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 1.1.0j and 1.1.0k [xx XXX xxxx]
12 *) Added SCA hardening for modular field inversion in EC_GROUP through
13 a new dedicated field_inv() pointer in EC_METHOD.
14 This also addresses a leakage affecting conversions from projective
15 to affine coordinates.
16 [Billy Bob Brumley, Nicola Tuveri]
18 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
19 re-used X509_PUBKEY object if the second PUBKEY is malformed.
22 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
25 *) Remove the 'dist' target and add a tarball building script. The
26 'dist' target has fallen out of use, and it shouldn't be
27 necessary to configure just to create a source distribution.
30 Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
32 *) Timing vulnerability in DSA signature generation
34 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
35 timing side channel attack. An attacker could use variations in the signing
36 algorithm to recover the private key.
38 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
42 *) Timing vulnerability in ECDSA signature generation
44 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
45 timing side channel attack. An attacker could use variations in the signing
46 algorithm to recover the private key.
48 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
52 *) Add coordinate blinding for EC_POINT and implement projective
53 coordinate blinding for generic prime curves as a countermeasure to
54 chosen point SCA attacks.
55 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
57 Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
59 *) Client DoS due to large DH parameter
61 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
62 malicious server can send a very large prime value to the client. This will
63 cause the client to spend an unreasonably long period of time generating a
64 key for this prime resulting in a hang until the client has finished. This
65 could be exploited in a Denial Of Service attack.
67 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
71 *) Cache timing vulnerability in RSA Key Generation
73 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
74 a cache timing side channel attack. An attacker with sufficient access to
75 mount cache timing attacks during the RSA key generation process could
76 recover the private key.
78 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
79 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
83 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
84 parameter is no longer accepted, as it leads to a corrupt table. NULL
85 pem_str is reserved for alias entries only.
88 *) Revert blinding in ECDSA sign and instead make problematic addition
89 length-invariant. Switch even to fixed-length Montgomery multiplication.
92 *) Change generating and checking of primes so that the error rate of not
93 being prime depends on the intended use based on the size of the input.
94 For larger primes this will result in more rounds of Miller-Rabin.
95 The maximal error rate for primes with more than 1080 bits is lowered
97 [Kurt Roeckx, Annie Yousar]
99 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
102 *) Add blinding to ECDSA and DSA signatures to protect against side channel
103 attacks discovered by Keegan Ryan (NCC Group).
106 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
107 now allow empty (zero character) pass phrases.
110 *) Certificate time validation (X509_cmp_time) enforces stricter
111 compliance with RFC 5280. Fractional seconds and timezone offsets
112 are no longer allowed.
115 *) Fixed a text canonicalisation bug in CMS
117 Where a CMS detached signature is used with text content the text goes
118 through a canonicalisation process first prior to signing or verifying a
119 signature. This process strips trailing space at the end of lines, converts
120 line terminators to CRLF and removes additional trailing line terminators
121 at the end of a file. A bug in the canonicalisation process meant that
122 some characters, such as form-feed, were incorrectly treated as whitespace
123 and removed. This is contrary to the specification (RFC5485). This fix
124 could mean that detached text data signed with an earlier version of
125 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
126 signed with a fixed OpenSSL may fail to verify with an earlier version of
127 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
128 and use the "-binary" flag (for the "cms" command line application) or set
129 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
132 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
134 *) Constructed ASN.1 types with a recursive definition could exceed the stack
136 Constructed ASN.1 types with a recursive definition (such as can be found
137 in PKCS7) could eventually exceed the stack given malicious input with
138 excessive recursion. This could result in a Denial Of Service attack. There
139 are no such structures used within SSL/TLS that come from untrusted sources
140 so this is considered safe.
142 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
147 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
149 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
150 effectively reduced to only comparing the least significant bit of each
151 byte. This allows an attacker to forge messages that would be considered as
152 authenticated in an amount of tries lower than that guaranteed by the
153 security claims of the scheme. The module can only be compiled by the
154 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
156 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
161 *) Add a build target 'build_all_generated', to build all generated files
162 and only that. This can be used to prepare everything that requires
163 things like perl for a system that lacks perl and then move everything
164 to that system and do the rest of the build there.
167 *) Backport SSL_OP_NO_RENGOTIATION
169 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
170 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
171 changes this is no longer possible in 1.1.0. Therefore the new
172 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
173 1.1.0 to provide equivalent functionality.
175 Note that if an application built against 1.1.0h headers (or above) is run
176 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
177 accepted but nothing will happen, i.e. renegotiation will not be prevented.
180 *) Removed the OS390-Unix config target. It relied on a script that doesn't
184 *) rsaz_1024_mul_avx2 overflow bug on x86_64
186 There is an overflow bug in the AVX2 Montgomery multiplication procedure
187 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
188 Analysis suggests that attacks against RSA and DSA as a result of this
189 defect would be very difficult to perform and are not believed likely.
190 Attacks against DH1024 are considered just feasible, because most of the
191 work necessary to deduce information about a private key may be performed
192 offline. The amount of resources required for such an attack would be
193 significant. However, for an attack on TLS to be meaningful, the server
194 would have to share the DH1024 private key among multiple clients, which is
195 no longer an option since CVE-2016-0701.
197 This only affects processors that support the AVX2 but not ADX extensions
198 like Intel Haswell (4th generation).
200 This issue was reported to OpenSSL by David Benjamin (Google). The issue
201 was originally found via the OSS-Fuzz project.
205 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
207 *) bn_sqrx8x_internal carry bug on x86_64
209 There is a carry propagating bug in the x86_64 Montgomery squaring
210 procedure. No EC algorithms are affected. Analysis suggests that attacks
211 against RSA and DSA as a result of this defect would be very difficult to
212 perform and are not believed likely. Attacks against DH are considered just
213 feasible (although very difficult) because most of the work necessary to
214 deduce information about a private key may be performed offline. The amount
215 of resources required for such an attack would be very significant and
216 likely only accessible to a limited number of attackers. An attacker would
217 additionally need online access to an unpatched system using the target
218 private key in a scenario with persistent DH parameters and a private
219 key that is shared between multiple clients.
221 This only affects processors that support the BMI1, BMI2 and ADX extensions
222 like Intel Broadwell (5th generation) and later or AMD Ryzen.
224 This issue was reported to OpenSSL by the OSS-Fuzz project.
228 *) Malformed X.509 IPAddressFamily could cause OOB read
230 If an X.509 certificate has a malformed IPAddressFamily extension,
231 OpenSSL could do a one-byte buffer overread. The most likely result
232 would be an erroneous display of the certificate in text format.
234 This issue was reported to OpenSSL by the OSS-Fuzz project.
238 *) Ignore the '-named_curve auto' value for compatibility of applications
240 [Tomas Mraz <tmraz@fedoraproject.org>]
242 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
245 Changes between 1.1.0e and 1.1.0f [25 May 2017]
247 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
248 platform rather than 'mingw'.
251 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
252 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
253 which is the minimum version we support.
256 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
258 *) Encrypt-Then-Mac renegotiation crash
260 During a renegotiation handshake if the Encrypt-Then-Mac extension is
261 negotiated where it was not in the original handshake (or vice-versa) then
262 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
263 and servers are affected.
265 This issue was reported to OpenSSL by Joe Orton (Red Hat).
269 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
271 *) Truncated packet could crash via OOB read
273 If one side of an SSL/TLS path is running on a 32-bit host and a specific
274 cipher is being used, then a truncated packet can cause that host to
275 perform an out-of-bounds read, usually resulting in a crash.
277 This issue was reported to OpenSSL by Robert Święcki of Google.
281 *) Bad (EC)DHE parameters cause a client crash
283 If a malicious server supplies bad parameters for a DHE or ECDHE key
284 exchange then this can result in the client attempting to dereference a
285 NULL pointer leading to a client crash. This could be exploited in a Denial
288 This issue was reported to OpenSSL by Guido Vranken.
292 *) BN_mod_exp may produce incorrect results on x86_64
294 There is a carry propagating bug in the x86_64 Montgomery squaring
295 procedure. No EC algorithms are affected. Analysis suggests that attacks
296 against RSA and DSA as a result of this defect would be very difficult to
297 perform and are not believed likely. Attacks against DH are considered just
298 feasible (although very difficult) because most of the work necessary to
299 deduce information about a private key may be performed offline. The amount
300 of resources required for such an attack would be very significant and
301 likely only accessible to a limited number of attackers. An attacker would
302 additionally need online access to an unpatched system using the target
303 private key in a scenario with persistent DH parameters and a private
304 key that is shared between multiple clients. For example this can occur by
305 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
306 similar to CVE-2015-3193 but must be treated as a separate problem.
308 This issue was reported to OpenSSL by the OSS-Fuzz project.
312 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
314 *) ChaCha20/Poly1305 heap-buffer-overflow
316 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
317 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
318 crash. This issue is not considered to be exploitable beyond a DoS.
320 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
324 *) CMS Null dereference
326 Applications parsing invalid CMS structures can crash with a NULL pointer
327 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
328 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
329 structure callback if an attempt is made to free certain invalid encodings.
330 Only CHOICE structures using a callback which do not handle NULL value are
333 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
337 *) Montgomery multiplication may produce incorrect results
339 There is a carry propagating bug in the Broadwell-specific Montgomery
340 multiplication procedure that handles input lengths divisible by, but
341 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
342 and DH private keys are impossible. This is because the subroutine in
343 question is not used in operations with the private key itself and an input
344 of the attacker's direct choice. Otherwise the bug can manifest itself as
345 transient authentication and key negotiation failures or reproducible
346 erroneous outcome of public-key operations with specially crafted input.
347 Among EC algorithms only Brainpool P-512 curves are affected and one
348 presumably can attack ECDH key negotiation. Impact was not analyzed in
349 detail, because pre-requisites for attack are considered unlikely. Namely
350 multiple clients have to choose the curve in question and the server has to
351 share the private key among them, neither of which is default behaviour.
352 Even then only clients that chose the curve will be affected.
354 This issue was publicly reported as transient failures and was not
355 initially recognized as a security issue. Thanks to Richard Morgan for
356 providing reproducible case.
360 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
361 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
362 prevent issues where no progress is being made and the peer continually
363 sends unrecognised record types, using up resources processing them.
366 *) Removed automatic addition of RPATH in shared libraries and executables,
367 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
370 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
372 *) Fix Use After Free for large message sizes
374 The patch applied to address CVE-2016-6307 resulted in an issue where if a
375 message larger than approx 16k is received then the underlying buffer to
376 store the incoming message is reallocated and moved. Unfortunately a
377 dangling pointer to the old location is left which results in an attempt to
378 write to the previously freed location. This is likely to result in a
379 crash, however it could potentially lead to execution of arbitrary code.
381 This issue only affects OpenSSL 1.1.0a.
383 This issue was reported to OpenSSL by Robert Święcki.
387 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
389 *) OCSP Status Request extension unbounded memory growth
391 A malicious client can send an excessively large OCSP Status Request
392 extension. If that client continually requests renegotiation, sending a
393 large OCSP Status Request extension each time, then there will be unbounded
394 memory growth on the server. This will eventually lead to a Denial Of
395 Service attack through memory exhaustion. Servers with a default
396 configuration are vulnerable even if they do not support OCSP. Builds using
397 the "no-ocsp" build time option are not affected.
399 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
403 *) SSL_peek() hang on empty record
405 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
406 sends an empty record. This could be exploited by a malicious peer in a
407 Denial Of Service attack.
409 This issue was reported to OpenSSL by Alex Gaynor.
413 *) Excessive allocation of memory in tls_get_message_header() and
414 dtls1_preprocess_fragment()
416 A (D)TLS message includes 3 bytes for its length in the header for the
417 message. This would allow for messages up to 16Mb in length. Messages of
418 this length are excessive and OpenSSL includes a check to ensure that a
419 peer is sending reasonably sized messages in order to avoid too much memory
420 being consumed to service a connection. A flaw in the logic of version
421 1.1.0 means that memory for the message is allocated too early, prior to
422 the excessive message length check. Due to way memory is allocated in
423 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
424 to service a connection. This could lead to a Denial of Service through
425 memory exhaustion. However, the excessive message length check still takes
426 place, and this would cause the connection to immediately fail. Assuming
427 that the application calls SSL_free() on the failed connection in a timely
428 manner then the 21Mb of allocated memory will then be immediately freed
429 again. Therefore the excessive memory allocation will be transitory in
430 nature. This then means that there is only a security impact if:
432 1) The application does not call SSL_free() in a timely manner in the event
433 that the connection fails
435 2) The application is working in a constrained environment where there is
436 very little free memory
438 3) The attacker initiates multiple connection attempts such that there are
439 multiple connections in a state where memory has been allocated for the
440 connection; SSL_free() has not yet been called; and there is insufficient
441 memory to service the multiple requests.
443 Except in the instance of (1) above any Denial Of Service is likely to be
444 transitory because as soon as the connection fails the memory is
445 subsequently freed again in the SSL_free() call. However there is an
446 increased risk during this period of application crashes due to the lack of
447 memory - which would then mean a more serious Denial of Service.
449 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
450 (CVE-2016-6307 and CVE-2016-6308)
453 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
454 had to be removed. Primary reason is that vendor assembler can't
455 assemble our modules with -KPIC flag. As result it, assembly
456 support, was not even available as option. But its lack means
457 lack of side-channel resistant code, which is incompatible with
458 security by todays standards. Fortunately gcc is readily available
459 prepackaged option, which we firmly point at...
462 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
464 *) Windows command-line tool supports UTF-8 opt-in option for arguments
465 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
466 (to any value) allows Windows user to access PKCS#12 file generated
467 with Windows CryptoAPI and protected with non-ASCII password, as well
468 as files generated under UTF-8 locale on Linux also protected with
472 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
473 have been disabled by default and removed from DEFAULT, just like RC4.
474 See the RC4 item below to re-enable both.
477 *) The method for finding the storage location for the Windows RAND seed file
478 has changed. First we check %RANDFILE%. If that is not set then we check
479 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
480 all else fails we fall back to C:\.
483 *) The EVP_EncryptUpdate() function has had its return type changed from void
484 to int. A return of 0 indicates and error while a return of 1 indicates
488 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
489 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
490 off the constant time implementation for RSA, DSA and DH have been made
491 no-ops and deprecated.
494 *) Windows RAND implementation was simplified to only get entropy by
495 calling CryptGenRandom(). Various other RAND-related tickets
497 [Joseph Wylie Yandle, Rich Salz]
499 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
500 and OPENSSL_LH_, respectively. The old names are available
501 with API compatibility. They new names are now completely documented.
504 *) Unify TYPE_up_ref(obj) methods signature.
505 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
506 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
507 int (instead of void) like all others TYPE_up_ref() methods.
508 So now these methods also check the return value of CRYPTO_atomic_add(),
509 and the validity of object reference counter.
510 [fdasilvayy@gmail.com]
512 *) With Windows Visual Studio builds, the .pdb files are installed
513 alongside the installed libraries and executables. For a static
514 library installation, ossl_static.pdb is the associate compiler
515 generated .pdb file to be used when linking programs.
518 *) Remove openssl.spec. Packaging files belong with the packagers.
521 *) Automatic Darwin/OSX configuration has had a refresh, it will now
522 recognise x86_64 architectures automatically. You can still decide
523 to build for a different bitness with the environment variable
524 KERNEL_BITS (can be 32 or 64), for example:
526 KERNEL_BITS=32 ./config
530 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
531 256 bit AES and HMAC with SHA256.
534 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
537 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
540 *) To enable users to have their own config files and build file templates,
541 Configure looks in the directory indicated by the environment variable
542 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
543 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
544 name and is used as is.
547 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
548 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
549 X509_CERT_FILE_CTX was removed.
552 *) "shared" builds are now the default. To create only static libraries use
553 the "no-shared" Configure option.
556 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
557 All of these option have not worked for some while and are fundamental
561 *) Make various cleanup routines no-ops and mark them as deprecated. Most
562 global cleanup functions are no longer required because they are handled
563 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
564 Explicitly de-initing can cause problems (e.g. where a library that uses
565 OpenSSL de-inits, but an application is still using it). The affected
566 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
567 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
568 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
572 *) --strict-warnings no longer enables runtime debugging options
573 such as REF_DEBUG. Instead, debug options are automatically
574 enabled with '--debug' builds.
575 [Andy Polyakov, Emilia Käsper]
577 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
578 have been moved out of the public header files. New functions for managing
579 these have been added.
582 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
583 objects have been moved out of the public header files. New
584 functions for managing these have been added.
587 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
588 have been moved out of the public header files. New functions for managing
589 these have been added.
592 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
593 moved out of the public header files. New functions for managing these
597 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
600 *) Removed the mk1mf build scripts.
603 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
604 it is always safe to #include a header now.
607 *) Removed the aged BC-32 config and all its supporting scripts
610 *) Removed support for Ultrix, Netware, and OS/2.
613 *) Add support for HKDF.
616 *) Add support for blake2b and blake2s
619 *) Added support for "pipelining". Ciphers that have the
620 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
621 encryptions/decryptions simultaneously. There are currently no built-in
622 ciphers with this property but the expectation is that engines will be able
623 to offer it to significantly improve throughput. Support has been extended
624 into libssl so that multiple records for a single connection can be
625 processed in one go (for >=TLS 1.1).
628 *) Added the AFALG engine. This is an async capable engine which is able to
629 offload work to the Linux kernel. In this initial version it only supports
630 AES128-CBC. The kernel must be version 4.1.0 or greater.
633 *) OpenSSL now uses a new threading API. It is no longer necessary to
634 set locking callbacks to use OpenSSL in a multi-threaded environment. There
635 are two supported threading models: pthreads and windows threads. It is
636 also possible to configure OpenSSL at compile time for "no-threads". The
637 old threading API should no longer be used. The functions have been
638 replaced with "no-op" compatibility macros.
639 [Alessandro Ghedini, Matt Caswell]
641 *) Modify behavior of ALPN to invoke callback after SNI/servername
642 callback, such that updates to the SSL_CTX affect ALPN.
645 *) Add SSL_CIPHER queries for authentication and key-exchange.
648 *) Changes to the DEFAULT cipherlist:
649 - Prefer (EC)DHE handshakes over plain RSA.
650 - Prefer AEAD ciphers over legacy ciphers.
651 - Prefer ECDSA over RSA when both certificates are available.
652 - Prefer TLSv1.2 ciphers/PRF.
653 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
657 *) Change the ECC default curve list to be this, in order: x25519,
658 secp256r1, secp521r1, secp384r1.
661 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
662 disabled by default. They can be re-enabled using the
663 enable-weak-ssl-ciphers option to Configure.
666 *) If the server has ALPN configured, but supports no protocols that the
667 client advertises, send a fatal "no_application_protocol" alert.
668 This behaviour is SHALL in RFC 7301, though it isn't universally
669 implemented by other servers.
672 *) Add X25519 support.
673 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
674 for public and private key encoding using the format documented in
675 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
676 key generation and key derivation.
678 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
682 *) Deprecate SRP_VBASE_get_by_user.
683 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
684 In order to fix an unavoidable memory leak (CVE-2016-0798),
685 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
686 seed, even if the seed is configured.
688 Users should use SRP_VBASE_get1_by_user instead. Note that in
689 SRP_VBASE_get1_by_user, caller must free the returned value. Note
690 also that even though configuring the SRP seed attempts to hide
691 invalid usernames by continuing the handshake with fake
692 credentials, this behaviour is not constant time and no strong
693 guarantees are made that the handshake is indistinguishable from
694 that of a valid user.
697 *) Configuration change; it's now possible to build dynamic engines
698 without having to build shared libraries and vice versa. This
699 only applies to the engines in engines/, those in crypto/engine/
700 will always be built into libcrypto (i.e. "static").
702 Building dynamic engines is enabled by default; to disable, use
703 the configuration option "disable-dynamic-engine".
705 The only requirements for building dynamic engines are the
706 presence of the DSO module and building with position independent
707 code, so they will also automatically be disabled if configuring
708 with "disable-dso" or "disable-pic".
710 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
711 are also taken away from openssl/opensslconf.h, as they are
715 *) Configuration change; if there is a known flag to compile
716 position independent code, it will always be applied on the
717 libcrypto and libssl object files, and never on the application
718 object files. This means other libraries that use routines from
719 libcrypto / libssl can be made into shared libraries regardless
720 of how OpenSSL was configured.
722 If this isn't desirable, the configuration options "disable-pic"
723 or "no-pic" can be used to disable the use of PIC. This will
724 also disable building shared libraries and dynamic engines.
727 *) Removed JPAKE code. It was experimental and has no wide use.
730 *) The INSTALL_PREFIX Makefile variable has been renamed to
731 DESTDIR. That makes for less confusion on what this variable
732 is for. Also, the configuration option --install_prefix is
736 *) Heartbeat for TLS has been removed and is disabled by default
737 for DTLS; configure with enable-heartbeats. Code that uses the
738 old #define's might need to be updated.
739 [Emilia Käsper, Rich Salz]
741 *) Rename REF_CHECK to REF_DEBUG.
744 *) New "unified" build system
746 The "unified" build system is aimed to be a common system for all
747 platforms we support. With it comes new support for VMS.
749 This system builds supports building in a different directory tree
750 than the source tree. It produces one Makefile (for unix family
751 or lookalikes), or one descrip.mms (for VMS).
753 The source of information to make the Makefile / descrip.mms is
754 small files called 'build.info', holding the necessary
755 information for each directory with source to compile, and a
756 template in Configurations, like unix-Makefile.tmpl or
759 With this change, the library names were also renamed on Windows
760 and on VMS. They now have names that are closer to the standard
761 on Unix, and include the major version number, and in certain
762 cases, the architecture they are built for. See "Notes on shared
763 libraries" in INSTALL.
765 We rely heavily on the perl module Text::Template.
768 *) Added support for auto-initialisation and de-initialisation of the library.
769 OpenSSL no longer requires explicit init or deinit routines to be called,
770 except in certain circumstances. See the OPENSSL_init_crypto() and
771 OPENSSL_init_ssl() man pages for further information.
774 *) The arguments to the DTLSv1_listen function have changed. Specifically the
775 "peer" argument is now expected to be a BIO_ADDR object.
777 *) Rewrite of BIO networking library. The BIO library lacked consistent
778 support of IPv6, and adding it required some more extensive
779 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
780 which hold all types of addresses and chains of address information.
781 It also introduces a new API, with functions like BIO_socket,
782 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
783 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
784 have been adapted accordingly.
787 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
791 *) CRIME protection: disable compression by default, even if OpenSSL is
792 compiled with zlib enabled. Applications can still enable compression
793 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
794 using the SSL_CONF library to configure compression.
797 *) The signature of the session callback configured with
798 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
799 was explicitly marked as 'const unsigned char*' instead of
803 *) Always DPURIFY. Remove the use of uninitialized memory in the
804 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
807 *) Removed many obsolete configuration items, including
808 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
809 MD2_CHAR, MD2_INT, MD2_LONG
811 IDEA_SHORT, IDEA_LONG
812 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
813 [Rich Salz, with advice from Andy Polyakov]
815 *) Many BN internals have been moved to an internal header file.
816 [Rich Salz with help from Andy Polyakov]
818 *) Configuration and writing out the results from it has changed.
819 Files such as Makefile include/openssl/opensslconf.h and are now
820 produced through general templates, such as Makefile.in and
821 crypto/opensslconf.h.in and some help from the perl module
824 Also, the center of configuration information is no longer
825 Makefile. Instead, Configure produces a perl module in
826 configdata.pm which holds most of the config data (in the hash
827 table %config), the target data that comes from the target
828 configuration in one of the Configurations/*.conf files (in
832 *) To clarify their intended purposes, the Configure options
833 --prefix and --openssldir change their semantics, and become more
834 straightforward and less interdependent.
836 --prefix shall be used exclusively to give the location INSTALLTOP
837 where programs, scripts, libraries, include files and manuals are
838 going to be installed. The default is now /usr/local.
840 --openssldir shall be used exclusively to give the default
841 location OPENSSLDIR where certificates, private keys, CRLs are
842 managed. This is also where the default openssl.cnf gets
844 If the directory given with this option is a relative path, the
845 values of both the --prefix value and the --openssldir value will
846 be combined to become OPENSSLDIR.
847 The default for --openssldir is INSTALLTOP/ssl.
849 Anyone who uses --openssldir to specify where OpenSSL is to be
850 installed MUST change to use --prefix instead.
853 *) The GOST engine was out of date and therefore it has been removed. An up
854 to date GOST engine is now being maintained in an external repository.
855 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
856 support for GOST ciphersuites (these are only activated if a GOST engine
860 *) EGD is no longer supported by default; use enable-egd when
862 [Ben Kaduk and Rich Salz]
864 *) The distribution now has Makefile.in files, which are used to
865 create Makefile's when Configure is run. *Configure must be run
866 before trying to build now.*
869 *) The return value for SSL_CIPHER_description() for error conditions
873 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
875 Obtaining and performing DNSSEC validation of TLSA records is
876 the application's responsibility. The application provides
877 the TLSA records of its choice to OpenSSL, and these are then
878 used to authenticate the peer.
880 The TLSA records need not even come from DNS. They can, for
881 example, be used to implement local end-entity certificate or
882 trust-anchor "pinning", where the "pin" data takes the form
883 of TLSA records, which can augment or replace verification
884 based on the usual WebPKI public certification authorities.
887 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
888 continues to support deprecated interfaces in default builds.
889 However, applications are strongly advised to compile their
890 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
891 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
892 or the 1.1.0 releases.
894 In environments in which all applications have been ported to
895 not use any deprecated interfaces OpenSSL's Configure script
896 should be used with the --api=1.1.0 option to entirely remove
897 support for the deprecated features from the library and
898 unconditionally disable them in the installed headers.
899 Essentially the same effect can be achieved with the "no-deprecated"
900 argument to Configure, except that this will always restrict
901 the build to just the latest API, rather than a fixed API
904 As applications are ported to future revisions of the API,
905 they should update their compile-time OPENSSL_API_COMPAT define
906 accordingly, but in most cases should be able to continue to
907 compile with later releases.
909 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
910 0x10000000L and 0x00908000L, respectively. However those
911 versions did not support the OPENSSL_API_COMPAT feature, and
912 so applications are not typically tested for explicit support
913 of just the undeprecated features of either release.
916 *) Add support for setting the minimum and maximum supported protocol.
917 It can bet set via the SSL_set_min_proto_version() and
918 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
919 MaxProtcol. It's recommended to use the new APIs to disable
920 protocols instead of disabling individual protocols using
921 SSL_set_options() or SSL_CONF's Protocol. This change also
922 removes support for disabling TLS 1.2 in the OpenSSL TLS
923 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
926 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
929 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
930 and integrates ECDSA and ECDH functionality into EC. Implementations can
931 now redirect key generation and no longer need to convert to or from
934 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
935 include the ec.h header file instead.
938 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
939 ciphers who are no longer supported and drops support the ephemeral RSA key
940 exchange. The LOW ciphers currently doesn't have any ciphers in it.
943 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
944 opaque. For HMAC_CTX, the following constructors and destructors
947 HMAC_CTX *HMAC_CTX_new(void);
948 void HMAC_CTX_free(HMAC_CTX *ctx);
950 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
951 destroy such methods has been added. See EVP_MD_meth_new(3) and
952 EVP_CIPHER_meth_new(3) for documentation.
955 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
956 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
957 EVP_MD_CTX_reset() should be called instead to reinitialise
958 an already created structure.
959 2) For consistency with the majority of our object creators and
960 destructors, EVP_MD_CTX_(create|destroy) were renamed to
961 EVP_MD_CTX_(new|free). The old names are retained as macros
962 for deprecated builds.
965 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
966 cryptographic operations to be performed asynchronously as long as an
967 asynchronous capable engine is used. See the ASYNC_start_job() man page for
968 further details. Libssl has also had this capability integrated with the
969 introduction of the new mode SSL_MODE_ASYNC and associated error
970 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
971 pages. This work was developed in partnership with Intel Corp.
974 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
975 always enabled now. If you want to disable the support you should
976 exclude it using the list of supported ciphers. This also means that the
977 "-no_ecdhe" option has been removed from s_server.
980 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
981 SSL_{CTX_}set1_curves() which can set a list.
984 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
985 curve you want to support using SSL_{CTX_}set1_curves().
988 *) State machine rewrite. The state machine code has been significantly
989 refactored in order to remove much duplication of code and solve issues
990 with the old code (see ssl/statem/README for further details). This change
991 does have some associated API changes. Notably the SSL_state() function
992 has been removed and replaced by SSL_get_state which now returns an
993 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
994 altogether. The previous handshake states defined in ssl.h and ssl3.h have
998 *) All instances of the string "ssleay" in the public API were replaced
999 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1000 Some error codes related to internal RSA_eay API's were renamed.
1003 *) The demo files in crypto/threads were moved to demo/threads.
1006 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1008 [Matt Caswell, Rich Salz]
1010 *) New ASN.1 embed macro.
1012 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1013 structure is not allocated: it is part of the parent. That is instead of
1021 This reduces memory fragmentation and make it impossible to accidentally
1022 set a mandatory field to NULL.
1024 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1025 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1026 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1030 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1033 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1034 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1035 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1036 DES and RC4 ciphersuites.
1039 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1040 This changes the decoding behaviour for some invalid messages,
1041 though the change is mostly in the more lenient direction, and
1042 legacy behaviour is preserved as much as possible.
1045 *) Fix no-stdio build.
1046 [ David Woodhouse <David.Woodhouse@intel.com> and also
1047 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1049 *) New testing framework
1050 The testing framework has been largely rewritten and is now using
1051 perl and the perl modules Test::Harness and an extended variant of
1052 Test::More called OpenSSL::Test to do its work. All test scripts in
1053 test/ have been rewritten into test recipes, and all direct calls to
1054 executables in test/Makefile have become individual recipes using the
1055 simplified testing OpenSSL::Test::Simple.
1057 For documentation on our testing modules, do:
1059 perldoc test/testlib/OpenSSL/Test/Simple.pm
1060 perldoc test/testlib/OpenSSL/Test.pm
1064 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1065 are used; the latter aborts on memory leaks (usually checked on exit).
1066 Some undocumented "set malloc, etc., hooks" functions were removed
1067 and others were changed. All are now documented.
1070 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1072 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1074 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1075 from RFC4279, RFC4785, RFC5487, RFC5489.
1077 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1078 original RSA_PSK patch.
1081 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1082 era flag was never set throughout the codebase (only read). Also removed
1083 SSL3_FLAGS_POP_BUFFER which was only used if
1084 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1087 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1088 to be "oneline" instead of "compat".
1091 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1092 not aware of clients that still exhibit this bug, and the workaround
1093 hasn't been working properly for a while.
1096 *) The return type of BIO_number_read() and BIO_number_written() as well as
1097 the corresponding num_read and num_write members in the BIO structure has
1098 changed from unsigned long to uint64_t. On platforms where an unsigned
1099 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1103 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1104 OpenSSL without support for them. It also means that maintaining
1105 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1106 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1109 *) Removed support for the two export grade static DH ciphersuites
1110 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1111 were newly added (along with a number of other static DH ciphersuites) to
1112 1.0.2. However the two export ones have *never* worked since they were
1113 introduced. It seems strange in any case to be adding new export
1114 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1117 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1118 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1119 and turned into macros which simply call the new preferred function names
1120 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1121 should use the new names instead. Also as part of this change the ssl23.h
1122 header file has been removed.
1125 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1126 code and the associated standard is no longer considered fit-for-purpose.
1129 *) RT2547 was closed. When generating a private key, try to make the
1130 output file readable only by the owner. This behavior change might
1131 be noticeable when interacting with other software.
1133 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1137 *) Added HTTP GET support to the ocsp command.
1140 *) Changed default digest for the dgst and enc commands from MD5 to
1144 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1147 *) Added support for TLS extended master secret from
1148 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1149 initial patch which was a great help during development.
1152 *) All libssl internal structures have been removed from the public header
1153 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1154 now redundant). Users should not attempt to access internal structures
1155 directly. Instead they should use the provided API functions.
1158 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1159 Access to deprecated functions can be re-enabled by running config with
1160 "enable-deprecated". In addition applications wishing to use deprecated
1161 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1162 will, by default, disable some transitive includes that previously existed
1163 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1166 *) Added support for OCB mode. OpenSSL has been granted a patent license
1167 compatible with the OpenSSL license for use of OCB. Details are available
1168 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1169 for OCB can be removed by calling config with no-ocb.
1172 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1173 compatible client hello.
1176 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1177 done while fixing the error code for the key-too-small case.
1178 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1180 *) CA.sh has been removed; use CA.pl instead.
1183 *) Removed old DES API.
1186 *) Remove various unsupported platforms:
1192 Sinix/ReliantUNIX RM400
1197 16-bit platforms such as WIN16
1200 *) Clean up OPENSSL_NO_xxx #define's
1201 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1202 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1203 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1204 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1205 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1206 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1207 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1208 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1209 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1210 Remove MS_STATIC; it's a relic from platforms <32 bits.
1213 *) Cleaned up dead code
1214 Remove all but one '#ifdef undef' which is to be looked at.
1217 *) Clean up calling of xxx_free routines.
1218 Just like free(), fix most of the xxx_free routines to accept
1219 NULL. Remove the non-null checks from callers. Save much code.
1222 *) Add secure heap for storage of private keys (when possible).
1223 Add BIO_s_secmem(), CBIGNUM, etc.
1224 Contributed by Akamai Technologies under our Corporate CLA.
1227 *) Experimental support for a new, fast, unbiased prime candidate generator,
1228 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1229 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1231 *) New output format NSS in the sess_id command line tool. This allows
1232 exporting the session id and the master key in NSS keylog format.
1233 [Martin Kaiser <martin@kaiser.cx>]
1235 *) Harmonize version and its documentation. -f flag is used to display
1237 [mancha <mancha1@zoho.com>]
1239 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1240 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1241 [mancha <mancha1@zoho.com>]
1243 *) Fix some double frees. These are not thought to be exploitable.
1244 [mancha <mancha1@zoho.com>]
1246 *) A missing bounds check in the handling of the TLS heartbeat extension
1247 can be used to reveal up to 64k of memory to a connected client or
1250 Thanks for Neel Mehta of Google Security for discovering this bug and to
1251 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1252 preparing the fix (CVE-2014-0160)
1253 [Adam Langley, Bodo Moeller]
1255 *) Fix for the attack described in the paper "Recovering OpenSSL
1256 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1257 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1258 http://eprint.iacr.org/2014/140
1260 Thanks to Yuval Yarom and Naomi Benger for discovering this
1261 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1262 [Yuval Yarom and Naomi Benger]
1264 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1265 this fixes a limitation in previous versions of OpenSSL.
1268 *) Experimental encrypt-then-mac support.
1270 Experimental support for encrypt then mac from
1271 draft-gutmann-tls-encrypt-then-mac-02.txt
1273 To enable it set the appropriate extension number (0x42 for the test
1274 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1276 For non-compliant peers (i.e. just about everything) this should have no
1279 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1283 *) Add EVP support for key wrapping algorithms, to avoid problems with
1284 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1285 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1286 algorithms and include tests cases.
1289 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1293 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1294 MGF1 digest and OAEP label.
1297 *) Make openssl verify return errors.
1298 [Chris Palmer <palmer@google.com> and Ben Laurie]
1300 *) New function ASN1_TIME_diff to calculate the difference between two
1301 ASN1_TIME structures or one structure and the current time.
1304 *) Update fips_test_suite to support multiple command line options. New
1305 test to induce all self test errors in sequence and check expected
1309 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1310 sign or verify all in one operation.
1313 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1314 test programs and fips_test_suite. Includes functionality to parse
1315 the minimal script output of fipsalgest.pl directly.
1318 *) Add authorisation parameter to FIPS_module_mode_set().
1321 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1324 *) Use separate DRBG fields for internal and external flags. New function
1325 FIPS_drbg_health_check() to perform on demand health checking. Add
1326 generation tests to fips_test_suite with reduced health check interval to
1327 demonstrate periodic health checking. Add "nodh" option to
1328 fips_test_suite to skip very slow DH test.
1331 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1335 *) More extensive health check for DRBG checking many more failure modes.
1336 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1337 combination: call this in fips_test_suite.
1340 *) Add support for canonical generation of DSA parameter 'g'. See
1343 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1344 POST to handle HMAC cases.
1347 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1348 to return numerical and string versions of the FIPS module number.
1351 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1352 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1353 outside the validated module in the FIPS capable OpenSSL.
1356 *) Minor change to DRBG entropy callback semantics. In some cases
1357 there is no multiple of the block length between min_len and
1358 max_len. Allow the callback to return more than max_len bytes
1359 of entropy but discard any extra: it is the callback's responsibility
1360 to ensure that the extra data discarded does not impact the
1361 requested amount of entropy.
1364 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1365 information in FIPS186-3, SP800-57 and SP800-131A.
1368 *) CCM support via EVP. Interface is very similar to GCM case except we
1369 must supply all data in one chunk (i.e. no update, final) and the
1370 message length must be supplied if AAD is used. Add algorithm test
1374 *) Initial version of POST overhaul. Add POST callback to allow the status
1375 of POST to be monitored and/or failures induced. Modify fips_test_suite
1376 to use callback. Always run all selftests even if one fails.
1379 *) XTS support including algorithm test driver in the fips_gcmtest program.
1380 Note: this does increase the maximum key length from 32 to 64 bytes but
1381 there should be no binary compatibility issues as existing applications
1382 will never use XTS mode.
1385 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1386 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1387 performs algorithm blocking for unapproved PRNG types. Also do not
1388 set PRNG type in FIPS_mode_set(): leave this to the application.
1389 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1390 the standard OpenSSL PRNG: set additional data to a date time vector.
1393 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1394 This shouldn't present any incompatibility problems because applications
1395 shouldn't be using these directly and any that are will need to rethink
1396 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1399 *) Extensive self tests and health checking required by SP800-90 DRBG.
1400 Remove strength parameter from FIPS_drbg_instantiate and always
1401 instantiate at maximum supported strength.
1404 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1407 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1410 *) New function DH_compute_key_padded() to compute a DH key and pad with
1411 leading zeroes if needed: this complies with SP800-56A et al.
1414 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1415 anything, incomplete, subject to change and largely untested at present.
1418 *) Modify fipscanisteronly build option to only build the necessary object
1419 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1422 *) Add experimental option FIPSSYMS to give all symbols in
1423 fipscanister.o and FIPS or fips prefix. This will avoid
1424 conflicts with future versions of OpenSSL. Add perl script
1425 util/fipsas.pl to preprocess assembly language source files
1426 and rename any affected symbols.
1429 *) Add selftest checks and algorithm block of non-fips algorithms in
1430 FIPS mode. Remove DES2 from selftests.
1433 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1434 return internal method without any ENGINE dependencies. Add new
1435 tiny fips sign and verify functions.
1438 *) New build option no-ec2m to disable characteristic 2 code.
1441 *) New build option "fipscanisteronly". This only builds fipscanister.o
1442 and (currently) associated fips utilities. Uses the file Makefile.fips
1443 instead of Makefile.org as the prototype.
1446 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1447 Update fips_gcmtest to use IV generator.
1450 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1451 setting output buffer to NULL. The *Final function must be
1452 called although it will not retrieve any additional data. The tag
1453 can be set or retrieved with a ctrl. The IV length is by default 12
1454 bytes (96 bits) but can be set to an alternative value. If the IV
1455 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1459 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1460 underlying do_cipher function handles all cipher semantics itself
1461 including padding and finalisation. This is useful if (for example)
1462 an ENGINE cipher handles block padding itself. The behaviour of
1463 do_cipher is subtly changed if this flag is set: the return value
1464 is the number of characters written to the output buffer (zero is
1465 no longer an error code) or a negative error code. Also if the
1466 input buffer is NULL and length 0 finalisation should be performed.
1469 *) If a candidate issuer certificate is already part of the constructed
1470 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
1473 *) Improve forward-security support: add functions
1475 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
1476 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
1478 for use by SSL/TLS servers; the callback function will be called whenever a
1479 new session is created, and gets to decide whether the session may be
1480 cached to make it resumable (return 0) or not (return 1). (As by the
1481 SSL/TLS protocol specifications, the session_id sent by the server will be
1482 empty to indicate that the session is not resumable; also, the server will
1483 not generate RFC 4507 (RFC 5077) session tickets.)
1485 A simple reasonable callback implementation is to return is_forward_secure.
1486 This parameter will be set to 1 or 0 depending on the ciphersuite selected
1487 by the SSL/TLS server library, indicating whether it can provide forward
1489 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1491 *) New -verify_name option in command line utilities to set verification
1495 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
1496 Add CMAC pkey methods.
1499 *) Experimental renegotiation in s_server -www mode. If the client
1500 browses /reneg connection is renegotiated. If /renegcert it is
1501 renegotiated requesting a certificate.
1504 *) Add an "external" session cache for debugging purposes to s_server. This
1505 should help trace issues which normally are only apparent in deployed
1506 multi-process servers.
1509 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
1510 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
1511 BIO_set_cipher() and some obscure PEM functions were changed so they
1512 can now return an error. The RAND changes required a change to the
1513 RAND_METHOD structure.
1516 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
1517 a gcc attribute to warn if the result of a function is ignored. This
1518 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
1519 whose return value is often ignored.
1522 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
1523 These allow SCTs (signed certificate timestamps) to be requested and
1524 validated when establishing a connection.
1525 [Rob Percival <robpercival@google.com>]
1527 Changes between 1.0.2g and 1.0.2h [3 May 2016]
1529 *) Prevent padding oracle in AES-NI CBC MAC check
1531 A MITM attacker can use a padding oracle attack to decrypt traffic
1532 when the connection uses an AES CBC cipher and the server support
1535 This issue was introduced as part of the fix for Lucky 13 padding
1536 attack (CVE-2013-0169). The padding check was rewritten to be in
1537 constant time by making sure that always the same bytes are read and
1538 compared against either the MAC or padding bytes. But it no longer
1539 checked that there was enough data to have both the MAC and padding
1542 This issue was reported by Juraj Somorovsky using TLS-Attacker.
1546 *) Fix EVP_EncodeUpdate overflow
1548 An overflow can occur in the EVP_EncodeUpdate() function which is used for
1549 Base64 encoding of binary data. If an attacker is able to supply very large
1550 amounts of input data then a length check can overflow resulting in a heap
1553 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
1554 the PEM_write_bio* family of functions. These are mainly used within the
1555 OpenSSL command line applications, so any application which processes data
1556 from an untrusted source and outputs it as a PEM file should be considered
1557 vulnerable to this issue. User applications that call these APIs directly
1558 with large amounts of untrusted data may also be vulnerable.
1560 This issue was reported by Guido Vranken.
1564 *) Fix EVP_EncryptUpdate overflow
1566 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
1567 is able to supply very large amounts of input data after a previous call to
1568 EVP_EncryptUpdate() with a partial block then a length check can overflow
1569 resulting in a heap corruption. Following an analysis of all OpenSSL
1570 internal usage of the EVP_EncryptUpdate() function all usage is one of two
1571 forms. The first form is where the EVP_EncryptUpdate() call is known to be
1572 the first called function after an EVP_EncryptInit(), and therefore that
1573 specific call must be safe. The second form is where the length passed to
1574 EVP_EncryptUpdate() can be seen from the code to be some small value and
1575 therefore there is no possibility of an overflow. Since all instances are
1576 one of these two forms, it is believed that there can be no overflows in
1577 internal code due to this problem. It should be noted that
1578 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
1579 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
1580 of these calls have also been analysed too and it is believed there are no
1581 instances in internal usage where an overflow could occur.
1583 This issue was reported by Guido Vranken.
1587 *) Prevent ASN.1 BIO excessive memory allocation
1589 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
1590 a short invalid encoding can cause allocation of large amounts of memory
1591 potentially consuming excessive resources or exhausting memory.
1593 Any application parsing untrusted data through d2i BIO functions is
1594 affected. The memory based functions such as d2i_X509() are *not* affected.
1595 Since the memory based functions are used by the TLS library, TLS
1596 applications are not affected.
1598 This issue was reported by Brian Carpenter.
1604 ASN1 Strings that are over 1024 bytes can cause an overread in applications
1605 using the X509_NAME_oneline() function on EBCDIC systems. This could result
1606 in arbitrary stack data being returned in the buffer.
1608 This issue was reported by Guido Vranken.
1612 *) Modify behavior of ALPN to invoke callback after SNI/servername
1613 callback, such that updates to the SSL_CTX affect ALPN.
1616 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
1620 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
1621 methods are enabled and ssl2 is disabled the methods return NULL.
1624 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
1626 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
1627 Builds that are not configured with "enable-weak-ssl-ciphers" will not
1628 provide any "EXPORT" or "LOW" strength ciphers.
1631 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
1632 is by default disabled at build-time. Builds that are not configured with
1633 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
1634 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
1635 will need to explicitly call either of:
1637 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
1639 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
1641 as appropriate. Even if either of those is used, or the application
1642 explicitly uses the version-specific SSLv2_method() or its client and
1643 server variants, SSLv2 ciphers vulnerable to exhaustive search key
1644 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
1645 ciphers, and SSLv2 56-bit DES are no longer available.
1649 *) Fix a double-free in DSA code
1651 A double free bug was discovered when OpenSSL parses malformed DSA private
1652 keys and could lead to a DoS attack or memory corruption for applications
1653 that receive DSA private keys from untrusted sources. This scenario is
1656 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
1661 *) Disable SRP fake user seed to address a server memory leak.
1663 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
1665 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1666 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
1667 was changed to ignore the "fake user" SRP seed, even if the seed
1670 Users should use SRP_VBASE_get1_by_user instead. Note that in
1671 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1672 also that even though configuring the SRP seed attempts to hide
1673 invalid usernames by continuing the handshake with fake
1674 credentials, this behaviour is not constant time and no strong
1675 guarantees are made that the handshake is indistinguishable from
1676 that of a valid user.
1680 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
1682 In the BN_hex2bn function the number of hex digits is calculated using an
1683 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
1684 large values of |i| this can result in |bn_expand| not allocating any
1685 memory because |i * 4| is negative. This can leave the internal BIGNUM data
1686 field as NULL leading to a subsequent NULL ptr deref. For very large values
1687 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
1688 In this case memory is allocated to the internal BIGNUM data field, but it
1689 is insufficiently sized leading to heap corruption. A similar issue exists
1690 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
1691 is ever called by user applications with very large untrusted hex/dec data.
1692 This is anticipated to be a rare occurrence.
1694 All OpenSSL internal usage of these functions use data that is not expected
1695 to be untrusted, e.g. config file data or application command line
1696 arguments. If user developed applications generate config file data based
1697 on untrusted data then it is possible that this could also lead to security
1698 consequences. This is also anticipated to be rare.
1700 This issue was reported to OpenSSL by Guido Vranken.
1704 *) Fix memory issues in BIO_*printf functions
1706 The internal |fmtstr| function used in processing a "%s" format string in
1707 the BIO_*printf functions could overflow while calculating the length of a
1708 string and cause an OOB read when printing very long strings.
1710 Additionally the internal |doapr_outch| function can attempt to write to an
1711 OOB memory location (at an offset from the NULL pointer) in the event of a
1712 memory allocation failure. In 1.0.2 and below this could be caused where
1713 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
1714 could be in processing a very long "%s" format string. Memory leaks can
1717 The first issue may mask the second issue dependent on compiler behaviour.
1718 These problems could enable attacks where large amounts of untrusted data
1719 is passed to the BIO_*printf functions. If applications use these functions
1720 in this way then they could be vulnerable. OpenSSL itself uses these
1721 functions when printing out human-readable dumps of ASN.1 data. Therefore
1722 applications that print this data could be vulnerable if the data is from
1723 untrusted sources. OpenSSL command line applications could also be
1724 vulnerable where they print out ASN.1 data, or if untrusted data is passed
1725 as command line arguments.
1727 Libssl is not considered directly vulnerable. Additionally certificates etc
1728 received via remote connections via libssl are also unlikely to be able to
1729 trigger these issues because of message size limits enforced within libssl.
1731 This issue was reported to OpenSSL Guido Vranken.
1735 *) Side channel attack on modular exponentiation
1737 A side-channel attack was found which makes use of cache-bank conflicts on
1738 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
1739 of RSA keys. The ability to exploit this issue is limited as it relies on
1740 an attacker who has control of code in a thread running on the same
1741 hyper-threaded core as the victim thread which is performing decryptions.
1743 This issue was reported to OpenSSL by Yuval Yarom, The University of
1744 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
1745 Nadia Heninger, University of Pennsylvania with more information at
1746 http://cachebleed.info.
1750 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
1751 if no keysize is specified with default_bits. This fixes an
1752 omission in an earlier change that changed all RSA/DSA key generation
1753 apps to use 2048 bits by default.
1756 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
1757 *) DH small subgroups
1759 Historically OpenSSL only ever generated DH parameters based on "safe"
1760 primes. More recently (in version 1.0.2) support was provided for
1761 generating X9.42 style parameter files such as those required for RFC 5114
1762 support. The primes used in such files may not be "safe". Where an
1763 application is using DH configured with parameters based on primes that are
1764 not "safe" then an attacker could use this fact to find a peer's private
1765 DH exponent. This attack requires that the attacker complete multiple
1766 handshakes in which the peer uses the same private DH exponent. For example
1767 this could be used to discover a TLS server's private DH exponent if it's
1768 reusing the private DH exponent or it's using a static DH ciphersuite.
1770 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
1771 TLS. It is not on by default. If the option is not set then the server
1772 reuses the same private DH exponent for the life of the server process and
1773 would be vulnerable to this attack. It is believed that many popular
1774 applications do set this option and would therefore not be at risk.
1776 The fix for this issue adds an additional check where a "q" parameter is
1777 available (as is the case in X9.42 based parameters). This detects the
1778 only known attack, and is the only possible defense for static DH
1779 ciphersuites. This could have some performance impact.
1781 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
1782 default and cannot be disabled. This could have some performance impact.
1784 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
1788 *) SSLv2 doesn't block disabled ciphers
1790 A malicious client can negotiate SSLv2 ciphers that have been disabled on
1791 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
1792 been disabled, provided that the SSLv2 protocol was not also disabled via
1795 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
1796 and Sebastian Schinzel.
1800 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
1802 *) BN_mod_exp may produce incorrect results on x86_64
1804 There is a carry propagating bug in the x86_64 Montgomery squaring
1805 procedure. No EC algorithms are affected. Analysis suggests that attacks
1806 against RSA and DSA as a result of this defect would be very difficult to
1807 perform and are not believed likely. Attacks against DH are considered just
1808 feasible (although very difficult) because most of the work necessary to
1809 deduce information about a private key may be performed offline. The amount
1810 of resources required for such an attack would be very significant and
1811 likely only accessible to a limited number of attackers. An attacker would
1812 additionally need online access to an unpatched system using the target
1813 private key in a scenario with persistent DH parameters and a private
1814 key that is shared between multiple clients. For example this can occur by
1815 default in OpenSSL DHE based SSL/TLS ciphersuites.
1817 This issue was reported to OpenSSL by Hanno Böck.
1821 *) Certificate verify crash with missing PSS parameter
1823 The signature verification routines will crash with a NULL pointer
1824 dereference if presented with an ASN.1 signature using the RSA PSS
1825 algorithm and absent mask generation function parameter. Since these
1826 routines are used to verify certificate signature algorithms this can be
1827 used to crash any certificate verification operation and exploited in a
1828 DoS attack. Any application which performs certificate verification is
1829 vulnerable including OpenSSL clients and servers which enable client
1832 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
1836 *) X509_ATTRIBUTE memory leak
1838 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
1839 memory. This structure is used by the PKCS#7 and CMS routines so any
1840 application which reads PKCS#7 or CMS data from untrusted sources is
1841 affected. SSL/TLS is not affected.
1843 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
1848 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1849 This changes the decoding behaviour for some invalid messages,
1850 though the change is mostly in the more lenient direction, and
1851 legacy behaviour is preserved as much as possible.
1854 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1856 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1858 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
1860 *) Alternate chains certificate forgery
1862 During certificate verification, OpenSSL will attempt to find an
1863 alternative certificate chain if the first attempt to build such a chain
1864 fails. An error in the implementation of this logic can mean that an
1865 attacker could cause certain checks on untrusted certificates to be
1866 bypassed, such as the CA flag, enabling them to use a valid leaf
1867 certificate to act as a CA and "issue" an invalid certificate.
1869 This issue was reported to OpenSSL by Adam Langley/David Benjamin
1873 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
1875 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
1876 incompatibility in the handling of HMAC. The previous ABI has now been
1880 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
1882 *) Malformed ECParameters causes infinite loop
1884 When processing an ECParameters structure OpenSSL enters an infinite loop
1885 if the curve specified is over a specially malformed binary polynomial
1888 This can be used to perform denial of service against any
1889 system which processes public keys, certificate requests or
1890 certificates. This includes TLS clients and TLS servers with
1891 client authentication enabled.
1893 This issue was reported to OpenSSL by Joseph Barr-Pixton.
1897 *) Exploitable out-of-bounds read in X509_cmp_time
1899 X509_cmp_time does not properly check the length of the ASN1_TIME
1900 string and can read a few bytes out of bounds. In addition,
1901 X509_cmp_time accepts an arbitrary number of fractional seconds in the
1904 An attacker can use this to craft malformed certificates and CRLs of
1905 various sizes and potentially cause a segmentation fault, resulting in
1906 a DoS on applications that verify certificates or CRLs. TLS clients
1907 that verify CRLs are affected. TLS clients and servers with client
1908 authentication enabled may be affected if they use custom verification
1911 This issue was reported to OpenSSL by Robert Swiecki (Google), and
1912 independently by Hanno Böck.
1916 *) PKCS7 crash with missing EnvelopedContent
1918 The PKCS#7 parsing code does not handle missing inner EncryptedContent
1919 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
1920 with missing content and trigger a NULL pointer dereference on parsing.
1922 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
1923 structures from untrusted sources are affected. OpenSSL clients and
1924 servers are not affected.
1926 This issue was reported to OpenSSL by Michal Zalewski (Google).
1930 *) CMS verify infinite loop with unknown hash function
1932 When verifying a signedData message the CMS code can enter an infinite loop
1933 if presented with an unknown hash function OID. This can be used to perform
1934 denial of service against any system which verifies signedData messages using
1936 This issue was reported to OpenSSL by Johannes Bauer.
1940 *) Race condition handling NewSessionTicket
1942 If a NewSessionTicket is received by a multi-threaded client when attempting to
1943 reuse a previous ticket then a race condition can occur potentially leading to
1944 a double free of the ticket data.
1948 *) Only support 256-bit or stronger elliptic curves with the
1949 'ecdh_auto' setting (server) or by default (client). Of supported
1950 curves, prefer P-256 (both).
1953 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
1955 *) ClientHello sigalgs DoS fix
1957 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
1958 invalid signature algorithms extension a NULL pointer dereference will
1959 occur. This can be exploited in a DoS attack against the server.
1961 This issue was was reported to OpenSSL by David Ramos of Stanford
1964 [Stephen Henson and Matt Caswell]
1966 *) Multiblock corrupted pointer fix
1968 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
1969 feature only applies on 64 bit x86 architecture platforms that support AES
1970 NI instructions. A defect in the implementation of "multiblock" can cause
1971 OpenSSL's internal write buffer to become incorrectly set to NULL when
1972 using non-blocking IO. Typically, when the user application is using a
1973 socket BIO for writing, this will only result in a failed connection.
1974 However if some other BIO is used then it is likely that a segmentation
1975 fault will be triggered, thus enabling a potential DoS attack.
1977 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
1981 *) Segmentation fault in DTLSv1_listen fix
1983 The DTLSv1_listen function is intended to be stateless and processes the
1984 initial ClientHello from many peers. It is common for user code to loop
1985 over the call to DTLSv1_listen until a valid ClientHello is received with
1986 an associated cookie. A defect in the implementation of DTLSv1_listen means
1987 that state is preserved in the SSL object from one invocation to the next
1988 that can lead to a segmentation fault. Errors processing the initial
1989 ClientHello can trigger this scenario. An example of such an error could be
1990 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
1993 This issue was reported to OpenSSL by Per Allansson.
1997 *) Segmentation fault in ASN1_TYPE_cmp fix
1999 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2000 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2001 certificate signature algorithm consistency this can be used to crash any
2002 certificate verification operation and exploited in a DoS attack. Any
2003 application which performs certificate verification is vulnerable including
2004 OpenSSL clients and servers which enable client authentication.
2008 *) Segmentation fault for invalid PSS parameters fix
2010 The signature verification routines will crash with a NULL pointer
2011 dereference if presented with an ASN.1 signature using the RSA PSS
2012 algorithm and invalid parameters. Since these routines are used to verify
2013 certificate signature algorithms this can be used to crash any
2014 certificate verification operation and exploited in a DoS attack. Any
2015 application which performs certificate verification is vulnerable including
2016 OpenSSL clients and servers which enable client authentication.
2018 This issue was was reported to OpenSSL by Brian Carpenter.
2022 *) ASN.1 structure reuse memory corruption fix
2024 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2025 memory corruption via an invalid write. Such reuse is and has been
2026 strongly discouraged and is believed to be rare.
2028 Applications that parse structures containing CHOICE or ANY DEFINED BY
2029 components may be affected. Certificate parsing (d2i_X509 and related
2030 functions) are however not affected. OpenSSL clients and servers are
2035 *) PKCS7 NULL pointer dereferences fix
2037 The PKCS#7 parsing code does not handle missing outer ContentInfo
2038 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2039 missing content and trigger a NULL pointer dereference on parsing.
2041 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2042 otherwise parse PKCS#7 structures from untrusted sources are
2043 affected. OpenSSL clients and servers are not affected.
2045 This issue was reported to OpenSSL by Michal Zalewski (Google).
2049 *) DoS via reachable assert in SSLv2 servers fix
2051 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2052 servers that both support SSLv2 and enable export cipher suites by sending
2053 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2055 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2056 (OpenSSL development team).
2060 *) Empty CKE with client auth and DHE fix
2062 If client auth is used then a server can seg fault in the event of a DHE
2063 ciphersuite being selected and a zero length ClientKeyExchange message
2064 being sent by the client. This could be exploited in a DoS attack.
2068 *) Handshake with unseeded PRNG fix
2070 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2071 with an unseeded PRNG. The conditions are:
2072 - The client is on a platform where the PRNG has not been seeded
2073 automatically, and the user has not seeded manually
2074 - A protocol specific client method version has been used (i.e. not
2075 SSL_client_methodv23)
2076 - A ciphersuite is used that does not require additional random data from
2077 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2079 If the handshake succeeds then the client random that has been used will
2080 have been generated from a PRNG with insufficient entropy and therefore the
2081 output may be predictable.
2083 For example using the following command with an unseeded openssl will
2084 succeed on an unpatched platform:
2086 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2090 *) Use After Free following d2i_ECPrivatekey error fix
2092 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2093 could cause a use after free condition. This, in turn, could cause a double
2094 free in several private key parsing functions (such as d2i_PrivateKey
2095 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2096 for applications that receive EC private keys from untrusted
2097 sources. This scenario is considered rare.
2099 This issue was discovered by the BoringSSL project and fixed in their
2104 *) X509_to_X509_REQ NULL pointer deref fix
2106 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2107 the certificate key is invalid. This function is rarely used in practice.
2109 This issue was discovered by Brian Carpenter.
2113 *) Removed the export ciphers from the DEFAULT ciphers
2116 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2118 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2119 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2120 So far those who have to target multiple platforms would compromise
2121 and argue that binary targeting say ARMv5 would still execute on
2122 ARMv8. "Universal" build resolves this compromise by providing
2123 near-optimal performance even on newer platforms.
2126 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2127 (other platforms pending).
2128 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2130 *) Add support for the SignedCertificateTimestampList certificate and
2131 OCSP response extensions from RFC6962.
2134 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2135 for corner cases. (Certain input points at infinity could lead to
2136 bogus results, with non-infinity inputs mapped to infinity too.)
2139 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2140 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2141 common cases are optimized and there still is room for further
2142 improvements. Vector Permutation AES for Altivec is also added.
2145 *) Add support for little-endian ppc64 Linux target.
2146 [Marcelo Cerri (IBM)]
2148 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2149 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2150 are optimized and there still is room for further improvements.
2151 Both 32- and 64-bit modes are supported.
2152 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2154 *) Improved ARMv7 NEON support.
2157 *) Support for SPARC Architecture 2011 crypto extensions, first
2158 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2159 SHA256/512, MD5, GHASH and modular exponentiation.
2160 [Andy Polyakov, David Miller]
2162 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2164 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2166 *) Support for new and upcoming Intel processors, including AVX2,
2167 BMI and SHA ISA extensions. This includes additional "stitched"
2168 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2171 This work was sponsored by Intel Corp.
2174 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2175 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2176 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2179 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2180 this fixes a limitation in previous versions of OpenSSL.
2183 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2184 MGF1 digest and OAEP label.
2187 *) Add EVP support for key wrapping algorithms, to avoid problems with
2188 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2189 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2190 algorithms and include tests cases.
2193 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2195 [Douglas E. Engert, Steve Henson]
2197 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2198 difference in days and seconds between two tm or ASN1_TIME structures.
2201 *) Add -rev test option to s_server to just reverse order of characters
2202 received by client and send back to server. Also prints an abbreviated
2203 summary of the connection parameters.
2206 *) New option -brief for s_client and s_server to print out a brief summary
2207 of connection parameters.
2210 *) Add callbacks for arbitrary TLS extensions.
2211 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2213 *) New option -crl_download in several openssl utilities to download CRLs
2214 from CRLDP extension in certificates.
2217 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2220 *) New function X509_CRL_diff to generate a delta CRL from the difference
2221 of two full CRLs. Add support to "crl" utility.
2224 *) New functions to set lookup_crls function and to retrieve
2225 X509_STORE from X509_STORE_CTX.
2228 *) Print out deprecated issuer and subject unique ID fields in
2232 *) Extend OCSP I/O functions so they can be used for simple general purpose
2233 HTTP as well as OCSP. New wrapper function which can be used to download
2234 CRLs using the OCSP API.
2237 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2240 *) SSL_CONF* functions. These provide a common framework for application
2241 configuration using configuration files or command lines.
2244 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2245 message callback and prints the results. Needs compile time option
2246 "enable-ssl-trace". New options to s_client and s_server to enable
2250 *) New ctrl and macro to retrieve supported points extensions.
2251 Print out extension in s_server and s_client.
2254 *) New functions to retrieve certificate signature and signature
2258 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2262 *) New Suite B modes for TLS code. These use and enforce the requirements
2263 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2264 only use Suite B curves. The Suite B modes can be set by using the
2265 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2268 *) New chain verification flags for Suite B levels of security. Check
2269 algorithms are acceptable when flags are set in X509_verify_cert.
2272 *) Make tls1_check_chain return a set of flags indicating checks passed
2273 by a certificate chain. Add additional tests to handle client
2274 certificates: checks for matching certificate type and issuer name
2278 *) If an attempt is made to use a signature algorithm not in the peer
2279 preference list abort the handshake. If client has no suitable
2280 signature algorithms in response to a certificate request do not
2281 use the certificate.
2284 *) If server EC tmp key is not in client preference list abort handshake.
2287 *) Add support for certificate stores in CERT structure. This makes it
2288 possible to have different stores per SSL structure or one store in
2289 the parent SSL_CTX. Include distinct stores for certificate chain
2290 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2291 to build and store a certificate chain in CERT structure: returning
2292 an error if the chain cannot be built: this will allow applications
2293 to test if a chain is correctly configured.
2295 Note: if the CERT based stores are not set then the parent SSL_CTX
2296 store is used to retain compatibility with existing behaviour.
2300 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2301 mask based on the current session, check mask when sending client
2302 hello and checking the requested ciphersuite.
2305 *) New ctrls to retrieve and set certificate types in a certificate
2306 request message. Print out received values in s_client. If certificate
2307 types is not set with custom values set sensible values based on
2308 supported signature algorithms.
2311 *) Support for distinct client and server supported signature algorithms.
2314 *) Add certificate callback. If set this is called whenever a certificate
2315 is required by client or server. An application can decide which
2316 certificate chain to present based on arbitrary criteria: for example
2317 supported signature algorithms. Add very simple example to s_server.
2318 This fixes many of the problems and restrictions of the existing client
2319 certificate callback: for example you can now clear an existing
2320 certificate and specify the whole chain.
2323 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2324 the certificate can be used for (if anything). Set valid_flags field
2325 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2326 to have similar checks in it.
2328 Add new "cert_flags" field to CERT structure and include a "strict mode".
2329 This enforces some TLS certificate requirements (such as only permitting
2330 certificate signature algorithms contained in the supported algorithms
2331 extension) which some implementations ignore: this option should be used
2332 with caution as it could cause interoperability issues.
2335 *) Update and tidy signature algorithm extension processing. Work out
2336 shared signature algorithms based on preferences and peer algorithms
2337 and print them out in s_client and s_server. Abort handshake if no
2338 shared signature algorithms.
2341 *) Add new functions to allow customised supported signature algorithms
2342 for SSL and SSL_CTX structures. Add options to s_client and s_server
2346 *) New function SSL_certs_clear() to delete all references to certificates
2347 from an SSL structure. Before this once a certificate had been added
2348 it couldn't be removed.
2351 *) Integrate hostname, email address and IP address checking with certificate
2352 verification. New verify options supporting checking in openssl utility.
2355 *) Fixes and wildcard matching support to hostname and email checking
2356 functions. Add manual page.
2357 [Florian Weimer (Red Hat Product Security Team)]
2359 *) New functions to check a hostname email or IP address against a
2360 certificate. Add options x509 utility to print results of checks against
2364 *) Fix OCSP checking.
2365 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2367 *) Initial experimental support for explicitly trusted non-root CAs.
2368 OpenSSL still tries to build a complete chain to a root but if an
2369 intermediate CA has a trust setting included that is used. The first
2370 setting is used: whether to trust (e.g., -addtrust option to the x509
2374 *) Add -trusted_first option which attempts to find certificates in the
2375 trusted store even if an untrusted chain is also supplied.
2378 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2379 platform support for Linux and Android.
2382 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2385 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2386 When in FIPS mode the approved implementations are used as normal,
2387 when not in FIPS mode the internal unapproved versions are used instead.
2388 This means that the FIPS capable OpenSSL isn't forced to use the
2389 (often lower performance) FIPS implementations outside FIPS mode.
2392 *) Transparently support X9.42 DH parameters when calling
2393 PEM_read_bio_DHparameters. This means existing applications can handle
2394 the new parameter format automatically.
2397 *) Initial experimental support for X9.42 DH parameter format: mainly
2398 to support use of 'q' parameter for RFC5114 parameters.
2401 *) Add DH parameters from RFC5114 including test data to dhtest.
2404 *) Support for automatic EC temporary key parameter selection. If enabled
2405 the most preferred EC parameters are automatically used instead of
2406 hardcoded fixed parameters. Now a server just has to call:
2407 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2408 support ECDH and use the most appropriate parameters.
2411 *) Enhance and tidy EC curve and point format TLS extension code. Use
2412 static structures instead of allocation if default values are used.
2413 New ctrls to set curves we wish to support and to retrieve shared curves.
2414 Print out shared curves in s_server. New options to s_server and s_client
2415 to set list of supported curves.
2418 *) New ctrls to retrieve supported signature algorithms and
2419 supported curve values as an array of NIDs. Extend openssl utility
2420 to print out received values.
2423 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2424 between NIDs and the more common NIST names such as "P-256". Enhance
2425 ecparam utility and ECC method to recognise the NIST names for curves.
2428 *) Enhance SSL/TLS certificate chain handling to support different
2429 chains for each certificate instead of one chain in the parent SSL_CTX.
2432 *) Support for fixed DH ciphersuite client authentication: where both
2433 server and client use DH certificates with common parameters.
2436 *) Support for fixed DH ciphersuites: those requiring DH server
2440 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2442 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2443 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2444 X509_CINF_get_signature were reverted post internal team review.
2446 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2448 *) Build fixes for the Windows and OpenVMS platforms
2449 [Matt Caswell and Richard Levitte]
2451 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2453 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2454 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2455 dereference. This could lead to a Denial Of Service attack. Thanks to
2456 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2460 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2461 dtls1_buffer_record function under certain conditions. In particular this
2462 could occur if an attacker sent repeated DTLS records with the same
2463 sequence number but for the next epoch. The memory leak could be exploited
2464 by an attacker in a Denial of Service attack through memory exhaustion.
2465 Thanks to Chris Mueller for reporting this issue.
2469 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
2470 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
2471 method would be set to NULL which could later result in a NULL pointer
2472 dereference. Thanks to Frank Schmirler for reporting this issue.
2476 *) Abort handshake if server key exchange message is omitted for ephemeral
2479 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
2480 reporting this issue.
2484 *) Remove non-export ephemeral RSA code on client and server. This code
2485 violated the TLS standard by allowing the use of temporary RSA keys in
2486 non-export ciphersuites and could be used by a server to effectively
2487 downgrade the RSA key length used to a value smaller than the server
2488 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
2489 INRIA or reporting this issue.
2493 *) Fixed issue where DH client certificates are accepted without verification.
2494 An OpenSSL server will accept a DH certificate for client authentication
2495 without the certificate verify message. This effectively allows a client to
2496 authenticate without the use of a private key. This only affects servers
2497 which trust a client certificate authority which issues certificates
2498 containing DH keys: these are extremely rare and hardly ever encountered.
2499 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
2504 *) Ensure that the session ID context of an SSL is updated when its
2505 SSL_CTX is updated via SSL_set_SSL_CTX.
2507 The session ID context is typically set from the parent SSL_CTX,
2508 and can vary with the CTX.
2511 *) Fix various certificate fingerprint issues.
2513 By using non-DER or invalid encodings outside the signed portion of a
2514 certificate the fingerprint can be changed without breaking the signature.
2515 Although no details of the signed portion of the certificate can be changed
2516 this can cause problems with some applications: e.g. those using the
2517 certificate fingerprint for blacklists.
2519 1. Reject signatures with non zero unused bits.
2521 If the BIT STRING containing the signature has non zero unused bits reject
2522 the signature. All current signature algorithms require zero unused bits.
2524 2. Check certificate algorithm consistency.
2526 Check the AlgorithmIdentifier inside TBS matches the one in the
2527 certificate signature. NB: this will result in signature failure
2528 errors for some broken certificates.
2530 Thanks to Konrad Kraszewski from Google for reporting this issue.
2532 3. Check DSA/ECDSA signatures use DER.
2534 Re-encode DSA/ECDSA signatures and compare with the original received
2535 signature. Return an error if there is a mismatch.
2537 This will reject various cases including garbage after signature
2538 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
2539 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
2540 (negative or with leading zeroes).
2542 Further analysis was conducted and fixes were developed by Stephen Henson
2543 of the OpenSSL core team.
2548 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
2549 results on some platforms, including x86_64. This bug occurs at random
2550 with a very low probability, and is not known to be exploitable in any
2551 way, though its exact impact is difficult to determine. Thanks to Pieter
2552 Wuille (Blockstream) who reported this issue and also suggested an initial
2553 fix. Further analysis was conducted by the OpenSSL development team and
2554 Adam Langley of Google. The final fix was developed by Andy Polyakov of
2555 the OpenSSL core team.
2559 *) Do not resume sessions on the server if the negotiated protocol
2560 version does not match the session's version. Resuming with a different
2561 version, while not strictly forbidden by the RFC, is of questionable
2562 sanity and breaks all known clients.
2563 [David Benjamin, Emilia Käsper]
2565 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
2566 early CCS messages during renegotiation. (Note that because
2567 renegotiation is encrypted, this early CCS was not exploitable.)
2570 *) Tighten client-side session ticket handling during renegotiation:
2571 ensure that the client only accepts a session ticket if the server sends
2572 the extension anew in the ServerHello. Previously, a TLS client would
2573 reuse the old extension state and thus accept a session ticket if one was
2574 announced in the initial ServerHello.
2576 Similarly, ensure that the client requires a session ticket if one
2577 was advertised in the ServerHello. Previously, a TLS client would
2578 ignore a missing NewSessionTicket message.
2581 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
2583 *) SRTP Memory Leak.
2585 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
2586 sends a carefully crafted handshake message, to cause OpenSSL to fail
2587 to free up to 64k of memory causing a memory leak. This could be
2588 exploited in a Denial Of Service attack. This issue affects OpenSSL
2589 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
2590 whether SRTP is used or configured. Implementations of OpenSSL that
2591 have been compiled with OPENSSL_NO_SRTP defined are not affected.
2593 The fix was developed by the OpenSSL team.
2597 *) Session Ticket Memory Leak.
2599 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
2600 integrity of that ticket is first verified. In the event of a session
2601 ticket integrity check failing, OpenSSL will fail to free memory
2602 causing a memory leak. By sending a large number of invalid session
2603 tickets an attacker could exploit this issue in a Denial Of Service
2608 *) Build option no-ssl3 is incomplete.
2610 When OpenSSL is configured with "no-ssl3" as a build option, servers
2611 could accept and complete a SSL 3.0 handshake, and clients could be
2612 configured to send them.
2614 [Akamai and the OpenSSL team]
2616 *) Add support for TLS_FALLBACK_SCSV.
2617 Client applications doing fallback retries should call
2618 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
2620 [Adam Langley, Bodo Moeller]
2622 *) Add additional DigestInfo checks.
2624 Re-encode DigestInto in DER and check against the original when
2625 verifying RSA signature: this will reject any improperly encoded
2626 DigestInfo structures.
2628 Note: this is a precautionary measure and no attacks are currently known.
2632 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
2634 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
2635 SRP code can be overrun an internal buffer. Add sanity check that
2636 g, A, B < N to SRP code.
2638 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
2639 Group for discovering this issue.
2643 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
2644 TLS 1.0 instead of higher protocol versions when the ClientHello message
2645 is badly fragmented. This allows a man-in-the-middle attacker to force a
2646 downgrade to TLS 1.0 even if both the server and the client support a
2647 higher protocol version, by modifying the client's TLS records.
2649 Thanks to David Benjamin and Adam Langley (Google) for discovering and
2650 researching this issue.
2654 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
2655 to a denial of service attack. A malicious server can crash the client
2656 with a null pointer dereference (read) by specifying an anonymous (EC)DH
2657 ciphersuite and sending carefully crafted handshake messages.
2659 Thanks to Felix Gröbert (Google) for discovering and researching this
2664 *) By sending carefully crafted DTLS packets an attacker could cause openssl
2665 to leak memory. This can be exploited through a Denial of Service attack.
2666 Thanks to Adam Langley for discovering and researching this issue.
2670 *) An attacker can force openssl to consume large amounts of memory whilst
2671 processing DTLS handshake messages. This can be exploited through a
2672 Denial of Service attack.
2673 Thanks to Adam Langley for discovering and researching this issue.
2677 *) An attacker can force an error condition which causes openssl to crash
2678 whilst processing DTLS packets due to memory being freed twice. This
2679 can be exploited through a Denial of Service attack.
2680 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
2685 *) If a multithreaded client connects to a malicious server using a resumed
2686 session and the server sends an ec point format extension it could write
2687 up to 255 bytes to freed memory.
2689 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
2694 *) A malicious server can crash an OpenSSL client with a null pointer
2695 dereference (read) by specifying an SRP ciphersuite even though it was not
2696 properly negotiated with the client. This can be exploited through a
2697 Denial of Service attack.
2699 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
2700 discovering and researching this issue.
2704 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
2705 X509_name_oneline, X509_name_print_ex et al. to leak some information
2706 from the stack. Applications may be affected if they echo pretty printing
2707 output to the attacker.
2709 Thanks to Ivan Fratric (Google) for discovering this issue.
2711 [Emilia Käsper, and Steve Henson]
2713 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2714 for corner cases. (Certain input points at infinity could lead to
2715 bogus results, with non-infinity inputs mapped to infinity too.)
2718 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
2720 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
2721 handshake can force the use of weak keying material in OpenSSL
2722 SSL/TLS clients and servers.
2724 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
2725 researching this issue. (CVE-2014-0224)
2726 [KIKUCHI Masashi, Steve Henson]
2728 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
2729 OpenSSL DTLS client the code can be made to recurse eventually crashing
2732 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
2734 [Imre Rad, Steve Henson]
2736 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
2737 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
2738 client or server. This is potentially exploitable to run arbitrary
2739 code on a vulnerable client or server.
2741 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
2742 [Jüri Aedla, Steve Henson]
2744 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
2745 are subject to a denial of service attack.
2747 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
2748 this issue. (CVE-2014-3470)
2749 [Felix Gröbert, Ivan Fratric, Steve Henson]
2751 *) Harmonize version and its documentation. -f flag is used to display
2753 [mancha <mancha1@zoho.com>]
2755 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2756 in i2d_ECPrivateKey.
2757 [mancha <mancha1@zoho.com>]
2759 *) Fix some double frees. These are not thought to be exploitable.
2760 [mancha <mancha1@zoho.com>]
2762 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
2764 *) A missing bounds check in the handling of the TLS heartbeat extension
2765 can be used to reveal up to 64k of memory to a connected client or
2768 Thanks for Neel Mehta of Google Security for discovering this bug and to
2769 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2770 preparing the fix (CVE-2014-0160)
2771 [Adam Langley, Bodo Moeller]
2773 *) Fix for the attack described in the paper "Recovering OpenSSL
2774 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2775 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2776 http://eprint.iacr.org/2014/140
2778 Thanks to Yuval Yarom and Naomi Benger for discovering this
2779 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2780 [Yuval Yarom and Naomi Benger]
2782 *) TLS pad extension: draft-agl-tls-padding-03
2784 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
2785 TLS client Hello record length value would otherwise be > 255 and
2786 less that 512 pad with a dummy extension containing zeroes so it
2787 is at least 512 bytes long.
2789 [Adam Langley, Steve Henson]
2791 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
2793 *) Fix for TLS record tampering bug. A carefully crafted invalid
2794 handshake could crash OpenSSL with a NULL pointer exception.
2795 Thanks to Anton Johansson for reporting this issues.
2798 *) Keep original DTLS digest and encryption contexts in retransmission
2799 structures so we can use the previous session parameters if they need
2800 to be resent. (CVE-2013-6450)
2803 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
2804 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
2805 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
2806 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
2807 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
2808 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
2809 [Rob Stradling, Adam Langley]
2811 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
2813 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
2814 supporting platforms or when small records were transferred.
2815 [Andy Polyakov, Steve Henson]
2817 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
2819 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2821 This addresses the flaw in CBC record processing discovered by
2822 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2823 at: http://www.isg.rhul.ac.uk/tls/
2825 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2826 Security Group at Royal Holloway, University of London
2827 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2828 Emilia Käsper for the initial patch.
2830 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2832 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
2833 ciphersuites which can be exploited in a denial of service attack.
2834 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
2835 and detecting this bug and to Wolfgang Ettlinger
2836 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
2840 *) Return an error when checking OCSP signatures when key is NULL.
2841 This fixes a DoS attack. (CVE-2013-0166)
2844 *) Make openssl verify return errors.
2845 [Chris Palmer <palmer@google.com> and Ben Laurie]
2847 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2848 the right response is stapled. Also change SSL_get_certificate()
2849 so it returns the certificate actually sent.
2850 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2851 [Rob Stradling <rob.stradling@comodo.com>]
2853 *) Fix possible deadlock when decoding public keys.
2856 *) Don't use TLS 1.0 record version number in initial client hello
2860 Changes between 1.0.1b and 1.0.1c [10 May 2012]
2862 *) Sanity check record length before skipping explicit IV in TLS
2863 1.2, 1.1 and DTLS to fix DoS attack.
2865 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2866 fuzzing as a service testing platform.
2870 *) Initialise tkeylen properly when encrypting CMS messages.
2871 Thanks to Solar Designer of Openwall for reporting this issue.
2874 *) In FIPS mode don't try to use composite ciphers as they are not
2878 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
2880 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
2881 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
2882 mean any application compiled against OpenSSL 1.0.0 headers setting
2883 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
2884 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
2885 0x10000000L Any application which was previously compiled against
2886 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
2887 will need to be recompiled as a result. Letting be results in
2888 inability to disable specifically TLS 1.1 and in client context,
2889 in unlike event, limit maximum offered version to TLS 1.0 [see below].
2892 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
2893 disable just protocol X, but all protocols above X *if* there are
2894 protocols *below* X still enabled. In more practical terms it means
2895 that if application wants to disable TLS1.0 in favor of TLS1.1 and
2896 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
2897 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
2901 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
2903 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2904 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2905 in CRYPTO_realloc_clean.
2907 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2908 issue and to Adam Langley <agl@chromium.org> for fixing it.
2910 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2912 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
2915 *) Workarounds for some broken servers that "hang" if a client hello
2916 record length exceeds 255 bytes.
2918 1. Do not use record version number > TLS 1.0 in initial client
2919 hello: some (but not all) hanging servers will now work.
2920 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
2921 the number of ciphers sent in the client hello. This should be
2922 set to an even number, such as 50, for example by passing:
2923 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
2924 Most broken servers should now work.
2925 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
2926 TLS 1.2 client support entirely.
2929 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
2932 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
2934 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
2935 STRING form instead of a DigestInfo.
2938 *) The format used for MDC2 RSA signatures is inconsistent between EVP
2939 and the RSA_sign/RSA_verify functions. This was made more apparent when
2940 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
2941 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
2942 the correct format in RSA_verify so both forms transparently work.
2945 *) Some servers which support TLS 1.0 can choke if we initially indicate
2946 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
2947 encrypted premaster secret. As a workaround use the maximum permitted
2948 client version in client hello, this should keep such servers happy
2949 and still work with previous versions of OpenSSL.
2952 *) Add support for TLS/DTLS heartbeats.
2953 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2955 *) Add support for SCTP.
2956 [Robin Seggelmann <seggelmann@fh-muenster.de>]
2958 *) Improved PRNG seeding for VOS.
2959 [Paul Green <Paul.Green@stratus.com>]
2961 *) Extensive assembler packs updates, most notably:
2963 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
2964 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
2965 - x86_64: bit-sliced AES implementation;
2966 - ARM: NEON support, contemporary platforms optimizations;
2967 - s390x: z196 support;
2968 - *: GHASH and GF(2^m) multiplication implementations;
2972 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
2973 (removal of unnecessary code)
2974 [Peter Sylvester <peter.sylvester@edelweb.fr>]
2976 *) Add TLS key material exporter from RFC 5705.
2979 *) Add DTLS-SRTP negotiation from RFC 5764.
2982 *) Add Next Protocol Negotiation,
2983 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
2984 disabled with a no-npn flag to config or Configure. Code donated
2986 [Adam Langley <agl@google.com> and Ben Laurie]
2988 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
2989 NIST-P256, NIST-P521, with constant-time single point multiplication on
2990 typical inputs. Compiler support for the nonstandard type __uint128_t is
2991 required to use this (present in gcc 4.4 and later, for 64-bit builds).
2992 Code made available under Apache License version 2.0.
2994 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
2995 line to include this in your build of OpenSSL, and run "make depend" (or
2996 "make update"). This enables the following EC_METHODs:
2998 EC_GFp_nistp224_method()
2999 EC_GFp_nistp256_method()
3000 EC_GFp_nistp521_method()
3002 EC_GROUP_new_by_curve_name() will automatically use these (while
3003 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3005 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3007 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3008 all platforms. Move ssize_t definition from e_os.h to the public
3009 header file e_os2.h as it now appears in public header file cms.h
3012 *) New -sigopt option to the ca, req and x509 utilities. Additional
3013 signature parameters can be passed using this option and in
3017 *) Add RSA PSS signing function. This will generate and set the
3018 appropriate AlgorithmIdentifiers for PSS based on those in the
3019 corresponding EVP_MD_CTX structure. No application support yet.
3022 *) Support for companion algorithm specific ASN1 signing routines.
3023 New function ASN1_item_sign_ctx() signs a pre-initialised
3024 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3025 the appropriate parameters.
3028 *) Add new algorithm specific ASN1 verification initialisation function
3029 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3030 handling will be the same no matter what EVP_PKEY_METHOD is used.
3031 Add a PSS handler to support verification of PSS signatures: checked
3032 against a number of sample certificates.
3035 *) Add signature printing for PSS. Add PSS OIDs.
3036 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3038 *) Add algorithm specific signature printing. An individual ASN1 method
3039 can now print out signatures instead of the standard hex dump.
3041 More complex signatures (e.g. PSS) can print out more meaningful
3042 information. Include DSA version that prints out the signature
3046 *) Password based recipient info support for CMS library: implementing
3050 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3051 neatly separates the code into cipher and PBE sections and is required
3052 for some algorithms that split PBES2 into separate pieces (such as
3053 password based CMS).
3056 *) Session-handling fixes:
3057 - Fix handling of connections that are resuming with a session ID,
3058 but also support Session Tickets.
3059 - Fix a bug that suppressed issuing of a new ticket if the client
3060 presented a ticket with an expired session.
3061 - Try to set the ticket lifetime hint to something reasonable.
3062 - Make tickets shorter by excluding irrelevant information.
3063 - On the client side, don't ignore renewed tickets.
3064 [Adam Langley, Bodo Moeller (Google)]
3066 *) Fix PSK session representation.
3069 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3071 This work was sponsored by Intel.
3074 *) Add GCM support to TLS library. Some custom code is needed to split
3075 the IV between the fixed (from PRF) and explicit (from TLS record)
3076 portions. This adds all GCM ciphersuites supported by RFC5288 and
3077 RFC5289. Generalise some AES* cipherstrings to include GCM and
3078 add a special AESGCM string for GCM only.
3081 *) Expand range of ctrls for AES GCM. Permit setting invocation
3082 field on decrypt and retrieval of invocation field only on encrypt.
3085 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3086 As required by RFC5289 these ciphersuites cannot be used if for
3087 versions of TLS earlier than 1.2.
3090 *) For FIPS capable OpenSSL interpret a NULL default public key method
3091 as unset and return the appropriate default but do *not* set the default.
3092 This means we can return the appropriate method in applications that
3093 switch between FIPS and non-FIPS modes.
3096 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3097 ENGINE is used then we cannot handle that in the FIPS module so we
3098 keep original code iff non-FIPS operations are allowed.
3101 *) Add -attime option to openssl utilities.
3102 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3104 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3107 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3108 FIPS EC methods unconditionally for now.
3111 *) New build option no-ec2m to disable characteristic 2 code.
3114 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3115 all cases can be covered as some introduce binary incompatibilities.
3118 *) Redirect RSA operations to FIPS module including keygen,
3119 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3122 *) Add similar low level API blocking to ciphers.
3125 *) Low level digest APIs are not approved in FIPS mode: any attempt
3126 to use these will cause a fatal error. Applications that *really* want
3127 to use them can use the private_* version instead.
3130 *) Redirect cipher operations to FIPS module for FIPS builds.
3133 *) Redirect digest operations to FIPS module for FIPS builds.
3136 *) Update build system to add "fips" flag which will link in fipscanister.o
3137 for static and shared library builds embedding a signature if needed.
3140 *) Output TLS supported curves in preference order instead of numerical
3141 order. This is currently hardcoded for the highest order curves first.
3142 This should be configurable so applications can judge speed vs strength.
3145 *) Add TLS v1.2 server support for client authentication.
3148 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3152 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3153 FIPS modules versions.
3156 *) Add TLS v1.2 client side support for client authentication. Keep cache
3157 of handshake records longer as we don't know the hash algorithm to use
3158 until after the certificate request message is received.
3161 *) Initial TLS v1.2 client support. Add a default signature algorithms
3162 extension including all the algorithms we support. Parse new signature
3163 format in client key exchange. Relax some ECC signing restrictions for
3164 TLS v1.2 as indicated in RFC5246.
3167 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3168 to new signature format when needed using client digest preference.
3169 All server ciphersuites should now work correctly in TLS v1.2. No client
3170 support yet and no support for client certificates.
3173 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3174 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3175 ciphersuites. At present only RSA key exchange ciphersuites work with
3176 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3177 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3178 and version checking.
3181 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3182 with this defined it will not be affected by any changes to ssl internal
3183 structures. Add several utility functions to allow openssl application
3184 to work with OPENSSL_NO_SSL_INTERN defined.
3187 *) A long standing patch to add support for SRP from EdelWeb (Peter
3188 Sylvester and Christophe Renou) was integrated.
3189 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3190 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3193 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3196 *) Permit abbreviated handshakes when renegotiating using the function
3197 SSL_renegotiate_abbreviated().
3198 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3200 *) Add call to ENGINE_register_all_complete() to
3201 ENGINE_load_builtin_engines(), so some implementations get used
3202 automatically instead of needing explicit application support.
3205 *) Add support for TLS key exporter as described in RFC5705.
3206 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3208 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3209 a few changes are required:
3211 Add SSL_OP_NO_TLSv1_1 flag.
3212 Add TLSv1_1 methods.
3213 Update version checking logic to handle version 1.1.
3214 Add explicit IV handling (ported from DTLS code).
3215 Add command line options to s_client/s_server.
3218 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3220 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3221 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3222 content decryption and always return the same error. Note: this attack
3223 needs on average 2^20 messages so it only affects automated senders. The
3224 old behaviour can be re-enabled in the CMS code by setting the
3225 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3226 an MMA defence is not necessary.
3227 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3228 this issue. (CVE-2012-0884)
3231 *) Fix CVE-2011-4619: make sure we really are receiving a
3232 client hello before rejecting multiple SGC restarts. Thanks to
3233 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3236 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3238 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3239 Thanks to Antonio Martin, Enterprise Secure Access Research and
3240 Development, Cisco Systems, Inc. for discovering this bug and
3241 preparing a fix. (CVE-2012-0050)
3244 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3246 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3247 of the Vaudenay padding oracle attack on CBC mode encryption
3248 which enables an efficient plaintext recovery attack against
3249 the OpenSSL implementation of DTLS. Their attack exploits timing
3250 differences arising during decryption processing. A research
3251 paper describing this attack can be found at:
3252 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3253 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3254 Security Group at Royal Holloway, University of London
3255 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3256 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3257 for preparing the fix. (CVE-2011-4108)
3258 [Robin Seggelmann, Michael Tuexen]
3260 *) Clear bytes used for block padding of SSL 3.0 records.
3262 [Adam Langley (Google)]
3264 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3265 Kadianakis <desnacked@gmail.com> for discovering this issue and
3266 Adam Langley for preparing the fix. (CVE-2011-4619)
3267 [Adam Langley (Google)]
3269 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3270 [Andrey Kulikov <amdeich@gmail.com>]
3272 *) Prevent malformed RFC3779 data triggering an assertion failure.
3273 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3274 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3275 [Rob Austein <sra@hactrn.net>]
3277 *) Improved PRNG seeding for VOS.
3278 [Paul Green <Paul.Green@stratus.com>]
3280 *) Fix ssl_ciph.c set-up race.
3281 [Adam Langley (Google)]
3283 *) Fix spurious failures in ecdsatest.c.
3284 [Emilia Käsper (Google)]
3286 *) Fix the BIO_f_buffer() implementation (which was mixing different
3287 interpretations of the '..._len' fields).
3288 [Adam Langley (Google)]
3290 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3291 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3292 threads won't reuse the same blinding coefficients.
3294 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3295 lock to call BN_BLINDING_invert_ex, and avoids one use of
3296 BN_BLINDING_update for each BN_BLINDING structure (previously,
3297 the last update always remained unused).
3298 [Emilia Käsper (Google)]
3300 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3301 [Bob Buckholz (Google)]
3303 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3305 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3306 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3307 [Kaspar Brand <ossl@velox.ch>]
3309 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3310 for multi-threaded use of ECDH. (CVE-2011-3210)
3311 [Adam Langley (Google)]
3313 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3316 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3317 signature public key algorithm by using OID xref utilities instead.
3318 Before this you could only use some ECC ciphersuites with SHA1 only.
3321 *) Add protection against ECDSA timing attacks as mentioned in the paper
3322 by Billy Bob Brumley and Nicola Tuveri, see:
3324 http://eprint.iacr.org/2011/232.pdf
3326 [Billy Bob Brumley and Nicola Tuveri]
3328 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3330 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3331 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3333 *) Fix bug in string printing code: if *any* escaping is enabled we must
3334 escape the escape character (backslash) or the resulting string is
3338 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3340 *) Disable code workaround for ancient and obsolete Netscape browsers
3341 and servers: an attacker can use it in a ciphersuite downgrade attack.
3342 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3345 *) Fixed J-PAKE implementation error, originally discovered by
3346 Sebastien Martini, further info and confirmation from Stefan
3347 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3350 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3352 *) Fix extension code to avoid race conditions which can result in a buffer
3353 overrun vulnerability: resumed sessions must not be modified as they can
3354 be shared by multiple threads. CVE-2010-3864
3357 *) Fix WIN32 build system to correctly link an ENGINE directory into
3361 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3363 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3365 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3367 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3369 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3370 context. The operation can be customised via the ctrl mechanism in
3371 case ENGINEs want to include additional functionality.
3374 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3377 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3378 output hashes compatible with older versions of OpenSSL.
3379 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3381 *) Fix compression algorithm handling: if resuming a session use the
3382 compression algorithm of the resumed session instead of determining
3383 it from client hello again. Don't allow server to change algorithm.
3386 *) Add load_crls() function to apps tidying load_certs() too. Add option
3387 to verify utility to allow additional CRLs to be included.
3390 *) Update OCSP request code to permit adding custom headers to the request:
3391 some responders need this.
3394 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3396 [Julia Lawall <julia@diku.dk>]
3398 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3399 needlessly dereferenced structures, used obsolete functions and
3400 didn't handle all updated verify codes correctly.
3403 *) Disable MD2 in the default configuration.
3406 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3407 indicate the initial BIO being pushed or popped. This makes it possible
3408 to determine whether the BIO is the one explicitly called or as a result
3409 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3410 it handles reference counts correctly and doesn't zero out the I/O bio
3411 when it is not being explicitly popped. WARNING: applications which
3412 included workarounds for the old buggy behaviour will need to be modified
3413 or they could free up already freed BIOs.
3416 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3417 renaming to all platforms (within the 0.9.8 branch, this was
3418 done conditionally on Netware platforms to avoid a name clash).
3419 [Guenter <lists@gknw.net>]
3421 *) Add ECDHE and PSK support to DTLS.
3422 [Michael Tuexen <tuexen@fh-muenster.de>]
3424 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3428 *) Add "missing" function EVP_MD_flags() (without this the only way to
3429 retrieve a digest flags is by accessing the structure directly. Update
3430 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3431 or cipher is registered as in the "from" argument. Print out all
3432 registered digests in the dgst usage message instead of manually
3433 attempting to work them out.
3436 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3437 this allows the use of compression and extensions. Change default cipher
3438 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3439 by default unless an application cipher string requests it.
3442 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3443 key ids to find matching certificates and keys but some PKCS#12 files
3444 don't follow the (somewhat unwritten) rules and this strategy fails.
3445 Now just gather all certificates together and the first private key
3446 then look for the first certificate that matches the key.
3449 *) Support use of registered digest and cipher names for dgst and cipher
3450 commands instead of having to add each one as a special case. So now
3457 openssl dgst -sha256 foo
3459 and this works for ENGINE based algorithms too.
3463 *) Update Gost ENGINE to support parameter files.
3464 [Victor B. Wagner <vitus@cryptocom.ru>]
3466 *) Support GeneralizedTime in ca utility.
3467 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3469 *) Enhance the hash format used for certificate directory links. The new
3470 form uses the canonical encoding (meaning equivalent names will work
3471 even if they aren't identical) and uses SHA1 instead of MD5. This form
3472 is incompatible with the older format and as a result c_rehash should
3473 be used to rebuild symbolic links.
3476 *) Make PKCS#8 the default write format for private keys, replacing the
3477 traditional format. This form is standardised, more secure and doesn't
3478 include an implicit MD5 dependency.
3481 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
3482 committed to OpenSSL should pass this lot as a minimum.
3485 *) Add session ticket override functionality for use by EAP-FAST.
3486 [Jouni Malinen <j@w1.fi>]
3488 *) Modify HMAC functions to return a value. Since these can be implemented
3489 in an ENGINE errors can occur.
3492 *) Type-checked OBJ_bsearch_ex.
3495 *) Type-checked OBJ_bsearch. Also some constification necessitated
3496 by type-checking. Still to come: TXT_DB, bsearch(?),
3497 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
3501 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
3502 seconds to a tm structure directly, instead of going through OS
3503 specific date routines. This avoids any issues with OS routines such
3504 as the year 2038 bug. New *_adj() functions for ASN1 time structures
3505 and X509_time_adj_ex() to cover the extended range. The existing
3506 X509_time_adj() is still usable and will no longer have any date issues.
3509 *) Delta CRL support. New use deltas option which will attempt to locate
3510 and search any appropriate delta CRLs available.
3512 This work was sponsored by Google.
3515 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
3516 code and add additional score elements. Validate alternate CRL paths
3517 as part of the CRL checking and indicate a new error "CRL path validation
3518 error" in this case. Applications wanting additional details can use
3519 the verify callback and check the new "parent" field. If this is not
3520 NULL CRL path validation is taking place. Existing applications won't
3521 see this because it requires extended CRL support which is off by
3524 This work was sponsored by Google.
3527 *) Support for freshest CRL extension.
3529 This work was sponsored by Google.
3532 *) Initial indirect CRL support. Currently only supported in the CRLs
3533 passed directly and not via lookup. Process certificate issuer
3534 CRL entry extension and lookup CRL entries by bother issuer name
3535 and serial number. Check and process CRL issuer entry in IDP extension.
3537 This work was sponsored by Google.
3540 *) Add support for distinct certificate and CRL paths. The CRL issuer
3541 certificate is validated separately in this case. Only enabled if
3542 an extended CRL support flag is set: this flag will enable additional
3543 CRL functionality in future.
3545 This work was sponsored by Google.
3548 *) Add support for policy mappings extension.
3550 This work was sponsored by Google.
3553 *) Fixes to pathlength constraint, self issued certificate handling,
3554 policy processing to align with RFC3280 and PKITS tests.
3556 This work was sponsored by Google.
3559 *) Support for name constraints certificate extension. DN, email, DNS
3560 and URI types are currently supported.
3562 This work was sponsored by Google.
3565 *) To cater for systems that provide a pointer-based thread ID rather
3566 than numeric, deprecate the current numeric thread ID mechanism and
3567 replace it with a structure and associated callback type. This
3568 mechanism allows a numeric "hash" to be extracted from a thread ID in
3569 either case, and on platforms where pointers are larger than 'long',
3570 mixing is done to help ensure the numeric 'hash' is usable even if it
3571 can't be guaranteed unique. The default mechanism is to use "&errno"
3572 as a pointer-based thread ID to distinguish between threads.
3574 Applications that want to provide their own thread IDs should now use
3575 CRYPTO_THREADID_set_callback() to register a callback that will call
3576 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
3578 Note that ERR_remove_state() is now deprecated, because it is tied
3579 to the assumption that thread IDs are numeric. ERR_remove_state(0)
3580 to free the current thread's error state should be replaced by
3581 ERR_remove_thread_state(NULL).
3583 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
3584 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
3585 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
3586 application was previously providing a numeric thread callback that
3587 was inappropriate for distinguishing threads, then uniqueness might
3588 have been obtained with &errno that happened immediately in the
3589 intermediate development versions of OpenSSL; this is no longer the
3590 case, the numeric thread callback will now override the automatic use
3592 [Geoff Thorpe, with help from Bodo Moeller]
3594 *) Initial support for different CRL issuing certificates. This covers a
3595 simple case where the self issued certificates in the chain exist and
3596 the real CRL issuer is higher in the existing chain.
3598 This work was sponsored by Google.
3601 *) Removed effectively defunct crypto/store from the build.
3604 *) Revamp of STACK to provide stronger type-checking. Still to come:
3605 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
3606 ASN1_STRING, CONF_VALUE.
3609 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
3610 RAM on SSL connections. This option can save about 34k per idle SSL.
3613 *) Revamp of LHASH to provide stronger type-checking. Still to come:
3614 STACK, TXT_DB, bsearch, qsort.
3617 *) Initial support for Cryptographic Message Syntax (aka CMS) based
3618 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
3619 support for data, signedData, compressedData, digestedData and
3620 encryptedData, envelopedData types included. Scripts to check against
3621 RFC4134 examples draft and interop and consistency checks of many
3622 content types and variants.
3625 *) Add options to enc utility to support use of zlib compression BIO.
3628 *) Extend mk1mf to support importing of options and assembly language
3629 files from Configure script, currently only included in VC-WIN32.
projects / openssl.git / blob