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 1.1.0k and 1.1.0l [xx XXX xxxx]
12 *) Use Windows installation paths in the mingw builds
14 Mingw isn't a POSIX environment per se, which means that Windows
15 paths should be used for installation.
20 Changes between 1.1.0j and 1.1.0k [28 May 2019]
22 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
23 This changes the size when using the genpkey app when no size is given. It
24 fixes an omission in earlier changes that changed all RSA, DSA and DH
25 generation apps to use 2048 bits by default.
28 *) Prevent over long nonces in ChaCha20-Poly1305.
30 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
31 for every encryption operation. RFC 7539 specifies that the nonce value
32 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
33 and front pads the nonce with 0 bytes if it is less than 12
34 bytes. However it also incorrectly allows a nonce to be set of up to 16
35 bytes. In this case only the last 12 bytes are significant and any
36 additional leading bytes are ignored.
38 It is a requirement of using this cipher that nonce values are
39 unique. Messages encrypted using a reused nonce value are susceptible to
40 serious confidentiality and integrity attacks. If an application changes
41 the default nonce length to be longer than 12 bytes and then makes a
42 change to the leading bytes of the nonce expecting the new value to be a
43 new unique nonce then such an application could inadvertently encrypt
44 messages with a reused nonce.
46 Additionally the ignored bytes in a long nonce are not covered by the
47 integrity guarantee of this cipher. Any application that relies on the
48 integrity of these ignored leading bytes of a long nonce may be further
49 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
50 is safe because no such use sets such a long nonce value. However user
51 applications that use this cipher directly and set a non-default nonce
52 length to be longer than 12 bytes may be vulnerable.
54 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
59 *) Added SCA hardening for modular field inversion in EC_GROUP through
60 a new dedicated field_inv() pointer in EC_METHOD.
61 This also addresses a leakage affecting conversions from projective
62 to affine coordinates.
63 [Billy Bob Brumley, Nicola Tuveri]
65 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
66 re-used X509_PUBKEY object if the second PUBKEY is malformed.
69 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
72 *) Remove the 'dist' target and add a tarball building script. The
73 'dist' target has fallen out of use, and it shouldn't be
74 necessary to configure just to create a source distribution.
77 Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
79 *) Timing vulnerability in DSA signature generation
81 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
82 timing side channel attack. An attacker could use variations in the signing
83 algorithm to recover the private key.
85 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
89 *) Timing vulnerability in ECDSA signature generation
91 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
92 timing side channel attack. An attacker could use variations in the signing
93 algorithm to recover the private key.
95 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
99 *) Add coordinate blinding for EC_POINT and implement projective
100 coordinate blinding for generic prime curves as a countermeasure to
101 chosen point SCA attacks.
102 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
104 Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
106 *) Client DoS due to large DH parameter
108 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
109 malicious server can send a very large prime value to the client. This will
110 cause the client to spend an unreasonably long period of time generating a
111 key for this prime resulting in a hang until the client has finished. This
112 could be exploited in a Denial Of Service attack.
114 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
118 *) Cache timing vulnerability in RSA Key Generation
120 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
121 a cache timing side channel attack. An attacker with sufficient access to
122 mount cache timing attacks during the RSA key generation process could
123 recover the private key.
125 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
126 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
130 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
131 parameter is no longer accepted, as it leads to a corrupt table. NULL
132 pem_str is reserved for alias entries only.
135 *) Revert blinding in ECDSA sign and instead make problematic addition
136 length-invariant. Switch even to fixed-length Montgomery multiplication.
139 *) Change generating and checking of primes so that the error rate of not
140 being prime depends on the intended use based on the size of the input.
141 For larger primes this will result in more rounds of Miller-Rabin.
142 The maximal error rate for primes with more than 1080 bits is lowered
144 [Kurt Roeckx, Annie Yousar]
146 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
149 *) Add blinding to ECDSA and DSA signatures to protect against side channel
150 attacks discovered by Keegan Ryan (NCC Group).
153 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
154 now allow empty (zero character) pass phrases.
157 *) Certificate time validation (X509_cmp_time) enforces stricter
158 compliance with RFC 5280. Fractional seconds and timezone offsets
159 are no longer allowed.
162 *) Fixed a text canonicalisation bug in CMS
164 Where a CMS detached signature is used with text content the text goes
165 through a canonicalisation process first prior to signing or verifying a
166 signature. This process strips trailing space at the end of lines, converts
167 line terminators to CRLF and removes additional trailing line terminators
168 at the end of a file. A bug in the canonicalisation process meant that
169 some characters, such as form-feed, were incorrectly treated as whitespace
170 and removed. This is contrary to the specification (RFC5485). This fix
171 could mean that detached text data signed with an earlier version of
172 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
173 signed with a fixed OpenSSL may fail to verify with an earlier version of
174 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
175 and use the "-binary" flag (for the "cms" command line application) or set
176 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
179 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
181 *) Constructed ASN.1 types with a recursive definition could exceed the stack
183 Constructed ASN.1 types with a recursive definition (such as can be found
184 in PKCS7) could eventually exceed the stack given malicious input with
185 excessive recursion. This could result in a Denial Of Service attack. There
186 are no such structures used within SSL/TLS that come from untrusted sources
187 so this is considered safe.
189 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
194 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
196 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
197 effectively reduced to only comparing the least significant bit of each
198 byte. This allows an attacker to forge messages that would be considered as
199 authenticated in an amount of tries lower than that guaranteed by the
200 security claims of the scheme. The module can only be compiled by the
201 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
203 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
208 *) Add a build target 'build_all_generated', to build all generated files
209 and only that. This can be used to prepare everything that requires
210 things like perl for a system that lacks perl and then move everything
211 to that system and do the rest of the build there.
214 *) Backport SSL_OP_NO_RENGOTIATION
216 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
217 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
218 changes this is no longer possible in 1.1.0. Therefore the new
219 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
220 1.1.0 to provide equivalent functionality.
222 Note that if an application built against 1.1.0h headers (or above) is run
223 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
224 accepted but nothing will happen, i.e. renegotiation will not be prevented.
227 *) Removed the OS390-Unix config target. It relied on a script that doesn't
231 *) rsaz_1024_mul_avx2 overflow bug on x86_64
233 There is an overflow bug in the AVX2 Montgomery multiplication procedure
234 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
235 Analysis suggests that attacks against RSA and DSA as a result of this
236 defect would be very difficult to perform and are not believed likely.
237 Attacks against DH1024 are considered just feasible, because most of the
238 work necessary to deduce information about a private key may be performed
239 offline. The amount of resources required for such an attack would be
240 significant. However, for an attack on TLS to be meaningful, the server
241 would have to share the DH1024 private key among multiple clients, which is
242 no longer an option since CVE-2016-0701.
244 This only affects processors that support the AVX2 but not ADX extensions
245 like Intel Haswell (4th generation).
247 This issue was reported to OpenSSL by David Benjamin (Google). The issue
248 was originally found via the OSS-Fuzz project.
252 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
254 *) bn_sqrx8x_internal carry bug on x86_64
256 There is a carry propagating bug in the x86_64 Montgomery squaring
257 procedure. No EC algorithms are affected. Analysis suggests that attacks
258 against RSA and DSA as a result of this defect would be very difficult to
259 perform and are not believed likely. Attacks against DH are considered just
260 feasible (although very difficult) because most of the work necessary to
261 deduce information about a private key may be performed offline. The amount
262 of resources required for such an attack would be very significant and
263 likely only accessible to a limited number of attackers. An attacker would
264 additionally need online access to an unpatched system using the target
265 private key in a scenario with persistent DH parameters and a private
266 key that is shared between multiple clients.
268 This only affects processors that support the BMI1, BMI2 and ADX extensions
269 like Intel Broadwell (5th generation) and later or AMD Ryzen.
271 This issue was reported to OpenSSL by the OSS-Fuzz project.
275 *) Malformed X.509 IPAddressFamily could cause OOB read
277 If an X.509 certificate has a malformed IPAddressFamily extension,
278 OpenSSL could do a one-byte buffer overread. The most likely result
279 would be an erroneous display of the certificate in text format.
281 This issue was reported to OpenSSL by the OSS-Fuzz project.
285 *) Ignore the '-named_curve auto' value for compatibility of applications
287 [Tomas Mraz <tmraz@fedoraproject.org>]
289 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
292 Changes between 1.1.0e and 1.1.0f [25 May 2017]
294 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
295 platform rather than 'mingw'.
298 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
299 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
300 which is the minimum version we support.
303 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
305 *) Encrypt-Then-Mac renegotiation crash
307 During a renegotiation handshake if the Encrypt-Then-Mac extension is
308 negotiated where it was not in the original handshake (or vice-versa) then
309 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
310 and servers are affected.
312 This issue was reported to OpenSSL by Joe Orton (Red Hat).
316 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
318 *) Truncated packet could crash via OOB read
320 If one side of an SSL/TLS path is running on a 32-bit host and a specific
321 cipher is being used, then a truncated packet can cause that host to
322 perform an out-of-bounds read, usually resulting in a crash.
324 This issue was reported to OpenSSL by Robert Święcki of Google.
328 *) Bad (EC)DHE parameters cause a client crash
330 If a malicious server supplies bad parameters for a DHE or ECDHE key
331 exchange then this can result in the client attempting to dereference a
332 NULL pointer leading to a client crash. This could be exploited in a Denial
335 This issue was reported to OpenSSL by Guido Vranken.
339 *) BN_mod_exp may produce incorrect results on x86_64
341 There is a carry propagating bug in the x86_64 Montgomery squaring
342 procedure. No EC algorithms are affected. Analysis suggests that attacks
343 against RSA and DSA as a result of this defect would be very difficult to
344 perform and are not believed likely. Attacks against DH are considered just
345 feasible (although very difficult) because most of the work necessary to
346 deduce information about a private key may be performed offline. The amount
347 of resources required for such an attack would be very significant and
348 likely only accessible to a limited number of attackers. An attacker would
349 additionally need online access to an unpatched system using the target
350 private key in a scenario with persistent DH parameters and a private
351 key that is shared between multiple clients. For example this can occur by
352 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
353 similar to CVE-2015-3193 but must be treated as a separate problem.
355 This issue was reported to OpenSSL by the OSS-Fuzz project.
359 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
361 *) ChaCha20/Poly1305 heap-buffer-overflow
363 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
364 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
365 crash. This issue is not considered to be exploitable beyond a DoS.
367 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
371 *) CMS Null dereference
373 Applications parsing invalid CMS structures can crash with a NULL pointer
374 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
375 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
376 structure callback if an attempt is made to free certain invalid encodings.
377 Only CHOICE structures using a callback which do not handle NULL value are
380 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
384 *) Montgomery multiplication may produce incorrect results
386 There is a carry propagating bug in the Broadwell-specific Montgomery
387 multiplication procedure that handles input lengths divisible by, but
388 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
389 and DH private keys are impossible. This is because the subroutine in
390 question is not used in operations with the private key itself and an input
391 of the attacker's direct choice. Otherwise the bug can manifest itself as
392 transient authentication and key negotiation failures or reproducible
393 erroneous outcome of public-key operations with specially crafted input.
394 Among EC algorithms only Brainpool P-512 curves are affected and one
395 presumably can attack ECDH key negotiation. Impact was not analyzed in
396 detail, because pre-requisites for attack are considered unlikely. Namely
397 multiple clients have to choose the curve in question and the server has to
398 share the private key among them, neither of which is default behaviour.
399 Even then only clients that chose the curve will be affected.
401 This issue was publicly reported as transient failures and was not
402 initially recognized as a security issue. Thanks to Richard Morgan for
403 providing reproducible case.
407 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
408 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
409 prevent issues where no progress is being made and the peer continually
410 sends unrecognised record types, using up resources processing them.
413 *) Removed automatic addition of RPATH in shared libraries and executables,
414 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
417 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
419 *) Fix Use After Free for large message sizes
421 The patch applied to address CVE-2016-6307 resulted in an issue where if a
422 message larger than approx 16k is received then the underlying buffer to
423 store the incoming message is reallocated and moved. Unfortunately a
424 dangling pointer to the old location is left which results in an attempt to
425 write to the previously freed location. This is likely to result in a
426 crash, however it could potentially lead to execution of arbitrary code.
428 This issue only affects OpenSSL 1.1.0a.
430 This issue was reported to OpenSSL by Robert Święcki.
434 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
436 *) OCSP Status Request extension unbounded memory growth
438 A malicious client can send an excessively large OCSP Status Request
439 extension. If that client continually requests renegotiation, sending a
440 large OCSP Status Request extension each time, then there will be unbounded
441 memory growth on the server. This will eventually lead to a Denial Of
442 Service attack through memory exhaustion. Servers with a default
443 configuration are vulnerable even if they do not support OCSP. Builds using
444 the "no-ocsp" build time option are not affected.
446 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
450 *) SSL_peek() hang on empty record
452 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
453 sends an empty record. This could be exploited by a malicious peer in a
454 Denial Of Service attack.
456 This issue was reported to OpenSSL by Alex Gaynor.
460 *) Excessive allocation of memory in tls_get_message_header() and
461 dtls1_preprocess_fragment()
463 A (D)TLS message includes 3 bytes for its length in the header for the
464 message. This would allow for messages up to 16Mb in length. Messages of
465 this length are excessive and OpenSSL includes a check to ensure that a
466 peer is sending reasonably sized messages in order to avoid too much memory
467 being consumed to service a connection. A flaw in the logic of version
468 1.1.0 means that memory for the message is allocated too early, prior to
469 the excessive message length check. Due to way memory is allocated in
470 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
471 to service a connection. This could lead to a Denial of Service through
472 memory exhaustion. However, the excessive message length check still takes
473 place, and this would cause the connection to immediately fail. Assuming
474 that the application calls SSL_free() on the failed connection in a timely
475 manner then the 21Mb of allocated memory will then be immediately freed
476 again. Therefore the excessive memory allocation will be transitory in
477 nature. This then means that there is only a security impact if:
479 1) The application does not call SSL_free() in a timely manner in the event
480 that the connection fails
482 2) The application is working in a constrained environment where there is
483 very little free memory
485 3) The attacker initiates multiple connection attempts such that there are
486 multiple connections in a state where memory has been allocated for the
487 connection; SSL_free() has not yet been called; and there is insufficient
488 memory to service the multiple requests.
490 Except in the instance of (1) above any Denial Of Service is likely to be
491 transitory because as soon as the connection fails the memory is
492 subsequently freed again in the SSL_free() call. However there is an
493 increased risk during this period of application crashes due to the lack of
494 memory - which would then mean a more serious Denial of Service.
496 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
497 (CVE-2016-6307 and CVE-2016-6308)
500 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
501 had to be removed. Primary reason is that vendor assembler can't
502 assemble our modules with -KPIC flag. As result it, assembly
503 support, was not even available as option. But its lack means
504 lack of side-channel resistant code, which is incompatible with
505 security by todays standards. Fortunately gcc is readily available
506 prepackaged option, which we firmly point at...
509 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
511 *) Windows command-line tool supports UTF-8 opt-in option for arguments
512 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
513 (to any value) allows Windows user to access PKCS#12 file generated
514 with Windows CryptoAPI and protected with non-ASCII password, as well
515 as files generated under UTF-8 locale on Linux also protected with
519 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
520 have been disabled by default and removed from DEFAULT, just like RC4.
521 See the RC4 item below to re-enable both.
524 *) The method for finding the storage location for the Windows RAND seed file
525 has changed. First we check %RANDFILE%. If that is not set then we check
526 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
527 all else fails we fall back to C:\.
530 *) The EVP_EncryptUpdate() function has had its return type changed from void
531 to int. A return of 0 indicates and error while a return of 1 indicates
535 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
536 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
537 off the constant time implementation for RSA, DSA and DH have been made
538 no-ops and deprecated.
541 *) Windows RAND implementation was simplified to only get entropy by
542 calling CryptGenRandom(). Various other RAND-related tickets
544 [Joseph Wylie Yandle, Rich Salz]
546 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
547 and OPENSSL_LH_, respectively. The old names are available
548 with API compatibility. They new names are now completely documented.
551 *) Unify TYPE_up_ref(obj) methods signature.
552 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
553 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
554 int (instead of void) like all others TYPE_up_ref() methods.
555 So now these methods also check the return value of CRYPTO_atomic_add(),
556 and the validity of object reference counter.
557 [fdasilvayy@gmail.com]
559 *) With Windows Visual Studio builds, the .pdb files are installed
560 alongside the installed libraries and executables. For a static
561 library installation, ossl_static.pdb is the associate compiler
562 generated .pdb file to be used when linking programs.
565 *) Remove openssl.spec. Packaging files belong with the packagers.
568 *) Automatic Darwin/OSX configuration has had a refresh, it will now
569 recognise x86_64 architectures automatically. You can still decide
570 to build for a different bitness with the environment variable
571 KERNEL_BITS (can be 32 or 64), for example:
573 KERNEL_BITS=32 ./config
577 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
578 256 bit AES and HMAC with SHA256.
581 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
584 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
587 *) To enable users to have their own config files and build file templates,
588 Configure looks in the directory indicated by the environment variable
589 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
590 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
591 name and is used as is.
594 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
595 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
596 X509_CERT_FILE_CTX was removed.
599 *) "shared" builds are now the default. To create only static libraries use
600 the "no-shared" Configure option.
603 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
604 All of these option have not worked for some while and are fundamental
608 *) Make various cleanup routines no-ops and mark them as deprecated. Most
609 global cleanup functions are no longer required because they are handled
610 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
611 Explicitly de-initing can cause problems (e.g. where a library that uses
612 OpenSSL de-inits, but an application is still using it). The affected
613 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
614 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
615 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
619 *) --strict-warnings no longer enables runtime debugging options
620 such as REF_DEBUG. Instead, debug options are automatically
621 enabled with '--debug' builds.
622 [Andy Polyakov, Emilia Käsper]
624 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
625 have been moved out of the public header files. New functions for managing
626 these have been added.
629 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
630 objects have been moved out of the public header files. New
631 functions for managing these have been added.
634 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
635 have been moved out of the public header files. New functions for managing
636 these have been added.
639 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
640 moved out of the public header files. New functions for managing these
644 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
647 *) Removed the mk1mf build scripts.
650 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
651 it is always safe to #include a header now.
654 *) Removed the aged BC-32 config and all its supporting scripts
657 *) Removed support for Ultrix, Netware, and OS/2.
660 *) Add support for HKDF.
663 *) Add support for blake2b and blake2s
666 *) Added support for "pipelining". Ciphers that have the
667 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
668 encryptions/decryptions simultaneously. There are currently no built-in
669 ciphers with this property but the expectation is that engines will be able
670 to offer it to significantly improve throughput. Support has been extended
671 into libssl so that multiple records for a single connection can be
672 processed in one go (for >=TLS 1.1).
675 *) Added the AFALG engine. This is an async capable engine which is able to
676 offload work to the Linux kernel. In this initial version it only supports
677 AES128-CBC. The kernel must be version 4.1.0 or greater.
680 *) OpenSSL now uses a new threading API. It is no longer necessary to
681 set locking callbacks to use OpenSSL in a multi-threaded environment. There
682 are two supported threading models: pthreads and windows threads. It is
683 also possible to configure OpenSSL at compile time for "no-threads". The
684 old threading API should no longer be used. The functions have been
685 replaced with "no-op" compatibility macros.
686 [Alessandro Ghedini, Matt Caswell]
688 *) Modify behavior of ALPN to invoke callback after SNI/servername
689 callback, such that updates to the SSL_CTX affect ALPN.
692 *) Add SSL_CIPHER queries for authentication and key-exchange.
695 *) Changes to the DEFAULT cipherlist:
696 - Prefer (EC)DHE handshakes over plain RSA.
697 - Prefer AEAD ciphers over legacy ciphers.
698 - Prefer ECDSA over RSA when both certificates are available.
699 - Prefer TLSv1.2 ciphers/PRF.
700 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
704 *) Change the ECC default curve list to be this, in order: x25519,
705 secp256r1, secp521r1, secp384r1.
708 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
709 disabled by default. They can be re-enabled using the
710 enable-weak-ssl-ciphers option to Configure.
713 *) If the server has ALPN configured, but supports no protocols that the
714 client advertises, send a fatal "no_application_protocol" alert.
715 This behaviour is SHALL in RFC 7301, though it isn't universally
716 implemented by other servers.
719 *) Add X25519 support.
720 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
721 for public and private key encoding using the format documented in
722 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
723 key generation and key derivation.
725 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
729 *) Deprecate SRP_VBASE_get_by_user.
730 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
731 In order to fix an unavoidable memory leak (CVE-2016-0798),
732 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
733 seed, even if the seed is configured.
735 Users should use SRP_VBASE_get1_by_user instead. Note that in
736 SRP_VBASE_get1_by_user, caller must free the returned value. Note
737 also that even though configuring the SRP seed attempts to hide
738 invalid usernames by continuing the handshake with fake
739 credentials, this behaviour is not constant time and no strong
740 guarantees are made that the handshake is indistinguishable from
741 that of a valid user.
744 *) Configuration change; it's now possible to build dynamic engines
745 without having to build shared libraries and vice versa. This
746 only applies to the engines in engines/, those in crypto/engine/
747 will always be built into libcrypto (i.e. "static").
749 Building dynamic engines is enabled by default; to disable, use
750 the configuration option "disable-dynamic-engine".
752 The only requirements for building dynamic engines are the
753 presence of the DSO module and building with position independent
754 code, so they will also automatically be disabled if configuring
755 with "disable-dso" or "disable-pic".
757 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
758 are also taken away from openssl/opensslconf.h, as they are
762 *) Configuration change; if there is a known flag to compile
763 position independent code, it will always be applied on the
764 libcrypto and libssl object files, and never on the application
765 object files. This means other libraries that use routines from
766 libcrypto / libssl can be made into shared libraries regardless
767 of how OpenSSL was configured.
769 If this isn't desirable, the configuration options "disable-pic"
770 or "no-pic" can be used to disable the use of PIC. This will
771 also disable building shared libraries and dynamic engines.
774 *) Removed JPAKE code. It was experimental and has no wide use.
777 *) The INSTALL_PREFIX Makefile variable has been renamed to
778 DESTDIR. That makes for less confusion on what this variable
779 is for. Also, the configuration option --install_prefix is
783 *) Heartbeat for TLS has been removed and is disabled by default
784 for DTLS; configure with enable-heartbeats. Code that uses the
785 old #define's might need to be updated.
786 [Emilia Käsper, Rich Salz]
788 *) Rename REF_CHECK to REF_DEBUG.
791 *) New "unified" build system
793 The "unified" build system is aimed to be a common system for all
794 platforms we support. With it comes new support for VMS.
796 This system builds supports building in a different directory tree
797 than the source tree. It produces one Makefile (for unix family
798 or lookalikes), or one descrip.mms (for VMS).
800 The source of information to make the Makefile / descrip.mms is
801 small files called 'build.info', holding the necessary
802 information for each directory with source to compile, and a
803 template in Configurations, like unix-Makefile.tmpl or
806 With this change, the library names were also renamed on Windows
807 and on VMS. They now have names that are closer to the standard
808 on Unix, and include the major version number, and in certain
809 cases, the architecture they are built for. See "Notes on shared
810 libraries" in INSTALL.
812 We rely heavily on the perl module Text::Template.
815 *) Added support for auto-initialisation and de-initialisation of the library.
816 OpenSSL no longer requires explicit init or deinit routines to be called,
817 except in certain circumstances. See the OPENSSL_init_crypto() and
818 OPENSSL_init_ssl() man pages for further information.
821 *) The arguments to the DTLSv1_listen function have changed. Specifically the
822 "peer" argument is now expected to be a BIO_ADDR object.
824 *) Rewrite of BIO networking library. The BIO library lacked consistent
825 support of IPv6, and adding it required some more extensive
826 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
827 which hold all types of addresses and chains of address information.
828 It also introduces a new API, with functions like BIO_socket,
829 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
830 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
831 have been adapted accordingly.
834 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
838 *) CRIME protection: disable compression by default, even if OpenSSL is
839 compiled with zlib enabled. Applications can still enable compression
840 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
841 using the SSL_CONF library to configure compression.
844 *) The signature of the session callback configured with
845 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
846 was explicitly marked as 'const unsigned char*' instead of
850 *) Always DPURIFY. Remove the use of uninitialized memory in the
851 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
854 *) Removed many obsolete configuration items, including
855 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
856 MD2_CHAR, MD2_INT, MD2_LONG
858 IDEA_SHORT, IDEA_LONG
859 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
860 [Rich Salz, with advice from Andy Polyakov]
862 *) Many BN internals have been moved to an internal header file.
863 [Rich Salz with help from Andy Polyakov]
865 *) Configuration and writing out the results from it has changed.
866 Files such as Makefile include/openssl/opensslconf.h and are now
867 produced through general templates, such as Makefile.in and
868 crypto/opensslconf.h.in and some help from the perl module
871 Also, the center of configuration information is no longer
872 Makefile. Instead, Configure produces a perl module in
873 configdata.pm which holds most of the config data (in the hash
874 table %config), the target data that comes from the target
875 configuration in one of the Configurations/*.conf files (in
879 *) To clarify their intended purposes, the Configure options
880 --prefix and --openssldir change their semantics, and become more
881 straightforward and less interdependent.
883 --prefix shall be used exclusively to give the location INSTALLTOP
884 where programs, scripts, libraries, include files and manuals are
885 going to be installed. The default is now /usr/local.
887 --openssldir shall be used exclusively to give the default
888 location OPENSSLDIR where certificates, private keys, CRLs are
889 managed. This is also where the default openssl.cnf gets
891 If the directory given with this option is a relative path, the
892 values of both the --prefix value and the --openssldir value will
893 be combined to become OPENSSLDIR.
894 The default for --openssldir is INSTALLTOP/ssl.
896 Anyone who uses --openssldir to specify where OpenSSL is to be
897 installed MUST change to use --prefix instead.
900 *) The GOST engine was out of date and therefore it has been removed. An up
901 to date GOST engine is now being maintained in an external repository.
902 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
903 support for GOST ciphersuites (these are only activated if a GOST engine
907 *) EGD is no longer supported by default; use enable-egd when
909 [Ben Kaduk and Rich Salz]
911 *) The distribution now has Makefile.in files, which are used to
912 create Makefile's when Configure is run. *Configure must be run
913 before trying to build now.*
916 *) The return value for SSL_CIPHER_description() for error conditions
920 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
922 Obtaining and performing DNSSEC validation of TLSA records is
923 the application's responsibility. The application provides
924 the TLSA records of its choice to OpenSSL, and these are then
925 used to authenticate the peer.
927 The TLSA records need not even come from DNS. They can, for
928 example, be used to implement local end-entity certificate or
929 trust-anchor "pinning", where the "pin" data takes the form
930 of TLSA records, which can augment or replace verification
931 based on the usual WebPKI public certification authorities.
934 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
935 continues to support deprecated interfaces in default builds.
936 However, applications are strongly advised to compile their
937 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
938 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
939 or the 1.1.0 releases.
941 In environments in which all applications have been ported to
942 not use any deprecated interfaces OpenSSL's Configure script
943 should be used with the --api=1.1.0 option to entirely remove
944 support for the deprecated features from the library and
945 unconditionally disable them in the installed headers.
946 Essentially the same effect can be achieved with the "no-deprecated"
947 argument to Configure, except that this will always restrict
948 the build to just the latest API, rather than a fixed API
951 As applications are ported to future revisions of the API,
952 they should update their compile-time OPENSSL_API_COMPAT define
953 accordingly, but in most cases should be able to continue to
954 compile with later releases.
956 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
957 0x10000000L and 0x00908000L, respectively. However those
958 versions did not support the OPENSSL_API_COMPAT feature, and
959 so applications are not typically tested for explicit support
960 of just the undeprecated features of either release.
963 *) Add support for setting the minimum and maximum supported protocol.
964 It can bet set via the SSL_set_min_proto_version() and
965 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
966 MaxProtcol. It's recommended to use the new APIs to disable
967 protocols instead of disabling individual protocols using
968 SSL_set_options() or SSL_CONF's Protocol. This change also
969 removes support for disabling TLS 1.2 in the OpenSSL TLS
970 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
973 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
976 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
977 and integrates ECDSA and ECDH functionality into EC. Implementations can
978 now redirect key generation and no longer need to convert to or from
981 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
982 include the ec.h header file instead.
985 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
986 ciphers who are no longer supported and drops support the ephemeral RSA key
987 exchange. The LOW ciphers currently doesn't have any ciphers in it.
990 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
991 opaque. For HMAC_CTX, the following constructors and destructors
994 HMAC_CTX *HMAC_CTX_new(void);
995 void HMAC_CTX_free(HMAC_CTX *ctx);
997 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
998 destroy such methods has been added. See EVP_MD_meth_new(3) and
999 EVP_CIPHER_meth_new(3) for documentation.
1002 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1003 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1004 EVP_MD_CTX_reset() should be called instead to reinitialise
1005 an already created structure.
1006 2) For consistency with the majority of our object creators and
1007 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1008 EVP_MD_CTX_(new|free). The old names are retained as macros
1009 for deprecated builds.
1012 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1013 cryptographic operations to be performed asynchronously as long as an
1014 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1015 further details. Libssl has also had this capability integrated with the
1016 introduction of the new mode SSL_MODE_ASYNC and associated error
1017 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1018 pages. This work was developed in partnership with Intel Corp.
1021 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1022 always enabled now. If you want to disable the support you should
1023 exclude it using the list of supported ciphers. This also means that the
1024 "-no_ecdhe" option has been removed from s_server.
1027 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1028 SSL_{CTX_}set1_curves() which can set a list.
1031 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1032 curve you want to support using SSL_{CTX_}set1_curves().
1035 *) State machine rewrite. The state machine code has been significantly
1036 refactored in order to remove much duplication of code and solve issues
1037 with the old code (see ssl/statem/README for further details). This change
1038 does have some associated API changes. Notably the SSL_state() function
1039 has been removed and replaced by SSL_get_state which now returns an
1040 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1041 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1045 *) All instances of the string "ssleay" in the public API were replaced
1046 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1047 Some error codes related to internal RSA_eay API's were renamed.
1050 *) The demo files in crypto/threads were moved to demo/threads.
1053 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1055 [Matt Caswell, Rich Salz]
1057 *) New ASN.1 embed macro.
1059 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1060 structure is not allocated: it is part of the parent. That is instead of
1068 This reduces memory fragmentation and make it impossible to accidentally
1069 set a mandatory field to NULL.
1071 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1072 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1073 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1077 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1080 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1081 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1082 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1083 DES and RC4 ciphersuites.
1086 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1087 This changes the decoding behaviour for some invalid messages,
1088 though the change is mostly in the more lenient direction, and
1089 legacy behaviour is preserved as much as possible.
1092 *) Fix no-stdio build.
1093 [ David Woodhouse <David.Woodhouse@intel.com> and also
1094 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1096 *) New testing framework
1097 The testing framework has been largely rewritten and is now using
1098 perl and the perl modules Test::Harness and an extended variant of
1099 Test::More called OpenSSL::Test to do its work. All test scripts in
1100 test/ have been rewritten into test recipes, and all direct calls to
1101 executables in test/Makefile have become individual recipes using the
1102 simplified testing OpenSSL::Test::Simple.
1104 For documentation on our testing modules, do:
1106 perldoc test/testlib/OpenSSL/Test/Simple.pm
1107 perldoc test/testlib/OpenSSL/Test.pm
1111 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1112 are used; the latter aborts on memory leaks (usually checked on exit).
1113 Some undocumented "set malloc, etc., hooks" functions were removed
1114 and others were changed. All are now documented.
1117 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1119 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1121 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1122 from RFC4279, RFC4785, RFC5487, RFC5489.
1124 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1125 original RSA_PSK patch.
1128 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1129 era flag was never set throughout the codebase (only read). Also removed
1130 SSL3_FLAGS_POP_BUFFER which was only used if
1131 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1134 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1135 to be "oneline" instead of "compat".
1138 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1139 not aware of clients that still exhibit this bug, and the workaround
1140 hasn't been working properly for a while.
1143 *) The return type of BIO_number_read() and BIO_number_written() as well as
1144 the corresponding num_read and num_write members in the BIO structure has
1145 changed from unsigned long to uint64_t. On platforms where an unsigned
1146 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1150 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1151 OpenSSL without support for them. It also means that maintaining
1152 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1153 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1156 *) Removed support for the two export grade static DH ciphersuites
1157 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1158 were newly added (along with a number of other static DH ciphersuites) to
1159 1.0.2. However the two export ones have *never* worked since they were
1160 introduced. It seems strange in any case to be adding new export
1161 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1164 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1165 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1166 and turned into macros which simply call the new preferred function names
1167 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1168 should use the new names instead. Also as part of this change the ssl23.h
1169 header file has been removed.
1172 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1173 code and the associated standard is no longer considered fit-for-purpose.
1176 *) RT2547 was closed. When generating a private key, try to make the
1177 output file readable only by the owner. This behavior change might
1178 be noticeable when interacting with other software.
1180 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1184 *) Added HTTP GET support to the ocsp command.
1187 *) Changed default digest for the dgst and enc commands from MD5 to
1191 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1194 *) Added support for TLS extended master secret from
1195 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1196 initial patch which was a great help during development.
1199 *) All libssl internal structures have been removed from the public header
1200 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1201 now redundant). Users should not attempt to access internal structures
1202 directly. Instead they should use the provided API functions.
1205 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1206 Access to deprecated functions can be re-enabled by running config with
1207 "enable-deprecated". In addition applications wishing to use deprecated
1208 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1209 will, by default, disable some transitive includes that previously existed
1210 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1213 *) Added support for OCB mode. OpenSSL has been granted a patent license
1214 compatible with the OpenSSL license for use of OCB. Details are available
1215 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1216 for OCB can be removed by calling config with no-ocb.
1219 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1220 compatible client hello.
1223 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1224 done while fixing the error code for the key-too-small case.
1225 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1227 *) CA.sh has been removed; use CA.pl instead.
1230 *) Removed old DES API.
1233 *) Remove various unsupported platforms:
1239 Sinix/ReliantUNIX RM400
1244 16-bit platforms such as WIN16
1247 *) Clean up OPENSSL_NO_xxx #define's
1248 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1249 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1250 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1251 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1252 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1253 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1254 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1255 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1256 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1257 Remove MS_STATIC; it's a relic from platforms <32 bits.
1260 *) Cleaned up dead code
1261 Remove all but one '#ifdef undef' which is to be looked at.
1264 *) Clean up calling of xxx_free routines.
1265 Just like free(), fix most of the xxx_free routines to accept
1266 NULL. Remove the non-null checks from callers. Save much code.
1269 *) Add secure heap for storage of private keys (when possible).
1270 Add BIO_s_secmem(), CBIGNUM, etc.
1271 Contributed by Akamai Technologies under our Corporate CLA.
1274 *) Experimental support for a new, fast, unbiased prime candidate generator,
1275 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1276 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1278 *) New output format NSS in the sess_id command line tool. This allows
1279 exporting the session id and the master key in NSS keylog format.
1280 [Martin Kaiser <martin@kaiser.cx>]
1282 *) Harmonize version and its documentation. -f flag is used to display
1284 [mancha <mancha1@zoho.com>]
1286 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1287 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1288 [mancha <mancha1@zoho.com>]
1290 *) Fix some double frees. These are not thought to be exploitable.
1291 [mancha <mancha1@zoho.com>]
1293 *) A missing bounds check in the handling of the TLS heartbeat extension
1294 can be used to reveal up to 64k of memory to a connected client or
1297 Thanks for Neel Mehta of Google Security for discovering this bug and to
1298 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1299 preparing the fix (CVE-2014-0160)
1300 [Adam Langley, Bodo Moeller]
1302 *) Fix for the attack described in the paper "Recovering OpenSSL
1303 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1304 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1305 http://eprint.iacr.org/2014/140
1307 Thanks to Yuval Yarom and Naomi Benger for discovering this
1308 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1309 [Yuval Yarom and Naomi Benger]
1311 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1312 this fixes a limitation in previous versions of OpenSSL.
1315 *) Experimental encrypt-then-mac support.
1317 Experimental support for encrypt then mac from
1318 draft-gutmann-tls-encrypt-then-mac-02.txt
1320 To enable it set the appropriate extension number (0x42 for the test
1321 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1323 For non-compliant peers (i.e. just about everything) this should have no
1326 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1330 *) Add EVP support for key wrapping algorithms, to avoid problems with
1331 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1332 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1333 algorithms and include tests cases.
1336 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1340 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1341 MGF1 digest and OAEP label.
1344 *) Make openssl verify return errors.
1345 [Chris Palmer <palmer@google.com> and Ben Laurie]
1347 *) New function ASN1_TIME_diff to calculate the difference between two
1348 ASN1_TIME structures or one structure and the current time.
1351 *) Update fips_test_suite to support multiple command line options. New
1352 test to induce all self test errors in sequence and check expected
1356 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1357 sign or verify all in one operation.
1360 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1361 test programs and fips_test_suite. Includes functionality to parse
1362 the minimal script output of fipsalgest.pl directly.
1365 *) Add authorisation parameter to FIPS_module_mode_set().
1368 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1371 *) Use separate DRBG fields for internal and external flags. New function
1372 FIPS_drbg_health_check() to perform on demand health checking. Add
1373 generation tests to fips_test_suite with reduced health check interval to
1374 demonstrate periodic health checking. Add "nodh" option to
1375 fips_test_suite to skip very slow DH test.
1378 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1382 *) More extensive health check for DRBG checking many more failure modes.
1383 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1384 combination: call this in fips_test_suite.
1387 *) Add support for canonical generation of DSA parameter 'g'. See
1390 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1391 POST to handle HMAC cases.
1394 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1395 to return numerical and string versions of the FIPS module number.
1398 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1399 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1400 outside the validated module in the FIPS capable OpenSSL.
1403 *) Minor change to DRBG entropy callback semantics. In some cases
1404 there is no multiple of the block length between min_len and
1405 max_len. Allow the callback to return more than max_len bytes
1406 of entropy but discard any extra: it is the callback's responsibility
1407 to ensure that the extra data discarded does not impact the
1408 requested amount of entropy.
1411 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1412 information in FIPS186-3, SP800-57 and SP800-131A.
1415 *) CCM support via EVP. Interface is very similar to GCM case except we
1416 must supply all data in one chunk (i.e. no update, final) and the
1417 message length must be supplied if AAD is used. Add algorithm test
1421 *) Initial version of POST overhaul. Add POST callback to allow the status
1422 of POST to be monitored and/or failures induced. Modify fips_test_suite
1423 to use callback. Always run all selftests even if one fails.
1426 *) XTS support including algorithm test driver in the fips_gcmtest program.
1427 Note: this does increase the maximum key length from 32 to 64 bytes but
1428 there should be no binary compatibility issues as existing applications
1429 will never use XTS mode.
1432 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1433 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1434 performs algorithm blocking for unapproved PRNG types. Also do not
1435 set PRNG type in FIPS_mode_set(): leave this to the application.
1436 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1437 the standard OpenSSL PRNG: set additional data to a date time vector.
1440 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1441 This shouldn't present any incompatibility problems because applications
1442 shouldn't be using these directly and any that are will need to rethink
1443 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1446 *) Extensive self tests and health checking required by SP800-90 DRBG.
1447 Remove strength parameter from FIPS_drbg_instantiate and always
1448 instantiate at maximum supported strength.
1451 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1454 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1457 *) New function DH_compute_key_padded() to compute a DH key and pad with
1458 leading zeroes if needed: this complies with SP800-56A et al.
1461 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1462 anything, incomplete, subject to change and largely untested at present.
1465 *) Modify fipscanisteronly build option to only build the necessary object
1466 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1469 *) Add experimental option FIPSSYMS to give all symbols in
1470 fipscanister.o and FIPS or fips prefix. This will avoid
1471 conflicts with future versions of OpenSSL. Add perl script
1472 util/fipsas.pl to preprocess assembly language source files
1473 and rename any affected symbols.
1476 *) Add selftest checks and algorithm block of non-fips algorithms in
1477 FIPS mode. Remove DES2 from selftests.
1480 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1481 return internal method without any ENGINE dependencies. Add new
1482 tiny fips sign and verify functions.
1485 *) New build option no-ec2m to disable characteristic 2 code.
1488 *) New build option "fipscanisteronly". This only builds fipscanister.o
1489 and (currently) associated fips utilities. Uses the file Makefile.fips
1490 instead of Makefile.org as the prototype.
1493 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1494 Update fips_gcmtest to use IV generator.
1497 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1498 setting output buffer to NULL. The *Final function must be
1499 called although it will not retrieve any additional data. The tag
1500 can be set or retrieved with a ctrl. The IV length is by default 12
1501 bytes (96 bits) but can be set to an alternative value. If the IV
1502 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1506 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1507 underlying do_cipher function handles all cipher semantics itself
1508 including padding and finalisation. This is useful if (for example)
1509 an ENGINE cipher handles block padding itself. The behaviour of
1510 do_cipher is subtly changed if this flag is set: the return value
1511 is the number of characters written to the output buffer (zero is
1512 no longer an error code) or a negative error code. Also if the
1513 input buffer is NULL and length 0 finalisation should be performed.
1516 *) If a candidate issuer certificate is already part of the constructed
1517 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1520 *) Improve forward-security support: add functions
1522 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1523 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1525 for use by SSL/TLS servers; the callback function will be called whenever a
1526 new session is created, and gets to decide whether the session may be
1527 cached to make it resumable (return 0) or not (return 1). (As by the
1528 SSL/TLS protocol specifications, the session_id sent by the server will be
1529 empty to indicate that the session is not resumable; also, the server will
1530 not generate RFC 4507 (RFC 5077) session tickets.)
1532 A simple reasonable callback implementation is to return is_forward_secure.
1533 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1534 by the SSL/TLS server library, indicating whether it can provide forward
1536 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1538 *) New -verify_name option in command line utilities to set verification
1542 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1543 Add CMAC pkey methods.
1546 *) Experimental renegotiation in s_server -www mode. If the client
1547 browses /reneg connection is renegotiated. If /renegcert it is
1548 renegotiated requesting a certificate.
1551 *) Add an "external" session cache for debugging purposes to s_server. This
1552 should help trace issues which normally are only apparent in deployed
1553 multi-process servers.
1556 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1557 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1558 BIO_set_cipher() and some obscure PEM functions were changed so they
1559 can now return an error. The RAND changes required a change to the
1560 RAND_METHOD structure.
1563 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1564 a gcc attribute to warn if the result of a function is ignored. This
1565 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1566 whose return value is often ignored.
1569 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1570 These allow SCTs (signed certificate timestamps) to be requested and
1571 validated when establishing a connection.
1572 [Rob Percival <robpercival@google.com>]
1574 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1576 *) Prevent padding oracle in AES-NI CBC MAC check
1578 A MITM attacker can use a padding oracle attack to decrypt traffic
1579 when the connection uses an AES CBC cipher and the server support
1582 This issue was introduced as part of the fix for Lucky 13 padding
1583 attack (CVE-2013-0169). The padding check was rewritten to be in
1584 constant time by making sure that always the same bytes are read and
1585 compared against either the MAC or padding bytes. But it no longer
1586 checked that there was enough data to have both the MAC and padding
1589 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1593 *) Fix EVP_EncodeUpdate overflow
1595 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1596 Base64 encoding of binary data. If an attacker is able to supply very large
1597 amounts of input data then a length check can overflow resulting in a heap
1600 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1601 the PEM_write_bio* family of functions. These are mainly used within the
1602 OpenSSL command line applications, so any application which processes data
1603 from an untrusted source and outputs it as a PEM file should be considered
1604 vulnerable to this issue. User applications that call these APIs directly
1605 with large amounts of untrusted data may also be vulnerable.
1607 This issue was reported by Guido Vranken.
1611 *) Fix EVP_EncryptUpdate overflow
1613 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1614 is able to supply very large amounts of input data after a previous call to
1615 EVP_EncryptUpdate() with a partial block then a length check can overflow
1616 resulting in a heap corruption. Following an analysis of all OpenSSL
1617 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1618 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1619 the first called function after an EVP_EncryptInit(), and therefore that
1620 specific call must be safe. The second form is where the length passed to
1621 EVP_EncryptUpdate() can be seen from the code to be some small value and
1622 therefore there is no possibility of an overflow. Since all instances are
1623 one of these two forms, it is believed that there can be no overflows in
1624 internal code due to this problem. It should be noted that
1625 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1626 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1627 of these calls have also been analysed too and it is believed there are no
1628 instances in internal usage where an overflow could occur.
1630 This issue was reported by Guido Vranken.
1634 *) Prevent ASN.1 BIO excessive memory allocation
1636 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1637 a short invalid encoding can cause allocation of large amounts of memory
1638 potentially consuming excessive resources or exhausting memory.
1640 Any application parsing untrusted data through d2i BIO functions is
1641 affected. The memory based functions such as d2i_X509() are *not* affected.
1642 Since the memory based functions are used by the TLS library, TLS
1643 applications are not affected.
1645 This issue was reported by Brian Carpenter.
1651 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1652 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1653 in arbitrary stack data being returned in the buffer.
1655 This issue was reported by Guido Vranken.
1659 *) Modify behavior of ALPN to invoke callback after SNI/servername
1660 callback, such that updates to the SSL_CTX affect ALPN.
1663 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1667 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1668 methods are enabled and ssl2 is disabled the methods return NULL.
1671 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1673 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1674 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1675 provide any "EXPORT" or "LOW" strength ciphers.
1678 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1679 is by default disabled at build-time. Builds that are not configured with
1680 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1681 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1682 will need to explicitly call either of:
1684 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1686 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1688 as appropriate. Even if either of those is used, or the application
1689 explicitly uses the version-specific SSLv2_method() or its client and
1690 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1691 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1692 ciphers, and SSLv2 56-bit DES are no longer available.
1696 *) Fix a double-free in DSA code
1698 A double free bug was discovered when OpenSSL parses malformed DSA private
1699 keys and could lead to a DoS attack or memory corruption for applications
1700 that receive DSA private keys from untrusted sources. This scenario is
1703 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1708 *) Disable SRP fake user seed to address a server memory leak.
1710 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1712 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1713 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1714 was changed to ignore the "fake user" SRP seed, even if the seed
1717 Users should use SRP_VBASE_get1_by_user instead. Note that in
1718 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1719 also that even though configuring the SRP seed attempts to hide
1720 invalid usernames by continuing the handshake with fake
1721 credentials, this behaviour is not constant time and no strong
1722 guarantees are made that the handshake is indistinguishable from
1723 that of a valid user.
1727 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1729 In the BN_hex2bn function the number of hex digits is calculated using an
1730 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1731 large values of |i| this can result in |bn_expand| not allocating any
1732 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1733 field as NULL leading to a subsequent NULL ptr deref. For very large values
1734 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1735 In this case memory is allocated to the internal BIGNUM data field, but it
1736 is insufficiently sized leading to heap corruption. A similar issue exists
1737 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1738 is ever called by user applications with very large untrusted hex/dec data.
1739 This is anticipated to be a rare occurrence.
1741 All OpenSSL internal usage of these functions use data that is not expected
1742 to be untrusted, e.g. config file data or application command line
1743 arguments. If user developed applications generate config file data based
1744 on untrusted data then it is possible that this could also lead to security
1745 consequences. This is also anticipated to be rare.
1747 This issue was reported to OpenSSL by Guido Vranken.
1751 *) Fix memory issues in BIO_*printf functions
1753 The internal |fmtstr| function used in processing a "%s" format string in
1754 the BIO_*printf functions could overflow while calculating the length of a
1755 string and cause an OOB read when printing very long strings.
1757 Additionally the internal |doapr_outch| function can attempt to write to an
1758 OOB memory location (at an offset from the NULL pointer) in the event of a
1759 memory allocation failure. In 1.0.2 and below this could be caused where
1760 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1761 could be in processing a very long "%s" format string. Memory leaks can
1764 The first issue may mask the second issue dependent on compiler behaviour.
1765 These problems could enable attacks where large amounts of untrusted data
1766 is passed to the BIO_*printf functions. If applications use these functions
1767 in this way then they could be vulnerable. OpenSSL itself uses these
1768 functions when printing out human-readable dumps of ASN.1 data. Therefore
1769 applications that print this data could be vulnerable if the data is from
1770 untrusted sources. OpenSSL command line applications could also be
1771 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1772 as command line arguments.
1774 Libssl is not considered directly vulnerable. Additionally certificates etc
1775 received via remote connections via libssl are also unlikely to be able to
1776 trigger these issues because of message size limits enforced within libssl.
1778 This issue was reported to OpenSSL Guido Vranken.
1782 *) Side channel attack on modular exponentiation
1784 A side-channel attack was found which makes use of cache-bank conflicts on
1785 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1786 of RSA keys. The ability to exploit this issue is limited as it relies on
1787 an attacker who has control of code in a thread running on the same
1788 hyper-threaded core as the victim thread which is performing decryptions.
1790 This issue was reported to OpenSSL by Yuval Yarom, The University of
1791 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1792 Nadia Heninger, University of Pennsylvania with more information at
1793 http://cachebleed.info.
1797 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1798 if no keysize is specified with default_bits. This fixes an
1799 omission in an earlier change that changed all RSA/DSA key generation
1800 apps to use 2048 bits by default.
1803 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1804 *) DH small subgroups
1806 Historically OpenSSL only ever generated DH parameters based on "safe"
1807 primes. More recently (in version 1.0.2) support was provided for
1808 generating X9.42 style parameter files such as those required for RFC 5114
1809 support. The primes used in such files may not be "safe". Where an
1810 application is using DH configured with parameters based on primes that are
1811 not "safe" then an attacker could use this fact to find a peer's private
1812 DH exponent. This attack requires that the attacker complete multiple
1813 handshakes in which the peer uses the same private DH exponent. For example
1814 this could be used to discover a TLS server's private DH exponent if it's
1815 reusing the private DH exponent or it's using a static DH ciphersuite.
1817 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1818 TLS. It is not on by default. If the option is not set then the server
1819 reuses the same private DH exponent for the life of the server process and
1820 would be vulnerable to this attack. It is believed that many popular
1821 applications do set this option and would therefore not be at risk.
1823 The fix for this issue adds an additional check where a "q" parameter is
1824 available (as is the case in X9.42 based parameters). This detects the
1825 only known attack, and is the only possible defense for static DH
1826 ciphersuites. This could have some performance impact.
1828 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1829 default and cannot be disabled. This could have some performance impact.
1831 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1835 *) SSLv2 doesn't block disabled ciphers
1837 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1838 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1839 been disabled, provided that the SSLv2 protocol was not also disabled via
1842 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1843 and Sebastian Schinzel.
1847 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1849 *) BN_mod_exp may produce incorrect results on x86_64
1851 There is a carry propagating bug in the x86_64 Montgomery squaring
1852 procedure. No EC algorithms are affected. Analysis suggests that attacks
1853 against RSA and DSA as a result of this defect would be very difficult to
1854 perform and are not believed likely. Attacks against DH are considered just
1855 feasible (although very difficult) because most of the work necessary to
1856 deduce information about a private key may be performed offline. The amount
1857 of resources required for such an attack would be very significant and
1858 likely only accessible to a limited number of attackers. An attacker would
1859 additionally need online access to an unpatched system using the target
1860 private key in a scenario with persistent DH parameters and a private
1861 key that is shared between multiple clients. For example this can occur by
1862 default in OpenSSL DHE based SSL/TLS ciphersuites.
1864 This issue was reported to OpenSSL by Hanno Böck.
1868 *) Certificate verify crash with missing PSS parameter
1870 The signature verification routines will crash with a NULL pointer
1871 dereference if presented with an ASN.1 signature using the RSA PSS
1872 algorithm and absent mask generation function parameter. Since these
1873 routines are used to verify certificate signature algorithms this can be
1874 used to crash any certificate verification operation and exploited in a
1875 DoS attack. Any application which performs certificate verification is
1876 vulnerable including OpenSSL clients and servers which enable client
1879 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1883 *) X509_ATTRIBUTE memory leak
1885 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1886 memory. This structure is used by the PKCS#7 and CMS routines so any
1887 application which reads PKCS#7 or CMS data from untrusted sources is
1888 affected. SSL/TLS is not affected.
1890 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1895 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1896 This changes the decoding behaviour for some invalid messages,
1897 though the change is mostly in the more lenient direction, and
1898 legacy behaviour is preserved as much as possible.
1901 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1903 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1905 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1907 *) Alternate chains certificate forgery
1909 During certificate verification, OpenSSL will attempt to find an
1910 alternative certificate chain if the first attempt to build such a chain
1911 fails. An error in the implementation of this logic can mean that an
1912 attacker could cause certain checks on untrusted certificates to be
1913 bypassed, such as the CA flag, enabling them to use a valid leaf
1914 certificate to act as a CA and "issue" an invalid certificate.
1916 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1920 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1922 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1923 incompatibility in the handling of HMAC. The previous ABI has now been
1927 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1929 *) Malformed ECParameters causes infinite loop
1931 When processing an ECParameters structure OpenSSL enters an infinite loop
1932 if the curve specified is over a specially malformed binary polynomial
1935 This can be used to perform denial of service against any
1936 system which processes public keys, certificate requests or
1937 certificates. This includes TLS clients and TLS servers with
1938 client authentication enabled.
1940 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1944 *) Exploitable out-of-bounds read in X509_cmp_time
1946 X509_cmp_time does not properly check the length of the ASN1_TIME
1947 string and can read a few bytes out of bounds. In addition,
1948 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1951 An attacker can use this to craft malformed certificates and CRLs of
1952 various sizes and potentially cause a segmentation fault, resulting in
1953 a DoS on applications that verify certificates or CRLs. TLS clients
1954 that verify CRLs are affected. TLS clients and servers with client
1955 authentication enabled may be affected if they use custom verification
1958 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1959 independently by Hanno Böck.
1963 *) PKCS7 crash with missing EnvelopedContent
1965 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1966 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1967 with missing content and trigger a NULL pointer dereference on parsing.
1969 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1970 structures from untrusted sources are affected. OpenSSL clients and
1971 servers are not affected.
1973 This issue was reported to OpenSSL by Michal Zalewski (Google).
1977 *) CMS verify infinite loop with unknown hash function
1979 When verifying a signedData message the CMS code can enter an infinite loop
1980 if presented with an unknown hash function OID. This can be used to perform
1981 denial of service against any system which verifies signedData messages using
1983 This issue was reported to OpenSSL by Johannes Bauer.
1987 *) Race condition handling NewSessionTicket
1989 If a NewSessionTicket is received by a multi-threaded client when attempting to
1990 reuse a previous ticket then a race condition can occur potentially leading to
1991 a double free of the ticket data.
1995 *) Only support 256-bit or stronger elliptic curves with the
1996 'ecdh_auto' setting (server) or by default (client). Of supported
1997 curves, prefer P-256 (both).
2000 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2002 *) ClientHello sigalgs DoS fix
2004 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2005 invalid signature algorithms extension a NULL pointer dereference will
2006 occur. This can be exploited in a DoS attack against the server.
2008 This issue was was reported to OpenSSL by David Ramos of Stanford
2011 [Stephen Henson and Matt Caswell]
2013 *) Multiblock corrupted pointer fix
2015 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2016 feature only applies on 64 bit x86 architecture platforms that support AES
2017 NI instructions. A defect in the implementation of "multiblock" can cause
2018 OpenSSL's internal write buffer to become incorrectly set to NULL when
2019 using non-blocking IO. Typically, when the user application is using a
2020 socket BIO for writing, this will only result in a failed connection.
2021 However if some other BIO is used then it is likely that a segmentation
2022 fault will be triggered, thus enabling a potential DoS attack.
2024 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2028 *) Segmentation fault in DTLSv1_listen fix
2030 The DTLSv1_listen function is intended to be stateless and processes the
2031 initial ClientHello from many peers. It is common for user code to loop
2032 over the call to DTLSv1_listen until a valid ClientHello is received with
2033 an associated cookie. A defect in the implementation of DTLSv1_listen means
2034 that state is preserved in the SSL object from one invocation to the next
2035 that can lead to a segmentation fault. Errors processing the initial
2036 ClientHello can trigger this scenario. An example of such an error could be
2037 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2040 This issue was reported to OpenSSL by Per Allansson.
2044 *) Segmentation fault in ASN1_TYPE_cmp fix
2046 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2047 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2048 certificate signature algorithm consistency this can be used to crash any
2049 certificate verification operation and exploited in a DoS attack. Any
2050 application which performs certificate verification is vulnerable including
2051 OpenSSL clients and servers which enable client authentication.
2055 *) Segmentation fault for invalid PSS parameters fix
2057 The signature verification routines will crash with a NULL pointer
2058 dereference if presented with an ASN.1 signature using the RSA PSS
2059 algorithm and invalid parameters. Since these routines are used to verify
2060 certificate signature algorithms this can be used to crash any
2061 certificate verification operation and exploited in a DoS attack. Any
2062 application which performs certificate verification is vulnerable including
2063 OpenSSL clients and servers which enable client authentication.
2065 This issue was was reported to OpenSSL by Brian Carpenter.
2069 *) ASN.1 structure reuse memory corruption fix
2071 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2072 memory corruption via an invalid write. Such reuse is and has been
2073 strongly discouraged and is believed to be rare.
2075 Applications that parse structures containing CHOICE or ANY DEFINED BY
2076 components may be affected. Certificate parsing (d2i_X509 and related
2077 functions) are however not affected. OpenSSL clients and servers are
2082 *) PKCS7 NULL pointer dereferences fix
2084 The PKCS#7 parsing code does not handle missing outer ContentInfo
2085 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2086 missing content and trigger a NULL pointer dereference on parsing.
2088 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2089 otherwise parse PKCS#7 structures from untrusted sources are
2090 affected. OpenSSL clients and servers are not affected.
2092 This issue was reported to OpenSSL by Michal Zalewski (Google).
2096 *) DoS via reachable assert in SSLv2 servers fix
2098 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2099 servers that both support SSLv2 and enable export cipher suites by sending
2100 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2102 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2103 (OpenSSL development team).
2107 *) Empty CKE with client auth and DHE fix
2109 If client auth is used then a server can seg fault in the event of a DHE
2110 ciphersuite being selected and a zero length ClientKeyExchange message
2111 being sent by the client. This could be exploited in a DoS attack.
2115 *) Handshake with unseeded PRNG fix
2117 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2118 with an unseeded PRNG. The conditions are:
2119 - The client is on a platform where the PRNG has not been seeded
2120 automatically, and the user has not seeded manually
2121 - A protocol specific client method version has been used (i.e. not
2122 SSL_client_methodv23)
2123 - A ciphersuite is used that does not require additional random data from
2124 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2126 If the handshake succeeds then the client random that has been used will
2127 have been generated from a PRNG with insufficient entropy and therefore the
2128 output may be predictable.
2130 For example using the following command with an unseeded openssl will
2131 succeed on an unpatched platform:
2133 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2137 *) Use After Free following d2i_ECPrivatekey error fix
2139 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2140 could cause a use after free condition. This, in turn, could cause a double
2141 free in several private key parsing functions (such as d2i_PrivateKey
2142 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2143 for applications that receive EC private keys from untrusted
2144 sources. This scenario is considered rare.
2146 This issue was discovered by the BoringSSL project and fixed in their
2151 *) X509_to_X509_REQ NULL pointer deref fix
2153 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2154 the certificate key is invalid. This function is rarely used in practice.
2156 This issue was discovered by Brian Carpenter.
2160 *) Removed the export ciphers from the DEFAULT ciphers
2163 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2165 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2166 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2167 So far those who have to target multiple platforms would compromise
2168 and argue that binary targeting say ARMv5 would still execute on
2169 ARMv8. "Universal" build resolves this compromise by providing
2170 near-optimal performance even on newer platforms.
2173 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2174 (other platforms pending).
2175 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2177 *) Add support for the SignedCertificateTimestampList certificate and
2178 OCSP response extensions from RFC6962.
2181 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2182 for corner cases. (Certain input points at infinity could lead to
2183 bogus results, with non-infinity inputs mapped to infinity too.)
2186 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2187 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2188 common cases are optimized and there still is room for further
2189 improvements. Vector Permutation AES for Altivec is also added.
2192 *) Add support for little-endian ppc64 Linux target.
2193 [Marcelo Cerri (IBM)]
2195 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2196 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2197 are optimized and there still is room for further improvements.
2198 Both 32- and 64-bit modes are supported.
2199 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2201 *) Improved ARMv7 NEON support.
2204 *) Support for SPARC Architecture 2011 crypto extensions, first
2205 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2206 SHA256/512, MD5, GHASH and modular exponentiation.
2207 [Andy Polyakov, David Miller]
2209 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2211 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2213 *) Support for new and upcoming Intel processors, including AVX2,
2214 BMI and SHA ISA extensions. This includes additional "stitched"
2215 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2218 This work was sponsored by Intel Corp.
2221 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2222 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2223 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2226 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2227 this fixes a limitation in previous versions of OpenSSL.
2230 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2231 MGF1 digest and OAEP label.
2234 *) Add EVP support for key wrapping algorithms, to avoid problems with
2235 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2236 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2237 algorithms and include tests cases.
2240 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2242 [Douglas E. Engert, Steve Henson]
2244 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2245 difference in days and seconds between two tm or ASN1_TIME structures.
2248 *) Add -rev test option to s_server to just reverse order of characters
2249 received by client and send back to server. Also prints an abbreviated
2250 summary of the connection parameters.
2253 *) New option -brief for s_client and s_server to print out a brief summary
2254 of connection parameters.
2257 *) Add callbacks for arbitrary TLS extensions.
2258 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2260 *) New option -crl_download in several openssl utilities to download CRLs
2261 from CRLDP extension in certificates.
2264 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2267 *) New function X509_CRL_diff to generate a delta CRL from the difference
2268 of two full CRLs. Add support to "crl" utility.
2271 *) New functions to set lookup_crls function and to retrieve
2272 X509_STORE from X509_STORE_CTX.
2275 *) Print out deprecated issuer and subject unique ID fields in
2279 *) Extend OCSP I/O functions so they can be used for simple general purpose
2280 HTTP as well as OCSP. New wrapper function which can be used to download
2281 CRLs using the OCSP API.
2284 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2287 *) SSL_CONF* functions. These provide a common framework for application
2288 configuration using configuration files or command lines.
2291 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2292 message callback and prints the results. Needs compile time option
2293 "enable-ssl-trace". New options to s_client and s_server to enable
2297 *) New ctrl and macro to retrieve supported points extensions.
2298 Print out extension in s_server and s_client.
2301 *) New functions to retrieve certificate signature and signature
2305 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2309 *) New Suite B modes for TLS code. These use and enforce the requirements
2310 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2311 only use Suite B curves. The Suite B modes can be set by using the
2312 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2315 *) New chain verification flags for Suite B levels of security. Check
2316 algorithms are acceptable when flags are set in X509_verify_cert.
2319 *) Make tls1_check_chain return a set of flags indicating checks passed
2320 by a certificate chain. Add additional tests to handle client
2321 certificates: checks for matching certificate type and issuer name
2325 *) If an attempt is made to use a signature algorithm not in the peer
2326 preference list abort the handshake. If client has no suitable
2327 signature algorithms in response to a certificate request do not
2328 use the certificate.
2331 *) If server EC tmp key is not in client preference list abort handshake.
2334 *) Add support for certificate stores in CERT structure. This makes it
2335 possible to have different stores per SSL structure or one store in
2336 the parent SSL_CTX. Include distinct stores for certificate chain
2337 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2338 to build and store a certificate chain in CERT structure: returning
2339 an error if the chain cannot be built: this will allow applications
2340 to test if a chain is correctly configured.
2342 Note: if the CERT based stores are not set then the parent SSL_CTX
2343 store is used to retain compatibility with existing behaviour.
2347 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2348 mask based on the current session, check mask when sending client
2349 hello and checking the requested ciphersuite.
2352 *) New ctrls to retrieve and set certificate types in a certificate
2353 request message. Print out received values in s_client. If certificate
2354 types is not set with custom values set sensible values based on
2355 supported signature algorithms.
2358 *) Support for distinct client and server supported signature algorithms.
2361 *) Add certificate callback. If set this is called whenever a certificate
2362 is required by client or server. An application can decide which
2363 certificate chain to present based on arbitrary criteria: for example
2364 supported signature algorithms. Add very simple example to s_server.
2365 This fixes many of the problems and restrictions of the existing client
2366 certificate callback: for example you can now clear an existing
2367 certificate and specify the whole chain.
2370 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2371 the certificate can be used for (if anything). Set valid_flags field
2372 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2373 to have similar checks in it.
2375 Add new "cert_flags" field to CERT structure and include a "strict mode".
2376 This enforces some TLS certificate requirements (such as only permitting
2377 certificate signature algorithms contained in the supported algorithms
2378 extension) which some implementations ignore: this option should be used
2379 with caution as it could cause interoperability issues.
2382 *) Update and tidy signature algorithm extension processing. Work out
2383 shared signature algorithms based on preferences and peer algorithms
2384 and print them out in s_client and s_server. Abort handshake if no
2385 shared signature algorithms.
2388 *) Add new functions to allow customised supported signature algorithms
2389 for SSL and SSL_CTX structures. Add options to s_client and s_server
2393 *) New function SSL_certs_clear() to delete all references to certificates
2394 from an SSL structure. Before this once a certificate had been added
2395 it couldn't be removed.
2398 *) Integrate hostname, email address and IP address checking with certificate
2399 verification. New verify options supporting checking in openssl utility.
2402 *) Fixes and wildcard matching support to hostname and email checking
2403 functions. Add manual page.
2404 [Florian Weimer (Red Hat Product Security Team)]
2406 *) New functions to check a hostname email or IP address against a
2407 certificate. Add options x509 utility to print results of checks against
2411 *) Fix OCSP checking.
2412 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2414 *) Initial experimental support for explicitly trusted non-root CAs.
2415 OpenSSL still tries to build a complete chain to a root but if an
2416 intermediate CA has a trust setting included that is used. The first
2417 setting is used: whether to trust (e.g., -addtrust option to the x509
2421 *) Add -trusted_first option which attempts to find certificates in the
2422 trusted store even if an untrusted chain is also supplied.
2425 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2426 platform support for Linux and Android.
2429 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2432 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2433 When in FIPS mode the approved implementations are used as normal,
2434 when not in FIPS mode the internal unapproved versions are used instead.
2435 This means that the FIPS capable OpenSSL isn't forced to use the
2436 (often lower performance) FIPS implementations outside FIPS mode.
2439 *) Transparently support X9.42 DH parameters when calling
2440 PEM_read_bio_DHparameters. This means existing applications can handle
2441 the new parameter format automatically.
2444 *) Initial experimental support for X9.42 DH parameter format: mainly
2445 to support use of 'q' parameter for RFC5114 parameters.
2448 *) Add DH parameters from RFC5114 including test data to dhtest.
2451 *) Support for automatic EC temporary key parameter selection. If enabled
2452 the most preferred EC parameters are automatically used instead of
2453 hardcoded fixed parameters. Now a server just has to call:
2454 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2455 support ECDH and use the most appropriate parameters.
2458 *) Enhance and tidy EC curve and point format TLS extension code. Use
2459 static structures instead of allocation if default values are used.
2460 New ctrls to set curves we wish to support and to retrieve shared curves.
2461 Print out shared curves in s_server. New options to s_server and s_client
2462 to set list of supported curves.
2465 *) New ctrls to retrieve supported signature algorithms and
2466 supported curve values as an array of NIDs. Extend openssl utility
2467 to print out received values.
2470 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2471 between NIDs and the more common NIST names such as "P-256". Enhance
2472 ecparam utility and ECC method to recognise the NIST names for curves.
2475 *) Enhance SSL/TLS certificate chain handling to support different
2476 chains for each certificate instead of one chain in the parent SSL_CTX.
2479 *) Support for fixed DH ciphersuite client authentication: where both
2480 server and client use DH certificates with common parameters.
2483 *) Support for fixed DH ciphersuites: those requiring DH server
2487 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2489 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2490 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2491 X509_CINF_get_signature were reverted post internal team review.
2493 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2495 *) Build fixes for the Windows and OpenVMS platforms
2496 [Matt Caswell and Richard Levitte]
2498 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2500 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2501 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2502 dereference. This could lead to a Denial Of Service attack. Thanks to
2503 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2507 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2508 dtls1_buffer_record function under certain conditions. In particular this
2509 could occur if an attacker sent repeated DTLS records with the same
2510 sequence number but for the next epoch. The memory leak could be exploited
2511 by an attacker in a Denial of Service attack through memory exhaustion.
2512 Thanks to Chris Mueller for reporting this issue.
2516 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2517 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2518 method would be set to NULL which could later result in a NULL pointer
2519 dereference. Thanks to Frank Schmirler for reporting this issue.
2523 *) Abort handshake if server key exchange message is omitted for ephemeral
2526 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2527 reporting this issue.
2531 *) Remove non-export ephemeral RSA code on client and server. This code
2532 violated the TLS standard by allowing the use of temporary RSA keys in
2533 non-export ciphersuites and could be used by a server to effectively
2534 downgrade the RSA key length used to a value smaller than the server
2535 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2536 INRIA or reporting this issue.
2540 *) Fixed issue where DH client certificates are accepted without verification.
2541 An OpenSSL server will accept a DH certificate for client authentication
2542 without the certificate verify message. This effectively allows a client to
2543 authenticate without the use of a private key. This only affects servers
2544 which trust a client certificate authority which issues certificates
2545 containing DH keys: these are extremely rare and hardly ever encountered.
2546 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2551 *) Ensure that the session ID context of an SSL is updated when its
2552 SSL_CTX is updated via SSL_set_SSL_CTX.
2554 The session ID context is typically set from the parent SSL_CTX,
2555 and can vary with the CTX.
2558 *) Fix various certificate fingerprint issues.
2560 By using non-DER or invalid encodings outside the signed portion of a
2561 certificate the fingerprint can be changed without breaking the signature.
2562 Although no details of the signed portion of the certificate can be changed
2563 this can cause problems with some applications: e.g. those using the
2564 certificate fingerprint for blacklists.
2566 1. Reject signatures with non zero unused bits.
2568 If the BIT STRING containing the signature has non zero unused bits reject
2569 the signature. All current signature algorithms require zero unused bits.
2571 2. Check certificate algorithm consistency.
2573 Check the AlgorithmIdentifier inside TBS matches the one in the
2574 certificate signature. NB: this will result in signature failure
2575 errors for some broken certificates.
2577 Thanks to Konrad Kraszewski from Google for reporting this issue.
2579 3. Check DSA/ECDSA signatures use DER.
2581 Re-encode DSA/ECDSA signatures and compare with the original received
2582 signature. Return an error if there is a mismatch.
2584 This will reject various cases including garbage after signature
2585 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2586 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2587 (negative or with leading zeroes).
2589 Further analysis was conducted and fixes were developed by Stephen Henson
2590 of the OpenSSL core team.
2595 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2596 results on some platforms, including x86_64. This bug occurs at random
2597 with a very low probability, and is not known to be exploitable in any
2598 way, though its exact impact is difficult to determine. Thanks to Pieter
2599 Wuille (Blockstream) who reported this issue and also suggested an initial
2600 fix. Further analysis was conducted by the OpenSSL development team and
2601 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2602 the OpenSSL core team.
2606 *) Do not resume sessions on the server if the negotiated protocol
2607 version does not match the session's version. Resuming with a different
2608 version, while not strictly forbidden by the RFC, is of questionable
2609 sanity and breaks all known clients.
2610 [David Benjamin, Emilia Käsper]
2612 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2613 early CCS messages during renegotiation. (Note that because
2614 renegotiation is encrypted, this early CCS was not exploitable.)
2617 *) Tighten client-side session ticket handling during renegotiation:
2618 ensure that the client only accepts a session ticket if the server sends
2619 the extension anew in the ServerHello. Previously, a TLS client would
2620 reuse the old extension state and thus accept a session ticket if one was
2621 announced in the initial ServerHello.
2623 Similarly, ensure that the client requires a session ticket if one
2624 was advertised in the ServerHello. Previously, a TLS client would
2625 ignore a missing NewSessionTicket message.
2628 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2630 *) SRTP Memory Leak.
2632 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2633 sends a carefully crafted handshake message, to cause OpenSSL to fail
2634 to free up to 64k of memory causing a memory leak. This could be
2635 exploited in a Denial Of Service attack. This issue affects OpenSSL
2636 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2637 whether SRTP is used or configured. Implementations of OpenSSL that
2638 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2640 The fix was developed by the OpenSSL team.
2644 *) Session Ticket Memory Leak.
2646 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2647 integrity of that ticket is first verified. In the event of a session
2648 ticket integrity check failing, OpenSSL will fail to free memory
2649 causing a memory leak. By sending a large number of invalid session
2650 tickets an attacker could exploit this issue in a Denial Of Service
2655 *) Build option no-ssl3 is incomplete.
2657 When OpenSSL is configured with "no-ssl3" as a build option, servers
2658 could accept and complete a SSL 3.0 handshake, and clients could be
2659 configured to send them.
2661 [Akamai and the OpenSSL team]
2663 *) Add support for TLS_FALLBACK_SCSV.
2664 Client applications doing fallback retries should call
2665 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2667 [Adam Langley, Bodo Moeller]
2669 *) Add additional DigestInfo checks.
2671 Re-encode DigestInto in DER and check against the original when
2672 verifying RSA signature: this will reject any improperly encoded
2673 DigestInfo structures.
2675 Note: this is a precautionary measure and no attacks are currently known.
2679 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2681 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2682 SRP code can be overrun an internal buffer. Add sanity check that
2683 g, A, B < N to SRP code.
2685 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2686 Group for discovering this issue.
2690 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2691 TLS 1.0 instead of higher protocol versions when the ClientHello message
2692 is badly fragmented. This allows a man-in-the-middle attacker to force a
2693 downgrade to TLS 1.0 even if both the server and the client support a
2694 higher protocol version, by modifying the client's TLS records.
2696 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2697 researching this issue.
2701 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2702 to a denial of service attack. A malicious server can crash the client
2703 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2704 ciphersuite and sending carefully crafted handshake messages.
2706 Thanks to Felix Gröbert (Google) for discovering and researching this
2711 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2712 to leak memory. This can be exploited through a Denial of Service attack.
2713 Thanks to Adam Langley for discovering and researching this issue.
2717 *) An attacker can force openssl to consume large amounts of memory whilst
2718 processing DTLS handshake messages. This can be exploited through a
2719 Denial of Service attack.
2720 Thanks to Adam Langley for discovering and researching this issue.
2724 *) An attacker can force an error condition which causes openssl to crash
2725 whilst processing DTLS packets due to memory being freed twice. This
2726 can be exploited through a Denial of Service attack.
2727 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2732 *) If a multithreaded client connects to a malicious server using a resumed
2733 session and the server sends an ec point format extension it could write
2734 up to 255 bytes to freed memory.
2736 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2741 *) A malicious server can crash an OpenSSL client with a null pointer
2742 dereference (read) by specifying an SRP ciphersuite even though it was not
2743 properly negotiated with the client. This can be exploited through a
2744 Denial of Service attack.
2746 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2747 discovering and researching this issue.
2751 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2752 X509_name_oneline, X509_name_print_ex et al. to leak some information
2753 from the stack. Applications may be affected if they echo pretty printing
2754 output to the attacker.
2756 Thanks to Ivan Fratric (Google) for discovering this issue.
2758 [Emilia Käsper, and Steve Henson]
2760 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2761 for corner cases. (Certain input points at infinity could lead to
2762 bogus results, with non-infinity inputs mapped to infinity too.)
2765 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2767 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2768 handshake can force the use of weak keying material in OpenSSL
2769 SSL/TLS clients and servers.
2771 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2772 researching this issue. (CVE-2014-0224)
2773 [KIKUCHI Masashi, Steve Henson]
2775 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2776 OpenSSL DTLS client the code can be made to recurse eventually crashing
2779 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2781 [Imre Rad, Steve Henson]
2783 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2784 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2785 client or server. This is potentially exploitable to run arbitrary
2786 code on a vulnerable client or server.
2788 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2789 [Jüri Aedla, Steve Henson]
2791 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2792 are subject to a denial of service attack.
2794 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2795 this issue. (CVE-2014-3470)
2796 [Felix Gröbert, Ivan Fratric, Steve Henson]
2798 *) Harmonize version and its documentation. -f flag is used to display
2800 [mancha <mancha1@zoho.com>]
2802 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2803 in i2d_ECPrivateKey.
2804 [mancha <mancha1@zoho.com>]
2806 *) Fix some double frees. These are not thought to be exploitable.
2807 [mancha <mancha1@zoho.com>]
2809 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2811 *) A missing bounds check in the handling of the TLS heartbeat extension
2812 can be used to reveal up to 64k of memory to a connected client or
2815 Thanks for Neel Mehta of Google Security for discovering this bug and to
2816 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2817 preparing the fix (CVE-2014-0160)
2818 [Adam Langley, Bodo Moeller]
2820 *) Fix for the attack described in the paper "Recovering OpenSSL
2821 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2822 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2823 http://eprint.iacr.org/2014/140
2825 Thanks to Yuval Yarom and Naomi Benger for discovering this
2826 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2827 [Yuval Yarom and Naomi Benger]
2829 *) TLS pad extension: draft-agl-tls-padding-03
2831 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2832 TLS client Hello record length value would otherwise be > 255 and
2833 less that 512 pad with a dummy extension containing zeroes so it
2834 is at least 512 bytes long.
2836 [Adam Langley, Steve Henson]
2838 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2840 *) Fix for TLS record tampering bug. A carefully crafted invalid
2841 handshake could crash OpenSSL with a NULL pointer exception.
2842 Thanks to Anton Johansson for reporting this issues.
2845 *) Keep original DTLS digest and encryption contexts in retransmission
2846 structures so we can use the previous session parameters if they need
2847 to be resent. (CVE-2013-6450)
2850 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2851 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2852 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2853 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2854 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2855 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2856 [Rob Stradling, Adam Langley]
2858 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2860 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2861 supporting platforms or when small records were transferred.
2862 [Andy Polyakov, Steve Henson]
2864 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2866 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2868 This addresses the flaw in CBC record processing discovered by
2869 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2870 at: http://www.isg.rhul.ac.uk/tls/
2872 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2873 Security Group at Royal Holloway, University of London
2874 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2875 Emilia Käsper for the initial patch.
2877 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2879 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2880 ciphersuites which can be exploited in a denial of service attack.
2881 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2882 and detecting this bug and to Wolfgang Ettlinger
2883 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2887 *) Return an error when checking OCSP signatures when key is NULL.
2888 This fixes a DoS attack. (CVE-2013-0166)
2891 *) Make openssl verify return errors.
2892 [Chris Palmer <palmer@google.com> and Ben Laurie]
2894 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2895 the right response is stapled. Also change SSL_get_certificate()
2896 so it returns the certificate actually sent.
2897 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2898 [Rob Stradling <rob.stradling@comodo.com>]
2900 *) Fix possible deadlock when decoding public keys.
2903 *) Don't use TLS 1.0 record version number in initial client hello
2907 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2909 *) Sanity check record length before skipping explicit IV in TLS
2910 1.2, 1.1 and DTLS to fix DoS attack.
2912 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2913 fuzzing as a service testing platform.
2917 *) Initialise tkeylen properly when encrypting CMS messages.
2918 Thanks to Solar Designer of Openwall for reporting this issue.
2921 *) In FIPS mode don't try to use composite ciphers as they are not
2925 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2927 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2928 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2929 mean any application compiled against OpenSSL 1.0.0 headers setting
2930 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2931 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2932 0x10000000L Any application which was previously compiled against
2933 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2934 will need to be recompiled as a result. Letting be results in
2935 inability to disable specifically TLS 1.1 and in client context,
2936 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2939 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2940 disable just protocol X, but all protocols above X *if* there are
2941 protocols *below* X still enabled. In more practical terms it means
2942 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2943 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2944 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2948 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2950 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2951 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2952 in CRYPTO_realloc_clean.
2954 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2955 issue and to Adam Langley <agl@chromium.org> for fixing it.
2957 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2959 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2962 *) Workarounds for some broken servers that "hang" if a client hello
2963 record length exceeds 255 bytes.
2965 1. Do not use record version number > TLS 1.0 in initial client
2966 hello: some (but not all) hanging servers will now work.
2967 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2968 the number of ciphers sent in the client hello. This should be
2969 set to an even number, such as 50, for example by passing:
2970 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2971 Most broken servers should now work.
2972 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2973 TLS 1.2 client support entirely.
2976 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2979 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2981 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2982 STRING form instead of a DigestInfo.
2985 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2986 and the RSA_sign/RSA_verify functions. This was made more apparent when
2987 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2988 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2989 the correct format in RSA_verify so both forms transparently work.
2992 *) Some servers which support TLS 1.0 can choke if we initially indicate
2993 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2994 encrypted premaster secret. As a workaround use the maximum permitted
2995 client version in client hello, this should keep such servers happy
2996 and still work with previous versions of OpenSSL.
2999 *) Add support for TLS/DTLS heartbeats.
3000 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3002 *) Add support for SCTP.
3003 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3005 *) Improved PRNG seeding for VOS.
3006 [Paul Green <Paul.Green@stratus.com>]
3008 *) Extensive assembler packs updates, most notably:
3010 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3011 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3012 - x86_64: bit-sliced AES implementation;
3013 - ARM: NEON support, contemporary platforms optimizations;
3014 - s390x: z196 support;
3015 - *: GHASH and GF(2^m) multiplication implementations;
3019 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3020 (removal of unnecessary code)
3021 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3023 *) Add TLS key material exporter from RFC 5705.
3026 *) Add DTLS-SRTP negotiation from RFC 5764.
3029 *) Add Next Protocol Negotiation,
3030 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3031 disabled with a no-npn flag to config or Configure. Code donated
3033 [Adam Langley <agl@google.com> and Ben Laurie]
3035 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3036 NIST-P256, NIST-P521, with constant-time single point multiplication on
3037 typical inputs. Compiler support for the nonstandard type __uint128_t is
3038 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3039 Code made available under Apache License version 2.0.
3041 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3042 line to include this in your build of OpenSSL, and run "make depend" (or
3043 "make update"). This enables the following EC_METHODs:
3045 EC_GFp_nistp224_method()
3046 EC_GFp_nistp256_method()
3047 EC_GFp_nistp521_method()
3049 EC_GROUP_new_by_curve_name() will automatically use these (while
3050 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3052 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3054 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3055 all platforms. Move ssize_t definition from e_os.h to the public
3056 header file e_os2.h as it now appears in public header file cms.h
3059 *) New -sigopt option to the ca, req and x509 utilities. Additional
3060 signature parameters can be passed using this option and in
3064 *) Add RSA PSS signing function. This will generate and set the
3065 appropriate AlgorithmIdentifiers for PSS based on those in the
3066 corresponding EVP_MD_CTX structure. No application support yet.
3069 *) Support for companion algorithm specific ASN1 signing routines.
3070 New function ASN1_item_sign_ctx() signs a pre-initialised
3071 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3072 the appropriate parameters.
3075 *) Add new algorithm specific ASN1 verification initialisation function
3076 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3077 handling will be the same no matter what EVP_PKEY_METHOD is used.
3078 Add a PSS handler to support verification of PSS signatures: checked
3079 against a number of sample certificates.
3082 *) Add signature printing for PSS. Add PSS OIDs.
3083 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3085 *) Add algorithm specific signature printing. An individual ASN1 method
3086 can now print out signatures instead of the standard hex dump.
3088 More complex signatures (e.g. PSS) can print out more meaningful
3089 information. Include DSA version that prints out the signature
3093 *) Password based recipient info support for CMS library: implementing
3097 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3098 neatly separates the code into cipher and PBE sections and is required
3099 for some algorithms that split PBES2 into separate pieces (such as
3100 password based CMS).
3103 *) Session-handling fixes:
3104 - Fix handling of connections that are resuming with a session ID,
3105 but also support Session Tickets.
3106 - Fix a bug that suppressed issuing of a new ticket if the client
3107 presented a ticket with an expired session.
3108 - Try to set the ticket lifetime hint to something reasonable.
3109 - Make tickets shorter by excluding irrelevant information.
3110 - On the client side, don't ignore renewed tickets.
3111 [Adam Langley, Bodo Moeller (Google)]
3113 *) Fix PSK session representation.
3116 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3118 This work was sponsored by Intel.
3121 *) Add GCM support to TLS library. Some custom code is needed to split
3122 the IV between the fixed (from PRF) and explicit (from TLS record)
3123 portions. This adds all GCM ciphersuites supported by RFC5288 and
3124 RFC5289. Generalise some AES* cipherstrings to include GCM and
3125 add a special AESGCM string for GCM only.
3128 *) Expand range of ctrls for AES GCM. Permit setting invocation
3129 field on decrypt and retrieval of invocation field only on encrypt.
3132 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3133 As required by RFC5289 these ciphersuites cannot be used if for
3134 versions of TLS earlier than 1.2.
3137 *) For FIPS capable OpenSSL interpret a NULL default public key method
3138 as unset and return the appropriate default but do *not* set the default.
3139 This means we can return the appropriate method in applications that
3140 switch between FIPS and non-FIPS modes.
3143 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3144 ENGINE is used then we cannot handle that in the FIPS module so we
3145 keep original code iff non-FIPS operations are allowed.
3148 *) Add -attime option to openssl utilities.
3149 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3151 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3154 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3155 FIPS EC methods unconditionally for now.
3158 *) New build option no-ec2m to disable characteristic 2 code.
3161 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3162 all cases can be covered as some introduce binary incompatibilities.
3165 *) Redirect RSA operations to FIPS module including keygen,
3166 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3169 *) Add similar low level API blocking to ciphers.
3172 *) Low level digest APIs are not approved in FIPS mode: any attempt
3173 to use these will cause a fatal error. Applications that *really* want
3174 to use them can use the private_* version instead.
3177 *) Redirect cipher operations to FIPS module for FIPS builds.
3180 *) Redirect digest operations to FIPS module for FIPS builds.
3183 *) Update build system to add "fips" flag which will link in fipscanister.o
3184 for static and shared library builds embedding a signature if needed.
3187 *) Output TLS supported curves in preference order instead of numerical
3188 order. This is currently hardcoded for the highest order curves first.
3189 This should be configurable so applications can judge speed vs strength.
3192 *) Add TLS v1.2 server support for client authentication.
3195 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3199 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3200 FIPS modules versions.
3203 *) Add TLS v1.2 client side support for client authentication. Keep cache
3204 of handshake records longer as we don't know the hash algorithm to use
3205 until after the certificate request message is received.
3208 *) Initial TLS v1.2 client support. Add a default signature algorithms
3209 extension including all the algorithms we support. Parse new signature
3210 format in client key exchange. Relax some ECC signing restrictions for
3211 TLS v1.2 as indicated in RFC5246.
3214 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3215 to new signature format when needed using client digest preference.
3216 All server ciphersuites should now work correctly in TLS v1.2. No client
3217 support yet and no support for client certificates.
3220 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3221 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3222 ciphersuites. At present only RSA key exchange ciphersuites work with
3223 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3224 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3225 and version checking.
3228 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3229 with this defined it will not be affected by any changes to ssl internal
3230 structures. Add several utility functions to allow openssl application
3231 to work with OPENSSL_NO_SSL_INTERN defined.
3234 *) A long standing patch to add support for SRP from EdelWeb (Peter
3235 Sylvester and Christophe Renou) was integrated.
3236 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3237 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3240 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3243 *) Permit abbreviated handshakes when renegotiating using the function
3244 SSL_renegotiate_abbreviated().
3245 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3247 *) Add call to ENGINE_register_all_complete() to
3248 ENGINE_load_builtin_engines(), so some implementations get used
3249 automatically instead of needing explicit application support.
3252 *) Add support for TLS key exporter as described in RFC5705.
3253 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3255 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3256 a few changes are required:
3258 Add SSL_OP_NO_TLSv1_1 flag.
3259 Add TLSv1_1 methods.
3260 Update version checking logic to handle version 1.1.
3261 Add explicit IV handling (ported from DTLS code).
3262 Add command line options to s_client/s_server.
3265 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3267 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3268 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3269 content decryption and always return the same error. Note: this attack
3270 needs on average 2^20 messages so it only affects automated senders. The
3271 old behaviour can be re-enabled in the CMS code by setting the
3272 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3273 an MMA defence is not necessary.
3274 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3275 this issue. (CVE-2012-0884)
3278 *) Fix CVE-2011-4619: make sure we really are receiving a
3279 client hello before rejecting multiple SGC restarts. Thanks to
3280 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3283 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3285 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3286 Thanks to Antonio Martin, Enterprise Secure Access Research and
3287 Development, Cisco Systems, Inc. for discovering this bug and
3288 preparing a fix. (CVE-2012-0050)
3291 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3293 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3294 of the Vaudenay padding oracle attack on CBC mode encryption
3295 which enables an efficient plaintext recovery attack against
3296 the OpenSSL implementation of DTLS. Their attack exploits timing
3297 differences arising during decryption processing. A research
3298 paper describing this attack can be found at:
3299 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3300 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3301 Security Group at Royal Holloway, University of London
3302 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3303 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3304 for preparing the fix. (CVE-2011-4108)
3305 [Robin Seggelmann, Michael Tuexen]
3307 *) Clear bytes used for block padding of SSL 3.0 records.
3309 [Adam Langley (Google)]
3311 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3312 Kadianakis <desnacked@gmail.com> for discovering this issue and
3313 Adam Langley for preparing the fix. (CVE-2011-4619)
3314 [Adam Langley (Google)]
3316 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3317 [Andrey Kulikov <amdeich@gmail.com>]
3319 *) Prevent malformed RFC3779 data triggering an assertion failure.
3320 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3321 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3322 [Rob Austein <sra@hactrn.net>]
3324 *) Improved PRNG seeding for VOS.
3325 [Paul Green <Paul.Green@stratus.com>]
3327 *) Fix ssl_ciph.c set-up race.
3328 [Adam Langley (Google)]
3330 *) Fix spurious failures in ecdsatest.c.
3331 [Emilia Käsper (Google)]
3333 *) Fix the BIO_f_buffer() implementation (which was mixing different
3334 interpretations of the '..._len' fields).
3335 [Adam Langley (Google)]
3337 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3338 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3339 threads won't reuse the same blinding coefficients.
3341 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3342 lock to call BN_BLINDING_invert_ex, and avoids one use of
3343 BN_BLINDING_update for each BN_BLINDING structure (previously,
3344 the last update always remained unused).
3345 [Emilia Käsper (Google)]
3347 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3348 [Bob Buckholz (Google)]
3350 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3352 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3353 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3354 [Kaspar Brand <ossl@velox.ch>]
3356 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3357 for multi-threaded use of ECDH. (CVE-2011-3210)
3358 [Adam Langley (Google)]
3360 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3363 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3364 signature public key algorithm by using OID xref utilities instead.
3365 Before this you could only use some ECC ciphersuites with SHA1 only.
3368 *) Add protection against ECDSA timing attacks as mentioned in the paper
3369 by Billy Bob Brumley and Nicola Tuveri, see:
3371 http://eprint.iacr.org/2011/232.pdf
3373 [Billy Bob Brumley and Nicola Tuveri]
3375 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3377 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3378 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3380 *) Fix bug in string printing code: if *any* escaping is enabled we must
3381 escape the escape character (backslash) or the resulting string is
3385 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3387 *) Disable code workaround for ancient and obsolete Netscape browsers
3388 and servers: an attacker can use it in a ciphersuite downgrade attack.
3389 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3392 *) Fixed J-PAKE implementation error, originally discovered by
3393 Sebastien Martini, further info and confirmation from Stefan
3394 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3397 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3399 *) Fix extension code to avoid race conditions which can result in a buffer
3400 overrun vulnerability: resumed sessions must not be modified as they can
3401 be shared by multiple threads. CVE-2010-3864
3404 *) Fix WIN32 build system to correctly link an ENGINE directory into
3408 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3410 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3412 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3414 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3416 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3417 context. The operation can be customised via the ctrl mechanism in
3418 case ENGINEs want to include additional functionality.
3421 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3424 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3425 output hashes compatible with older versions of OpenSSL.
3426 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3428 *) Fix compression algorithm handling: if resuming a session use the
3429 compression algorithm of the resumed session instead of determining
3430 it from client hello again. Don't allow server to change algorithm.
3433 *) Add load_crls() function to apps tidying load_certs() too. Add option
3434 to verify utility to allow additional CRLs to be included.
3437 *) Update OCSP request code to permit adding custom headers to the request:
3438 some responders need this.
3441 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3443 [Julia Lawall <julia@diku.dk>]
3445 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3446 needlessly dereferenced structures, used obsolete functions and
3447 didn't handle all updated verify codes correctly.
3450 *) Disable MD2 in the default configuration.
3453 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3454 indicate the initial BIO being pushed or popped. This makes it possible
3455 to determine whether the BIO is the one explicitly called or as a result
3456 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3457 it handles reference counts correctly and doesn't zero out the I/O bio
3458 when it is not being explicitly popped. WARNING: applications which
3459 included workarounds for the old buggy behaviour will need to be modified
3460 or they could free up already freed BIOs.
3463 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3464 renaming to all platforms (within the 0.9.8 branch, this was
3465 done conditionally on Netware platforms to avoid a name clash).
3466 [Guenter <lists@gknw.net>]
3468 *) Add ECDHE and PSK support to DTLS.
3469 [Michael Tuexen <tuexen@fh-muenster.de>]
3471 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3475 *) Add "missing" function EVP_MD_flags() (without this the only way to
3476 retrieve a digest flags is by accessing the structure directly. Update
3477 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3478 or cipher is registered as in the "from" argument. Print out all
3479 registered digests in the dgst usage message instead of manually
3480 attempting to work them out.
3483 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3484 this allows the use of compression and extensions. Change default cipher
3485 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3486 by default unless an application cipher string requests it.
3489 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3490 key ids to find matching certificates and keys but some PKCS#12 files
3491 don't follow the (somewhat unwritten) rules and this strategy fails.
3492 Now just gather all certificates together and the first private key
3493 then look for the first certificate that matches the key.
3496 *) Support use of registered digest and cipher names for dgst and cipher
3497 commands instead of having to add each one as a special case. So now
3504 openssl dgst -sha256 foo
3506 and this works for ENGINE based algorithms too.
3510 *) Update Gost ENGINE to support parameter files.
3511 [Victor B. Wagner <vitus@cryptocom.ru>]
3513 *) Support GeneralizedTime in ca utility.
3514 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3516 *) Enhance the hash format used for certificate directory links. The new
3517 form uses the canonical encoding (meaning equivalent names will work
3518 even if they aren't identical) and uses SHA1 instead of MD5. This form
3519 is incompatible with the older format and as a result c_rehash should
3520 be used to rebuild symbolic links.
3523 *) Make PKCS#8 the default write format for private keys, replacing the
3524 traditional format. This form is standardised, more secure and doesn't
3525 include an implicit MD5 dependency.
3528 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3529 committed to OpenSSL should pass this lot as a minimum.
3532 *) Add session ticket override functionality for use by EAP-FAST.
3533 [Jouni Malinen <j@w1.fi>]
3535 *) Modify HMAC functions to return a value. Since these can be implemented
3536 in an ENGINE errors can occur.
3539 *) Type-checked OBJ_bsearch_ex.
3542 *) Type-checked OBJ_bsearch. Also some constification necessitated
3543 by type-checking. Still to come: TXT_DB, bsearch(?),
3544 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3548 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3549 seconds to a tm structure directly, instead of going through OS
3550 specific date routines. This avoids any issues with OS routines such
3551 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3552 and X509_time_adj_ex() to cover the extended range. The existing
3553 X509_time_adj() is still usable and will no longer have any date issues.
3556 *) Delta CRL support. New use deltas option which will attempt to locate
3557 and search any appropriate delta CRLs available.
3559 This work was sponsored by Google.
3562 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3563 code and add additional score elements. Validate alternate CRL paths
3564 as part of the CRL checking and indicate a new error "CRL path validation
3565 error" in this case. Applications wanting additional details can use
3566 the verify callback and check the new "parent" field. If this is not
3567 NULL CRL path validation is taking place. Existing applications won't
3568 see this because it requires extended CRL support which is off by
3571 This work was sponsored by Google.
3574 *) Support for freshest CRL extension.
3576 This work was sponsored by Google.
3579 *) Initial indirect CRL support. Currently only supported in the CRLs
3580 passed directly and not via lookup. Process certificate issuer
3581 CRL entry extension and lookup CRL entries by bother issuer name
3582 and serial number. Check and process CRL issuer entry in IDP extension.
3584 This work was sponsored by Google.
3587 *) Add support for distinct certificate and CRL paths. The CRL issuer
3588 certificate is validated separately in this case. Only enabled if
3589 an extended CRL support flag is set: this flag will enable additional
3590 CRL functionality in future.
3592 This work was sponsored by Google.
3595 *) Add support for policy mappings extension.
3597 This work was sponsored by Google.
3600 *) Fixes to pathlength constraint, self issued certificate handling,
3601 policy processing to align with RFC3280 and PKITS tests.
3603 This work was sponsored by Google.
3606 *) Support for name constraints certificate extension. DN, email, DNS
3607 and URI types are currently supported.
3609 This work was sponsored by Google.
3612 *) To cater for systems that provide a pointer-based thread ID rather
3613 than numeric, deprecate the current numeric thread ID mechanism and
3614 replace it with a structure and associated callback type. This
3615 mechanism allows a numeric "hash" to be extracted from a thread ID in
3616 either case, and on platforms where pointers are larger than 'long',
3617 mixing is done to help ensure the numeric 'hash' is usable even if it
3618 can't be guaranteed unique. The default mechanism is to use "&errno"
3619 as a pointer-based thread ID to distinguish between threads.
3621 Applications that want to provide their own thread IDs should now use
3622 CRYPTO_THREADID_set_callback() to register a callback that will call
3623 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3625 Note that ERR_remove_state() is now deprecated, because it is tied
3626 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3627 to free the current thread's error state should be replaced by
projects / openssl.git / blob