5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
10 Changes between 1.1.0h and 1.1.1 [xx XXX xxxx]
11 *) Add blinding to an ECDSA signature to protect against side channel attacks
12 discovered by Keegan Ryan (NCC Group).
15 *) Enforce checking in the pkeyutl command line app to ensure that the input
16 length does not exceed the maximum supported digest length when performing
17 a sign, verify or verifyrecover operation.
20 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
21 I/O in combination with something like select() or poll() will hang. This
22 can be turned off again using SSL_CTX_clear_mode().
23 Many applications do not properly handle non-application data records, and
24 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
25 around the problems in those applications, but can also break some.
26 It's recommended to read the manpages about SSL_read(), SSL_write(),
27 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
28 SSL_CTX_set_read_ahead() again.
31 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
32 now allow empty (zero character) pass phrases.
35 *) Apply blinding to binary field modular inversion and remove patent
36 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
39 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
40 binary and prime elliptic curves.
43 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
44 constant time fixed point multiplication.
47 *) Revise elliptic curve scalar multiplication with timing attack
48 defenses: ec_wNAF_mul redirects to a constant time implementation
49 when computing fixed point and variable point multiplication (which
50 in OpenSSL are mostly used with secret scalars in keygen, sign,
51 ECDH derive operations).
52 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
55 *) Updated CONTRIBUTING
58 *) Updated DRBG / RAND to request nonce and additional low entropy
59 randomness from the system.
62 *) Updated 'openssl rehash' to use OpenSSL consistent default.
65 *) Moved the load of the ssl_conf module to libcrypto, which helps
66 loading engines that libssl uses before libssl is initialised.
69 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
72 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
73 [Ingo Schwarze, Rich Salz]
75 *) Added output of accepting IP address and port for 'openssl s_server'
78 *) Added a new API for TLSv1.3 ciphersuites:
79 SSL_CTX_set_ciphersuites()
80 SSL_set_ciphersuites()
83 *) Memory allocation failures consistenly add an error to the error
87 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
88 in libcrypto when run as setuid/setgid.
91 *) Load any config file by default when libssl is used.
94 *) Added new public header file <openssl/rand_drbg.h> and documentation
95 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
98 *) QNX support removed (cannot find contributors to get their approval
99 for the license change).
102 *) TLSv1.3 replay protection for early data has been implemented. See the
103 SSL_read_early_data() man page for further details.
106 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
107 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
108 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
109 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
110 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
111 configuration has been separated out. See the ciphers man page or the
112 SSL_CTX_set_ciphersuites() man page for more information.
115 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
116 in responder mode now supports the new "-multi" option, which
117 spawns the specified number of child processes to handle OCSP
118 requests. The "-timeout" option now also limits the OCSP
119 responder's patience to wait to receive the full client request
120 on a newly accepted connection. Child processes are respawned
121 as needed, and the CA index file is automatically reloaded
122 when changed. This makes it possible to run the "ocsp" responder
123 as a long-running service, making the OpenSSL CA somewhat more
124 feature-complete. In this mode, most diagnostic messages logged
125 after entering the event loop are logged via syslog(3) rather than
129 *) Added support for X448 and Ed448. Heavily based on original work by
133 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
134 objects loaded. This adds the functions OSSL_STORE_expect() and
135 OSSL_STORE_find() as well as needed tools to construct searches and
136 get the search data out of them.
139 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
140 version of OpenSSL should review their configuration settings to ensure
141 that they are still appropriate for TLSv1.3. For further information see:
142 https://www.openssl.org/blog/blog/2018/02/08/tlsv1.3/
144 NOTE: In this pre-release of OpenSSL a draft version of the
145 TLSv1.3 standard has been implemented. Implementations of different draft
146 versions of the standard do not inter-operate, and this version will not
147 inter-operate with an implementation of the final standard when it is
148 eventually published. Different pre-release versions may implement
149 different versions of the draft. The final version of OpenSSL 1.1.1 will
150 implement the final version of the standard.
151 TODO(TLS1.3): Remove the above note before final release
154 *) Grand redesign of the OpenSSL random generator
156 The default RAND method now utilizes an AES-CTR DRBG according to
157 NIST standard SP 800-90Ar1. The new random generator is essentially
158 a port of the default random generator from the OpenSSL FIPS 2.0
159 object module. It is a hybrid deterministic random bit generator
160 using an AES-CTR bit stream and which seeds and reseeds itself
161 automatically using trusted system entropy sources.
163 Some of its new features are:
164 o Support for multiple DRBG instances with seed chaining.
165 o Add a public DRBG instance for the default RAND method.
166 o Add a dedicated DRBG instance for generating long term private keys.
167 o Make the DRBG instances fork-safe.
168 o Keep all global DRBG instances on the secure heap if it is enabled.
169 o Add a DRBG instance to every SSL instance for lock free operation
170 and to increase unpredictability.
171 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
173 *) Changed Configure so it only says what it does and doesn't dump
174 so much data. Instead, ./configdata.pm should be used as a script
175 to display all sorts of configuration data.
178 *) Added processing of "make variables" to Configure.
181 *) Added SHA512/224 and SHA512/256 algorithm support.
184 *) The last traces of Netware support, first removed in 1.1.0, have
188 *) Get rid of Makefile.shared, and in the process, make the processing
189 of certain files (rc.obj, or the .def/.map/.opt files produced from
190 the ordinal files) more visible and hopefully easier to trace and
191 debug (or make silent).
194 *) Make it possible to have environment variable assignments as
195 arguments to config / Configure.
198 *) Add multi-prime RSA (RFC 8017) support.
201 *) Add SM3 implemented according to GB/T 32905-2016
202 [ Jack Lloyd <jack.lloyd@ribose.com>,
203 Ronald Tse <ronald.tse@ribose.com>,
204 Erick Borsboom <erick.borsboom@ribose.com> ]
206 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
207 as documented in RFC6066.
208 Based on a patch from Tomasz Moń
209 [Filipe Raimundo da Silva]
211 *) Add SM4 implemented according to GB/T 32907-2016.
212 [ Jack Lloyd <jack.lloyd@ribose.com>,
213 Ronald Tse <ronald.tse@ribose.com>,
214 Erick Borsboom <erick.borsboom@ribose.com> ]
216 *) Reimplement -newreq-nodes and ERR_error_string_n; the
217 original author does not agree with the license change.
220 *) Add ARIA AEAD TLS support.
223 *) Some macro definitions to support VS6 have been removed. Visual
224 Studio 6 has not worked since 1.1.0
227 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
228 without clearing the errors.
231 *) Add "atfork" functions. If building on a system that without
232 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
233 requirements. The RAND facility now uses/requires this.
239 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
240 not possible to disable entirely. However, it's still possible to
241 disable the console reading UI method, UI_OpenSSL() (use UI_null()
244 To disable, configure with 'no-ui-console'. 'no-ui' is still
245 possible to use as an alias. Check at compile time with the
246 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
247 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
250 *) Add a STORE module, which implements a uniform and URI based reader of
251 stores that can contain keys, certificates, CRLs and numerous other
252 objects. The main API is loosely based on a few stdio functions,
253 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
254 OSSL_STORE_error and OSSL_STORE_close.
255 The implementation uses backends called "loaders" to implement arbitrary
256 URI schemes. There is one built in "loader" for the 'file' scheme.
259 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
260 then adjusted to work on FreeBSD 8.4 as well.
261 Enable by configuring with 'enable-devcryptoeng'. This is done by default
262 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
265 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
266 util/mkerr.pl, which is adapted to allow those prefixes, leading to
267 error code calls like this:
269 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
271 With this change, we claim the namespaces OSSL and OPENSSL in a manner
272 that can be encoded in C. For the foreseeable future, this will only
274 [Richard Levitte and Tim Hudson]
276 *) Removed BSD cryptodev engine.
279 *) Add a build target 'build_all_generated', to build all generated files
280 and only that. This can be used to prepare everything that requires
281 things like perl for a system that lacks perl and then move everything
282 to that system and do the rest of the build there.
285 *) In the UI interface, make it possible to duplicate the user data. This
286 can be used by engines that need to retain the data for a longer time
287 than just the call where this user data is passed.
290 *) Ignore the '-named_curve auto' value for compatibility of applications
292 [Tomas Mraz <tmraz@fedoraproject.org>]
294 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
295 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
296 alerts across multiple records (some of which could be empty). In practice
297 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
298 prohibts this altogether and other libraries (BoringSSL, NSS) do not
299 support this at all. Supporting it adds significant complexity to the
300 record layer, and its removal is unlikely to cause inter-operability
304 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
305 with Z. These are meant to replace LONG and ZLONG and to be size safe.
306 The use of LONG and ZLONG is discouraged and scheduled for deprecation
310 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
311 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
312 [Richard Levitte, Andy Polyakov]
314 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
318 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
319 platform rather than 'mingw'.
322 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
323 success if they are asked to add an object which already exists
324 in the store. This change cascades to other functions which load
325 certificates and CRLs.
328 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
329 facilitate stack unwinding even from assembly subroutines.
332 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
333 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
336 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
337 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
338 which is the minimum version we support.
341 *) Certificate time validation (X509_cmp_time) enforces stricter
342 compliance with RFC 5280. Fractional seconds and timezone offsets
343 are no longer allowed.
346 *) Add support for ARIA
349 *) s_client will now send the Server Name Indication (SNI) extension by
350 default unless the new "-noservername" option is used. The server name is
351 based on the host provided to the "-connect" option unless overridden by
355 *) Add support for SipHash
358 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
359 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
360 prevent issues where no progress is being made and the peer continually
361 sends unrecognised record types, using up resources processing them.
364 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
365 using the algorithm defined in
366 https://www.akkadia.org/drepper/SHA-crypt.txt
369 *) Heartbeat support has been removed; the ABI is changed for now.
370 [Richard Levitte, Rich Salz]
372 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
375 *) The RSA "null" method, which was partially supported to avoid patent
376 issues, has been replaced to always returns NULL.
380 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
382 *) Fixed a text canonicalisation bug in CMS
384 Where a CMS detached signature is used with text content the text goes
385 through a canonicalisation process first prior to signing or verifying a
386 signature. This process strips trailing space at the end of lines, converts
387 line terminators to CRLF and removes additional trailing line terminators
388 at the end of a file. A bug in the canonicalisation process meant that
389 some characters, such as form-feed, were incorrectly treated as whitespace
390 and removed. This is contrary to the specification (RFC5485). This fix
391 could mean that detached text data signed with an earlier version of
392 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
393 signed with a fixed OpenSSL may fail to verify with an earlier version of
394 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
395 and use the "-binary" flag (for the "cms" command line application) or set
396 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
399 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
401 *) Constructed ASN.1 types with a recursive definition could exceed the stack
403 Constructed ASN.1 types with a recursive definition (such as can be found
404 in PKCS7) could eventually exceed the stack given malicious input with
405 excessive recursion. This could result in a Denial Of Service attack. There
406 are no such structures used within SSL/TLS that come from untrusted sources
407 so this is considered safe.
409 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
414 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
416 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
417 effectively reduced to only comparing the least significant bit of each
418 byte. This allows an attacker to forge messages that would be considered as
419 authenticated in an amount of tries lower than that guaranteed by the
420 security claims of the scheme. The module can only be compiled by the
421 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
423 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
428 *) Add a build target 'build_all_generated', to build all generated files
429 and only that. This can be used to prepare everything that requires
430 things like perl for a system that lacks perl and then move everything
431 to that system and do the rest of the build there.
434 *) Backport SSL_OP_NO_RENGOTIATION
436 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
437 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
438 changes this is no longer possible in 1.1.0. Therefore the new
439 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
440 1.1.0 to provide equivalent functionality.
442 Note that if an application built against 1.1.0h headers (or above) is run
443 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
444 accepted but nothing will happen, i.e. renegotiation will not be prevented.
447 *) Removed the OS390-Unix config target. It relied on a script that doesn't
451 *) rsaz_1024_mul_avx2 overflow bug on x86_64
453 There is an overflow bug in the AVX2 Montgomery multiplication procedure
454 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
455 Analysis suggests that attacks against RSA and DSA as a result of this
456 defect would be very difficult to perform and are not believed likely.
457 Attacks against DH1024 are considered just feasible, because most of the
458 work necessary to deduce information about a private key may be performed
459 offline. The amount of resources required for such an attack would be
460 significant. However, for an attack on TLS to be meaningful, the server
461 would have to share the DH1024 private key among multiple clients, which is
462 no longer an option since CVE-2016-0701.
464 This only affects processors that support the AVX2 but not ADX extensions
465 like Intel Haswell (4th generation).
467 This issue was reported to OpenSSL by David Benjamin (Google). The issue
468 was originally found via the OSS-Fuzz project.
472 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
474 *) bn_sqrx8x_internal carry bug on x86_64
476 There is a carry propagating bug in the x86_64 Montgomery squaring
477 procedure. No EC algorithms are affected. Analysis suggests that attacks
478 against RSA and DSA as a result of this defect would be very difficult to
479 perform and are not believed likely. Attacks against DH are considered just
480 feasible (although very difficult) because most of the work necessary to
481 deduce information about a private key may be performed offline. The amount
482 of resources required for such an attack would be very significant and
483 likely only accessible to a limited number of attackers. An attacker would
484 additionally need online access to an unpatched system using the target
485 private key in a scenario with persistent DH parameters and a private
486 key that is shared between multiple clients.
488 This only affects processors that support the BMI1, BMI2 and ADX extensions
489 like Intel Broadwell (5th generation) and later or AMD Ryzen.
491 This issue was reported to OpenSSL by the OSS-Fuzz project.
495 *) Malformed X.509 IPAddressFamily could cause OOB read
497 If an X.509 certificate has a malformed IPAddressFamily extension,
498 OpenSSL could do a one-byte buffer overread. The most likely result
499 would be an erroneous display of the certificate in text format.
501 This issue was reported to OpenSSL by the OSS-Fuzz project.
505 Changes between 1.1.0e and 1.1.0f [25 May 2017]
507 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
508 platform rather than 'mingw'.
511 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
512 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
513 which is the minimum version we support.
516 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
518 *) Encrypt-Then-Mac renegotiation crash
520 During a renegotiation handshake if the Encrypt-Then-Mac extension is
521 negotiated where it was not in the original handshake (or vice-versa) then
522 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
523 and servers are affected.
525 This issue was reported to OpenSSL by Joe Orton (Red Hat).
529 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
531 *) Truncated packet could crash via OOB read
533 If one side of an SSL/TLS path is running on a 32-bit host and a specific
534 cipher is being used, then a truncated packet can cause that host to
535 perform an out-of-bounds read, usually resulting in a crash.
537 This issue was reported to OpenSSL by Robert Święcki of Google.
541 *) Bad (EC)DHE parameters cause a client crash
543 If a malicious server supplies bad parameters for a DHE or ECDHE key
544 exchange then this can result in the client attempting to dereference a
545 NULL pointer leading to a client crash. This could be exploited in a Denial
548 This issue was reported to OpenSSL by Guido Vranken.
552 *) BN_mod_exp may produce incorrect results on x86_64
554 There is a carry propagating bug in the x86_64 Montgomery squaring
555 procedure. No EC algorithms are affected. Analysis suggests that attacks
556 against RSA and DSA as a result of this defect would be very difficult to
557 perform and are not believed likely. Attacks against DH are considered just
558 feasible (although very difficult) because most of the work necessary to
559 deduce information about a private key may be performed offline. The amount
560 of resources required for such an attack would be very significant and
561 likely only accessible to a limited number of attackers. An attacker would
562 additionally need online access to an unpatched system using the target
563 private key in a scenario with persistent DH parameters and a private
564 key that is shared between multiple clients. For example this can occur by
565 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
566 similar to CVE-2015-3193 but must be treated as a separate problem.
568 This issue was reported to OpenSSL by the OSS-Fuzz project.
572 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
574 *) ChaCha20/Poly1305 heap-buffer-overflow
576 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
577 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
578 crash. This issue is not considered to be exploitable beyond a DoS.
580 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
584 *) CMS Null dereference
586 Applications parsing invalid CMS structures can crash with a NULL pointer
587 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
588 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
589 structure callback if an attempt is made to free certain invalid encodings.
590 Only CHOICE structures using a callback which do not handle NULL value are
593 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
597 *) Montgomery multiplication may produce incorrect results
599 There is a carry propagating bug in the Broadwell-specific Montgomery
600 multiplication procedure that handles input lengths divisible by, but
601 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
602 and DH private keys are impossible. This is because the subroutine in
603 question is not used in operations with the private key itself and an input
604 of the attacker's direct choice. Otherwise the bug can manifest itself as
605 transient authentication and key negotiation failures or reproducible
606 erroneous outcome of public-key operations with specially crafted input.
607 Among EC algorithms only Brainpool P-512 curves are affected and one
608 presumably can attack ECDH key negotiation. Impact was not analyzed in
609 detail, because pre-requisites for attack are considered unlikely. Namely
610 multiple clients have to choose the curve in question and the server has to
611 share the private key among them, neither of which is default behaviour.
612 Even then only clients that chose the curve will be affected.
614 This issue was publicly reported as transient failures and was not
615 initially recognized as a security issue. Thanks to Richard Morgan for
616 providing reproducible case.
620 *) Removed automatic addition of RPATH in shared libraries and executables,
621 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
624 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
626 *) Fix Use After Free for large message sizes
628 The patch applied to address CVE-2016-6307 resulted in an issue where if a
629 message larger than approx 16k is received then the underlying buffer to
630 store the incoming message is reallocated and moved. Unfortunately a
631 dangling pointer to the old location is left which results in an attempt to
632 write to the previously freed location. This is likely to result in a
633 crash, however it could potentially lead to execution of arbitrary code.
635 This issue only affects OpenSSL 1.1.0a.
637 This issue was reported to OpenSSL by Robert Święcki.
641 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
643 *) OCSP Status Request extension unbounded memory growth
645 A malicious client can send an excessively large OCSP Status Request
646 extension. If that client continually requests renegotiation, sending a
647 large OCSP Status Request extension each time, then there will be unbounded
648 memory growth on the server. This will eventually lead to a Denial Of
649 Service attack through memory exhaustion. Servers with a default
650 configuration are vulnerable even if they do not support OCSP. Builds using
651 the "no-ocsp" build time option are not affected.
653 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
657 *) SSL_peek() hang on empty record
659 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
660 sends an empty record. This could be exploited by a malicious peer in a
661 Denial Of Service attack.
663 This issue was reported to OpenSSL by Alex Gaynor.
667 *) Excessive allocation of memory in tls_get_message_header() and
668 dtls1_preprocess_fragment()
670 A (D)TLS message includes 3 bytes for its length in the header for the
671 message. This would allow for messages up to 16Mb in length. Messages of
672 this length are excessive and OpenSSL includes a check to ensure that a
673 peer is sending reasonably sized messages in order to avoid too much memory
674 being consumed to service a connection. A flaw in the logic of version
675 1.1.0 means that memory for the message is allocated too early, prior to
676 the excessive message length check. Due to way memory is allocated in
677 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
678 to service a connection. This could lead to a Denial of Service through
679 memory exhaustion. However, the excessive message length check still takes
680 place, and this would cause the connection to immediately fail. Assuming
681 that the application calls SSL_free() on the failed connection in a timely
682 manner then the 21Mb of allocated memory will then be immediately freed
683 again. Therefore the excessive memory allocation will be transitory in
684 nature. This then means that there is only a security impact if:
686 1) The application does not call SSL_free() in a timely manner in the event
687 that the connection fails
689 2) The application is working in a constrained environment where there is
690 very little free memory
692 3) The attacker initiates multiple connection attempts such that there are
693 multiple connections in a state where memory has been allocated for the
694 connection; SSL_free() has not yet been called; and there is insufficient
695 memory to service the multiple requests.
697 Except in the instance of (1) above any Denial Of Service is likely to be
698 transitory because as soon as the connection fails the memory is
699 subsequently freed again in the SSL_free() call. However there is an
700 increased risk during this period of application crashes due to the lack of
701 memory - which would then mean a more serious Denial of Service.
703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
704 (CVE-2016-6307 and CVE-2016-6308)
707 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
708 had to be removed. Primary reason is that vendor assembler can't
709 assemble our modules with -KPIC flag. As result it, assembly
710 support, was not even available as option. But its lack means
711 lack of side-channel resistant code, which is incompatible with
712 security by todays standards. Fortunately gcc is readily available
713 prepackaged option, which we firmly point at...
716 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
718 *) Windows command-line tool supports UTF-8 opt-in option for arguments
719 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
720 (to any value) allows Windows user to access PKCS#12 file generated
721 with Windows CryptoAPI and protected with non-ASCII password, as well
722 as files generated under UTF-8 locale on Linux also protected with
726 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
727 have been disabled by default and removed from DEFAULT, just like RC4.
728 See the RC4 item below to re-enable both.
731 *) The method for finding the storage location for the Windows RAND seed file
732 has changed. First we check %RANDFILE%. If that is not set then we check
733 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
734 all else fails we fall back to C:\.
737 *) The EVP_EncryptUpdate() function has had its return type changed from void
738 to int. A return of 0 indicates and error while a return of 1 indicates
742 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
743 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
744 off the constant time implementation for RSA, DSA and DH have been made
745 no-ops and deprecated.
748 *) Windows RAND implementation was simplified to only get entropy by
749 calling CryptGenRandom(). Various other RAND-related tickets
751 [Joseph Wylie Yandle, Rich Salz]
753 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
754 and OPENSSL_LH_, respectively. The old names are available
755 with API compatibility. They new names are now completely documented.
758 *) Unify TYPE_up_ref(obj) methods signature.
759 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
760 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
761 int (instead of void) like all others TYPE_up_ref() methods.
762 So now these methods also check the return value of CRYPTO_atomic_add(),
763 and the validity of object reference counter.
764 [fdasilvayy@gmail.com]
766 *) With Windows Visual Studio builds, the .pdb files are installed
767 alongside the installed libraries and executables. For a static
768 library installation, ossl_static.pdb is the associate compiler
769 generated .pdb file to be used when linking programs.
772 *) Remove openssl.spec. Packaging files belong with the packagers.
775 *) Automatic Darwin/OSX configuration has had a refresh, it will now
776 recognise x86_64 architectures automatically. You can still decide
777 to build for a different bitness with the environment variable
778 KERNEL_BITS (can be 32 or 64), for example:
780 KERNEL_BITS=32 ./config
784 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
785 256 bit AES and HMAC with SHA256.
788 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
791 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
794 *) To enable users to have their own config files and build file templates,
795 Configure looks in the directory indicated by the environment variable
796 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
797 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
798 name and is used as is.
801 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
802 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
803 X509_CERT_FILE_CTX was removed.
806 *) "shared" builds are now the default. To create only static libraries use
807 the "no-shared" Configure option.
810 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
811 All of these option have not worked for some while and are fundamental
815 *) Make various cleanup routines no-ops and mark them as deprecated. Most
816 global cleanup functions are no longer required because they are handled
817 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
818 Explicitly de-initing can cause problems (e.g. where a library that uses
819 OpenSSL de-inits, but an application is still using it). The affected
820 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
821 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
822 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
826 *) --strict-warnings no longer enables runtime debugging options
827 such as REF_DEBUG. Instead, debug options are automatically
828 enabled with '--debug' builds.
829 [Andy Polyakov, Emilia Käsper]
831 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
832 have been moved out of the public header files. New functions for managing
833 these have been added.
836 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
837 objects have been moved out of the public header files. New
838 functions for managing these have been added.
841 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
842 have been moved out of the public header files. New functions for managing
843 these have been added.
846 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
847 moved out of the public header files. New functions for managing these
851 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
854 *) Removed the mk1mf build scripts.
857 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
858 it is always safe to #include a header now.
861 *) Removed the aged BC-32 config and all its supporting scripts
864 *) Removed support for Ultrix, Netware, and OS/2.
867 *) Add support for HKDF.
870 *) Add support for blake2b and blake2s
873 *) Added support for "pipelining". Ciphers that have the
874 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
875 encryptions/decryptions simultaneously. There are currently no built-in
876 ciphers with this property but the expectation is that engines will be able
877 to offer it to significantly improve throughput. Support has been extended
878 into libssl so that multiple records for a single connection can be
879 processed in one go (for >=TLS 1.1).
882 *) Added the AFALG engine. This is an async capable engine which is able to
883 offload work to the Linux kernel. In this initial version it only supports
884 AES128-CBC. The kernel must be version 4.1.0 or greater.
887 *) OpenSSL now uses a new threading API. It is no longer necessary to
888 set locking callbacks to use OpenSSL in a multi-threaded environment. There
889 are two supported threading models: pthreads and windows threads. It is
890 also possible to configure OpenSSL at compile time for "no-threads". The
891 old threading API should no longer be used. The functions have been
892 replaced with "no-op" compatibility macros.
893 [Alessandro Ghedini, Matt Caswell]
895 *) Modify behavior of ALPN to invoke callback after SNI/servername
896 callback, such that updates to the SSL_CTX affect ALPN.
899 *) Add SSL_CIPHER queries for authentication and key-exchange.
902 *) Changes to the DEFAULT cipherlist:
903 - Prefer (EC)DHE handshakes over plain RSA.
904 - Prefer AEAD ciphers over legacy ciphers.
905 - Prefer ECDSA over RSA when both certificates are available.
906 - Prefer TLSv1.2 ciphers/PRF.
907 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
911 *) Change the ECC default curve list to be this, in order: x25519,
912 secp256r1, secp521r1, secp384r1.
915 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
916 disabled by default. They can be re-enabled using the
917 enable-weak-ssl-ciphers option to Configure.
920 *) If the server has ALPN configured, but supports no protocols that the
921 client advertises, send a fatal "no_application_protocol" alert.
922 This behaviour is SHALL in RFC 7301, though it isn't universally
923 implemented by other servers.
926 *) Add X25519 support.
927 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
928 for public and private key encoding using the format documented in
929 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
930 key generation and key derivation.
932 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
936 *) Deprecate SRP_VBASE_get_by_user.
937 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
938 In order to fix an unavoidable memory leak (CVE-2016-0798),
939 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
940 seed, even if the seed is configured.
942 Users should use SRP_VBASE_get1_by_user instead. Note that in
943 SRP_VBASE_get1_by_user, caller must free the returned value. Note
944 also that even though configuring the SRP seed attempts to hide
945 invalid usernames by continuing the handshake with fake
946 credentials, this behaviour is not constant time and no strong
947 guarantees are made that the handshake is indistinguishable from
948 that of a valid user.
951 *) Configuration change; it's now possible to build dynamic engines
952 without having to build shared libraries and vice versa. This
953 only applies to the engines in engines/, those in crypto/engine/
954 will always be built into libcrypto (i.e. "static").
956 Building dynamic engines is enabled by default; to disable, use
957 the configuration option "disable-dynamic-engine".
959 The only requirements for building dynamic engines are the
960 presence of the DSO module and building with position independent
961 code, so they will also automatically be disabled if configuring
962 with "disable-dso" or "disable-pic".
964 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
965 are also taken away from openssl/opensslconf.h, as they are
969 *) Configuration change; if there is a known flag to compile
970 position independent code, it will always be applied on the
971 libcrypto and libssl object files, and never on the application
972 object files. This means other libraries that use routines from
973 libcrypto / libssl can be made into shared libraries regardless
974 of how OpenSSL was configured.
976 If this isn't desirable, the configuration options "disable-pic"
977 or "no-pic" can be used to disable the use of PIC. This will
978 also disable building shared libraries and dynamic engines.
981 *) Removed JPAKE code. It was experimental and has no wide use.
984 *) The INSTALL_PREFIX Makefile variable has been renamed to
985 DESTDIR. That makes for less confusion on what this variable
986 is for. Also, the configuration option --install_prefix is
990 *) Heartbeat for TLS has been removed and is disabled by default
991 for DTLS; configure with enable-heartbeats. Code that uses the
992 old #define's might need to be updated.
993 [Emilia Käsper, Rich Salz]
995 *) Rename REF_CHECK to REF_DEBUG.
998 *) New "unified" build system
1000 The "unified" build system is aimed to be a common system for all
1001 platforms we support. With it comes new support for VMS.
1003 This system builds supports building in a different directory tree
1004 than the source tree. It produces one Makefile (for unix family
1005 or lookalikes), or one descrip.mms (for VMS).
1007 The source of information to make the Makefile / descrip.mms is
1008 small files called 'build.info', holding the necessary
1009 information for each directory with source to compile, and a
1010 template in Configurations, like unix-Makefile.tmpl or
1013 With this change, the library names were also renamed on Windows
1014 and on VMS. They now have names that are closer to the standard
1015 on Unix, and include the major version number, and in certain
1016 cases, the architecture they are built for. See "Notes on shared
1017 libraries" in INSTALL.
1019 We rely heavily on the perl module Text::Template.
1022 *) Added support for auto-initialisation and de-initialisation of the library.
1023 OpenSSL no longer requires explicit init or deinit routines to be called,
1024 except in certain circumstances. See the OPENSSL_init_crypto() and
1025 OPENSSL_init_ssl() man pages for further information.
1028 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1029 "peer" argument is now expected to be a BIO_ADDR object.
1031 *) Rewrite of BIO networking library. The BIO library lacked consistent
1032 support of IPv6, and adding it required some more extensive
1033 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1034 which hold all types of addresses and chains of address information.
1035 It also introduces a new API, with functions like BIO_socket,
1036 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1037 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1038 have been adapted accordingly.
1041 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1045 *) CRIME protection: disable compression by default, even if OpenSSL is
1046 compiled with zlib enabled. Applications can still enable compression
1047 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1048 using the SSL_CONF library to configure compression.
1051 *) The signature of the session callback configured with
1052 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1053 was explicitly marked as 'const unsigned char*' instead of
1057 *) Always DPURIFY. Remove the use of uninitialized memory in the
1058 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1061 *) Removed many obsolete configuration items, including
1062 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1063 MD2_CHAR, MD2_INT, MD2_LONG
1065 IDEA_SHORT, IDEA_LONG
1066 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1067 [Rich Salz, with advice from Andy Polyakov]
1069 *) Many BN internals have been moved to an internal header file.
1070 [Rich Salz with help from Andy Polyakov]
1072 *) Configuration and writing out the results from it has changed.
1073 Files such as Makefile include/openssl/opensslconf.h and are now
1074 produced through general templates, such as Makefile.in and
1075 crypto/opensslconf.h.in and some help from the perl module
1078 Also, the center of configuration information is no longer
1079 Makefile. Instead, Configure produces a perl module in
1080 configdata.pm which holds most of the config data (in the hash
1081 table %config), the target data that comes from the target
1082 configuration in one of the Configurations/*.conf files (in
1086 *) To clarify their intended purposes, the Configure options
1087 --prefix and --openssldir change their semantics, and become more
1088 straightforward and less interdependent.
1090 --prefix shall be used exclusively to give the location INSTALLTOP
1091 where programs, scripts, libraries, include files and manuals are
1092 going to be installed. The default is now /usr/local.
1094 --openssldir shall be used exclusively to give the default
1095 location OPENSSLDIR where certificates, private keys, CRLs are
1096 managed. This is also where the default openssl.cnf gets
1098 If the directory given with this option is a relative path, the
1099 values of both the --prefix value and the --openssldir value will
1100 be combined to become OPENSSLDIR.
1101 The default for --openssldir is INSTALLTOP/ssl.
1103 Anyone who uses --openssldir to specify where OpenSSL is to be
1104 installed MUST change to use --prefix instead.
1107 *) The GOST engine was out of date and therefore it has been removed. An up
1108 to date GOST engine is now being maintained in an external repository.
1109 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1110 support for GOST ciphersuites (these are only activated if a GOST engine
1114 *) EGD is no longer supported by default; use enable-egd when
1116 [Ben Kaduk and Rich Salz]
1118 *) The distribution now has Makefile.in files, which are used to
1119 create Makefile's when Configure is run. *Configure must be run
1120 before trying to build now.*
1123 *) The return value for SSL_CIPHER_description() for error conditions
1127 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1129 Obtaining and performing DNSSEC validation of TLSA records is
1130 the application's responsibility. The application provides
1131 the TLSA records of its choice to OpenSSL, and these are then
1132 used to authenticate the peer.
1134 The TLSA records need not even come from DNS. They can, for
1135 example, be used to implement local end-entity certificate or
1136 trust-anchor "pinning", where the "pin" data takes the form
1137 of TLSA records, which can augment or replace verification
1138 based on the usual WebPKI public certification authorities.
1141 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1142 continues to support deprecated interfaces in default builds.
1143 However, applications are strongly advised to compile their
1144 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1145 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1146 or the 1.1.0 releases.
1148 In environments in which all applications have been ported to
1149 not use any deprecated interfaces OpenSSL's Configure script
1150 should be used with the --api=1.1.0 option to entirely remove
1151 support for the deprecated features from the library and
1152 unconditionally disable them in the installed headers.
1153 Essentially the same effect can be achieved with the "no-deprecated"
1154 argument to Configure, except that this will always restrict
1155 the build to just the latest API, rather than a fixed API
1158 As applications are ported to future revisions of the API,
1159 they should update their compile-time OPENSSL_API_COMPAT define
1160 accordingly, but in most cases should be able to continue to
1161 compile with later releases.
1163 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1164 0x10000000L and 0x00908000L, respectively. However those
1165 versions did not support the OPENSSL_API_COMPAT feature, and
1166 so applications are not typically tested for explicit support
1167 of just the undeprecated features of either release.
1170 *) Add support for setting the minimum and maximum supported protocol.
1171 It can bet set via the SSL_set_min_proto_version() and
1172 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1173 MaxProtocol. It's recommended to use the new APIs to disable
1174 protocols instead of disabling individual protocols using
1175 SSL_set_options() or SSL_CONF's Protocol. This change also
1176 removes support for disabling TLS 1.2 in the OpenSSL TLS
1177 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1180 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1183 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1184 and integrates ECDSA and ECDH functionality into EC. Implementations can
1185 now redirect key generation and no longer need to convert to or from
1188 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1189 include the ec.h header file instead.
1192 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1193 ciphers who are no longer supported and drops support the ephemeral RSA key
1194 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1197 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1198 opaque. For HMAC_CTX, the following constructors and destructors
1201 HMAC_CTX *HMAC_CTX_new(void);
1202 void HMAC_CTX_free(HMAC_CTX *ctx);
1204 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1205 destroy such methods has been added. See EVP_MD_meth_new(3) and
1206 EVP_CIPHER_meth_new(3) for documentation.
1209 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1210 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1211 EVP_MD_CTX_reset() should be called instead to reinitialise
1212 an already created structure.
1213 2) For consistency with the majority of our object creators and
1214 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1215 EVP_MD_CTX_(new|free). The old names are retained as macros
1216 for deprecated builds.
1219 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1220 cryptographic operations to be performed asynchronously as long as an
1221 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1222 further details. Libssl has also had this capability integrated with the
1223 introduction of the new mode SSL_MODE_ASYNC and associated error
1224 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1225 pages. This work was developed in partnership with Intel Corp.
1228 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1229 always enabled now. If you want to disable the support you should
1230 exclude it using the list of supported ciphers. This also means that the
1231 "-no_ecdhe" option has been removed from s_server.
1234 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1235 SSL_{CTX_}set1_curves() which can set a list.
1238 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1239 curve you want to support using SSL_{CTX_}set1_curves().
1242 *) State machine rewrite. The state machine code has been significantly
1243 refactored in order to remove much duplication of code and solve issues
1244 with the old code (see ssl/statem/README for further details). This change
1245 does have some associated API changes. Notably the SSL_state() function
1246 has been removed and replaced by SSL_get_state which now returns an
1247 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1248 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1252 *) All instances of the string "ssleay" in the public API were replaced
1253 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1254 Some error codes related to internal RSA_eay API's were renamed.
1257 *) The demo files in crypto/threads were moved to demo/threads.
1260 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1262 [Matt Caswell, Rich Salz]
1264 *) New ASN.1 embed macro.
1266 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1267 structure is not allocated: it is part of the parent. That is instead of
1275 This reduces memory fragmentation and make it impossible to accidentally
1276 set a mandatory field to NULL.
1278 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1279 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1280 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1284 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1287 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1288 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1289 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1290 DES and RC4 ciphersuites.
1293 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1294 This changes the decoding behaviour for some invalid messages,
1295 though the change is mostly in the more lenient direction, and
1296 legacy behaviour is preserved as much as possible.
1299 *) Fix no-stdio build.
1300 [ David Woodhouse <David.Woodhouse@intel.com> and also
1301 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1303 *) New testing framework
1304 The testing framework has been largely rewritten and is now using
1305 perl and the perl modules Test::Harness and an extended variant of
1306 Test::More called OpenSSL::Test to do its work. All test scripts in
1307 test/ have been rewritten into test recipes, and all direct calls to
1308 executables in test/Makefile have become individual recipes using the
1309 simplified testing OpenSSL::Test::Simple.
1311 For documentation on our testing modules, do:
1313 perldoc test/testlib/OpenSSL/Test/Simple.pm
1314 perldoc test/testlib/OpenSSL/Test.pm
1318 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1319 are used; the latter aborts on memory leaks (usually checked on exit).
1320 Some undocumented "set malloc, etc., hooks" functions were removed
1321 and others were changed. All are now documented.
1324 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1326 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1328 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1329 from RFC4279, RFC4785, RFC5487, RFC5489.
1331 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1332 original RSA_PSK patch.
1335 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1336 era flag was never set throughout the codebase (only read). Also removed
1337 SSL3_FLAGS_POP_BUFFER which was only used if
1338 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1341 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1342 to be "oneline" instead of "compat".
1345 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1346 not aware of clients that still exhibit this bug, and the workaround
1347 hasn't been working properly for a while.
1350 *) The return type of BIO_number_read() and BIO_number_written() as well as
1351 the corresponding num_read and num_write members in the BIO structure has
1352 changed from unsigned long to uint64_t. On platforms where an unsigned
1353 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1357 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1358 OpenSSL without support for them. It also means that maintaining
1359 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1360 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1363 *) Removed support for the two export grade static DH ciphersuites
1364 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1365 were newly added (along with a number of other static DH ciphersuites) to
1366 1.0.2. However the two export ones have *never* worked since they were
1367 introduced. It seems strange in any case to be adding new export
1368 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1371 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1372 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1373 and turned into macros which simply call the new preferred function names
1374 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1375 should use the new names instead. Also as part of this change the ssl23.h
1376 header file has been removed.
1379 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1380 code and the associated standard is no longer considered fit-for-purpose.
1383 *) RT2547 was closed. When generating a private key, try to make the
1384 output file readable only by the owner. This behavior change might
1385 be noticeable when interacting with other software.
1387 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1391 *) Added HTTP GET support to the ocsp command.
1394 *) Changed default digest for the dgst and enc commands from MD5 to
1398 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1401 *) Added support for TLS extended master secret from
1402 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1403 initial patch which was a great help during development.
1406 *) All libssl internal structures have been removed from the public header
1407 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1408 now redundant). Users should not attempt to access internal structures
1409 directly. Instead they should use the provided API functions.
1412 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1413 Access to deprecated functions can be re-enabled by running config with
1414 "enable-deprecated". In addition applications wishing to use deprecated
1415 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1416 will, by default, disable some transitive includes that previously existed
1417 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1420 *) Added support for OCB mode. OpenSSL has been granted a patent license
1421 compatible with the OpenSSL license for use of OCB. Details are available
1422 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1423 for OCB can be removed by calling config with no-ocb.
1426 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1427 compatible client hello.
1430 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1431 done while fixing the error code for the key-too-small case.
1432 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1434 *) CA.sh has been removed; use CA.pl instead.
1437 *) Removed old DES API.
1440 *) Remove various unsupported platforms:
1446 Sinix/ReliantUNIX RM400
1451 16-bit platforms such as WIN16
1454 *) Clean up OPENSSL_NO_xxx #define's
1455 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1456 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1457 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1458 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1459 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1460 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1461 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1462 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1463 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1464 Remove MS_STATIC; it's a relic from platforms <32 bits.
1467 *) Cleaned up dead code
1468 Remove all but one '#ifdef undef' which is to be looked at.
1471 *) Clean up calling of xxx_free routines.
1472 Just like free(), fix most of the xxx_free routines to accept
1473 NULL. Remove the non-null checks from callers. Save much code.
1476 *) Add secure heap for storage of private keys (when possible).
1477 Add BIO_s_secmem(), CBIGNUM, etc.
1478 Contributed by Akamai Technologies under our Corporate CLA.
1481 *) Experimental support for a new, fast, unbiased prime candidate generator,
1482 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1483 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1485 *) New output format NSS in the sess_id command line tool. This allows
1486 exporting the session id and the master key in NSS keylog format.
1487 [Martin Kaiser <martin@kaiser.cx>]
1489 *) Harmonize version and its documentation. -f flag is used to display
1491 [mancha <mancha1@zoho.com>]
1493 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1494 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1495 [mancha <mancha1@zoho.com>]
1497 *) Fix some double frees. These are not thought to be exploitable.
1498 [mancha <mancha1@zoho.com>]
1500 *) A missing bounds check in the handling of the TLS heartbeat extension
1501 can be used to reveal up to 64k of memory to a connected client or
1504 Thanks for Neel Mehta of Google Security for discovering this bug and to
1505 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1506 preparing the fix (CVE-2014-0160)
1507 [Adam Langley, Bodo Moeller]
1509 *) Fix for the attack described in the paper "Recovering OpenSSL
1510 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1511 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1512 http://eprint.iacr.org/2014/140
1514 Thanks to Yuval Yarom and Naomi Benger for discovering this
1515 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1516 [Yuval Yarom and Naomi Benger]
1518 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1519 this fixes a limitation in previous versions of OpenSSL.
1522 *) Experimental encrypt-then-mac support.
1524 Experimental support for encrypt then mac from
1525 draft-gutmann-tls-encrypt-then-mac-02.txt
1527 To enable it set the appropriate extension number (0x42 for the test
1528 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1530 For non-compliant peers (i.e. just about everything) this should have no
1533 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1537 *) Add EVP support for key wrapping algorithms, to avoid problems with
1538 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1539 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1540 algorithms and include tests cases.
1543 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1547 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1548 MGF1 digest and OAEP label.
1551 *) Make openssl verify return errors.
1552 [Chris Palmer <palmer@google.com> and Ben Laurie]
1554 *) New function ASN1_TIME_diff to calculate the difference between two
1555 ASN1_TIME structures or one structure and the current time.
1558 *) Update fips_test_suite to support multiple command line options. New
1559 test to induce all self test errors in sequence and check expected
1563 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1564 sign or verify all in one operation.
1567 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1568 test programs and fips_test_suite. Includes functionality to parse
1569 the minimal script output of fipsalgest.pl directly.
1572 *) Add authorisation parameter to FIPS_module_mode_set().
1575 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1578 *) Use separate DRBG fields for internal and external flags. New function
1579 FIPS_drbg_health_check() to perform on demand health checking. Add
1580 generation tests to fips_test_suite with reduced health check interval to
1581 demonstrate periodic health checking. Add "nodh" option to
1582 fips_test_suite to skip very slow DH test.
1585 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1589 *) More extensive health check for DRBG checking many more failure modes.
1590 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1591 combination: call this in fips_test_suite.
1594 *) Add support for canonical generation of DSA parameter 'g'. See
1597 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1598 POST to handle HMAC cases.
1601 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1602 to return numerical and string versions of the FIPS module number.
1605 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1606 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1607 outside the validated module in the FIPS capable OpenSSL.
1610 *) Minor change to DRBG entropy callback semantics. In some cases
1611 there is no multiple of the block length between min_len and
1612 max_len. Allow the callback to return more than max_len bytes
1613 of entropy but discard any extra: it is the callback's responsibility
1614 to ensure that the extra data discarded does not impact the
1615 requested amount of entropy.
1618 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1619 information in FIPS186-3, SP800-57 and SP800-131A.
1622 *) CCM support via EVP. Interface is very similar to GCM case except we
1623 must supply all data in one chunk (i.e. no update, final) and the
1624 message length must be supplied if AAD is used. Add algorithm test
1628 *) Initial version of POST overhaul. Add POST callback to allow the status
1629 of POST to be monitored and/or failures induced. Modify fips_test_suite
1630 to use callback. Always run all selftests even if one fails.
1633 *) XTS support including algorithm test driver in the fips_gcmtest program.
1634 Note: this does increase the maximum key length from 32 to 64 bytes but
1635 there should be no binary compatibility issues as existing applications
1636 will never use XTS mode.
1639 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1640 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1641 performs algorithm blocking for unapproved PRNG types. Also do not
1642 set PRNG type in FIPS_mode_set(): leave this to the application.
1643 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1644 the standard OpenSSL PRNG: set additional data to a date time vector.
1647 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1648 This shouldn't present any incompatibility problems because applications
1649 shouldn't be using these directly and any that are will need to rethink
1650 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1653 *) Extensive self tests and health checking required by SP800-90 DRBG.
1654 Remove strength parameter from FIPS_drbg_instantiate and always
1655 instantiate at maximum supported strength.
1658 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1661 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1664 *) New function DH_compute_key_padded() to compute a DH key and pad with
1665 leading zeroes if needed: this complies with SP800-56A et al.
1668 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1669 anything, incomplete, subject to change and largely untested at present.
1672 *) Modify fipscanisteronly build option to only build the necessary object
1673 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1676 *) Add experimental option FIPSSYMS to give all symbols in
1677 fipscanister.o and FIPS or fips prefix. This will avoid
1678 conflicts with future versions of OpenSSL. Add perl script
1679 util/fipsas.pl to preprocess assembly language source files
1680 and rename any affected symbols.
1683 *) Add selftest checks and algorithm block of non-fips algorithms in
1684 FIPS mode. Remove DES2 from selftests.
1687 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1688 return internal method without any ENGINE dependencies. Add new
1689 tiny fips sign and verify functions.
1692 *) New build option no-ec2m to disable characteristic 2 code.
1695 *) New build option "fipscanisteronly". This only builds fipscanister.o
1696 and (currently) associated fips utilities. Uses the file Makefile.fips
1697 instead of Makefile.org as the prototype.
1700 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1701 Update fips_gcmtest to use IV generator.
1704 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1705 setting output buffer to NULL. The *Final function must be
1706 called although it will not retrieve any additional data. The tag
1707 can be set or retrieved with a ctrl. The IV length is by default 12
1708 bytes (96 bits) but can be set to an alternative value. If the IV
1709 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1713 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1714 underlying do_cipher function handles all cipher semantics itself
1715 including padding and finalisation. This is useful if (for example)
1716 an ENGINE cipher handles block padding itself. The behaviour of
1717 do_cipher is subtly changed if this flag is set: the return value
1718 is the number of characters written to the output buffer (zero is
1719 no longer an error code) or a negative error code. Also if the
1720 input buffer is NULL and length 0 finalisation should be performed.
1723 *) If a candidate issuer certificate is already part of the constructed
1724 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1727 *) Improve forward-security support: add functions
1729 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1730 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1732 for use by SSL/TLS servers; the callback function will be called whenever a
1733 new session is created, and gets to decide whether the session may be
1734 cached to make it resumable (return 0) or not (return 1). (As by the
1735 SSL/TLS protocol specifications, the session_id sent by the server will be
1736 empty to indicate that the session is not resumable; also, the server will
1737 not generate RFC 4507 (RFC 5077) session tickets.)
1739 A simple reasonable callback implementation is to return is_forward_secure.
1740 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1741 by the SSL/TLS server library, indicating whether it can provide forward
1743 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1745 *) New -verify_name option in command line utilities to set verification
1749 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1750 Add CMAC pkey methods.
1753 *) Experimental renegotiation in s_server -www mode. If the client
1754 browses /reneg connection is renegotiated. If /renegcert it is
1755 renegotiated requesting a certificate.
1758 *) Add an "external" session cache for debugging purposes to s_server. This
1759 should help trace issues which normally are only apparent in deployed
1760 multi-process servers.
1763 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1764 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1765 BIO_set_cipher() and some obscure PEM functions were changed so they
1766 can now return an error. The RAND changes required a change to the
1767 RAND_METHOD structure.
1770 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1771 a gcc attribute to warn if the result of a function is ignored. This
1772 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1773 whose return value is often ignored.
1776 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1777 These allow SCTs (signed certificate timestamps) to be requested and
1778 validated when establishing a connection.
1779 [Rob Percival <robpercival@google.com>]
1781 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1783 *) Prevent padding oracle in AES-NI CBC MAC check
1785 A MITM attacker can use a padding oracle attack to decrypt traffic
1786 when the connection uses an AES CBC cipher and the server support
1789 This issue was introduced as part of the fix for Lucky 13 padding
1790 attack (CVE-2013-0169). The padding check was rewritten to be in
1791 constant time by making sure that always the same bytes are read and
1792 compared against either the MAC or padding bytes. But it no longer
1793 checked that there was enough data to have both the MAC and padding
1796 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1800 *) Fix EVP_EncodeUpdate overflow
1802 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1803 Base64 encoding of binary data. If an attacker is able to supply very large
1804 amounts of input data then a length check can overflow resulting in a heap
1807 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1808 the PEM_write_bio* family of functions. These are mainly used within the
1809 OpenSSL command line applications, so any application which processes data
1810 from an untrusted source and outputs it as a PEM file should be considered
1811 vulnerable to this issue. User applications that call these APIs directly
1812 with large amounts of untrusted data may also be vulnerable.
1814 This issue was reported by Guido Vranken.
1818 *) Fix EVP_EncryptUpdate overflow
1820 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1821 is able to supply very large amounts of input data after a previous call to
1822 EVP_EncryptUpdate() with a partial block then a length check can overflow
1823 resulting in a heap corruption. Following an analysis of all OpenSSL
1824 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1825 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1826 the first called function after an EVP_EncryptInit(), and therefore that
1827 specific call must be safe. The second form is where the length passed to
1828 EVP_EncryptUpdate() can be seen from the code to be some small value and
1829 therefore there is no possibility of an overflow. Since all instances are
1830 one of these two forms, it is believed that there can be no overflows in
1831 internal code due to this problem. It should be noted that
1832 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1833 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1834 of these calls have also been analysed too and it is believed there are no
1835 instances in internal usage where an overflow could occur.
1837 This issue was reported by Guido Vranken.
1841 *) Prevent ASN.1 BIO excessive memory allocation
1843 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1844 a short invalid encoding can cause allocation of large amounts of memory
1845 potentially consuming excessive resources or exhausting memory.
1847 Any application parsing untrusted data through d2i BIO functions is
1848 affected. The memory based functions such as d2i_X509() are *not* affected.
1849 Since the memory based functions are used by the TLS library, TLS
1850 applications are not affected.
1852 This issue was reported by Brian Carpenter.
1858 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1859 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1860 in arbitrary stack data being returned in the buffer.
1862 This issue was reported by Guido Vranken.
1866 *) Modify behavior of ALPN to invoke callback after SNI/servername
1867 callback, such that updates to the SSL_CTX affect ALPN.
1870 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1874 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1875 methods are enabled and ssl2 is disabled the methods return NULL.
1878 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1880 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1881 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1882 provide any "EXPORT" or "LOW" strength ciphers.
1885 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1886 is by default disabled at build-time. Builds that are not configured with
1887 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1888 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1889 will need to explicitly call either of:
1891 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1893 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1895 as appropriate. Even if either of those is used, or the application
1896 explicitly uses the version-specific SSLv2_method() or its client and
1897 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1898 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1899 ciphers, and SSLv2 56-bit DES are no longer available.
1903 *) Fix a double-free in DSA code
1905 A double free bug was discovered when OpenSSL parses malformed DSA private
1906 keys and could lead to a DoS attack or memory corruption for applications
1907 that receive DSA private keys from untrusted sources. This scenario is
1910 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1915 *) Disable SRP fake user seed to address a server memory leak.
1917 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1919 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1920 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1921 was changed to ignore the "fake user" SRP seed, even if the seed
1924 Users should use SRP_VBASE_get1_by_user instead. Note that in
1925 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1926 also that even though configuring the SRP seed attempts to hide
1927 invalid usernames by continuing the handshake with fake
1928 credentials, this behaviour is not constant time and no strong
1929 guarantees are made that the handshake is indistinguishable from
1930 that of a valid user.
1934 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1936 In the BN_hex2bn function the number of hex digits is calculated using an
1937 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1938 large values of |i| this can result in |bn_expand| not allocating any
1939 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1940 field as NULL leading to a subsequent NULL ptr deref. For very large values
1941 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1942 In this case memory is allocated to the internal BIGNUM data field, but it
1943 is insufficiently sized leading to heap corruption. A similar issue exists
1944 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1945 is ever called by user applications with very large untrusted hex/dec data.
1946 This is anticipated to be a rare occurrence.
1948 All OpenSSL internal usage of these functions use data that is not expected
1949 to be untrusted, e.g. config file data or application command line
1950 arguments. If user developed applications generate config file data based
1951 on untrusted data then it is possible that this could also lead to security
1952 consequences. This is also anticipated to be rare.
1954 This issue was reported to OpenSSL by Guido Vranken.
1958 *) Fix memory issues in BIO_*printf functions
1960 The internal |fmtstr| function used in processing a "%s" format string in
1961 the BIO_*printf functions could overflow while calculating the length of a
1962 string and cause an OOB read when printing very long strings.
1964 Additionally the internal |doapr_outch| function can attempt to write to an
1965 OOB memory location (at an offset from the NULL pointer) in the event of a
1966 memory allocation failure. In 1.0.2 and below this could be caused where
1967 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1968 could be in processing a very long "%s" format string. Memory leaks can
1971 The first issue may mask the second issue dependent on compiler behaviour.
1972 These problems could enable attacks where large amounts of untrusted data
1973 is passed to the BIO_*printf functions. If applications use these functions
1974 in this way then they could be vulnerable. OpenSSL itself uses these
1975 functions when printing out human-readable dumps of ASN.1 data. Therefore
1976 applications that print this data could be vulnerable if the data is from
1977 untrusted sources. OpenSSL command line applications could also be
1978 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1979 as command line arguments.
1981 Libssl is not considered directly vulnerable. Additionally certificates etc
1982 received via remote connections via libssl are also unlikely to be able to
1983 trigger these issues because of message size limits enforced within libssl.
1985 This issue was reported to OpenSSL Guido Vranken.
1989 *) Side channel attack on modular exponentiation
1991 A side-channel attack was found which makes use of cache-bank conflicts on
1992 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1993 of RSA keys. The ability to exploit this issue is limited as it relies on
1994 an attacker who has control of code in a thread running on the same
1995 hyper-threaded core as the victim thread which is performing decryptions.
1997 This issue was reported to OpenSSL by Yuval Yarom, The University of
1998 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1999 Nadia Heninger, University of Pennsylvania with more information at
2000 http://cachebleed.info.
2004 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2005 if no keysize is specified with default_bits. This fixes an
2006 omission in an earlier change that changed all RSA/DSA key generation
2007 apps to use 2048 bits by default.
2010 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2011 *) DH small subgroups
2013 Historically OpenSSL only ever generated DH parameters based on "safe"
2014 primes. More recently (in version 1.0.2) support was provided for
2015 generating X9.42 style parameter files such as those required for RFC 5114
2016 support. The primes used in such files may not be "safe". Where an
2017 application is using DH configured with parameters based on primes that are
2018 not "safe" then an attacker could use this fact to find a peer's private
2019 DH exponent. This attack requires that the attacker complete multiple
2020 handshakes in which the peer uses the same private DH exponent. For example
2021 this could be used to discover a TLS server's private DH exponent if it's
2022 reusing the private DH exponent or it's using a static DH ciphersuite.
2024 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2025 TLS. It is not on by default. If the option is not set then the server
2026 reuses the same private DH exponent for the life of the server process and
2027 would be vulnerable to this attack. It is believed that many popular
2028 applications do set this option and would therefore not be at risk.
2030 The fix for this issue adds an additional check where a "q" parameter is
2031 available (as is the case in X9.42 based parameters). This detects the
2032 only known attack, and is the only possible defense for static DH
2033 ciphersuites. This could have some performance impact.
2035 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2036 default and cannot be disabled. This could have some performance impact.
2038 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2042 *) SSLv2 doesn't block disabled ciphers
2044 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2045 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2046 been disabled, provided that the SSLv2 protocol was not also disabled via
2049 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2050 and Sebastian Schinzel.
2054 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2056 *) BN_mod_exp may produce incorrect results on x86_64
2058 There is a carry propagating bug in the x86_64 Montgomery squaring
2059 procedure. No EC algorithms are affected. Analysis suggests that attacks
2060 against RSA and DSA as a result of this defect would be very difficult to
2061 perform and are not believed likely. Attacks against DH are considered just
2062 feasible (although very difficult) because most of the work necessary to
2063 deduce information about a private key may be performed offline. The amount
2064 of resources required for such an attack would be very significant and
2065 likely only accessible to a limited number of attackers. An attacker would
2066 additionally need online access to an unpatched system using the target
2067 private key in a scenario with persistent DH parameters and a private
2068 key that is shared between multiple clients. For example this can occur by
2069 default in OpenSSL DHE based SSL/TLS ciphersuites.
2071 This issue was reported to OpenSSL by Hanno Böck.
2075 *) Certificate verify crash with missing PSS parameter
2077 The signature verification routines will crash with a NULL pointer
2078 dereference if presented with an ASN.1 signature using the RSA PSS
2079 algorithm and absent mask generation function parameter. Since these
2080 routines are used to verify certificate signature algorithms this can be
2081 used to crash any certificate verification operation and exploited in a
2082 DoS attack. Any application which performs certificate verification is
2083 vulnerable including OpenSSL clients and servers which enable client
2086 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2090 *) X509_ATTRIBUTE memory leak
2092 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2093 memory. This structure is used by the PKCS#7 and CMS routines so any
2094 application which reads PKCS#7 or CMS data from untrusted sources is
2095 affected. SSL/TLS is not affected.
2097 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2102 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2103 This changes the decoding behaviour for some invalid messages,
2104 though the change is mostly in the more lenient direction, and
2105 legacy behaviour is preserved as much as possible.
2108 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2110 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2112 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2114 *) Alternate chains certificate forgery
2116 During certificate verification, OpenSSL will attempt to find an
2117 alternative certificate chain if the first attempt to build such a chain
2118 fails. An error in the implementation of this logic can mean that an
2119 attacker could cause certain checks on untrusted certificates to be
2120 bypassed, such as the CA flag, enabling them to use a valid leaf
2121 certificate to act as a CA and "issue" an invalid certificate.
2123 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2127 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2129 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2130 incompatibility in the handling of HMAC. The previous ABI has now been
2134 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2136 *) Malformed ECParameters causes infinite loop
2138 When processing an ECParameters structure OpenSSL enters an infinite loop
2139 if the curve specified is over a specially malformed binary polynomial
2142 This can be used to perform denial of service against any
2143 system which processes public keys, certificate requests or
2144 certificates. This includes TLS clients and TLS servers with
2145 client authentication enabled.
2147 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2151 *) Exploitable out-of-bounds read in X509_cmp_time
2153 X509_cmp_time does not properly check the length of the ASN1_TIME
2154 string and can read a few bytes out of bounds. In addition,
2155 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2158 An attacker can use this to craft malformed certificates and CRLs of
2159 various sizes and potentially cause a segmentation fault, resulting in
2160 a DoS on applications that verify certificates or CRLs. TLS clients
2161 that verify CRLs are affected. TLS clients and servers with client
2162 authentication enabled may be affected if they use custom verification
2165 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2166 independently by Hanno Böck.
2170 *) PKCS7 crash with missing EnvelopedContent
2172 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2173 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2174 with missing content and trigger a NULL pointer dereference on parsing.
2176 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2177 structures from untrusted sources are affected. OpenSSL clients and
2178 servers are not affected.
2180 This issue was reported to OpenSSL by Michal Zalewski (Google).
2184 *) CMS verify infinite loop with unknown hash function
2186 When verifying a signedData message the CMS code can enter an infinite loop
2187 if presented with an unknown hash function OID. This can be used to perform
2188 denial of service against any system which verifies signedData messages using
2190 This issue was reported to OpenSSL by Johannes Bauer.
2194 *) Race condition handling NewSessionTicket
2196 If a NewSessionTicket is received by a multi-threaded client when attempting to
2197 reuse a previous ticket then a race condition can occur potentially leading to
2198 a double free of the ticket data.
2202 *) Only support 256-bit or stronger elliptic curves with the
2203 'ecdh_auto' setting (server) or by default (client). Of supported
2204 curves, prefer P-256 (both).
2207 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2209 *) ClientHello sigalgs DoS fix
2211 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2212 invalid signature algorithms extension a NULL pointer dereference will
2213 occur. This can be exploited in a DoS attack against the server.
2215 This issue was was reported to OpenSSL by David Ramos of Stanford
2218 [Stephen Henson and Matt Caswell]
2220 *) Multiblock corrupted pointer fix
2222 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2223 feature only applies on 64 bit x86 architecture platforms that support AES
2224 NI instructions. A defect in the implementation of "multiblock" can cause
2225 OpenSSL's internal write buffer to become incorrectly set to NULL when
2226 using non-blocking IO. Typically, when the user application is using a
2227 socket BIO for writing, this will only result in a failed connection.
2228 However if some other BIO is used then it is likely that a segmentation
2229 fault will be triggered, thus enabling a potential DoS attack.
2231 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2235 *) Segmentation fault in DTLSv1_listen fix
2237 The DTLSv1_listen function is intended to be stateless and processes the
2238 initial ClientHello from many peers. It is common for user code to loop
2239 over the call to DTLSv1_listen until a valid ClientHello is received with
2240 an associated cookie. A defect in the implementation of DTLSv1_listen means
2241 that state is preserved in the SSL object from one invocation to the next
2242 that can lead to a segmentation fault. Errors processing the initial
2243 ClientHello can trigger this scenario. An example of such an error could be
2244 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2247 This issue was reported to OpenSSL by Per Allansson.
2251 *) Segmentation fault in ASN1_TYPE_cmp fix
2253 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2254 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2255 certificate signature algorithm consistency this can be used to crash any
2256 certificate verification operation and exploited in a DoS attack. Any
2257 application which performs certificate verification is vulnerable including
2258 OpenSSL clients and servers which enable client authentication.
2262 *) Segmentation fault for invalid PSS parameters fix
2264 The signature verification routines will crash with a NULL pointer
2265 dereference if presented with an ASN.1 signature using the RSA PSS
2266 algorithm and invalid parameters. Since these routines are used to verify
2267 certificate signature algorithms this can be used to crash any
2268 certificate verification operation and exploited in a DoS attack. Any
2269 application which performs certificate verification is vulnerable including
2270 OpenSSL clients and servers which enable client authentication.
2272 This issue was was reported to OpenSSL by Brian Carpenter.
2276 *) ASN.1 structure reuse memory corruption fix
2278 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2279 memory corruption via an invalid write. Such reuse is and has been
2280 strongly discouraged and is believed to be rare.
2282 Applications that parse structures containing CHOICE or ANY DEFINED BY
2283 components may be affected. Certificate parsing (d2i_X509 and related
2284 functions) are however not affected. OpenSSL clients and servers are
2289 *) PKCS7 NULL pointer dereferences fix
2291 The PKCS#7 parsing code does not handle missing outer ContentInfo
2292 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2293 missing content and trigger a NULL pointer dereference on parsing.
2295 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2296 otherwise parse PKCS#7 structures from untrusted sources are
2297 affected. OpenSSL clients and servers are not affected.
2299 This issue was reported to OpenSSL by Michal Zalewski (Google).
2303 *) DoS via reachable assert in SSLv2 servers fix
2305 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2306 servers that both support SSLv2 and enable export cipher suites by sending
2307 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2309 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2310 (OpenSSL development team).
2314 *) Empty CKE with client auth and DHE fix
2316 If client auth is used then a server can seg fault in the event of a DHE
2317 ciphersuite being selected and a zero length ClientKeyExchange message
2318 being sent by the client. This could be exploited in a DoS attack.
2322 *) Handshake with unseeded PRNG fix
2324 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2325 with an unseeded PRNG. The conditions are:
2326 - The client is on a platform where the PRNG has not been seeded
2327 automatically, and the user has not seeded manually
2328 - A protocol specific client method version has been used (i.e. not
2329 SSL_client_methodv23)
2330 - A ciphersuite is used that does not require additional random data from
2331 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2333 If the handshake succeeds then the client random that has been used will
2334 have been generated from a PRNG with insufficient entropy and therefore the
2335 output may be predictable.
2337 For example using the following command with an unseeded openssl will
2338 succeed on an unpatched platform:
2340 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2344 *) Use After Free following d2i_ECPrivatekey error fix
2346 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2347 could cause a use after free condition. This, in turn, could cause a double
2348 free in several private key parsing functions (such as d2i_PrivateKey
2349 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2350 for applications that receive EC private keys from untrusted
2351 sources. This scenario is considered rare.
2353 This issue was discovered by the BoringSSL project and fixed in their
2358 *) X509_to_X509_REQ NULL pointer deref fix
2360 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2361 the certificate key is invalid. This function is rarely used in practice.
2363 This issue was discovered by Brian Carpenter.
2367 *) Removed the export ciphers from the DEFAULT ciphers
2370 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2372 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2373 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2374 So far those who have to target multiple platforms would compromise
2375 and argue that binary targeting say ARMv5 would still execute on
2376 ARMv8. "Universal" build resolves this compromise by providing
2377 near-optimal performance even on newer platforms.
2380 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2381 (other platforms pending).
2382 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2384 *) Add support for the SignedCertificateTimestampList certificate and
2385 OCSP response extensions from RFC6962.
2388 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2389 for corner cases. (Certain input points at infinity could lead to
2390 bogus results, with non-infinity inputs mapped to infinity too.)
2393 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2394 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2395 common cases are optimized and there still is room for further
2396 improvements. Vector Permutation AES for Altivec is also added.
2399 *) Add support for little-endian ppc64 Linux target.
2400 [Marcelo Cerri (IBM)]
2402 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2403 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2404 are optimized and there still is room for further improvements.
2405 Both 32- and 64-bit modes are supported.
2406 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2408 *) Improved ARMv7 NEON support.
2411 *) Support for SPARC Architecture 2011 crypto extensions, first
2412 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2413 SHA256/512, MD5, GHASH and modular exponentiation.
2414 [Andy Polyakov, David Miller]
2416 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2418 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2420 *) Support for new and upcoming Intel processors, including AVX2,
2421 BMI and SHA ISA extensions. This includes additional "stitched"
2422 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2425 This work was sponsored by Intel Corp.
2428 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2429 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2430 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2433 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2434 this fixes a limitation in previous versions of OpenSSL.
2437 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2438 MGF1 digest and OAEP label.
2441 *) Add EVP support for key wrapping algorithms, to avoid problems with
2442 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2443 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2444 algorithms and include tests cases.
2447 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2449 [Douglas E. Engert, Steve Henson]
2451 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2452 difference in days and seconds between two tm or ASN1_TIME structures.
2455 *) Add -rev test option to s_server to just reverse order of characters
2456 received by client and send back to server. Also prints an abbreviated
2457 summary of the connection parameters.
2460 *) New option -brief for s_client and s_server to print out a brief summary
2461 of connection parameters.
2464 *) Add callbacks for arbitrary TLS extensions.
2465 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2467 *) New option -crl_download in several openssl utilities to download CRLs
2468 from CRLDP extension in certificates.
2471 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2474 *) New function X509_CRL_diff to generate a delta CRL from the difference
2475 of two full CRLs. Add support to "crl" utility.
2478 *) New functions to set lookup_crls function and to retrieve
2479 X509_STORE from X509_STORE_CTX.
2482 *) Print out deprecated issuer and subject unique ID fields in
2486 *) Extend OCSP I/O functions so they can be used for simple general purpose
2487 HTTP as well as OCSP. New wrapper function which can be used to download
2488 CRLs using the OCSP API.
2491 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2494 *) SSL_CONF* functions. These provide a common framework for application
2495 configuration using configuration files or command lines.
2498 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2499 message callback and prints the results. Needs compile time option
2500 "enable-ssl-trace". New options to s_client and s_server to enable
2504 *) New ctrl and macro to retrieve supported points extensions.
2505 Print out extension in s_server and s_client.
2508 *) New functions to retrieve certificate signature and signature
2512 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2516 *) New Suite B modes for TLS code. These use and enforce the requirements
2517 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2518 only use Suite B curves. The Suite B modes can be set by using the
2519 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2522 *) New chain verification flags for Suite B levels of security. Check
2523 algorithms are acceptable when flags are set in X509_verify_cert.
2526 *) Make tls1_check_chain return a set of flags indicating checks passed
2527 by a certificate chain. Add additional tests to handle client
2528 certificates: checks for matching certificate type and issuer name
2532 *) If an attempt is made to use a signature algorithm not in the peer
2533 preference list abort the handshake. If client has no suitable
2534 signature algorithms in response to a certificate request do not
2535 use the certificate.
2538 *) If server EC tmp key is not in client preference list abort handshake.
2541 *) Add support for certificate stores in CERT structure. This makes it
2542 possible to have different stores per SSL structure or one store in
2543 the parent SSL_CTX. Include distinct stores for certificate chain
2544 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2545 to build and store a certificate chain in CERT structure: returning
2546 an error if the chain cannot be built: this will allow applications
2547 to test if a chain is correctly configured.
2549 Note: if the CERT based stores are not set then the parent SSL_CTX
2550 store is used to retain compatibility with existing behaviour.
2554 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2555 mask based on the current session, check mask when sending client
2556 hello and checking the requested ciphersuite.
2559 *) New ctrls to retrieve and set certificate types in a certificate
2560 request message. Print out received values in s_client. If certificate
2561 types is not set with custom values set sensible values based on
2562 supported signature algorithms.
2565 *) Support for distinct client and server supported signature algorithms.
2568 *) Add certificate callback. If set this is called whenever a certificate
2569 is required by client or server. An application can decide which
2570 certificate chain to present based on arbitrary criteria: for example
2571 supported signature algorithms. Add very simple example to s_server.
2572 This fixes many of the problems and restrictions of the existing client
2573 certificate callback: for example you can now clear an existing
2574 certificate and specify the whole chain.
2577 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2578 the certificate can be used for (if anything). Set valid_flags field
2579 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2580 to have similar checks in it.
2582 Add new "cert_flags" field to CERT structure and include a "strict mode".
2583 This enforces some TLS certificate requirements (such as only permitting
2584 certificate signature algorithms contained in the supported algorithms
2585 extension) which some implementations ignore: this option should be used
2586 with caution as it could cause interoperability issues.
2589 *) Update and tidy signature algorithm extension processing. Work out
2590 shared signature algorithms based on preferences and peer algorithms
2591 and print them out in s_client and s_server. Abort handshake if no
2592 shared signature algorithms.
2595 *) Add new functions to allow customised supported signature algorithms
2596 for SSL and SSL_CTX structures. Add options to s_client and s_server
2600 *) New function SSL_certs_clear() to delete all references to certificates
2601 from an SSL structure. Before this once a certificate had been added
2602 it couldn't be removed.
2605 *) Integrate hostname, email address and IP address checking with certificate
2606 verification. New verify options supporting checking in openssl utility.
2609 *) Fixes and wildcard matching support to hostname and email checking
2610 functions. Add manual page.
2611 [Florian Weimer (Red Hat Product Security Team)]
2613 *) New functions to check a hostname email or IP address against a
2614 certificate. Add options x509 utility to print results of checks against
2618 *) Fix OCSP checking.
2619 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2621 *) Initial experimental support for explicitly trusted non-root CAs.
2622 OpenSSL still tries to build a complete chain to a root but if an
2623 intermediate CA has a trust setting included that is used. The first
2624 setting is used: whether to trust (e.g., -addtrust option to the x509
2628 *) Add -trusted_first option which attempts to find certificates in the
2629 trusted store even if an untrusted chain is also supplied.
2632 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2633 platform support for Linux and Android.
2636 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2639 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2640 When in FIPS mode the approved implementations are used as normal,
2641 when not in FIPS mode the internal unapproved versions are used instead.
2642 This means that the FIPS capable OpenSSL isn't forced to use the
2643 (often lower performance) FIPS implementations outside FIPS mode.
2646 *) Transparently support X9.42 DH parameters when calling
2647 PEM_read_bio_DHparameters. This means existing applications can handle
2648 the new parameter format automatically.
2651 *) Initial experimental support for X9.42 DH parameter format: mainly
2652 to support use of 'q' parameter for RFC5114 parameters.
2655 *) Add DH parameters from RFC5114 including test data to dhtest.
2658 *) Support for automatic EC temporary key parameter selection. If enabled
2659 the most preferred EC parameters are automatically used instead of
2660 hardcoded fixed parameters. Now a server just has to call:
2661 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2662 support ECDH and use the most appropriate parameters.
2665 *) Enhance and tidy EC curve and point format TLS extension code. Use
2666 static structures instead of allocation if default values are used.
2667 New ctrls to set curves we wish to support and to retrieve shared curves.
2668 Print out shared curves in s_server. New options to s_server and s_client
2669 to set list of supported curves.
2672 *) New ctrls to retrieve supported signature algorithms and
2673 supported curve values as an array of NIDs. Extend openssl utility
2674 to print out received values.
2677 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2678 between NIDs and the more common NIST names such as "P-256". Enhance
2679 ecparam utility and ECC method to recognise the NIST names for curves.
2682 *) Enhance SSL/TLS certificate chain handling to support different
2683 chains for each certificate instead of one chain in the parent SSL_CTX.
2686 *) Support for fixed DH ciphersuite client authentication: where both
2687 server and client use DH certificates with common parameters.
2690 *) Support for fixed DH ciphersuites: those requiring DH server
2694 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2696 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2697 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2698 X509_CINF_get_signature were reverted post internal team review.
2700 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2702 *) Build fixes for the Windows and OpenVMS platforms
2703 [Matt Caswell and Richard Levitte]
2705 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2707 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2708 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2709 dereference. This could lead to a Denial Of Service attack. Thanks to
2710 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2714 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2715 dtls1_buffer_record function under certain conditions. In particular this
2716 could occur if an attacker sent repeated DTLS records with the same
2717 sequence number but for the next epoch. The memory leak could be exploited
2718 by an attacker in a Denial of Service attack through memory exhaustion.
2719 Thanks to Chris Mueller for reporting this issue.
2723 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2724 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2725 method would be set to NULL which could later result in a NULL pointer
2726 dereference. Thanks to Frank Schmirler for reporting this issue.
2730 *) Abort handshake if server key exchange message is omitted for ephemeral
2733 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2734 reporting this issue.
2738 *) Remove non-export ephemeral RSA code on client and server. This code
2739 violated the TLS standard by allowing the use of temporary RSA keys in
2740 non-export ciphersuites and could be used by a server to effectively
2741 downgrade the RSA key length used to a value smaller than the server
2742 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2743 INRIA or reporting this issue.
2747 *) Fixed issue where DH client certificates are accepted without verification.
2748 An OpenSSL server will accept a DH certificate for client authentication
2749 without the certificate verify message. This effectively allows a client to
2750 authenticate without the use of a private key. This only affects servers
2751 which trust a client certificate authority which issues certificates
2752 containing DH keys: these are extremely rare and hardly ever encountered.
2753 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2758 *) Ensure that the session ID context of an SSL is updated when its
2759 SSL_CTX is updated via SSL_set_SSL_CTX.
2761 The session ID context is typically set from the parent SSL_CTX,
2762 and can vary with the CTX.
2765 *) Fix various certificate fingerprint issues.
2767 By using non-DER or invalid encodings outside the signed portion of a
2768 certificate the fingerprint can be changed without breaking the signature.
2769 Although no details of the signed portion of the certificate can be changed
2770 this can cause problems with some applications: e.g. those using the
2771 certificate fingerprint for blacklists.
2773 1. Reject signatures with non zero unused bits.
2775 If the BIT STRING containing the signature has non zero unused bits reject
2776 the signature. All current signature algorithms require zero unused bits.
2778 2. Check certificate algorithm consistency.
2780 Check the AlgorithmIdentifier inside TBS matches the one in the
2781 certificate signature. NB: this will result in signature failure
2782 errors for some broken certificates.
2784 Thanks to Konrad Kraszewski from Google for reporting this issue.
2786 3. Check DSA/ECDSA signatures use DER.
2788 Re-encode DSA/ECDSA signatures and compare with the original received
2789 signature. Return an error if there is a mismatch.
2791 This will reject various cases including garbage after signature
2792 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2793 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2794 (negative or with leading zeroes).
2796 Further analysis was conducted and fixes were developed by Stephen Henson
2797 of the OpenSSL core team.
2802 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2803 results on some platforms, including x86_64. This bug occurs at random
2804 with a very low probability, and is not known to be exploitable in any
2805 way, though its exact impact is difficult to determine. Thanks to Pieter
2806 Wuille (Blockstream) who reported this issue and also suggested an initial
2807 fix. Further analysis was conducted by the OpenSSL development team and
2808 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2809 the OpenSSL core team.
2813 *) Do not resume sessions on the server if the negotiated protocol
2814 version does not match the session's version. Resuming with a different
2815 version, while not strictly forbidden by the RFC, is of questionable
2816 sanity and breaks all known clients.
2817 [David Benjamin, Emilia Käsper]
2819 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2820 early CCS messages during renegotiation. (Note that because
2821 renegotiation is encrypted, this early CCS was not exploitable.)
2824 *) Tighten client-side session ticket handling during renegotiation:
2825 ensure that the client only accepts a session ticket if the server sends
2826 the extension anew in the ServerHello. Previously, a TLS client would
2827 reuse the old extension state and thus accept a session ticket if one was
2828 announced in the initial ServerHello.
2830 Similarly, ensure that the client requires a session ticket if one
2831 was advertised in the ServerHello. Previously, a TLS client would
2832 ignore a missing NewSessionTicket message.
2835 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2837 *) SRTP Memory Leak.
2839 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2840 sends a carefully crafted handshake message, to cause OpenSSL to fail
2841 to free up to 64k of memory causing a memory leak. This could be
2842 exploited in a Denial Of Service attack. This issue affects OpenSSL
2843 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2844 whether SRTP is used or configured. Implementations of OpenSSL that
2845 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2847 The fix was developed by the OpenSSL team.
2851 *) Session Ticket Memory Leak.
2853 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2854 integrity of that ticket is first verified. In the event of a session
2855 ticket integrity check failing, OpenSSL will fail to free memory
2856 causing a memory leak. By sending a large number of invalid session
2857 tickets an attacker could exploit this issue in a Denial Of Service
2862 *) Build option no-ssl3 is incomplete.
2864 When OpenSSL is configured with "no-ssl3" as a build option, servers
2865 could accept and complete a SSL 3.0 handshake, and clients could be
2866 configured to send them.
2868 [Akamai and the OpenSSL team]
2870 *) Add support for TLS_FALLBACK_SCSV.
2871 Client applications doing fallback retries should call
2872 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2874 [Adam Langley, Bodo Moeller]
2876 *) Add additional DigestInfo checks.
2878 Re-encode DigestInto in DER and check against the original when
2879 verifying RSA signature: this will reject any improperly encoded
2880 DigestInfo structures.
2882 Note: this is a precautionary measure and no attacks are currently known.
2886 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2888 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2889 SRP code can be overrun an internal buffer. Add sanity check that
2890 g, A, B < N to SRP code.
2892 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2893 Group for discovering this issue.
2897 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2898 TLS 1.0 instead of higher protocol versions when the ClientHello message
2899 is badly fragmented. This allows a man-in-the-middle attacker to force a
2900 downgrade to TLS 1.0 even if both the server and the client support a
2901 higher protocol version, by modifying the client's TLS records.
2903 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2904 researching this issue.
2908 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2909 to a denial of service attack. A malicious server can crash the client
2910 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2911 ciphersuite and sending carefully crafted handshake messages.
2913 Thanks to Felix Gröbert (Google) for discovering and researching this
2918 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2919 to leak memory. This can be exploited through a Denial of Service attack.
2920 Thanks to Adam Langley for discovering and researching this issue.
2924 *) An attacker can force openssl to consume large amounts of memory whilst
2925 processing DTLS handshake messages. This can be exploited through a
2926 Denial of Service attack.
2927 Thanks to Adam Langley for discovering and researching this issue.
2931 *) An attacker can force an error condition which causes openssl to crash
2932 whilst processing DTLS packets due to memory being freed twice. This
2933 can be exploited through a Denial of Service attack.
2934 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2939 *) If a multithreaded client connects to a malicious server using a resumed
2940 session and the server sends an ec point format extension it could write
2941 up to 255 bytes to freed memory.
2943 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2948 *) A malicious server can crash an OpenSSL client with a null pointer
2949 dereference (read) by specifying an SRP ciphersuite even though it was not
2950 properly negotiated with the client. This can be exploited through a
2951 Denial of Service attack.
2953 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2954 discovering and researching this issue.
2958 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2959 X509_name_oneline, X509_name_print_ex et al. to leak some information
2960 from the stack. Applications may be affected if they echo pretty printing
2961 output to the attacker.
2963 Thanks to Ivan Fratric (Google) for discovering this issue.
2965 [Emilia Käsper, and Steve Henson]
2967 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2968 for corner cases. (Certain input points at infinity could lead to
2969 bogus results, with non-infinity inputs mapped to infinity too.)
2972 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2974 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2975 handshake can force the use of weak keying material in OpenSSL
2976 SSL/TLS clients and servers.
2978 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2979 researching this issue. (CVE-2014-0224)
2980 [KIKUCHI Masashi, Steve Henson]
2982 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2983 OpenSSL DTLS client the code can be made to recurse eventually crashing
2986 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2988 [Imre Rad, Steve Henson]
2990 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2991 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2992 client or server. This is potentially exploitable to run arbitrary
2993 code on a vulnerable client or server.
2995 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2996 [Jüri Aedla, Steve Henson]
2998 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2999 are subject to a denial of service attack.
3001 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3002 this issue. (CVE-2014-3470)
3003 [Felix Gröbert, Ivan Fratric, Steve Henson]
3005 *) Harmonize version and its documentation. -f flag is used to display
3007 [mancha <mancha1@zoho.com>]
3009 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3010 in i2d_ECPrivateKey.
3011 [mancha <mancha1@zoho.com>]
3013 *) Fix some double frees. These are not thought to be exploitable.
3014 [mancha <mancha1@zoho.com>]
3016 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3018 *) A missing bounds check in the handling of the TLS heartbeat extension
3019 can be used to reveal up to 64k of memory to a connected client or
3022 Thanks for Neel Mehta of Google Security for discovering this bug and to
3023 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3024 preparing the fix (CVE-2014-0160)
3025 [Adam Langley, Bodo Moeller]
3027 *) Fix for the attack described in the paper "Recovering OpenSSL
3028 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3029 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3030 http://eprint.iacr.org/2014/140
3032 Thanks to Yuval Yarom and Naomi Benger for discovering this
3033 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3034 [Yuval Yarom and Naomi Benger]
3036 *) TLS pad extension: draft-agl-tls-padding-03
3038 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3039 TLS client Hello record length value would otherwise be > 255 and
3040 less that 512 pad with a dummy extension containing zeroes so it
3041 is at least 512 bytes long.
3043 [Adam Langley, Steve Henson]
3045 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3047 *) Fix for TLS record tampering bug. A carefully crafted invalid
3048 handshake could crash OpenSSL with a NULL pointer exception.
3049 Thanks to Anton Johansson for reporting this issues.
3052 *) Keep original DTLS digest and encryption contexts in retransmission
3053 structures so we can use the previous session parameters if they need
3054 to be resent. (CVE-2013-6450)
3057 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3058 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3059 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3060 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3061 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3062 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3063 [Rob Stradling, Adam Langley]
3065 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3067 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3068 supporting platforms or when small records were transferred.
3069 [Andy Polyakov, Steve Henson]
3071 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3073 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3075 This addresses the flaw in CBC record processing discovered by
3076 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3077 at: http://www.isg.rhul.ac.uk/tls/
3079 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3080 Security Group at Royal Holloway, University of London
3081 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3082 Emilia Käsper for the initial patch.
3084 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3086 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3087 ciphersuites which can be exploited in a denial of service attack.
3088 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3089 and detecting this bug and to Wolfgang Ettlinger
3090 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3094 *) Return an error when checking OCSP signatures when key is NULL.
3095 This fixes a DoS attack. (CVE-2013-0166)
3098 *) Make openssl verify return errors.
3099 [Chris Palmer <palmer@google.com> and Ben Laurie]
3101 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3102 the right response is stapled. Also change SSL_get_certificate()
3103 so it returns the certificate actually sent.
3104 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3105 [Rob Stradling <rob.stradling@comodo.com>]
3107 *) Fix possible deadlock when decoding public keys.
3110 *) Don't use TLS 1.0 record version number in initial client hello
3114 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3116 *) Sanity check record length before skipping explicit IV in TLS
3117 1.2, 1.1 and DTLS to fix DoS attack.
3119 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3120 fuzzing as a service testing platform.
3124 *) Initialise tkeylen properly when encrypting CMS messages.
3125 Thanks to Solar Designer of Openwall for reporting this issue.
3128 *) In FIPS mode don't try to use composite ciphers as they are not
3132 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3134 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3135 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3136 mean any application compiled against OpenSSL 1.0.0 headers setting
3137 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3138 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3139 0x10000000L Any application which was previously compiled against
3140 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3141 will need to be recompiled as a result. Letting be results in
3142 inability to disable specifically TLS 1.1 and in client context,
3143 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3146 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3147 disable just protocol X, but all protocols above X *if* there are
3148 protocols *below* X still enabled. In more practical terms it means
3149 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3150 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3151 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3155 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3157 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3158 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3159 in CRYPTO_realloc_clean.
3161 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3162 issue and to Adam Langley <agl@chromium.org> for fixing it.
3164 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3166 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3169 *) Workarounds for some broken servers that "hang" if a client hello
3170 record length exceeds 255 bytes.
3172 1. Do not use record version number > TLS 1.0 in initial client
3173 hello: some (but not all) hanging servers will now work.
3174 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3175 the number of ciphers sent in the client hello. This should be
3176 set to an even number, such as 50, for example by passing:
3177 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3178 Most broken servers should now work.
3179 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3180 TLS 1.2 client support entirely.
3183 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3186 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3188 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3189 STRING form instead of a DigestInfo.
3192 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3193 and the RSA_sign/RSA_verify functions. This was made more apparent when
3194 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3195 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3196 the correct format in RSA_verify so both forms transparently work.
3199 *) Some servers which support TLS 1.0 can choke if we initially indicate
3200 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3201 encrypted premaster secret. As a workaround use the maximum permitted
3202 client version in client hello, this should keep such servers happy
3203 and still work with previous versions of OpenSSL.
3206 *) Add support for TLS/DTLS heartbeats.
3207 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3209 *) Add support for SCTP.
3210 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3212 *) Improved PRNG seeding for VOS.
3213 [Paul Green <Paul.Green@stratus.com>]
3215 *) Extensive assembler packs updates, most notably:
3217 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3218 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3219 - x86_64: bit-sliced AES implementation;
3220 - ARM: NEON support, contemporary platforms optimizations;
3221 - s390x: z196 support;
3222 - *: GHASH and GF(2^m) multiplication implementations;
3226 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3227 (removal of unnecessary code)
3228 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3230 *) Add TLS key material exporter from RFC 5705.
3233 *) Add DTLS-SRTP negotiation from RFC 5764.
3236 *) Add Next Protocol Negotiation,
3237 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3238 disabled with a no-npn flag to config or Configure. Code donated
3240 [Adam Langley <agl@google.com> and Ben Laurie]
3242 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3243 NIST-P256, NIST-P521, with constant-time single point multiplication on
3244 typical inputs. Compiler support for the nonstandard type __uint128_t is
3245 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3246 Code made available under Apache License version 2.0.
3248 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3249 line to include this in your build of OpenSSL, and run "make depend" (or
3250 "make update"). This enables the following EC_METHODs:
3252 EC_GFp_nistp224_method()
3253 EC_GFp_nistp256_method()
3254 EC_GFp_nistp521_method()
3256 EC_GROUP_new_by_curve_name() will automatically use these (while
3257 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3259 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3261 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3262 all platforms. Move ssize_t definition from e_os.h to the public
3263 header file e_os2.h as it now appears in public header file cms.h
3266 *) New -sigopt option to the ca, req and x509 utilities. Additional
3267 signature parameters can be passed using this option and in
3271 *) Add RSA PSS signing function. This will generate and set the
3272 appropriate AlgorithmIdentifiers for PSS based on those in the
3273 corresponding EVP_MD_CTX structure. No application support yet.
3276 *) Support for companion algorithm specific ASN1 signing routines.
3277 New function ASN1_item_sign_ctx() signs a pre-initialised
3278 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3279 the appropriate parameters.
3282 *) Add new algorithm specific ASN1 verification initialisation function
3283 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3284 handling will be the same no matter what EVP_PKEY_METHOD is used.
3285 Add a PSS handler to support verification of PSS signatures: checked
3286 against a number of sample certificates.
3289 *) Add signature printing for PSS. Add PSS OIDs.
3290 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3292 *) Add algorithm specific signature printing. An individual ASN1 method
3293 can now print out signatures instead of the standard hex dump.
3295 More complex signatures (e.g. PSS) can print out more meaningful
3296 information. Include DSA version that prints out the signature
3300 *) Password based recipient info support for CMS library: implementing
3304 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3305 neatly separates the code into cipher and PBE sections and is required
3306 for some algorithms that split PBES2 into separate pieces (such as
3307 password based CMS).
3310 *) Session-handling fixes:
3311 - Fix handling of connections that are resuming with a session ID,
3312 but also support Session Tickets.
3313 - Fix a bug that suppressed issuing of a new ticket if the client
3314 presented a ticket with an expired session.
3315 - Try to set the ticket lifetime hint to something reasonable.
3316 - Make tickets shorter by excluding irrelevant information.
3317 - On the client side, don't ignore renewed tickets.
3318 [Adam Langley, Bodo Moeller (Google)]
3320 *) Fix PSK session representation.
3323 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3325 This work was sponsored by Intel.
3328 *) Add GCM support to TLS library. Some custom code is needed to split
3329 the IV between the fixed (from PRF) and explicit (from TLS record)
3330 portions. This adds all GCM ciphersuites supported by RFC5288 and
3331 RFC5289. Generalise some AES* cipherstrings to include GCM and
3332 add a special AESGCM string for GCM only.
3335 *) Expand range of ctrls for AES GCM. Permit setting invocation
3336 field on decrypt and retrieval of invocation field only on encrypt.
3339 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3340 As required by RFC5289 these ciphersuites cannot be used if for
3341 versions of TLS earlier than 1.2.
3344 *) For FIPS capable OpenSSL interpret a NULL default public key method
3345 as unset and return the appropriate default but do *not* set the default.
3346 This means we can return the appropriate method in applications that
3347 switch between FIPS and non-FIPS modes.
3350 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3351 ENGINE is used then we cannot handle that in the FIPS module so we
3352 keep original code iff non-FIPS operations are allowed.
3355 *) Add -attime option to openssl utilities.
3356 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3358 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3361 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3362 FIPS EC methods unconditionally for now.
3365 *) New build option no-ec2m to disable characteristic 2 code.
3368 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3369 all cases can be covered as some introduce binary incompatibilities.
3372 *) Redirect RSA operations to FIPS module including keygen,
3373 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3376 *) Add similar low level API blocking to ciphers.
3379 *) Low level digest APIs are not approved in FIPS mode: any attempt
3380 to use these will cause a fatal error. Applications that *really* want
3381 to use them can use the private_* version instead.
3384 *) Redirect cipher operations to FIPS module for FIPS builds.
3387 *) Redirect digest operations to FIPS module for FIPS builds.
3390 *) Update build system to add "fips" flag which will link in fipscanister.o
3391 for static and shared library builds embedding a signature if needed.
3394 *) Output TLS supported curves in preference order instead of numerical
3395 order. This is currently hardcoded for the highest order curves first.
3396 This should be configurable so applications can judge speed vs strength.
3399 *) Add TLS v1.2 server support for client authentication.
3402 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3406 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3407 FIPS modules versions.
3410 *) Add TLS v1.2 client side support for client authentication. Keep cache
3411 of handshake records longer as we don't know the hash algorithm to use
3412 until after the certificate request message is received.
3415 *) Initial TLS v1.2 client support. Add a default signature algorithms
3416 extension including all the algorithms we support. Parse new signature
3417 format in client key exchange. Relax some ECC signing restrictions for
3418 TLS v1.2 as indicated in RFC5246.
3421 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3422 to new signature format when needed using client digest preference.
3423 All server ciphersuites should now work correctly in TLS v1.2. No client
3424 support yet and no support for client certificates.
3427 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3428 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3429 ciphersuites. At present only RSA key exchange ciphersuites work with
3430 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3431 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3432 and version checking.
3435 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3436 with this defined it will not be affected by any changes to ssl internal
3437 structures. Add several utility functions to allow openssl application
3438 to work with OPENSSL_NO_SSL_INTERN defined.
3441 *) A long standing patch to add support for SRP from EdelWeb (Peter
3442 Sylvester and Christophe Renou) was integrated.
3443 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3444 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3447 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3450 *) Permit abbreviated handshakes when renegotiating using the function
3451 SSL_renegotiate_abbreviated().
3452 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3454 *) Add call to ENGINE_register_all_complete() to
3455 ENGINE_load_builtin_engines(), so some implementations get used
3456 automatically instead of needing explicit application support.
3459 *) Add support for TLS key exporter as described in RFC5705.
3460 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3462 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3463 a few changes are required:
3465 Add SSL_OP_NO_TLSv1_1 flag.
3466 Add TLSv1_1 methods.
3467 Update version checking logic to handle version 1.1.
3468 Add explicit IV handling (ported from DTLS code).
3469 Add command line options to s_client/s_server.
3472 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3474 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3475 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3476 content decryption and always return the same error. Note: this attack
3477 needs on average 2^20 messages so it only affects automated senders. The
3478 old behaviour can be re-enabled in the CMS code by setting the
3479 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3480 an MMA defence is not necessary.
3481 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3482 this issue. (CVE-2012-0884)
3485 *) Fix CVE-2011-4619: make sure we really are receiving a
3486 client hello before rejecting multiple SGC restarts. Thanks to
3487 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3490 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3492 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3493 Thanks to Antonio Martin, Enterprise Secure Access Research and
3494 Development, Cisco Systems, Inc. for discovering this bug and
3495 preparing a fix. (CVE-2012-0050)
3498 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3500 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3501 of the Vaudenay padding oracle attack on CBC mode encryption
3502 which enables an efficient plaintext recovery attack against
3503 the OpenSSL implementation of DTLS. Their attack exploits timing
3504 differences arising during decryption processing. A research
3505 paper describing this attack can be found at:
3506 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3507 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3508 Security Group at Royal Holloway, University of London
3509 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3510 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3511 for preparing the fix. (CVE-2011-4108)
3512 [Robin Seggelmann, Michael Tuexen]
3514 *) Clear bytes used for block padding of SSL 3.0 records.
3516 [Adam Langley (Google)]
3518 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3519 Kadianakis <desnacked@gmail.com> for discovering this issue and
3520 Adam Langley for preparing the fix. (CVE-2011-4619)
3521 [Adam Langley (Google)]
3523 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3524 [Andrey Kulikov <amdeich@gmail.com>]
3526 *) Prevent malformed RFC3779 data triggering an assertion failure.
3527 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3528 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3529 [Rob Austein <sra@hactrn.net>]
3531 *) Improved PRNG seeding for VOS.
3532 [Paul Green <Paul.Green@stratus.com>]
3534 *) Fix ssl_ciph.c set-up race.
3535 [Adam Langley (Google)]
3537 *) Fix spurious failures in ecdsatest.c.
3538 [Emilia Käsper (Google)]
3540 *) Fix the BIO_f_buffer() implementation (which was mixing different
3541 interpretations of the '..._len' fields).
3542 [Adam Langley (Google)]
3544 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3545 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3546 threads won't reuse the same blinding coefficients.
3548 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3549 lock to call BN_BLINDING_invert_ex, and avoids one use of
3550 BN_BLINDING_update for each BN_BLINDING structure (previously,
3551 the last update always remained unused).
3552 [Emilia Käsper (Google)]
3554 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3555 [Bob Buckholz (Google)]
3557 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3559 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3560 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3561 [Kaspar Brand <ossl@velox.ch>]
3563 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3564 for multi-threaded use of ECDH. (CVE-2011-3210)
3565 [Adam Langley (Google)]
3567 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3570 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3571 signature public key algorithm by using OID xref utilities instead.
3572 Before this you could only use some ECC ciphersuites with SHA1 only.
3575 *) Add protection against ECDSA timing attacks as mentioned in the paper
3576 by Billy Bob Brumley and Nicola Tuveri, see:
3578 http://eprint.iacr.org/2011/232.pdf
3580 [Billy Bob Brumley and Nicola Tuveri]
3582 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3584 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3585 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3587 *) Fix bug in string printing code: if *any* escaping is enabled we must
3588 escape the escape character (backslash) or the resulting string is
3592 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3594 *) Disable code workaround for ancient and obsolete Netscape browsers
3595 and servers: an attacker can use it in a ciphersuite downgrade attack.
3596 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3599 *) Fixed J-PAKE implementation error, originally discovered by
3600 Sebastien Martini, further info and confirmation from Stefan
3601 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3604 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3606 *) Fix extension code to avoid race conditions which can result in a buffer
3607 overrun vulnerability: resumed sessions must not be modified as they can
3608 be shared by multiple threads. CVE-2010-3864
3611 *) Fix WIN32 build system to correctly link an ENGINE directory into
3615 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3617 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3619 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3621 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3623 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3624 context. The operation can be customised via the ctrl mechanism in
3625 case ENGINEs want to include additional functionality.
3628 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3631 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3632 output hashes compatible with older versions of OpenSSL.
3633 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3635 *) Fix compression algorithm handling: if resuming a session use the
3636 compression algorithm of the resumed session instead of determining
3637 it from client hello again. Don't allow server to change algorithm.
3640 *) Add load_crls() function to apps tidying load_certs() too. Add option
3641 to verify utility to allow additional CRLs to be included.
3644 *) Update OCSP request code to permit adding custom headers to the request:
3645 some responders need this.
3648 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3650 [Julia Lawall <julia@diku.dk>]
3652 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3653 needlessly dereferenced structures, used obsolete functions and
3654 didn't handle all updated verify codes correctly.
3657 *) Disable MD2 in the default configuration.
3660 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3661 indicate the initial BIO being pushed or popped. This makes it possible
3662 to determine whether the BIO is the one explicitly called or as a result
3663 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3664 it handles reference counts correctly and doesn't zero out the I/O bio
3665 when it is not being explicitly popped. WARNING: applications which
3666 included workarounds for the old buggy behaviour will need to be modified
3667 or they could free up already freed BIOs.
3670 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3671 renaming to all platforms (within the 0.9.8 branch, this was
3672 done conditionally on Netware platforms to avoid a name clash).
3673 [Guenter <lists@gknw.net>]
3675 *) Add ECDHE and PSK support to DTLS.
3676 [Michael Tuexen <tuexen@fh-muenster.de>]
3678 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3682 *) Add "missing" function EVP_MD_flags() (without this the only way to
3683 retrieve a digest flags is by accessing the structure directly. Update
3684 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3685 or cipher is registered as in the "from" argument. Print out all
3686 registered digests in the dgst usage message instead of manually
3687 attempting to work them out.
3690 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3691 this allows the use of compression and extensions. Change default cipher
3692 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3693 by default unless an application cipher string requests it.
3696 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3697 key ids to find matching certificates and keys but some PKCS#12 files
3698 don't follow the (somewhat unwritten) rules and this strategy fails.
3699 Now just gather all certificates together and the first private key
3700 then look for the first certificate that matches the key.
3703 *) Support use of registered digest and cipher names for dgst and cipher
3704 commands instead of having to add each one as a special case. So now
3711 openssl dgst -sha256 foo
3713 and this works for ENGINE based algorithms too.
3717 *) Update Gost ENGINE to support parameter files.
3718 [Victor B. Wagner <vitus@cryptocom.ru>]
3720 *) Support GeneralizedTime in ca utility.
3721 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3723 *) Enhance the hash format used for certificate directory links. The new
3724 form uses the canonical encoding (meaning equivalent names will work
3725 even if they aren't identical) and uses SHA1 instead of MD5. This form
3726 is incompatible with the older format and as a result c_rehash should
3727 be used to rebuild symbolic links.
3730 *) Make PKCS#8 the default write format for private keys, replacing the
3731 traditional format. This form is standardised, more secure and doesn't
3732 include an implicit MD5 dependency.
3735 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3736 committed to OpenSSL should pass this lot as a minimum.
3739 *) Add session ticket override functionality for use by EAP-FAST.
3740 [Jouni Malinen <j@w1.fi>]
3742 *) Modify HMAC functions to return a value. Since these can be implemented
3743 in an ENGINE errors can occur.
3746 *) Type-checked OBJ_bsearch_ex.
3749 *) Type-checked OBJ_bsearch. Also some constification necessitated
3750 by type-checking. Still to come: TXT_DB, bsearch(?),
3751 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3755 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3756 seconds to a tm structure directly, instead of going through OS
3757 specific date routines. This avoids any issues with OS routines such
3758 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3759 and X509_time_adj_ex() to cover the extended range. The existing
3760 X509_time_adj() is still usable and will no longer have any date issues.
3763 *) Delta CRL support. New use deltas option which will attempt to locate
3764 and search any appropriate delta CRLs available.
3766 This work was sponsored by Google.
3769 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3770 code and add additional score elements. Validate alternate CRL paths
3771 as part of the CRL checking and indicate a new error "CRL path validation
3772 error" in this case. Applications wanting additional details can use
3773 the verify callback and check the new "parent" field. If this is not
3774 NULL CRL path validation is taking place. Existing applications won't
3775 see this because it requires extended CRL support which is off by
3778 This work was sponsored by Google.
3781 *) Support for freshest CRL extension.
3783 This work was sponsored by Google.
3786 *) Initial indirect CRL support. Currently only supported in the CRLs
3787 passed directly and not via lookup. Process certificate issuer
3788 CRL entry extension and lookup CRL entries by bother issuer name
3789 and serial number. Check and process CRL issuer entry in IDP extension.
3791 This work was sponsored by Google.
3794 *) Add support for distinct certificate and CRL paths. The CRL issuer
3795 certificate is validated separately in this case. Only enabled if
3796 an extended CRL support flag is set: this flag will enable additional
3797 CRL functionality in future.
3799 This work was sponsored by Google.
3802 *) Add support for policy mappings extension.
3804 This work was sponsored by Google.
3807 *) Fixes to pathlength constraint, self issued certificate handling,
3808 policy processing to align with RFC3280 and PKITS tests.
3810 This work was sponsored by Google.
3813 *) Support for name constraints certificate extension. DN, email, DNS
3814 and URI types are currently supported.
3816 This work was sponsored by Google.
3819 *) To cater for systems that provide a pointer-based thread ID rather
3820 than numeric, deprecate the current numeric thread ID mechanism and
3821 replace it with a structure and associated callback type. This
3822 mechanism allows a numeric "hash" to be extracted from a thread ID in
3823 either case, and on platforms where pointers are larger than 'long',
3824 mixing is done to help ensure the numeric 'hash' is usable even if it
3825 can't be guaranteed unique. The default mechanism is to use "&errno"
3826 as a pointer-based thread ID to distinguish between threads.
3828 Applications that want to provide their own thread IDs should now use
3829 CRYPTO_THREADID_set_callback() to register a callback that will call
3830 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3832 Note that ERR_remove_state() is now deprecated, because it is tied
3833 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3834 to free the current thread's error state should be replaced by
3835 ERR_remove_thread_state(NULL).
3837 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3838 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3839 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3840 application was previously providing a numeric thread callback that
3841 was inappropriate for distinguishing threads, then uniqueness might
3842 have been obtained with &errno that happened immediately in the
3843 intermediate development versions of OpenSSL; this is no longer the
3844 case, the numeric thread callback will now override the automatic use
3846 [Geoff Thorpe, with help from Bodo Moeller]
3848 *) Initial support for different CRL issuing certificates. This covers a
3849 simple case where the self issued certificates in the chain exist and
3850 the real CRL issuer is higher in the existing chain.
3852 This work was sponsored by Google.
3855 *) Removed effectively defunct crypto/store from the build.
3858 *) Revamp of STACK to provide stronger type-checking. Still to come:
3859 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3860 ASN1_STRING, CONF_VALUE.
3863 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3864 RAM on SSL connections. This option can save about 34k per idle SSL.
3867 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3868 STACK, TXT_DB, bsearch, qsort.
3871 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3872 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3873 support for data, signedData, compressedData, digestedData and
3874 encryptedData, envelopedData types included. Scripts to check against
3875 RFC4134 examples draft and interop and consistency checks of many
3876 content types and variants.
3879 *) Add options to enc utility to support use of zlib compression BIO.
3882 *) Extend mk1mf to support importing of options and assembly language
3883 files from Configure script, currently only included in VC-WIN32.
3884 The assembly language rules can now optionally generate the source
3885 files from the associated perl scripts.
3888 *) Implement remaining functionality needed to support GOST ciphersuites.
3889 Interop testing has been performed using CryptoPro implementations.
3890 [Victor B. Wagner <vitus@cryptocom.ru>]
3892 *) s390x assembler pack.
3895 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
3899 *) Implement Opaque PRF Input TLS extension as specified in
3900 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
3901 official specification yet and no extension type assignment by
3902 IANA exists, this extension (for now) will have to be explicitly
3903 enabled when building OpenSSL by providing the extension number
3904 to use. For example, specify an option
3906 -DTLSEXT_TYPE_opaque_prf_input=0x9527
3908 to the "config" or "Configure" script to enable the extension,
3909 assuming extension number 0x9527 (which is a completely arbitrary
3910 and unofficial assignment based on the MD5 hash of the Internet
3911 Draft). Note that by doing so, you potentially lose
3912 interoperability with other TLS implementations since these might
3913 be using the same extension number for other purposes.
3915 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
3916 opaque PRF input value to use in the handshake. This will create
3917 an internal copy of the length-'len' string at 'src', and will
3918 return non-zero for success.
3920 To get more control and flexibility, provide a callback function
3923 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
3924 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
3928 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
3931 Callback function 'cb' will be called in handshakes, and is
3932 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
3933 Argument 'arg' is for application purposes (the value as given to
3934 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
3935 be provided to the callback function). The callback function
3936 has to return non-zero to report success: usually 1 to use opaque
3937 PRF input just if possible, or 2 to enforce use of the opaque PRF
3938 input. In the latter case, the library will abort the handshake
3939 if opaque PRF input is not successfully negotiated.
3941 Arguments 'peerinput' and 'len' given to the callback function
3942 will always be NULL and 0 in the case of a client. A server will
3943 see the client's opaque PRF input through these variables if
3944 available (NULL and 0 otherwise). Note that if the server
3945 provides an opaque PRF input, the length must be the same as the
3946 length of the client's opaque PRF input.
3948 Note that the callback function will only be called when creating
3949 a new session (session resumption can resume whatever was
3950 previously negotiated), and will not be called in SSL 2.0
3951 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
3952 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
3953 for applications that need to enforce opaque PRF input.
3957 *) Update ssl code to support digests other than SHA1+MD5 for handshake
3960 [Victor B. Wagner <vitus@cryptocom.ru>]
3962 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3963 RFC4507bis. The encrypted ticket format is an encrypted encoded
3964 SSL_SESSION structure, that way new session features are automatically
3967 If a client application caches session in an SSL_SESSION structure
3968 support is transparent because tickets are now stored in the encoded
3971 The SSL_CTX structure automatically generates keys for ticket
3972 protection in servers so again support should be possible
3973 with no application modification.
3975 If a client or server wishes to disable RFC4507 support then the option
3976 SSL_OP_NO_TICKET can be set.
3978 Add a TLS extension debugging callback to allow the contents of any client
3979 or server extensions to be examined.
3981 This work was sponsored by Google.
3984 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
3985 OpenSSL should now compile cleanly on gcc 4.2
3986 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
3988 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
3989 support including streaming MAC support: this is required for GOST
3990 ciphersuite support.
3991 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
3993 *) Add option -stream to use PKCS#7 streaming in smime utility. New
3994 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
3995 to output in BER and PEM format.
3998 *) Experimental support for use of HMAC via EVP_PKEY interface. This
3999 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4000 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4001 ENGINE support for HMAC keys which are unextractable. New -mac and
4002 -macopt options to dgst utility.
4005 *) New option -sigopt to dgst utility. Update dgst to use
4006 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4007 alternative signing parameters such as X9.31 or PSS in the dgst
4011 *) Change ssl_cipher_apply_rule(), the internal function that does
4012 the work each time a ciphersuite string requests enabling
4013 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4014 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4015 the order of disabled ciphersuites such that those ciphersuites
4016 that most recently went from enabled to disabled not only stay
4017 in order with respect to each other, but also have higher priority
4018 than other disabled ciphersuites the next time ciphersuites are
4021 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4022 the same ciphersuites as with "HIGH" alone, but in a specific
4023 order where the PSK ciphersuites come first (since they are the
4024 most recently disabled ciphersuites when "HIGH" is parsed).
4026 Also, change ssl_create_cipher_list() (using this new
4027 functionality) such that between otherwise identical
4028 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4032 *) Change ssl_create_cipher_list() so that it automatically
4033 arranges the ciphersuites in reasonable order before starting
4034 to process the rule string. Thus, the definition for "DEFAULT"
4035 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4036 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4037 This makes it much easier to arrive at a reasonable default order
4038 in applications for which anonymous ciphers are OK (meaning
4039 that you can't actually use DEFAULT).
4040 [Bodo Moeller; suggested by Victor Duchovni]
4042 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4043 processing) into multiple integers instead of setting
4044 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4045 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4046 (These masks as well as the individual bit definitions are hidden
4047 away into the non-exported interface ssl/ssl_locl.h, so this
4048 change to the definition of the SSL_CIPHER structure shouldn't
4049 affect applications.) This give us more bits for each of these
4050 categories, so there is no longer a need to coagulate AES128 and
4051 AES256 into a single algorithm bit, and to coagulate Camellia128
4052 and Camellia256 into a single algorithm bit, which has led to all
4055 Thus, among other things, the kludge introduced in 0.9.7m and
4056 0.9.8e for masking out AES256 independently of AES128 or masking
4057 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4059 With the change, we also introduce new ciphersuite aliases that
4060 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4064 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4065 Use the leftmost N bytes of the signature input if the input is
4066 larger than the prime q (with N being the size in bytes of q).
4069 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4070 it yet and it is largely untested.
4073 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4076 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4077 some compilers (gcc 4.2 and later) reject their use. Safestack is
4078 reimplemented. Update ASN1 to avoid use of legacy functions.
4081 *) Win32/64 targets are linked with Winsock2.
4084 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4085 to external functions. This can be used to increase CRL handling
4086 efficiency especially when CRLs are very large by (for example) storing
4087 the CRL revoked certificates in a database.
4090 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4091 new CRLs added to a directory can be used. New command line option
4092 -verify_return_error to s_client and s_server. This causes real errors
4093 to be returned by the verify callback instead of carrying on no matter
4094 what. This reflects the way a "real world" verify callback would behave.
4097 *) GOST engine, supporting several GOST algorithms and public key formats.
4098 Kindly donated by Cryptocom.
4101 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4102 partitioned by DP are handled but no indirect CRL or reason partitioning
4103 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4104 selected via a scoring technique which handles IDP and AKID in CRLs.
4107 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4108 will ultimately be used for all verify operations: this will remove the
4109 X509_STORE dependency on certificate verification and allow alternative
4110 lookup methods. X509_STORE based implementations of these two callbacks.
4113 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4114 Modify get_crl() to find a valid (unexpired) CRL if possible.
4117 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4118 this would be called X509_CRL_cmp() but that name is already used by
4119 a function that just compares CRL issuer names. Cache several CRL
4120 extensions in X509_CRL structure and cache CRLDP in X509.
4123 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4124 this maps equivalent X509_NAME structures into a consistent structure.
4125 Name comparison can then be performed rapidly using memcmp().
4128 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4132 *) Allow digests to supply their own micalg string for S/MIME type using
4133 the ctrl EVP_MD_CTRL_MICALG.
4136 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4137 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4138 ctrl. It can then customise the structure before and/or after signing
4142 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4143 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4144 to free up any added signature OIDs.
4147 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4148 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4149 digest and cipher tables. New options added to openssl utility:
4150 list-message-digest-algorithms and list-cipher-algorithms.
4153 *) Change the array representation of binary polynomials: the list
4154 of degrees of non-zero coefficients is now terminated with -1.
4155 Previously it was terminated with 0, which was also part of the
4156 value; thus, the array representation was not applicable to
4157 polynomials where t^0 has coefficient zero. This change makes
4158 the array representation useful in a more general context.
4161 *) Various modifications and fixes to SSL/TLS cipher string
4162 handling. For ECC, the code now distinguishes between fixed ECDH
4163 with RSA certificates on the one hand and with ECDSA certificates
4164 on the other hand, since these are separate ciphersuites. The
4165 unused code for Fortezza ciphersuites has been removed.
4167 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4168 (not "ECDHE"). For consistency with the code for DH
4169 certificates, use of ECDH certificates is now considered ECDH
4170 authentication, not RSA or ECDSA authentication (the latter is
4171 merely the CA's signing algorithm and not actively used in the
4174 The temporary ciphersuite alias "ECCdraft" is no longer
4175 available, and ECC ciphersuites are no longer excluded from "ALL"
4176 and "DEFAULT". The following aliases now exist for RFC 4492
4177 ciphersuites, most of these by analogy with the DH case:
4179 kECDHr - ECDH cert, signed with RSA
4180 kECDHe - ECDH cert, signed with ECDSA
4181 kECDH - ECDH cert (signed with either RSA or ECDSA)
4182 kEECDH - ephemeral ECDH
4183 ECDH - ECDH cert or ephemeral ECDH
4189 AECDH - anonymous ECDH
4190 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4194 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4195 Use correct micalg parameters depending on digest(s) in signed message.
4198 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4199 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4202 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4203 an engine to register a method. Add ENGINE lookups for methods and
4204 functional reference processing.
4207 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4208 EVP_{Sign,Verify}* which allow an application to customise the signature
4212 *) New -resign option to smime utility. This adds one or more signers
4213 to an existing PKCS#7 signedData structure. Also -md option to use an
4214 alternative message digest algorithm for signing.
4217 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4218 create PKCS7 structures containing multiple signers. Update smime
4219 application to support multiple signers.
4222 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4226 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4227 Reorganize PBE internals to lookup from a static table using NIDs,
4228 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4229 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4230 PRF which will be automatically used with PBES2.
4233 *) Replace the algorithm specific calls to generate keys in "req" with the
4237 *) Update PKCS#7 enveloped data routines to use new API. This is now
4238 supported by any public key method supporting the encrypt operation. A
4239 ctrl is added to allow the public key algorithm to examine or modify
4240 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4244 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4245 a default digest type to use. In most cases this will be SHA1 but some
4246 algorithms (such as GOST) need to specify an alternative digest. The
4247 return value indicates how strong the preference is 1 means optional and
4248 2 is mandatory (that is it is the only supported type). Modify
4249 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4250 use the default md. Update openssl utilities to use the default digest
4251 type for signing if it is not explicitly indicated.
4254 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4255 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4256 signing method from the key type. This effectively removes the link
4257 between digests and public key types.
4260 *) Add an OID cross reference table and utility functions. Its purpose is to
4261 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4262 rsaEncryption. This will allow some of the algorithm specific hackery
4263 needed to use the correct OID to be removed.
4266 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4267 structures for PKCS7_sign(). They are now set up by the relevant public
4271 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4274 *) Add support for key derivation (agreement) in the API, DH method and
4278 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4279 public and private key formats. As a side effect these add additional
4280 command line functionality not previously available: DSA signatures can be