshake.
2575 *) Add support for certificate stores in CERT structure. This makes it
2576 possible to have different stores per SSL structure or one store in
2577 the parent SSL_CTX. Include distinct stores for certificate chain
2578 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2579 to build and store a certificate chain in CERT structure: returning
2580 an error if the chain cannot be built: this will allow applications
2581 to test if a chain is correctly configured.
2583 Note: if the CERT based stores are not set then the parent SSL_CTX
2584 store is used to retain compatibility with existing behaviour.
2588 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2589 mask based on the current session, check mask when sending client
2590 hello and checking the requested ciphersuite.
2593 *) New ctrls to retrieve and set certificate types in a certificate
2594 request message. Print out received values in s_client. If certificate
2595 types is not set with custom values set sensible values based on
2596 supported signature algorithms.
2599 *) Support for distinct client and server supported signature algorithms.
2602 *) Add certificate callback. If set this is called whenever a certificate
2603 is required by client or server. An application can decide which
2604 certificate chain to present based on arbitrary criteria: for example
2605 supported signature algorithms. Add very simple example to s_server.
2606 This fixes many of the problems and restrictions of the existing client
2607 certificate callback: for example you can now clear an existing
2608 certificate and specify the whole chain.
2611 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2612 the certificate can be used for (if anything). Set valid_flags field
2613 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2614 to have similar checks in it.
2616 Add new "cert_flags" field to CERT structure and include a "strict mode".
2617 This enforces some TLS certificate requirements (such as only permitting
2618 certificate signature algorithms contained in the supported algorithms
2619 extension) which some implementations ignore: this option should be used
2620 with caution as it could cause interoperability issues.
2623 *) Update and tidy signature algorithm extension processing. Work out
2624 shared signature algorithms based on preferences and peer algorithms
2625 and print them out in s_client and s_server. Abort handshake if no
2626 shared signature algorithms.
2629 *) Add new functions to allow customised supported signature algorithms
2630 for SSL and SSL_CTX structures. Add options to s_client and s_server
2634 *) New function SSL_certs_clear() to delete all references to certificates
2635 from an SSL structure. Before this once a certificate had been added
2636 it couldn't be removed.
2639 *) Integrate hostname, email address and IP address checking with certificate
2640 verification. New verify options supporting checking in openssl utility.
2643 *) Fixes and wildcard matching support to hostname and email checking
2644 functions. Add manual page.
2645 [Florian Weimer (Red Hat Product Security Team)]
2647 *) New functions to check a hostname email or IP address against a
2648 certificate. Add options x509 utility to print results of checks against
2652 *) Fix OCSP checking.
2653 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2655 *) Initial experimental support for explicitly trusted non-root CAs.
2656 OpenSSL still tries to build a complete chain to a root but if an
2657 intermediate CA has a trust setting included that is used. The first
2658 setting is used: whether to trust (e.g., -addtrust option to the x509
2662 *) Add -trusted_first option which attempts to find certificates in the
2663 trusted store even if an untrusted chain is also supplied.
2666 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2667 platform support for Linux and Android.
2670 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2673 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2674 When in FIPS mode the approved implementations are used as normal,
2675 when not in FIPS mode the internal unapproved versions are used instead.
2676 This means that the FIPS capable OpenSSL isn't forced to use the
2677 (often lower performance) FIPS implementations outside FIPS mode.
2680 *) Transparently support X9.42 DH parameters when calling
2681 PEM_read_bio_DHparameters. This means existing applications can handle
2682 the new parameter format automatically.
2685 *) Initial experimental support for X9.42 DH parameter format: mainly
2686 to support use of 'q' parameter for RFC5114 parameters.
2689 *) Add DH parameters from RFC5114 including test data to dhtest.
2692 *) Support for automatic EC temporary key parameter selection. If enabled
2693 the most preferred EC parameters are automatically used instead of
2694 hardcoded fixed parameters. Now a server just has to call:
2695 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2696 support ECDH and use the most appropriate parameters.
2699 *) Enhance and tidy EC curve and point format TLS extension code. Use
2700 static structures instead of allocation if default values are used.
2701 New ctrls to set curves we wish to support and to retrieve shared curves.
2702 Print out shared curves in s_server. New options to s_server and s_client
2703 to set list of supported curves.
2706 *) New ctrls to retrieve supported signature algorithms and
2707 supported curve values as an array of NIDs. Extend openssl utility
2708 to print out received values.
2711 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2712 between NIDs and the more common NIST names such as "P-256". Enhance
2713 ecparam utility and ECC method to recognise the NIST names for curves.
2716 *) Enhance SSL/TLS certificate chain handling to support different
2717 chains for each certificate instead of one chain in the parent SSL_CTX.
2720 *) Support for fixed DH ciphersuite client authentication: where both
2721 server and client use DH certificates with common parameters.
2724 *) Support for fixed DH ciphersuites: those requiring DH server
2728 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2730 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2731 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2732 X509_CINF_get_signature were reverted post internal team review.
2734 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2736 *) Build fixes for the Windows and OpenVMS platforms
2737 [Matt Caswell and Richard Levitte]
2739 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2741 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2742 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2743 dereference. This could lead to a Denial Of Service attack. Thanks to
2744 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2748 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2749 dtls1_buffer_record function under certain conditions. In particular this
2750 could occur if an attacker sent repeated DTLS records with the same
2751 sequence number but for the next epoch. The memory leak could be exploited
2752 by an attacker in a Denial of Service attack through memory exhaustion.
2753 Thanks to Chris Mueller for reporting this issue.
2757 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2758 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2759 method would be set to NULL which could later result in a NULL pointer
2760 dereference. Thanks to Frank Schmirler for reporting this issue.
2764 *) Abort handshake if server key exchange message is omitted for ephemeral
2767 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2768 reporting this issue.
2772 *) Remove non-export ephemeral RSA code on client and server. This code
2773 violated the TLS standard by allowing the use of temporary RSA keys in
2774 non-export ciphersuites and could be used by a server to effectively
2775 downgrade the RSA key length used to a value smaller than the server
2776 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2777 INRIA or reporting this issue.
2781 *) Fixed issue where DH client certificates are accepted without verification.
2782 An OpenSSL server will accept a DH certificate for client authentication
2783 without the certificate verify message. This effectively allows a client to
2784 authenticate without the use of a private key. This only affects servers
2785 which trust a client certificate authority which issues certificates
2786 containing DH keys: these are extremely rare and hardly ever encountered.
2787 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2792 *) Ensure that the session ID context of an SSL is updated when its
2793 SSL_CTX is updated via SSL_set_SSL_CTX.
2795 The session ID context is typically set from the parent SSL_CTX,
2796 and can vary with the CTX.
2799 *) Fix various certificate fingerprint issues.
2801 By using non-DER or invalid encodings outside the signed portion of a
2802 certificate the fingerprint can be changed without breaking the signature.
2803 Although no details of the signed portion of the certificate can be changed
2804 this can cause problems with some applications: e.g. those using the
2805 certificate fingerprint for blacklists.
2807 1. Reject signatures with non zero unused bits.
2809 If the BIT STRING containing the signature has non zero unused bits reject
2810 the signature. All current signature algorithms require zero unused bits.
2812 2. Check certificate algorithm consistency.
2814 Check the AlgorithmIdentifier inside TBS matches the one in the
2815 certificate signature. NB: this will result in signature failure
2816 errors for some broken certificates.
2818 Thanks to Konrad Kraszewski from Google for reporting this issue.
2820 3. Check DSA/ECDSA signatures use DER.
2822 Re-encode DSA/ECDSA signatures and compare with the original received
2823 signature. Return an error if there is a mismatch.
2825 This will reject various cases including garbage after signature
2826 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2827 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2828 (negative or with leading zeroes).
2830 Further analysis was conducted and fixes were developed by Stephen Henson
2831 of the OpenSSL core team.
2836 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2837 results on some platforms, including x86_64. This bug occurs at random
2838 with a very low probability, and is not known to be exploitable in any
2839 way, though its exact impact is difficult to determine. Thanks to Pieter
2840 Wuille (Blockstream) who reported this issue and also suggested an initial
2841 fix. Further analysis was conducted by the OpenSSL development team and
2842 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2843 the OpenSSL core team.
2847 *) Do not resume sessions on the server if the negotiated protocol
2848 version does not match the session's version. Resuming with a different
2849 version, while not strictly forbidden by the RFC, is of questionable
2850 sanity and breaks all known clients.
2851 [David Benjamin, Emilia Käsper]
2853 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2854 early CCS messages during renegotiation. (Note that because
2855 renegotiation is encrypted, this early CCS was not exploitable.)
2858 *) Tighten client-side session ticket handling during renegotiation:
2859 ensure that the client only accepts a session ticket if the server sends
2860 the extension anew in the ServerHello. Previously, a TLS client would
2861 reuse the old extension state and thus accept a session ticket if one was
2862 announced in the initial ServerHello.
2864 Similarly, ensure that the client requires a session ticket if one
2865 was advertised in the ServerHello. Previously, a TLS client would
2866 ignore a missing NewSessionTicket message.
2869 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2871 *) SRTP Memory Leak.
2873 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2874 sends a carefully crafted handshake message, to cause OpenSSL to fail
2875 to free up to 64k of memory causing a memory leak. This could be
2876 exploited in a Denial Of Service attack. This issue affects OpenSSL
2877 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2878 whether SRTP is used or configured. Implementations of OpenSSL that
2879 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2881 The fix was developed by the OpenSSL team.
2885 *) Session Ticket Memory Leak.
2887 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2888 integrity of that ticket is first verified. In the event of a session
2889 ticket integrity check failing, OpenSSL will fail to free memory
2890 causing a memory leak. By sending a large number of invalid session
2891 tickets an attacker could exploit this issue in a Denial Of Service
2896 *) Build option no-ssl3 is incomplete.
2898 When OpenSSL is configured with "no-ssl3" as a build option, servers
2899 could accept and complete a SSL 3.0 handshake, and clients could be
2900 configured to send them.
2902 [Akamai and the OpenSSL team]
2904 *) Add support for TLS_FALLBACK_SCSV.
2905 Client applications doing fallback retries should call
2906 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2908 [Adam Langley, Bodo Moeller]
2910 *) Add additional DigestInfo checks.
2912 Re-encode DigestInto in DER and check against the original when
2913 verifying RSA signature: this will reject any improperly encoded
2914 DigestInfo structures.
2916 Note: this is a precautionary measure and no attacks are currently known.
2920 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2922 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2923 SRP code can be overrun an internal buffer. Add sanity check that
2924 g, A, B < N to SRP code.
2926 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2927 Group for discovering this issue.
2931 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2932 TLS 1.0 instead of higher protocol versions when the ClientHello message
2933 is badly fragmented. This allows a man-in-the-middle attacker to force a
2934 downgrade to TLS 1.0 even if both the server and the client support a
2935 higher protocol version, by modifying the client's TLS records.
2937 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2938 researching this issue.
2942 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2943 to a denial of service attack. A malicious server can crash the client
2944 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2945 ciphersuite and sending carefully crafted handshake messages.
2947 Thanks to Felix Gröbert (Google) for discovering and researching this
2952 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2953 to leak memory. This can be exploited through a Denial of Service attack.
2954 Thanks to Adam Langley for discovering and researching this issue.
2958 *) An attacker can force openssl to consume large amounts of memory whilst
2959 processing DTLS handshake messages. This can be exploited through a
2960 Denial of Service attack.
2961 Thanks to Adam Langley for discovering and researching this issue.
2965 *) An attacker can force an error condition which causes openssl to crash
2966 whilst processing DTLS packets due to memory being freed twice. This
2967 can be exploited through a Denial of Service attack.
2968 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2973 *) If a multithreaded client connects to a malicious server using a resumed
2974 session and the server sends an ec point format extension it could write
2975 up to 255 bytes to freed memory.
2977 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2982 *) A malicious server can crash an OpenSSL client with a null pointer
2983 dereference (read) by specifying an SRP ciphersuite even though it was not
2984 properly negotiated with the client. This can be exploited through a
2985 Denial of Service attack.
2987 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2988 discovering and researching this issue.
2992 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2993 X509_name_oneline, X509_name_print_ex et al. to leak some information
2994 from the stack. Applications may be affected if they echo pretty printing
2995 output to the attacker.
2997 Thanks to Ivan Fratric (Google) for discovering this issue.
2999 [Emilia Käsper, and Steve Henson]
3001 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3002 for corner cases. (Certain input points at infinity could lead to
3003 bogus results, with non-infinity inputs mapped to infinity too.)
3006 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3008 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3009 handshake can force the use of weak keying material in OpenSSL
3010 SSL/TLS clients and servers.
3012 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3013 researching this issue. (CVE-2014-0224)
3014 [KIKUCHI Masashi, Steve Henson]
3016 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3017 OpenSSL DTLS client the code can be made to recurse eventually crashing
3020 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3022 [Imre Rad, Steve Henson]
3024 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3025 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3026 client or server. This is potentially exploitable to run arbitrary
3027 code on a vulnerable client or server.
3029 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3030 [Jüri Aedla, Steve Henson]
3032 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3033 are subject to a denial of service attack.
3035 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3036 this issue. (CVE-2014-3470)
3037 [Felix Gröbert, Ivan Fratric, Steve Henson]
3039 *) Harmonize version and its documentation. -f flag is used to display
3041 [mancha <mancha1@zoho.com>]
3043 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3044 in i2d_ECPrivateKey.
3045 [mancha <mancha1@zoho.com>]
3047 *) Fix some double frees. These are not thought to be exploitable.
3048 [mancha <mancha1@zoho.com>]
3050 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3052 *) A missing bounds check in the handling of the TLS heartbeat extension
3053 can be used to reveal up to 64k of memory to a connected client or
3056 Thanks for Neel Mehta of Google Security for discovering this bug and to
3057 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3058 preparing the fix (CVE-2014-0160)
3059 [Adam Langley, Bodo Moeller]
3061 *) Fix for the attack described in the paper "Recovering OpenSSL
3062 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3063 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3064 http://eprint.iacr.org/2014/140
3066 Thanks to Yuval Yarom and Naomi Benger for discovering this
3067 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3068 [Yuval Yarom and Naomi Benger]
3070 *) TLS pad extension: draft-agl-tls-padding-03
3072 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3073 TLS client Hello record length value would otherwise be > 255 and
3074 less that 512 pad with a dummy extension containing zeroes so it
3075 is at least 512 bytes long.
3077 [Adam Langley, Steve Henson]
3079 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3081 *) Fix for TLS record tampering bug. A carefully crafted invalid
3082 handshake could crash OpenSSL with a NULL pointer exception.
3083 Thanks to Anton Johansson for reporting this issues.
3086 *) Keep original DTLS digest and encryption contexts in retransmission
3087 structures so we can use the previous session parameters if they need
3088 to be resent. (CVE-2013-6450)
3091 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3092 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3093 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3094 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3095 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3096 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3097 [Rob Stradling, Adam Langley]
3099 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3101 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3102 supporting platforms or when small records were transferred.
3103 [Andy Polyakov, Steve Henson]
3105 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3107 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3109 This addresses the flaw in CBC record processing discovered by
3110 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3111 at: http://www.isg.rhul.ac.uk/tls/
3113 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3114 Security Group at Royal Holloway, University of London
3115 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3116 Emilia Käsper for the initial patch.
3118 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3120 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3121 ciphersuites which can be exploited in a denial of service attack.
3122 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3123 and detecting this bug and to Wolfgang Ettlinger
3124 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3128 *) Return an error when checking OCSP signatures when key is NULL.
3129 This fixes a DoS attack. (CVE-2013-0166)
3132 *) Make openssl verify return errors.
3133 [Chris Palmer <palmer@google.com> and Ben Laurie]
3135 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3136 the right response is stapled. Also change SSL_get_certificate()
3137 so it returns the certificate actually sent.
3138 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3139 [Rob Stradling <rob.stradling@comodo.com>]
3141 *) Fix possible deadlock when decoding public keys.
3144 *) Don't use TLS 1.0 record version number in initial client hello
3148 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3150 *) Sanity check record length before skipping explicit IV in TLS
3151 1.2, 1.1 and DTLS to fix DoS attack.
3153 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3154 fuzzing as a service testing platform.
3158 *) Initialise tkeylen properly when encrypting CMS messages.
3159 Thanks to Solar Designer of Openwall for reporting this issue.
3162 *) In FIPS mode don't try to use composite ciphers as they are not
3166 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3168 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3169 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3170 mean any application compiled against OpenSSL 1.0.0 headers setting
3171 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3172 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3173 0x10000000L Any application which was previously compiled against
3174 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3175 will need to be recompiled as a result. Letting be results in
3176 inability to disable specifically TLS 1.1 and in client context,
3177 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3180 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3181 disable just protocol X, but all protocols above X *if* there are
3182 protocols *below* X still enabled. In more practical terms it means
3183 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3184 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3185 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3189 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3191 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3192 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3193 in CRYPTO_realloc_clean.
3195 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3196 issue and to Adam Langley <agl@chromium.org> for fixing it.
3198 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3200 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3203 *) Workarounds for some broken servers that "hang" if a client hello
3204 record length exceeds 255 bytes.
3206 1. Do not use record version number > TLS 1.0 in initial client
3207 hello: some (but not all) hanging servers will now work.
3208 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3209 the number of ciphers sent in the client hello. This should be
3210 set to an even number, such as 50, for example by passing:
3211 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3212 Most broken servers should now work.
3213 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3214 TLS 1.2 client support entirely.
3217 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3220 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3222 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3223 STRING form instead of a DigestInfo.
3226 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3227 and the RSA_sign/RSA_verify functions. This was made more apparent when
3228 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3229 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3230 the correct format in RSA_verify so both forms transparently work.
3233 *) Some servers which support TLS 1.0 can choke if we initially indicate
3234 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3235 encrypted premaster secret. As a workaround use the maximum permitted
3236 client version in client hello, this should keep such servers happy
3237 and still work with previous versions of OpenSSL.
3240 *) Add support for TLS/DTLS heartbeats.
3241 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3243 *) Add support for SCTP.
3244 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3246 *) Improved PRNG seeding for VOS.
3247 [Paul Green <Paul.Green@stratus.com>]
3249 *) Extensive assembler packs updates, most notably:
3251 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3252 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3253 - x86_64: bit-sliced AES implementation;
3254 - ARM: NEON support, contemporary platforms optimizations;
3255 - s390x: z196 support;
3256 - *: GHASH and GF(2^m) multiplication implementations;
3260 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3261 (removal of unnecessary code)
3262 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3264 *) Add TLS key material exporter from RFC 5705.
3267 *) Add DTLS-SRTP negotiation from RFC 5764.
3270 *) Add Next Protocol Negotiation,
3271 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3272 disabled with a no-npn flag to config or Configure. Code donated
3274 [Adam Langley <agl@google.com> and Ben Laurie]
3276 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3277 NIST-P256, NIST-P521, with constant-time single point multiplication on
3278 typical inputs. Compiler support for the nonstandard type __uint128_t is
3279 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3280 Code made available under Apache License version 2.0.
3282 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3283 line to include this in your build of OpenSSL, and run "make depend" (or
3284 "make update"). This enables the following EC_METHODs:
3286 EC_GFp_nistp224_method()
3287 EC_GFp_nistp256_method()
3288 EC_GFp_nistp521_method()
3290 EC_GROUP_new_by_curve_name() will automatically use these (while
3291 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3293 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3295 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3296 all platforms. Move ssize_t definition from e_os.h to the public
3297 header file e_os2.h as it now appears in public header file cms.h
3300 *) New -sigopt option to the ca, req and x509 utilities. Additional
3301 signature parameters can be passed using this option and in
3305 *) Add RSA PSS signing function. This will generate and set the
3306 appropriate AlgorithmIdentifiers for PSS based on those in the
3307 corresponding EVP_MD_CTX structure. No application support yet.
3310 *) Support for companion algorithm specific ASN1 signing routines.
3311 New function ASN1_item_sign_ctx() signs a pre-initialised
3312 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3313 the appropriate parameters.
3316 *) Add new algorithm specific ASN1 verification initialisation function
3317 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3318 handling will be the same no matter what EVP_PKEY_METHOD is used.
3319 Add a PSS handler to support verification of PSS signatures: checked
3320 against a number of sample certificates.
3323 *) Add signature printing for PSS. Add PSS OIDs.
3324 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3326 *) Add algorithm specific signature printing. An individual ASN1 method
3327 can now print out signatures instead of the standard hex dump.
3329 More complex signatures (e.g. PSS) can print out more meaningful
3330 information. Include DSA version that prints out the signature
3334 *) Password based recipient info support for CMS library: implementing
3338 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3339 neatly separates the code into cipher and PBE sections and is required
3340 for some algorithms that split PBES2 into separate pieces (such as
3341 password based CMS).
3344 *) Session-handling fixes:
3345 - Fix handling of connections that are resuming with a session ID,
3346 but also support Session Tickets.
3347 - Fix a bug that suppressed issuing of a new ticket if the client
3348 presented a ticket with an expired session.
3349 - Try to set the ticket lifetime hint to something reasonable.
3350 - Make tickets shorter by excluding irrelevant information.
3351 - On the client side, don't ignore renewed tickets.
3352 [Adam Langley, Bodo Moeller (Google)]
3354 *) Fix PSK session representation.
3357 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3359 This work was sponsored by Intel.
3362 *) Add GCM support to TLS library. Some custom code is needed to split
3363 the IV between the fixed (from PRF) and explicit (from TLS record)
3364 portions. This adds all GCM ciphersuites supported by RFC5288 and
3365 RFC5289. Generalise some AES* cipherstrings to include GCM and
3366 add a special AESGCM string for GCM only.
3369 *) Expand range of ctrls for AES GCM. Permit setting invocation
3370 field on decrypt and retrieval of invocation field only on encrypt.
3373 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3374 As required by RFC5289 these ciphersuites cannot be used if for
3375 versions of TLS earlier than 1.2.
3378 *) For FIPS capable OpenSSL interpret a NULL default public key method
3379 as unset and return the appropriate default but do *not* set the default.
3380 This means we can return the appropriate method in applications that
3381 switch between FIPS and non-FIPS modes.
3384 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3385 ENGINE is used then we cannot handle that in the FIPS module so we
3386 keep original code iff non-FIPS operations are allowed.
3389 *) Add -attime option to openssl utilities.
3390 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3392 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3395 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3396 FIPS EC methods unconditionally for now.
3399 *) New build option no-ec2m to disable characteristic 2 code.
3402 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3403 all cases can be covered as some introduce binary incompatibilities.
3406 *) Redirect RSA operations to FIPS module including keygen,
3407 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3410 *) Add similar low level API blocking to ciphers.
3413 *) Low level digest APIs are not approved in FIPS mode: any attempt
3414 to use these will cause a fatal error. Applications that *really* want
3415 to use them can use the private_* version instead.
3418 *) Redirect cipher operations to FIPS module for FIPS builds.
3421 *) Redirect digest operations to FIPS module for FIPS builds.
3424 *) Update build system to add "fips" flag which will link in fipscanister.o
3425 for static and shared library builds embedding a signature if needed.
3428 *) Output TLS supported curves in preference order instead of numerical
3429 order. This is currently hardcoded for the highest order curves first.
3430 This should be configurable so applications can judge speed vs strength.
3433 *) Add TLS v1.2 server support for client authentication.
3436 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3440 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3441 FIPS modules versions.
3444 *) Add TLS v1.2 client side support for client authentication. Keep cache
3445 of handshake records longer as we don't know the hash algorithm to use
3446 until after the certificate request message is received.
3449 *) Initial TLS v1.2 client support. Add a default signature algorithms
3450 extension including all the algorithms we support. Parse new signature
3451 format in client key exchange. Relax some ECC signing restrictions for
3452 TLS v1.2 as indicated in RFC5246.
3455 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3456 to new signature format when needed using client digest preference.
3457 All server ciphersuites should now work correctly in TLS v1.2. No client
3458 support yet and no support for client certificates.
3461 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3462 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3463 ciphersuites. At present only RSA key exchange ciphersuites work with
3464 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3465 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3466 and version checking.
3469 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3470 with this defined it will not be affected by any changes to ssl internal
3471 structures. Add several utility functions to allow openssl application
3472 to work with OPENSSL_NO_SSL_INTERN defined.
3475 *) A long standing patch to add support for SRP from EdelWeb (Peter
3476 Sylvester and Christophe Renou) was integrated.
3477 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3478 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3481 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3484 *) Permit abbreviated handshakes when renegotiating using the function
3485 SSL_renegotiate_abbreviated().
3486 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3488 *) Add call to ENGINE_register_all_complete() to
3489 ENGINE_load_builtin_engines(), so some implementations get used
3490 automatically instead of needing explicit application support.
3493 *) Add support for TLS key exporter as described in RFC5705.
3494 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3496 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3497 a few changes are required:
3499 Add SSL_OP_NO_TLSv1_1 flag.
3500 Add TLSv1_1 methods.
3501 Update version checking logic to handle version 1.1.
3502 Add explicit IV handling (ported from DTLS code).
3503 Add command line options to s_client/s_server.
3506 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3508 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3509 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3510 content decryption and always return the same error. Note: this attack
3511 needs on average 2^20 messages so it only affects automated senders. The
3512 old behaviour can be re-enabled in the CMS code by setting the
3513 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3514 an MMA defence is not necessary.
3515 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3516 this issue. (CVE-2012-0884)
3519 *) Fix CVE-2011-4619: make sure we really are receiving a
3520 client hello before rejecting multiple SGC restarts. Thanks to
3521 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3524 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3526 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3527 Thanks to Antonio Martin, Enterprise Secure Access Research and
3528 Development, Cisco Systems, Inc. for discovering this bug and
3529 preparing a fix. (CVE-2012-0050)
3532 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3534 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3535 of the Vaudenay padding oracle attack on CBC mode encryption
3536 which enables an efficient plaintext recovery attack against
3537 the OpenSSL implementation of DTLS. Their attack exploits timing
3538 differences arising during decryption processing. A research
3539 paper describing this attack can be found at:
3540 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3541 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3542 Security Group at Royal Holloway, University of London
3543 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3544 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3545 for preparing the fix. (CVE-2011-4108)
3546 [Robin Seggelmann, Michael Tuexen]
3548 *) Clear bytes used for block padding of SSL 3.0 records.
3550 [Adam Langley (Google)]
3552 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3553 Kadianakis <desnacked@gmail.com> for discovering this issue and
3554 Adam Langley for preparing the fix. (CVE-2011-4619)
3555 [Adam Langley (Google)]
3557 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3558 [Andrey Kulikov <amdeich@gmail.com>]
3560 *) Prevent malformed RFC3779 data triggering an assertion failure.
3561 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3562 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3563 [Rob Austein <sra@hactrn.net>]
3565 *) Improved PRNG seeding for VOS.
3566 [Paul Green <Paul.Green@stratus.com>]
3568 *) Fix ssl_ciph.c set-up race.
3569 [Adam Langley (Google)]
3571 *) Fix spurious failures in ecdsatest.c.
3572 [Emilia Käsper (Google)]
3574 *) Fix the BIO_f_buffer() implementation (which was mixing different
3575 interpretations of the '..._len' fields).
3576 [Adam Langley (Google)]
3578 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3579 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3580 threads won't reuse the same blinding coefficients.
3582 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3583 lock to call BN_BLINDING_invert_ex, and avoids one use of
3584 BN_BLINDING_update for each BN_BLINDING structure (previously,
3585 the last update always remained unused).
3586 [Emilia Käsper (Google)]
3588 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3589 [Bob Buckholz (Google)]
3591 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3593 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3594 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3595 [Kaspar Brand <ossl@velox.ch>]
3597 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3598 for multi-threaded use of ECDH. (CVE-2011-3210)
3599 [Adam Langley (Google)]
3601 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3604 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3605 signature public key algorithm by using OID xref utilities instead.
3606 Before this you could only use some ECC ciphersuites with SHA1 only.
3609 *) Add protection against ECDSA timing attacks as mentioned in the paper
3610 by Billy Bob Brumley and Nicola Tuveri, see:
3612 http://eprint.iacr.org/2011/232.pdf
3614 [Billy Bob Brumley and Nicola Tuveri]
3616 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3618 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3619 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3621 *) Fix bug in string printing code: if *any* escaping is enabled we must
3622 escape the escape character (backslash) or the resulting string is
3626 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3628 *) Disable code workaround for ancient and obsolete Netscape browsers
3629 and servers: an attacker can use it in a ciphersuite downgrade attack.
3630 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3633 *) Fixed J-PAKE implementation error, originally discovered by
3634 Sebastien Martini, further info and confirmation from Stefan
3635 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3638 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3640 *) Fix extension code to avoid race conditions which can result in a buffer
3641 overrun vulnerability: resumed sessions must not be modified as they can
3642 be shared by multiple threads. CVE-2010-3864
3645 *) Fix WIN32 build system to correctly link an ENGINE directory into
3649 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3651 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3653 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3655 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3657 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3658 context. The operation can be customised via the ctrl mechanism in
3659 case ENGINEs want to include additional functionality.
3662 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3665 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3666 output hashes compatible with older versions of OpenSSL.
3667 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3669 *) Fix compression algorithm handling: if resuming a session use the
3670 compression algorithm of the resumed session instead of determining
3671 it from client hello again. Don't allow server to change algorithm.
3674 *) Add load_crls() function to apps tidying load_certs() too. Add option
3675 to verify utility to allow additional CRLs to be included.
3678 *) Update OCSP request code to permit adding custom headers to the request:
3679 some responders need this.
3682 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3684 [Julia Lawall <julia@diku.dk>]
3686 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3687 needlessly dereferenced structures, used obsolete functions and
3688 didn't handle all updated verify codes correctly.
3691 *) Disable MD2 in the default configuration.
3694 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3695 indicate the initial BIO being pushed or popped. This makes it possible
3696 to determine whether the BIO is the one explicitly called or as a result
3697 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3698 it handles reference counts correctly and doesn't zero out the I/O bio
3699 when it is not being explicitly popped. WARNING: applications which
3700 included workarounds for the old buggy behaviour will need to be modified
3701 or they could free up already freed BIOs.
3704 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3705 renaming to all platforms (within the 0.9.8 branch, this was
3706 done conditionally on Netware platforms to avoid a name clash).
3707 [Guenter <lists@gknw.net>]
3709 *) Add ECDHE and PSK support to DTLS.
3710 [Michael Tuexen <tuexen@fh-muenster.de>]
3712 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3716 *) Add "missing" function EVP_MD_flags() (without this the only way to
3717 retrieve a digest flags is by accessing the structure directly. Update
3718 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3719 or cipher is registered as in the "from" argument. Print out all
projects / openssl.git / blob