4 This is a detailed breakdown of significant changes. For a high-level overview
5 of changes in each release, see [NEWS.md](./NEWS.md).
7 For a full list of changes, see the [git commit log][log] and pick the
8 appropriate release branch.
10 [log]: https://github.com/openssl/openssl/commits/
15 - [OpenSSL 3.3](#openssl-33)
16 - [OpenSSL 3.2](#openssl-32)
17 - [OpenSSL 3.1](#openssl-31)
18 - [OpenSSL 3.0](#openssl-30)
19 - [OpenSSL 1.1.1](#openssl-111)
20 - [OpenSSL 1.1.0](#openssl-110)
21 - [OpenSSL 1.0.2](#openssl-102)
22 - [OpenSSL 1.0.1](#openssl-101)
23 - [OpenSSL 1.0.0](#openssl-100)
24 - [OpenSSL 0.9.x](#openssl-09x)
29 ### Changes between 3.2 and 3.3 [xx XXX xxxx]
31 * Added API functions SSL_SESSION_get_time_ex(), SSL_SESSION_set_time_ex()
32 using time_t which is Y2038 safe on 32 bit systems when 64 bit time
33 is enabled (e.g via setting glibc macro _TIME_BITS=64).
37 * The d2i_ASN1_GENERALIZEDTIME(), d2i_ASN1_UTCTIME(), ASN1_TIME_check(), and
38 related functions have been augmented to check for a minimum length of
39 the input string, in accordance with ITU-T X.690 section 11.7 and 11.8.
43 * Unknown entries in TLS SignatureAlgorithms, ClientSignatureAlgorithms
44 config options and the respective calls to SSL[_CTX]_set1_sigalgs() and
45 SSL[_CTX]_set1_client_sigalgs() that start with `?` character are
46 ignored and the configuration will still be used.
48 Similarly unknown entries that start with `?` character in a TLS
49 Groups config option or set with SSL[_CTX]_set1_groups_list() are ignored
50 and the configuration will still be used.
52 In both cases if the resulting list is empty, an error is returned.
56 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
57 of CRT (Chinese Remainder Theorem) parameters when requested. See the
58 OSSL_PKEY_PARAM_RSA_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
62 * The activate and soft_load configuration settings for providers in
63 openssl.cnf have been updated to require a value of [1|yes|true|on]
64 (in lower or UPPER case) to enable the setting. Conversely a value
65 of [0|no|false|off] will disable the setting. All other values, or the
66 omission of a value for these settings will result in an error.
70 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
71 override the Issuer and Subject when creating a certificate. The `-subj`
72 option now is an alias for `-set_subject`.
74 *Job Snijders, George Michaelson*
76 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
77 if called with a NULL stack argument.
81 * In `openssl speed`, changed the default hash function used with `hmac` from
86 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
87 - `certProfile` request message header and respective `-profile` CLI option
88 - support for delayed delivery of all types of response messages
92 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
93 be less hard coded in the build file templates, and to allow easier
94 addition of more exporters. With that, an exporter for CMake is also
99 * The BLAKE2s hash algorithm matches BLAKE2b's support
100 for configurable output length.
102 *Ahelenia Ziemiańska*
104 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
105 server to prefer session resumption using PSK-only key exchange over PSK
106 with DHE, if both are available.
108 *Markus Minichmayr, Tapkey GmbH*
110 * New API `SSL_write_ex2`, which can be used to send an end-of-stream (FIN)
111 condition in an optimised way when using QUIC.
115 * New atexit configuration switch, which controls whether the OPENSSL_cleanup
116 is registered when libcrypto is unloaded. This is turned off on NonStop
117 configurations because of loader differences on that platform compared to
122 * Support for qlog for tracing QUIC connections has been added.
124 The qlog output from OpenSSL currently uses a pre-standard draft version of
125 qlog. The output from OpenSSL will change in incompatible ways in future
126 releases, and is not subject to any format stability or compatibility
127 guarantees at this time. This functionality can be
128 disabled with the build-time option `no-unstable-qlog`. See the
129 openssl-qlog(7) manpage for details.
133 * Added APIs to allow configuring the negotiated idle timeout for QUIC
134 connections, and to allow determining the number of additional streams
135 that can currently be created for a QUIC connection.
139 * Added APIs to allow disabling implicit QUIC event processing for
140 QUIC SSL objects, allowing applications to control when event handling
141 occurs. Refer to the SSL_get_value_uint(3) manpage for details.
145 * Added APIs to allow querying the size and utilisation of a QUIC stream's
146 write buffer. Refer to the SSL_get_value_uint(3) manpage for details.
153 ### Changes between 3.2.1 and 3.2.2 [xx XXX xxxx]
155 * Fixed bug where SSL_export_keying_material() could not be used with QUIC
156 connections. (#23560)
160 ### Changes between 3.2.0 and 3.2.1 [30 Jan 2024]
162 * A file in PKCS12 format can contain certificates and keys and may come from
163 an untrusted source. The PKCS12 specification allows certain fields to be
164 NULL, but OpenSSL did not correctly check for this case. A fix has been
165 applied to prevent a NULL pointer dereference that results in OpenSSL
166 crashing. If an application processes PKCS12 files from an untrusted source
167 using the OpenSSL APIs then that application will be vulnerable to this
168 issue prior to this fix.
170 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
171 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
172 and PKCS12_newpass().
174 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
175 function is related to writing data we do not consider it security
182 * When function EVP_PKEY_public_check() is called on RSA public keys,
183 a computation is done to confirm that the RSA modulus, n, is composite.
184 For valid RSA keys, n is a product of two or more large primes and this
185 computation completes quickly. However, if n is an overly large prime,
186 then this computation would take a long time.
188 An application that calls EVP_PKEY_public_check() and supplies an RSA key
189 obtained from an untrusted source could be vulnerable to a Denial of Service
192 The function EVP_PKEY_public_check() is not called from other OpenSSL
193 functions however it is called from the OpenSSL pkey command line
194 application. For that reason that application is also vulnerable if used
195 with the "-pubin" and "-check" options on untrusted data.
197 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
198 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
204 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
205 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
210 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
211 for PowerPC CPUs saves the contents of vector registers in different
212 order than they are restored. Thus the contents of some of these vector
213 registers is corrupted when returning to the caller. The vulnerable code is
214 used only on newer PowerPC processors supporting the PowerISA 2.07
217 The consequences of this kind of internal application state corruption can
218 be various - from no consequences, if the calling application does not
219 depend on the contents of non-volatile XMM registers at all, to the worst
220 consequences, where the attacker could get complete control of the
221 application process. However unless the compiler uses the vector registers
222 for storing pointers, the most likely consequence, if any, would be an
223 incorrect result of some application dependent calculations or a crash
224 leading to a denial of service.
230 * Disable building QUIC server utility when OpenSSL is configured with
235 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
237 * Fix excessive time spent in DH check / generation with large Q parameter
240 Applications that use the functions DH_generate_key() to generate an
241 X9.42 DH key may experience long delays. Likewise, applications that use
242 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
243 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
244 Where the key or parameters that are being checked have been obtained from
245 an untrusted source this may lead to a Denial of Service.
251 * The BLAKE2b hash algorithm supports a configurable output length
252 by setting the "size" parameter.
254 *Čestmír Kalina and Tomáš Mráz*
256 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
260 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
261 and the corresponding provider-storemgmt API function
262 OSSL_FUNC_store_delete().
266 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
267 a passphrase callback when opening a store.
271 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
272 from 8 bytes to 16 bytes.
273 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
274 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
275 requires a salt length of 128 bits. This affects OpenSSL command line
276 applications such as "genrsa" and "pkcs8" and API's such as
277 PEM_write_bio_PrivateKey() that are reliant on the default value.
278 The additional commandline option 'saltlen' has been added to the
279 OpenSSL command line applications for "pkcs8" and "enc" to allow the
280 salt length to be set to a non default value.
284 * Changed the default value of the `ess_cert_id_alg` configuration
285 option which is used to calculate the TSA's public key certificate
286 identifier. The default algorithm is updated to be sha256 instead
289 *Małgorzata Olszówka*
291 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
292 table for point multiplication of the base point, which increases the size of
293 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
294 been added to disable the precomputed table.
298 * Added client side support for QUIC
300 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
302 * Added multiple tutorials on the OpenSSL library and in particular
303 on writing various clients (using TLS and QUIC protocols) with libssl.
307 * Added secp384r1 implementation using Solinas' reduction to improve
308 speed of the NIST P-384 elliptic curve. To enable the implementation
309 the build option `enable-ec_nistp_64_gcc_128` must be used.
313 * Improved RFC7468 compliance of the asn1parse command.
315 *Matthias St. Pierre*
317 * Added SHA256/192 algorithm support.
321 * Added support for securely getting root CA certificate update in
326 * Improved contention on global write locks by using more read locks where
331 * Improved performance of OSSL_PARAM lookups in performance critical
336 * Added the SSL_get0_group_name() function to provide access to the
337 name of the group used for the TLS key exchange.
341 * Provide a new configure option `no-http` that can be used to disable the
342 HTTP support. Provide new configure options `no-apps` and `no-docs` to
343 disable building the openssl command line application and the documentation.
347 * Provide a new configure option `no-ecx` that can be used to disable the
348 X25519, X448, and EdDSA support.
352 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
353 the EVP_KDF_CTX_set_params() function they are now concatenated not just
354 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
358 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
359 the provider context as a parameter.
363 * TLS round-trip time calculation was added by a Brigham Young University
364 Capstone team partnering with Sandia National Laboratories. A new function
365 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
370 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
371 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
372 option. Use of the "advanced" s_client command command via the "-adv" option
377 * Added an "advanced" command mode to s_client. Use this with the "-adv"
378 option. The old "basic" command mode recognises certain letters that must
379 always appear at the start of a line and cannot be escaped. The advanced
380 command mode enables commands to be entered anywhere and there is an
381 escaping mechanism. After starting s_client with "-adv" type "{help}"
382 to show a list of available commands.
386 * Add Raw Public Key (RFC7250) support. Authentication is supported
387 by matching keys against either local policy (TLSA records synthesised
388 from the expected keys) or DANE (TLSA records obtained by the
389 application from DNS). TLSA records will also match the same key in
390 the server certificate, should RPK use not happen to be negotiated.
394 * Added support for modular exponentiation and CRT offloading for the
399 * Added further assembler code for the RISC-V architecture.
403 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
404 from a given EC_GROUP.
408 * Improved support for non-default library contexts and property queries
409 when parsing PKCS#12 files.
413 * Implemented support for all five instances of EdDSA from RFC8032:
414 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
415 The streaming is not yet supported for the HashEdDSA variants
416 (Ed25519ph and Ed448ph).
420 * Added SM4 optimization for ARM processors using ASIMD and AES HW
425 * Implemented SM4-XTS support.
429 * Added platform-agnostic OSSL_sleep() function.
433 * Implemented deterministic ECDSA signatures (RFC6979) support.
437 * Implemented AES-GCM-SIV (RFC8452) support.
441 * Added support for pluggable (provider-based) TLS signature algorithms.
442 This enables TLS 1.3 authentication operations with algorithms embedded
443 in providers not included by default in OpenSSL. In combination with
444 the already available pluggable KEM and X.509 support, this enables
445 for example suitable providers to deliver post-quantum or quantum-safe
446 cryptography to OpenSSL users.
450 * Added support for pluggable (provider-based) CMS signature algorithms.
451 This enables CMS sign and verify operations with algorithms embedded
452 in providers not included by default in OpenSSL.
456 * Added support for Hybrid Public Key Encryption (HPKE) as defined
457 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
458 Message Layer Security (MLS) and other IETF specifications.
459 HPKE can also be used by other applications that require
460 encrypting "to" an ECDH public key. External APIs are defined in
461 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
465 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
470 * Add support for certificate compression (RFC8879), including
471 library support for Brotli and Zstandard compression.
475 * Add the ability to add custom attributes to PKCS12 files. Add a new API
476 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
477 for a user specified callback and optional argument.
478 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
479 added to the existing STACK_OF attrs.
483 * Major refactor of the libssl record layer.
487 * Add a mac salt length option for the pkcs12 command.
491 * Add more SRTP protection profiles from RFC8723 and RFC8269.
495 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
497 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
499 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
500 supported and enabled.
504 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
505 to the list of ciphersuites providing Perfect Forward Secrecy as
506 required by SECLEVEL >= 3.
508 *Dmitry Belyavskiy, Nicola Tuveri*
510 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
511 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
512 SSL_get0_iana_groups() function-like macro, retrieves the list of
513 supported groups sent by the peer.
514 The function SSL_client_hello_get_extension_order() populates
515 a caller-supplied array with the list of extension types present in the
516 ClientHello, in order of appearance.
520 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
521 to make it possible to use empty passphrase strings.
525 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
529 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
530 to change functions used for allocating the memory of asynchronous call stack.
534 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
538 * A failure exit code is returned when using the openssl x509 command to check
539 certificate attributes and the checks fail.
543 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
544 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
545 of 160 bits and above and less than 224 bits were previously accepted by
546 default but are now no longer allowed. By default TLS compression was
547 already disabled in previous OpenSSL versions. At security level 2 it cannot
552 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
557 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
558 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
559 will need to load the legacy crypto provider.
563 * CCM8 cipher suites in TLS have been downgraded to security level zero
564 because they use a short authentication tag which lowers their strength.
568 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
569 by default. Also spaces surrounding `=` in DN output are removed.
573 * Add X.509 certificate codeSigning purpose and related checks on key usage and
574 extended key usage of the leaf certificate according to the CA/Browser Forum.
578 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
579 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
580 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
581 X.509 version 3 if the certificate information includes X.509 extensions.
585 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
586 such as adding a trace facility for debugging certificate chain building.
590 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
591 in particular supporting requests for central key generation, generalized
592 polling, and various types of genm/genp exchanges defined in CMP Updates.
596 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
597 like correcting the TLS and proxy support and adding tracing for debugging.
601 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
605 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
606 a certificate to be added is already present. `CMS_sign_ex()` and
607 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
608 and no longer throw an error for them.
612 * Fixed and extended `util/check-format.pl` for checking adherence to the
613 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
614 The checks are meanwhile more complete and yield fewer false positives.
618 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
619 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
620 calls. They can be used as the transport BIOs for QUIC.
622 *Hugo Landau, Matt Caswell and Tomáš Mráz*
624 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
625 sending and receiving multiple messages in a single call. An implementation
626 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
630 * Support for loading root certificates from the Windows certificate store
631 has been added. The support is in the form of a store which recognises the
632 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
633 arguments. This store is built by default and can be disabled using the new
634 compile-time option `no-winstore`. This store is not currently used by
635 default and must be loaded explicitly using the above store URI. It is
636 expected to be loaded by default in the future.
640 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
641 kernel versions that support KTLS have a known bug in CCM processing. That
642 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
643 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
648 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
653 * Zerocopy KTLS sendfile() support on Linux.
657 * The OBJ_ calls are now thread safe using a global lock.
661 * New parameter `-digest` for openssl cms command allowing signing
662 pre-computed digests and new CMS API functions supporting that
667 * OPENSSL_malloc() and other allocation functions now raise errors on
668 allocation failures. The callers do not need to explicitly raise errors
669 unless they want to for tracing purposes.
673 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
674 decryption as a protection against Bleichenbacher-like attacks.
675 The RSA decryption API will now return a randomly generated deterministic
676 message instead of an error in case it detects an error when checking
677 padding during PKCS#1 v1.5 decryption. This is a general protection against
678 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
680 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
681 on the RSA decryption context.
685 * Added support for Brainpool curves in TLS-1.3.
687 *Bernd Edlinger and Matt Caswell*
689 * Added OpenBSD specific build targets.
693 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
694 a basic thread pool implementation for select platforms.
701 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
703 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
704 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
705 that alter the key or IV length ([CVE-2023-5363]).
709 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
711 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
713 The POLY1305 MAC (message authentication code) implementation in OpenSSL
714 does not save the contents of non-volatile XMM registers on Windows 64
715 platform when calculating the MAC of data larger than 64 bytes. Before
716 returning to the caller all the XMM registers are set to zero rather than
717 restoring their previous content. The vulnerable code is used only on newer
718 x86_64 processors supporting the AVX512-IFMA instructions.
720 The consequences of this kind of internal application state corruption can
721 be various - from no consequences, if the calling application does not
722 depend on the contents of non-volatile XMM registers at all, to the worst
723 consequences, where the attacker could get complete control of the
724 application process. However given the contents of the registers are just
725 zeroized so the attacker cannot put arbitrary values inside, the most likely
726 consequence, if any, would be an incorrect result of some application
727 dependent calculations or a crash leading to a denial of service.
733 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
735 * Fix excessive time spent checking DH q parameter value.
737 The function DH_check() performs various checks on DH parameters. After
738 fixing CVE-2023-3446 it was discovered that a large q parameter value can
739 also trigger an overly long computation during some of these checks.
740 A correct q value, if present, cannot be larger than the modulus p
741 parameter, thus it is unnecessary to perform these checks if q is larger
744 If DH_check() is called with such q parameter value,
745 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
746 intensive checks are skipped.
752 * Fix DH_check() excessive time with over sized modulus.
754 The function DH_check() performs various checks on DH parameters. One of
755 those checks confirms that the modulus ("p" parameter) is not too large.
756 Trying to use a very large modulus is slow and OpenSSL will not normally use
757 a modulus which is over 10,000 bits in length.
759 However the DH_check() function checks numerous aspects of the key or
760 parameters that have been supplied. Some of those checks use the supplied
761 modulus value even if it has already been found to be too large.
763 A new limit has been added to DH_check of 32,768 bits. Supplying a
764 key/parameters with a modulus over this size will simply cause DH_check() to
771 * Do not ignore empty associated data entries with AES-SIV.
773 The AES-SIV algorithm allows for authentication of multiple associated
774 data entries along with the encryption. To authenticate empty data the
775 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
776 with NULL pointer as the output buffer and 0 as the input buffer length.
777 The AES-SIV implementation in OpenSSL just returns success for such call
778 instead of performing the associated data authentication operation.
779 The empty data thus will not be authenticated. ([CVE-2023-2975])
781 Thanks to Juerg Wullschleger (Google) for discovering the issue.
783 The fix changes the authentication tag value and the ciphertext for
784 applications that use empty associated data entries with AES-SIV.
785 To decrypt data encrypted with previous versions of OpenSSL the application
786 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
791 * When building with the `enable-fips` option and using the resulting
792 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
793 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
794 not operate with truncated digests (FIPS 140-3 IG G.R).
798 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
800 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
801 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
803 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
804 numeric text form. For gigantic sub-identifiers, this would take a very
805 long time, the time complexity being O(n^2) where n is the size of that
806 sub-identifier. ([CVE-2023-2650])
808 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
809 IDENTIFIER to canonical numeric text form if the size of that OBJECT
810 IDENTIFIER is 586 bytes or less, and fail otherwise.
812 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
813 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
814 most 128 sub-identifiers, and that the maximum value that each sub-
815 identifier may have is 2^32-1 (4294967295 decimal).
817 For each byte of every sub-identifier, only the 7 lower bits are part of
818 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
819 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
824 * Multiple algorithm implementation fixes for ARM BE platforms.
828 * Added a -pedantic option to fipsinstall that adjusts the various
829 settings to ensure strict FIPS compliance rather than backwards
834 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
835 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
836 trigger a crash of an application using AES-XTS decryption if the memory
837 just after the buffer being decrypted is not mapped.
838 Thanks to Anton Romanov (Amazon) for discovering the issue.
843 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
844 The previous fix for this timing side channel turned out to cause
845 a severe 2-3x performance regression in the typical use case
846 compared to 3.0.7. The new fix uses existing constant time
847 code paths, and restores the previous performance level while
848 fully eliminating all existing timing side channels.
849 The fix was developed by Bernd Edlinger with testing support
854 * Add FIPS provider configuration option to disallow the use of
855 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
856 The option '-no_drbg_truncated_digests' can optionally be
857 supplied to 'openssl fipsinstall'.
861 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
862 that it does not enable policy checking. Thanks to David Benjamin for
863 discovering this issue.
868 * Fixed an issue where invalid certificate policies in leaf certificates are
869 silently ignored by OpenSSL and other certificate policy checks are skipped
870 for that certificate. A malicious CA could use this to deliberately assert
871 invalid certificate policies in order to circumvent policy checking on the
872 certificate altogether.
877 * Limited the number of nodes created in a policy tree to mitigate
878 against CVE-2023-0464. The default limit is set to 1000 nodes, which
879 should be sufficient for most installations. If required, the limit
880 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
881 time define to a desired maximum number of nodes or zero to allow
887 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
889 * Add FIPS provider configuration option to enforce the
890 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
891 The option '-ems-check' can optionally be supplied to
892 'openssl fipsinstall'.
896 * The FIPS provider includes a few non-approved algorithms for
897 backward compatibility purposes and the "fips=yes" property query
898 must be used for all algorithm fetches to ensure FIPS compliance.
900 The algorithms that are included but not approved are Triple DES ECB,
901 Triple DES CBC and EdDSA.
905 * Added support for KMAC in KBKDF.
909 * RNDR and RNDRRS support in provider functions to provide
910 random number generation for Arm CPUs (aarch64).
914 * s_client and s_server apps now explicitly say when the TLS version
915 does not include the renegotiation mechanism. This avoids confusion
916 between that scenario versus when the TLS version includes secure
917 renegotiation but the peer lacks support for it.
921 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
923 *Tomasz Kantecki, Andrey Matyukov*
925 * The various OBJ_* functions have been made thread safe.
929 * Parallel dual-prime 1536/2048-bit modular exponentiation for
930 AVX512_IFMA capable processors.
932 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
934 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
935 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
936 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
937 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
938 `OPENSSL_NO_DEPRECATED_3_1`.
940 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
941 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
942 definitions for these functions regardless of whether
943 `OPENSSL_NO_DEPRECATED_3_1` is defined.
945 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
946 functions regardless of whether they are using them. It is recommended that
947 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
951 * When generating safe-prime DH parameters set the recommended private key
952 length equivalent to minimum key lengths as in RFC 7919.
956 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
957 maximum size that is smaller or equal to the digest length to comply with
958 FIPS 186-4 section 5. This is implemented by a new option
959 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
960 `rsa_pss_saltlen` parameter, which is now the default. Signature
961 verification is not affected by this change and continues to work as before.
968 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
969 listed here are only a brief description.
970 The migration guide contains more detailed information related to new features,
971 breaking changes, and mappings for the large list of deprecated functions.
973 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
975 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
977 * Fixed NULL dereference during PKCS7 data verification.
979 A NULL pointer can be dereferenced when signatures are being
980 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
981 algorithm used for the signature is known to the OpenSSL library but
982 the implementation of the hash algorithm is not available the digest
983 initialization will fail. There is a missing check for the return
984 value from the initialization function which later leads to invalid
985 usage of the digest API most likely leading to a crash.
988 PKCS7 data is processed by the SMIME library calls and also by the
989 time stamp (TS) library calls. The TLS implementation in OpenSSL does
990 not call these functions however third party applications would be
991 affected if they call these functions to verify signatures on untrusted
996 * Fixed X.400 address type confusion in X.509 GeneralName.
998 There is a type confusion vulnerability relating to X.400 address processing
999 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
1000 but the public structure definition for GENERAL_NAME incorrectly specified
1001 the type of the x400Address field as ASN1_TYPE. This field is subsequently
1002 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
1003 than an ASN1_STRING.
1005 When CRL checking is enabled (i.e. the application sets the
1006 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
1007 pass arbitrary pointers to a memcmp call, enabling them to read memory
1008 contents or enact a denial of service.
1013 * Fixed NULL dereference validating DSA public key.
1015 An invalid pointer dereference on read can be triggered when an
1016 application tries to check a malformed DSA public key by the
1017 EVP_PKEY_public_check() function. This will most likely lead
1018 to an application crash. This function can be called on public
1019 keys supplied from untrusted sources which could allow an attacker
1020 to cause a denial of service attack.
1022 The TLS implementation in OpenSSL does not call this function
1023 but applications might call the function if there are additional
1024 security requirements imposed by standards such as FIPS 140-3.
1027 *Shane Lontis, Tomáš Mráz*
1029 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
1031 An invalid pointer dereference on read can be triggered when an
1032 application tries to load malformed PKCS7 data with the
1033 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
1035 The result of the dereference is an application crash which could
1036 lead to a denial of service attack. The TLS implementation in OpenSSL
1037 does not call this function however third party applications might
1038 call these functions on untrusted data.
1043 * Fixed Use-after-free following BIO_new_NDEF.
1045 The public API function BIO_new_NDEF is a helper function used for
1046 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
1047 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
1048 be called directly by end user applications.
1050 The function receives a BIO from the caller, prepends a new BIO_f_asn1
1051 filter BIO onto the front of it to form a BIO chain, and then returns
1052 the new head of the BIO chain to the caller. Under certain conditions,
1053 for example if a CMS recipient public key is invalid, the new filter BIO
1054 is freed and the function returns a NULL result indicating a failure.
1055 However, in this case, the BIO chain is not properly cleaned up and the
1056 BIO passed by the caller still retains internal pointers to the previously
1057 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
1058 then a use-after-free will occur. This will most likely result in a crash.
1061 *Viktor Dukhovni, Matt Caswell*
1063 * Fixed Double free after calling PEM_read_bio_ex.
1065 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
1066 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1067 data. If the function succeeds then the "name_out", "header" and "data"
1068 arguments are populated with pointers to buffers containing the relevant
1069 decoded data. The caller is responsible for freeing those buffers. It is
1070 possible to construct a PEM file that results in 0 bytes of payload data.
1071 In this case PEM_read_bio_ex() will return a failure code but will populate
1072 the header argument with a pointer to a buffer that has already been freed.
1073 If the caller also frees this buffer then a double free will occur. This
1074 will most likely lead to a crash.
1076 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1077 PEM_read_bio_ex() and therefore these functions are also directly affected.
1079 These functions are also called indirectly by a number of other OpenSSL
1080 functions including PEM_X509_INFO_read_bio_ex() and
1081 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1082 internal uses of these functions are not vulnerable because the caller does
1083 not free the header argument if PEM_read_bio_ex() returns a failure code.
1086 *Kurt Roeckx, Matt Caswell*
1088 * Fixed Timing Oracle in RSA Decryption.
1090 A timing based side channel exists in the OpenSSL RSA Decryption
1091 implementation which could be sufficient to recover a plaintext across
1092 a network in a Bleichenbacher style attack. To achieve a successful
1093 decryption an attacker would have to be able to send a very large number
1094 of trial messages for decryption. The vulnerability affects all RSA padding
1095 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1098 *Dmitry Belyavsky, Hubert Kario*
1100 * Fixed X.509 Name Constraints Read Buffer Overflow.
1102 A read buffer overrun can be triggered in X.509 certificate verification,
1103 specifically in name constraint checking. The read buffer overrun might
1104 result in a crash which could lead to a denial of service attack.
1105 In a TLS client, this can be triggered by connecting to a malicious
1106 server. In a TLS server, this can be triggered if the server requests
1107 client authentication and a malicious client connects.
1112 * Fixed X.509 Policy Constraints Double Locking security issue.
1114 If an X.509 certificate contains a malformed policy constraint and
1115 policy processing is enabled, then a write lock will be taken twice
1116 recursively. On some operating systems (most widely: Windows) this
1117 results in a denial of service when the affected process hangs. Policy
1118 processing being enabled on a publicly facing server is not considered
1119 to be a common setup.
1124 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1125 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1126 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1127 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1128 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1129 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1130 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1131 for legacy EC and SM2 keys is also changed similarly to honor the
1132 equivalent conversion format flag as specified in the underlying
1133 `EC_KEY` object being exported to a provider, when this function is
1134 called through `EVP_PKEY_export()`.
1138 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1140 * Fixed two buffer overflows in punycode decoding functions.
1142 A buffer overrun can be triggered in X.509 certificate verification,
1143 specifically in name constraint checking. Note that this occurs after
1144 certificate chain signature verification and requires either a CA to
1145 have signed the malicious certificate or for the application to continue
1146 certificate verification despite failure to construct a path to a trusted
1149 In a TLS client, this can be triggered by connecting to a malicious
1150 server. In a TLS server, this can be triggered if the server requests
1151 client authentication and a malicious client connects.
1153 An attacker can craft a malicious email address to overflow
1154 an arbitrary number of bytes containing the `.` character (decimal 46)
1155 on the stack. This buffer overflow could result in a crash (causing a
1159 An attacker can craft a malicious email address to overflow four
1160 attacker-controlled bytes on the stack. This buffer overflow could
1161 result in a crash (causing a denial of service) or potentially remote code
1162 execution depending on stack layout for any given platform/compiler.
1167 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1168 parameters in OpenSSL code.
1169 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1170 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1171 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1172 Using these invalid names may cause algorithms to use slower methods
1173 that ignore the CRT parameters.
1177 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1182 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1183 data to be signed before signing the certificate.
1187 * Added RIPEMD160 to the default provider.
1191 * Ensured that the key share group sent or accepted for the key exchange
1192 is allowed for the protocol version.
1196 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1198 * OpenSSL supports creating a custom cipher via the legacy
1199 EVP_CIPHER_meth_new() function and associated function calls. This function
1200 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1201 to use the new provider mechanism in order to implement custom ciphers.
1203 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1204 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1205 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1206 and decryption initialisation functions). Instead of using the custom cipher
1207 directly it incorrectly tries to fetch an equivalent cipher from the
1208 available providers. An equivalent cipher is found based on the NID passed to
1209 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1210 given cipher. However it is possible for an application to incorrectly pass
1211 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1212 is used in this way the OpenSSL encryption/decryption initialisation function
1213 will match the NULL cipher as being equivalent and will fetch this from the
1214 available providers. This will succeed if the default provider has been
1215 loaded (or if a third party provider has been loaded that offers this
1216 cipher). Using the NULL cipher means that the plaintext is emitted as the
1219 Applications are only affected by this issue if they call
1220 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1221 encryption/decryption initialisation function. Applications that only use
1222 SSL/TLS are not impacted by this issue.
1227 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1232 * Fixed the linux-mips64 Configure target which was missing the
1233 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1238 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1243 * Correctly handle a retransmitted ClientHello in DTLS
1247 * Fixed detection of ktls support in cross-compile environment on Linux
1251 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1256 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1257 report correct results in some cases
1261 * Fix UWP builds by defining VirtualLock
1265 * For known safe primes use the minimum key length according to RFC 7919.
1266 Longer private key sizes unnecessarily raise the cycles needed to compute the
1267 shared secret without any increase of the real security. This fixes a
1268 regression from 1.1.1 where these shorter keys were generated for the known
1273 * Added the loongarch64 target
1277 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1278 only passed to the FIPS provider and not to the default or legacy provider.
1282 * Fixed reported performance degradation on aarch64. Restored the
1283 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1284 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1285 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1286 The new algorithm is still used for 32 bit targets.
1290 * Added a missing header for memcmp that caused compilation failure on some
1295 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1297 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1298 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1299 This issue makes the RSA implementation with 2048 bit private keys
1300 incorrect on such machines and memory corruption will happen during
1301 the computation. As a consequence of the memory corruption an attacker
1302 may be able to trigger a remote code execution on the machine performing
1305 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1306 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1307 are affected by this issue.
1312 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1313 implementation would not encrypt the entirety of the data under some
1314 circumstances. This could reveal sixteen bytes of data that was
1315 preexisting in the memory that wasn't written. In the special case of
1316 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1318 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1319 they are both unaffected.
1322 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1324 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1326 * In addition to the c_rehash shell command injection identified in
1327 CVE-2022-1292, further bugs where the c_rehash script does not
1328 properly sanitise shell metacharacters to prevent command injection have been
1331 When the CVE-2022-1292 was fixed it was not discovered that there
1332 are other places in the script where the file names of certificates
1333 being hashed were possibly passed to a command executed through the shell.
1335 This script is distributed by some operating systems in a manner where
1336 it is automatically executed. On such operating systems, an attacker
1337 could execute arbitrary commands with the privileges of the script.
1339 Use of the c_rehash script is considered obsolete and should be replaced
1340 by the OpenSSL rehash command line tool.
1343 *Daniel Fiala, Tomáš Mráz*
1345 * Case insensitive string comparison no longer uses locales. It has instead
1346 been directly implemented.
1350 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1352 * Case insensitive string comparison is reimplemented via new locale-agnostic
1353 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1354 comparison. The previous implementation had problems when the Turkish locale
1359 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1360 metacharacters to prevent command injection. This script is distributed by
1361 some operating systems in a manner where it is automatically executed. On
1362 such operating systems, an attacker could execute arbitrary commands with the
1363 privileges of the script.
1365 Use of the c_rehash script is considered obsolete and should be replaced
1366 by the OpenSSL rehash command line tool.
1371 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1372 certificate on an OCSP response. The bug caused the function in the case
1373 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1374 response (meaning a successful verification) even in the case where the
1375 response signing certificate fails to verify.
1377 It is anticipated that most users of `OCSP_basic_verify` will not use the
1378 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1379 a negative value (indicating a fatal error) in the case of a certificate
1380 verification failure. The normal expected return value in this case would be
1383 This issue also impacts the command line OpenSSL "ocsp" application. When
1384 verifying an ocsp response with the "-no_cert_checks" option the command line
1385 application will report that the verification is successful even though it
1386 has in fact failed. In this case the incorrect successful response will also
1387 be accompanied by error messages showing the failure and contradicting the
1388 apparently successful result.
1393 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1394 AAD data as the MAC key. This made the MAC key trivially predictable.
1396 An attacker could exploit this issue by performing a man-in-the-middle attack
1397 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1398 that the modified data would still pass the MAC integrity check.
1400 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1401 endpoint will always be rejected by the recipient and the connection will
1402 fail at that point. Many application protocols require data to be sent from
1403 the client to the server first. Therefore, in such a case, only an OpenSSL
1404 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1406 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1407 sent in both directions. In this case both clients and servers could be
1408 affected, regardless of the application protocol.
1410 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1411 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1412 the handshake when using this ciphersuite.
1414 The confidentiality of data is not impacted by this issue, i.e. an attacker
1415 cannot decrypt data that has been encrypted using this ciphersuite - they can
1418 In order for this attack to work both endpoints must legitimately negotiate
1419 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1420 OpenSSL 3.0, and is not available within the default provider or the default
1421 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1422 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1423 following must have occurred:
1425 1) OpenSSL must have been compiled with the (non-default) compile time option
1426 enable-weak-ssl-ciphers
1428 2) OpenSSL must have had the legacy provider explicitly loaded (either
1429 through application code or via configuration)
1431 3) The ciphersuite must have been explicitly added to the ciphersuite list
1433 4) The libssl security level must have been set to 0 (default is 1)
1435 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1437 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1438 others that both endpoints have in common
1443 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1444 occupied by the removed hash table entries.
1446 This function is used when decoding certificates or keys. If a long lived
1447 process periodically decodes certificates or keys its memory usage will
1448 expand without bounds and the process might be terminated by the operating
1449 system causing a denial of service. Also traversing the empty hash table
1450 entries will take increasingly more time.
1452 Typically such long lived processes might be TLS clients or TLS servers
1453 configured to accept client certificate authentication.
1456 *Hugo Landau, Aliaksei Levin*
1458 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1459 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1460 statistics are no longer supported. For compatibility, these statistics are
1461 still listed in the output but are now always reported as zero.
1465 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1467 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1468 for non-prime moduli.
1470 Internally this function is used when parsing certificates that contain
1471 elliptic curve public keys in compressed form or explicit elliptic curve
1472 parameters with a base point encoded in compressed form.
1474 It is possible to trigger the infinite loop by crafting a certificate that
1475 has invalid explicit curve parameters.
1477 Since certificate parsing happens prior to verification of the certificate
1478 signature, any process that parses an externally supplied certificate may thus
1479 be subject to a denial of service attack. The infinite loop can also be
1480 reached when parsing crafted private keys as they can contain explicit
1481 elliptic curve parameters.
1483 Thus vulnerable situations include:
1485 - TLS clients consuming server certificates
1486 - TLS servers consuming client certificates
1487 - Hosting providers taking certificates or private keys from customers
1488 - Certificate authorities parsing certification requests from subscribers
1489 - Anything else which parses ASN.1 elliptic curve parameters
1491 Also any other applications that use the BN_mod_sqrt() where the attacker
1492 can control the parameter values are vulnerable to this DoS issue.
1497 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1498 to the list of ciphersuites providing Perfect Forward Secrecy as
1499 required by SECLEVEL >= 3.
1501 *Dmitry Belyavskiy, Nicola Tuveri*
1503 * Made the AES constant time code for no-asm configurations
1504 optional due to the resulting 95% performance degradation.
1505 The AES constant time code can be enabled, for no assembly
1506 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1510 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1515 * The negative return value handling of the certificate verification callback
1516 was reverted. The replacement is to set the verification retry state with
1517 the SSL_set_retry_verify() function.
1521 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1523 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1524 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1525 verify a certificate supplied by a server. That function may return a
1526 negative return value to indicate an internal error (for example out of
1527 memory). Such a negative return value is mishandled by OpenSSL and will cause
1528 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1529 success and a subsequent call to SSL_get_error() to return the value
1530 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1531 returned by OpenSSL if the application has previously called
1532 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1533 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1534 totally unexpected and applications may not behave correctly as a result. The
1535 exact behaviour will depend on the application but it could result in
1536 crashes, infinite loops or other similar incorrect responses.
1538 This issue is made more serious in combination with a separate bug in OpenSSL
1539 3.0 that will cause X509_verify_cert() to indicate an internal error when
1540 processing a certificate chain. This will occur where a certificate does not
1541 include the Subject Alternative Name extension but where a Certificate
1542 Authority has enforced name constraints. This issue can occur even with valid
1548 * Corrected a few file name and file reference bugs in the build,
1549 installation and setup scripts, which lead to installation verification
1550 failures. Slightly enhanced the installation verification script.
1554 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1559 * Fixed PVK encoder to properly query for the passphrase.
1563 * Multiple fixes in the OSSL_HTTP API functions.
1567 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1568 OSSL_PARAM_INTEGER data type and return error on negative numbers
1569 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1570 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1574 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1578 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1582 * Multiple threading fixes.
1586 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1590 * Allow fetching an operation from the provider that owns an unexportable key
1591 as a fallback if that is still allowed by the property query.
1595 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1597 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1602 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1603 S390X capability vector to zero. This simplifies testing of different code
1604 paths on S390X architecture.
1608 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1609 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1610 SP 800-38D". The communication will fail at this point.
1614 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1615 confidential in EC_GROUP data.
1619 * The byte order mark (BOM) character is ignored if encountered at the
1620 beginning of a PEM-formatted file.
1624 * Added CMS support for the Russian GOST algorithms.
1628 * Due to move of the implementation of cryptographic operations
1629 to the providers, validation of various operation parameters can
1630 be postponed until the actual operation is executed where previously
1631 it happened immediately when an operation parameter was set.
1633 For example when setting an unsupported curve with
1634 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1635 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1637 *OpenSSL team members and many third party contributors*
1639 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1640 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1641 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1642 instead to retrieve these algorithms from a provider.
1646 * On build targets where the multilib postfix is set in the build
1647 configuration the libdir directory was changing based on whether
1648 the lib directory with the multilib postfix exists on the system
1649 or not. This unpredictable behavior was removed and eventual
1650 multilib postfix is now always added to the default libdir. Use
1651 `--libdir=lib` to override the libdir if adding the postfix is
1656 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1657 no longer interoperable with OpenSSL 1.1.1.
1661 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1662 function codes, this function can only cause problems for calling
1667 * Add a configurable flag to output date formats as ISO 8601. Does not
1668 change the default date format.
1672 * Version of MSVC earlier than 1300 could get link warnings, which could
1673 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1674 Support for this flag has been removed.
1678 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1679 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1680 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1681 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1682 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1686 * The signatures of the functions to get and set options on SSL and
1687 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1688 Some source code changes may be required.
1692 * The public definitions of conf_method_st and conf_st have been
1693 deprecated. They will be made opaque in a future release.
1695 *Rich Salz and Tomáš Mráz*
1697 * Client-initiated renegotiation is disabled by default. To allow it, use
1698 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1699 flag, or the "ClientRenegotiation" config parameter as appropriate.
1703 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1704 or modify relative pathname inclusion.
1708 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1709 validated. Please consult the README-FIPS and
1710 README-PROVIDERS files, as well as the migration guide.
1712 *OpenSSL team members and many third party contributors*
1714 * For the key types DH and DHX the allowed settable parameters are now different.
1718 * The openssl commands that read keys, certificates, and CRLs now
1719 automatically detect the PEM or DER format of the input files.
1721 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1723 * Added enhanced PKCS#12 APIs which accept a library context.
1727 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1731 * Added support for Kernel TLS (KTLS).
1733 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1735 * Support for RFC 5746 secure renegotiation is now required by default for
1736 SSL or TLS connections to succeed.
1740 * The signature of the `copy` functional parameter of the
1741 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1742 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1743 the signature of the `pub_decode` functional parameter of the
1744 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1745 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1749 * The error return values from some control calls (ctrl) have changed.
1753 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1757 * Many functions in the EVP_ namespace that are getters of values from
1758 implementations or contexts were renamed to include get or get0 in their
1759 names. Old names are provided as macro aliases for compatibility and
1764 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1765 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1766 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1771 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1774 * The output from the command line applications may have minor
1779 * The output from numerous "printing" may have minor changes.
1783 * Windows thread synchronization uses read/write primitives (SRWLock) when
1784 supported by the OS, otherwise CriticalSection continues to be used.
1788 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1789 work on read only BIO source/sinks that do not support these functions.
1790 This allows piping or redirection of a file BIO using stdin to be buffered
1791 into memory. This is used internally in OSSL_DECODER_from_bio().
1795 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1796 this function would return one of the values OSSL_STORE_INFO_NAME,
1797 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1798 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1799 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1800 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1801 using this function should be amended to handle the changed return value.
1805 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1806 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1807 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1808 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1809 contains more than one certificate identifier: This means that all
1810 certificates referenced there MUST be part of the validation chain.
1814 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1815 RC5, DESX and DES have been moved to the legacy provider.
1819 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1820 RIPEMD-160 have been moved to the legacy provider.
1824 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1829 * The deprecated functions EVP_PKEY_get0_RSA(),
1830 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1831 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1832 well as the similarly named "get1" functions behave differently in
1837 * A number of functions handling low-level keys or engines were deprecated
1838 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1839 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1840 EVP_PKEY_get0_siphash().
1844 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1845 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1846 will need to load the legacy crypto provider. This includes these PBE
1847 algorithms which use this KDF:
1848 - NID_pbeWithMD2AndDES_CBC
1849 - NID_pbeWithMD5AndDES_CBC
1850 - NID_pbeWithSHA1AndRC2_CBC
1851 - NID_pbeWithMD2AndRC2_CBC
1852 - NID_pbeWithMD5AndRC2_CBC
1853 - NID_pbeWithSHA1AndDES_CBC
1857 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1858 BIO_debug_callback() functions.
1862 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1863 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1867 * The RAND_METHOD APIs have been deprecated.
1871 * The SRP APIs have been deprecated.
1875 * Add a compile time option to prevent the caching of provider fetched
1876 algorithms. This is enabled by including the no-cached-fetch option
1877 at configuration time.
1881 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1882 count of PKCS12_DEFAULT_ITER.
1884 *Tomáš Mráz and Sahana Prasad*
1886 * The openssl speed command does not use low-level API calls anymore.
1890 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1893 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1895 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1899 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1900 providers may supply their own group implementations (using either the "key
1901 exchange" or the "key encapsulation" methods) which will automatically be
1902 detected and used by libssl.
1904 *Matt Caswell, Nicola Tuveri*
1906 * The undocumented function X509_certificate_type() has been deprecated;
1910 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1914 * Removed RSA padding mode for SSLv23 (which was only used for
1915 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1916 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1921 * Deprecated the obsolete X9.31 RSA key generation related functions.
1923 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1924 is not allowed to return a value > 1, this is no more taken as failure.
1926 *Viktor Dukhovni and David von Oheimb*
1928 * Deprecated the obsolete X9.31 RSA key generation related functions
1929 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1930 BN_X931_generate_prime_ex().
1934 * The default key generation method for the regular 2-prime RSA keys was
1935 changed to the FIPS 186-4 B.3.6 method.
1939 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1943 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1947 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1948 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1950 *Rich Salz, Richard Levitte, and David von Oheimb*
1952 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1956 * Deprecated `OCSP_parse_url()`.
1960 * Validation of SM2 keys has been separated from the validation of regular EC
1965 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1966 switches: a validation failure triggers an early exit, returning a failure
1967 exit status to the parent process.
1971 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1972 to ignore unknown ciphers.
1976 * The `-cipher-commands` and `-digest-commands` options
1977 of the command line utility `list` have been deprecated.
1978 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1982 * Added convenience functions for generating asymmetric key pairs:
1983 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1984 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1988 * All of the low-level EC_KEY functions have been deprecated.
1990 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1992 * Deprecated all the libcrypto and libssl error string loading
1997 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1998 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
2003 * The `-crypt` option to the `passwd` command line tool has been removed.
2007 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
2012 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
2016 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
2017 EVP_PKEY_get1_tls_encodedpoint().
2021 * The security callback, which can be customised by application code, supports
2022 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
2023 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
2027 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
2028 interface. Their functionality remains unchanged.
2032 * Added new option for 'openssl list', '-providers', which will display the
2033 list of loaded providers, their names, version and status. It optionally
2034 displays their gettable parameters.
2038 * Removed EVP_PKEY_set_alias_type().
2042 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
2043 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
2047 * Changed all "STACK" functions to be macros instead of inline functions. Macro
2048 parameters are still checked for type safety at compile time via helper
2053 * Remove the RAND_DRBG API
2055 *Paul Dale and Matthias St. Pierre*
2057 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
2058 as well as actual hostnames.
2062 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2063 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2064 conversely, silently ignore DTLS protocol version bounds when configuring
2065 TLS-based contexts. The commands can be repeated to set bounds of both
2066 types. The same applies with the corresponding "min_protocol" and
2067 "max_protocol" command-line switches, in case some application uses both TLS
2070 SSL_CTX instances that are created for a fixed protocol version (e.g.
2071 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2072 attempts to apply bounds to these protocol versions would result in an
2073 error. Now only the "version-flexible" SSL_CTX instances are subject to
2074 limits in configuration files in command-line options.
2078 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2083 * Reworked the recorded ERR codes to make better space for system errors.
2084 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2085 given code is a system error (true) or an OpenSSL error (false).
2089 * Reworked the test perl framework to better allow parallel testing.
2091 *Nicola Tuveri and David von Oheimb*
2093 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2094 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2098 * 'Configure' has been changed to figure out the configuration target if
2099 none is given on the command line. Consequently, the 'config' script is
2100 now only a mere wrapper. All documentation is changed to only mention
2103 *Rich Salz and Richard Levitte*
2105 * Added a library context `OSSL_LIB_CTX` that applications as well as
2106 other libraries can use to form a separate context within which
2107 libcrypto operations are performed.
2111 * Added various `_ex` functions to the OpenSSL API that support using
2112 a non-default `OSSL_LIB_CTX`.
2116 * Handshake now fails if Extended Master Secret extension is dropped
2121 * Dropped interactive mode from the `openssl` program.
2125 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2127 *David von Oheimb and Shane Lontis*
2129 * Deprecated `EC_METHOD_get_field_type()`.
2133 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2134 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2135 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2139 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2143 * Add CAdES-BES signature verification support, mostly derived
2144 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2146 *Filipe Raimundo da Silva*
2148 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2152 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2153 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2157 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2161 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2162 EC_KEY_precompute_mult().
2166 * Deprecated EC_POINTs_mul().
2170 * Removed FIPS_mode() and FIPS_mode_set().
2174 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2178 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2179 EC_POINT_get_Jprojective_coordinates_GFp().
2183 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2184 arrays to be more easily constructed via a series of utility functions.
2185 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2186 the various push functions and finally convert to a passable OSSL_PARAM
2187 array using OSSL_PARAM_BLD_to_param().
2191 * The security strength of SHA1 and MD5 based signatures in TLS has been
2196 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2197 contain a provider side internal key.
2201 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2205 * Project text documents not yet having a proper file name extension
2206 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2207 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2211 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2212 have been converted to Markdown with the goal to produce documents
2213 which not only look pretty when viewed online in the browser, but
2214 remain well readable inside a plain text editor.
2216 To achieve this goal, a 'minimalistic' Markdown style has been applied
2217 which avoids formatting elements that interfere too much with the
2218 reading flow in the text file. For example, it
2220 * avoids [ATX headings][] and uses [setext headings][] instead
2221 (which works for `<h1>` and `<h2>` headings only).
2222 * avoids [inline links][] and uses [reference links][] instead.
2223 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2225 [ATX headings]: https://github.github.com/gfm/#atx-headings
2226 [setext headings]: https://github.github.com/gfm/#setext-headings
2227 [inline links]: https://github.github.com/gfm/#inline-link
2228 [reference links]: https://github.github.com/gfm/#reference-link
2229 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2230 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2232 *Matthias St. Pierre*
2234 * The test suite is changed to preserve results of each test recipe.
2235 A new directory test-runs/ with subdirectories named like the
2236 test recipes are created in the build tree for this purpose.
2240 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2241 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2242 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2244 *David von Oheimb, Martin Peylo*
2246 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2247 It supports arbitrary request and response content types, GET redirection,
2248 TLS, connections via HTTP(S) proxies, connections and exchange via
2249 user-defined BIOs (allowing implicit connections), persistent connections,
2250 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2251 The legacy OCSP-focused (and only partly documented) API
2252 is retained for backward compatibility, while most of it is deprecated.
2256 * Added `util/check-format.pl`, a tool for checking adherence to the
2257 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2258 The checks performed are incomplete and yield some false positives.
2259 Still the tool should be useful for detecting most typical glitches.
2263 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2264 If domain name resolution yields multiple IP addresses all of them are tried
2265 after `connect()` failures.
2269 * All of the low-level RSA functions have been deprecated.
2273 * X509 certificates signed using SHA1 are no longer allowed at security
2278 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2279 modified to use PKEY APIs. These commands are now in maintenance mode
2280 and no new features will be added to them.
2284 * The command line utility rsautl has been deprecated.
2288 * The command line utilities genrsa and rsa have been modified to use PKEY
2289 APIs. They now write PKCS#8 keys by default. These commands are now in
2290 maintenance mode and no new features will be added to them.
2294 * All of the low-level DH functions have been deprecated.
2296 *Paul Dale and Matt Caswell*
2298 * All of the low-level DSA functions have been deprecated.
2302 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2303 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2307 * Deprecated low-level ECDH and ECDSA functions.
2311 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2315 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2316 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2317 a new formulation to include all the things it can be used for,
2318 as well as words of caution.
2322 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2326 * All of the low-level HMAC functions have been deprecated.
2328 *Paul Dale and David von Oheimb*
2330 * Over two thousand fixes were made to the documentation, including:
2331 - Common options (such as -rand/-writerand, TLS version control, etc)
2332 were refactored and point to newly-enhanced descriptions in openssl.pod.
2333 - Added style conformance for all options (with help from Richard Levitte),
2334 documented all reported missing options, added a CI build to check
2335 that all options are documented and that no unimplemented options
2337 - Documented some internals, such as all use of environment variables.
2338 - Addressed all internal broken L<> references.
2342 * All of the low-level CMAC functions have been deprecated.
2346 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2347 functions have been deprecated.
2349 *Paul Dale and David von Oheimb*
2351 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2352 set of functions. The documentation mentioned negative values for some
2353 errors, but this was never the case, so the mention of negative values
2356 Code that followed the documentation and thereby check with something
2357 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2361 * All of the low-level cipher functions have been deprecated.
2363 *Matt Caswell and Paul Dale*
2365 * Removed include/openssl/opensslconf.h.in and replaced it with
2366 include/openssl/configuration.h.in, which differs in not including
2367 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2368 was added to include both.
2370 This allows internal hacks where one might need to modify the set
2371 of configured macros, for example this if deprecated symbols are
2372 still supposed to be available internally:
2374 #include <openssl/configuration.h>
2376 #undef OPENSSL_NO_DEPRECATED
2377 #define OPENSSL_SUPPRESS_DEPRECATED
2379 #include <openssl/macros.h>
2381 This should not be used by applications that use the exported
2382 symbols, as that will lead to linking errors.
2386 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2387 used in exponentiation with 512-bit moduli. No EC algorithms are
2388 affected. Analysis suggests that attacks against 2-prime RSA1024,
2389 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2390 difficult to perform and are not believed likely. Attacks against DH512
2391 are considered just feasible. However, for an attack the target would
2392 have to reuse the DH512 private key, which is not recommended anyway.
2393 Also applications directly using the low-level API BN_mod_exp may be
2394 affected if they use BN_FLG_CONSTTIME.
2399 * Most memory-debug features have been deprecated, and the functionality
2400 replaced with no-ops.
2404 * Added documentation for the STACK API.
2408 * Introduced a new method type and API, OSSL_ENCODER, to represent
2409 generic encoders. These do the same sort of job that PEM writers
2410 and d2i functions do, but with support for methods supplied by
2411 providers, and the possibility for providers to support other
2416 * Introduced a new method type and API, OSSL_DECODER, to represent
2417 generic decoders. These do the same sort of job that PEM readers
2418 and i2d functions do, but with support for methods supplied by
2419 providers, and the possibility for providers to support other
2424 * Added a .pragma directive to the syntax of configuration files, to
2425 allow varying behavior in a supported and predictable manner.
2426 Currently added pragma:
2430 This allows dollar signs to be a keyword character unless it's
2431 followed by a opening brace or parenthesis. This is useful for
2432 platforms where dollar signs are commonly used in names, such as
2433 volume names and system directory names on VMS.
2437 * Added functionality to create an EVP_PKEY from user data.
2441 * Change the interpretation of the '--api' configuration option to
2442 mean that this is a desired API compatibility level with no
2443 further meaning. The previous interpretation, that this would
2444 also mean to remove all deprecated symbols up to and including
2445 the given version, no requires that 'no-deprecated' is also used
2446 in the configuration.
2448 When building applications, the desired API compatibility level
2449 can be set with the OPENSSL_API_COMPAT macro like before. For
2450 API compatibility version below 3.0, the old style numerical
2451 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2452 For version 3.0 and on, the value is expected to be the decimal
2453 value calculated from the major and minor version like this:
2455 MAJOR * 10000 + MINOR * 100
2459 -DOPENSSL_API_COMPAT=30000 For 3.0
2460 -DOPENSSL_API_COMPAT=30200 For 3.2
2462 To hide declarations that are deprecated up to and including the
2463 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2464 given when building the application as well.
2468 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2469 access to certificate and CRL stores via URIs and OSSL_STORE
2472 This adds the following functions:
2474 - X509_LOOKUP_store()
2475 - X509_STORE_load_file()
2476 - X509_STORE_load_path()
2477 - X509_STORE_load_store()
2478 - SSL_add_store_cert_subjects_to_stack()
2479 - SSL_CTX_set_default_verify_store()
2480 - SSL_CTX_load_verify_file()
2481 - SSL_CTX_load_verify_dir()
2482 - SSL_CTX_load_verify_store()
2486 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2487 The presence of this system service is determined at run-time.
2491 * Added functionality to create an EVP_PKEY context based on data
2492 for methods from providers. This takes an algorithm name and a
2493 property query string and simply stores them, with the intent
2494 that any operation that uses this context will use those strings
2495 to fetch the needed methods implicitly, thereby making the port
2496 of application written for pre-3.0 OpenSSL easier.
2500 * The undocumented function NCONF_WIN32() has been deprecated; for
2501 conversion details see the HISTORY section of doc/man5/config.pod
2505 * Introduced the new functions EVP_DigestSignInit_ex() and
2506 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2507 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2508 pages for further details.
2512 * Over two thousand fixes were made to the documentation, including:
2513 adding missing command flags, better style conformance, documentation
2516 *Rich Salz, Richard Levitte*
2518 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2519 X25519, X448, Ed25519 and Ed448.
2523 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2528 * Deprecated the public definition of `ERR_STATE` as well as the function
2529 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2534 * Added ERR functionality to give callers access to the stored function
2535 names that have replaced the older function code based functions.
2537 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2538 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2539 ERR_peek_error_all() and ERR_peek_last_error_all().
2541 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2542 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2543 ERR_func_error_string().
2547 * Extended testing to be verbose for failing tests only. The make variables
2548 VERBOSE_FAILURE or VF can be used to enable this:
2550 $ make VF=1 test # Unix
2551 $ mms /macro=(VF=1) test ! OpenVMS
2552 $ nmake VF=1 test # Windows
2556 * Added the `-copy_extensions` option to the `x509` command for use with
2557 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2558 all extensions in the request are copied to the certificate or vice versa.
2560 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2562 * Added the `-copy_extensions` option to the `req` command for use with
2563 `-x509`. When given with the `copy` or `copyall` argument,
2564 all extensions in the certification request are copied to the certificate.
2568 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2569 they generate are by default RFC 5280 compliant in the following sense:
2570 There is a subjectKeyIdentifier extension with a hash value of the public key
2571 and for not self-signed certs there is an authorityKeyIdentifier extension
2572 with a keyIdentifier field or issuer information identifying the signing key.
2573 This is done unless some configuration overrides the new default behavior,
2574 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2578 * Added several checks to `X509_verify_cert()` according to requirements in
2579 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2580 (which may be done by using the CLI option `-x509_strict`):
2581 * The basicConstraints of CA certificates must be marked critical.
2582 * CA certificates must explicitly include the keyUsage extension.
2583 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2584 * The issuer name of any certificate must not be empty.
2585 * The subject name of CA certs, certs with keyUsage crlSign,
2586 and certs without subjectAlternativeName must not be empty.
2587 * If a subjectAlternativeName extension is given it must not be empty.
2588 * The signatureAlgorithm field and the cert signature must be consistent.
2589 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2590 must not be marked critical.
2591 * The authorityKeyIdentifier must be given for X.509v3 certs
2592 unless they are self-signed.
2593 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2597 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2598 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2602 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2603 used even when parsing explicit parameters, when loading a encoded key
2604 or calling `EC_GROUP_new_from_ecpkparameters()`/
2605 `EC_GROUP_new_from_ecparameters()`.
2606 This prevents bypass of security hardening and performance gains,
2607 especially for curves with specialized EC_METHODs.
2608 By default, if a key encoded with explicit parameters is loaded and later
2609 encoded, the output is still encoded with explicit parameters, even if
2610 internally a "named" EC_GROUP is used for computation.
2614 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2615 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2616 NULL. After this change, only the cofactor parameter can be NULL. It also
2617 does some minimal sanity checks on the passed order.
2622 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2623 An attack is simple, if the first CMS_recipientInfo is valid but the
2624 second CMS_recipientInfo is chosen ciphertext. If the second
2625 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2626 encryption key will be replaced by garbage, and the message cannot be
2627 decoded, but if the RSA decryption fails, the correct encryption key is
2628 used and the recipient will not notice the attack.
2629 As a work around for this potential attack the length of the decrypted
2630 key must be equal to the cipher default key length, in case the
2631 certificate is not given and all recipientInfo are tried out.
2632 The old behaviour can be re-enabled in the CMS code by setting the
2633 CMS_DEBUG_DECRYPT flag.
2637 * Early start up entropy quality from the DEVRANDOM seed source has been
2638 improved for older Linux systems. The RAND subsystem will wait for
2639 /dev/random to be producing output before seeding from /dev/urandom.
2640 The seeded state is stored for future library initialisations using
2641 a system global shared memory segment. The shared memory identifier
2642 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2643 the desired value. The default identifier is 114.
2647 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2648 when primes for RSA keys are computed.
2649 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2650 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2651 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2652 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2653 This avoids possible fingerprinting of newly generated RSA modules.
2657 * Correct the extended master secret constant on EBCDIC systems. Without this
2658 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2659 negotiate EMS will fail. Unfortunately this also means that TLS connections
2660 between EBCDIC systems with this fix, and EBCDIC systems without this
2661 fix will fail if they negotiate EMS.
2665 * Changed the library initialisation so that the config file is now loaded
2666 by default. This was already the case for libssl. It now occurs for both
2667 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2668 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2672 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2673 where the former acts as a replacement for `ERR_put_error()`, and the
2674 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2675 `ERR_raise_data()` adds more flexibility by taking a format string and
2676 an arbitrary number of arguments following it, to be processed with
2681 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2682 to check if a named provider is loaded and available. When called, it
2683 will also activate all fallback providers if such are still present.
2687 * Enforce a minimum DH modulus size of 512 bits.
2691 * Changed DH parameters to generate the order q subgroup instead of 2q.
2692 Previously generated DH parameters are still accepted by DH_check
2693 but DH_generate_key works around that by clearing bit 0 of the
2694 private key for those. This avoids leaking bit 0 of the private key.
2698 * Significantly reduce secure memory usage by the randomness pools.
2702 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2707 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2708 algorithms. An implementation of a key exchange algorithm can be obtained
2709 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2710 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2711 the older EVP_PKEY_derive_init() function. See the man pages for the new
2712 functions for further details.
2716 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2720 * Removed the function names from error messages and deprecated the
2725 * Removed NextStep support and the macro OPENSSL_UNISTD
2729 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2730 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2731 Also removed "export var as function" capability; we do not export
2732 variables, only functions.
2736 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2737 an error and 1 indicating success. In previous versions of OpenSSL this
2738 was a void type. If a key was set longer than the maximum possible this
2743 * Support SM2 signing and verification schemes with X509 certificate.
2747 * Use SHA256 as the default digest for TS query in the `ts` app.
2751 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2755 * Default cipher lists/suites are now available via a function, the
2756 #defines are deprecated.
2760 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2761 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2762 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2766 * Join the directories crypto/x509 and crypto/x509v3
2770 * Added command 'openssl kdf' that uses the EVP_KDF API.
2774 * Added command 'openssl mac' that uses the EVP_MAC API.
2778 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2779 as default directories. Also added the command 'openssl info'
2780 for scripting purposes.
2784 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2789 * Add prediction resistance to the DRBG reseeding process.
2793 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2794 mandated by IEEE Std 1619-2018.
2798 * Added newline escaping functionality to a filename when using openssl dgst.
2799 This output format is to replicate the output format found in the `*sum`
2800 checksum programs. This aims to preserve backward compatibility.
2802 *Matt Eaton, Richard Levitte, and Paul Dale*
2804 * Removed the heartbeat message in DTLS feature, as it has very
2805 little usage and doesn't seem to fulfill a valuable purpose.
2806 The configuration option is now deprecated.
2810 * Changed the output of 'openssl {digestname} < file' to display the
2811 digest name in its output.
2815 * Added a new generic trace API which provides support for enabling
2816 instrumentation through trace output.
2818 *Richard Levitte & Matthias St. Pierre*
2820 * Added build tests for C++. These are generated files that only do one
2821 thing, to include one public OpenSSL head file each. This tests that
2822 the public header files can be usefully included in a C++ application.
2824 This test isn't enabled by default. It can be enabled with the option
2825 'enable-buildtest-c++'.
2829 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2833 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2837 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2841 * Added KMAC to EVP_MAC.
2845 * Added property based algorithm implementation selection framework to
2850 * Added SCA hardening for modular field inversion in EC_GROUP through
2851 a new dedicated field_inv() pointer in EC_METHOD.
2852 This also addresses a leakage affecting conversions from projective
2853 to affine coordinates.
2855 *Billy Bob Brumley, Nicola Tuveri*
2857 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2858 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2859 those algorithms that were already supported through the EVP_PKEY API
2860 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2861 and scrypt are now wrappers that call EVP_KDF.
2865 * Build devcrypto engine as a dynamic engine.
2867 *Eneas U de Queiroz*
2869 * Add keyed BLAKE2 to EVP_MAC.
2873 * Fix a bug in the computation of the endpoint-pair shared secret used
2874 by DTLS over SCTP. This breaks interoperability with older versions
2875 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2876 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2877 interoperability with such broken implementations. However, enabling
2878 this switch breaks interoperability with correct implementations.
2880 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2881 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2885 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2889 * Changed the license to the Apache License v2.0.
2893 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2895 - Major releases (indicated by incrementing the MAJOR release number)
2896 may introduce incompatible API/ABI changes.
2897 - Minor releases (indicated by incrementing the MINOR release number)
2898 may introduce new features but retain API/ABI compatibility.
2899 - Patch releases (indicated by incrementing the PATCH number)
2900 are intended for bug fixes and other improvements of existing
2901 features only (like improving performance or adding documentation)
2902 and retain API/ABI compatibility.
2906 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2910 * Remove the 'dist' target and add a tarball building script. The
2911 'dist' target has fallen out of use, and it shouldn't be
2912 necessary to configure just to create a source distribution.
2916 * Recreate the OS390-Unix config target. It no longer relies on a
2917 special script like it did for OpenSSL pre-1.1.0.
2921 * Instead of having the source directories listed in Configure, add
2922 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2927 * Add GMAC to EVP_MAC.
2931 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2935 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2936 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2937 to facilitate the continued use of MACs through raw private keys in
2938 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2942 * Deprecate ECDH_KDF_X9_62().
2946 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2947 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2948 are retained for backwards compatibility.
2952 * AES-XTS mode now enforces that its two keys are different to mitigate
2953 the attacked described in "Efficient Instantiations of Tweakable
2954 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2955 Details of this attack can be obtained from:
2956 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2960 * Rename the object files, i.e. give them other names than in previous
2961 versions. Their names now include the name of the final product, as
2962 well as its type mnemonic (bin, lib, shlib).
2966 * Added new option for 'openssl list', '-objects', which will display the
2967 list of built in objects, i.e. OIDs with names.
2971 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2972 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2977 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2978 improves application performance by removing data copies and providing
2979 applications with zero-copy system calls such as sendfile and splice.
2983 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2987 * `PKCS12_parse` now maintains the order of the parsed certificates
2988 when outputting them via `*ca` (rather than reversing it).
2992 * Deprecated pthread fork support methods.
2996 * Added support for FFDHE key exchange in TLS 1.3.
3000 * Added a new concept for OpenSSL plugability: providers. This
3001 functionality is designed to replace the ENGINE API and ENGINE
3002 implementations, and to be much more dynamic, allowing provider
3003 authors to introduce new algorithms among other things, as long as
3004 there's an API that supports the algorithm type.
3006 With this concept comes a new core API for interaction between
3007 libcrypto and provider implementations. Public libcrypto functions
3008 that want to use providers do so through this core API.
3010 The main documentation for this core API is found in
3011 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
3012 refer to other manuals describing the API specific for supported
3013 algorithm types (also called operations).
3020 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
3022 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
3024 * Avoid loading of a dynamic engine twice.
3028 * Prioritise DANE TLSA issuer certs over peer certs
3032 * Fixed random API for MacOS prior to 10.12
3034 These MacOS versions don't support the CommonCrypto APIs
3038 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
3040 * Fixed an SM2 Decryption Buffer Overflow.
3042 In order to decrypt SM2 encrypted data an application is expected to
3043 call the API function EVP_PKEY_decrypt(). Typically an application will
3044 call this function twice. The first time, on entry, the "out" parameter
3045 can be NULL and, on exit, the "outlen" parameter is populated with the
3046 buffer size required to hold the decrypted plaintext. The application
3047 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
3048 again, but this time passing a non-NULL value for the "out" parameter.
3050 A bug in the implementation of the SM2 decryption code means that the
3051 calculation of the buffer size required to hold the plaintext returned
3052 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
3053 size required by the second call. This can lead to a buffer overflow
3054 when EVP_PKEY_decrypt() is called by the application a second time with
3055 a buffer that is too small.
3057 A malicious attacker who is able present SM2 content for decryption to
3058 an application could cause attacker chosen data to overflow the buffer
3059 by up to a maximum of 62 bytes altering the contents of other data held
3060 after the buffer, possibly changing application behaviour or causing
3061 the application to crash. The location of the buffer is application
3062 dependent but is typically heap allocated.
3067 * Fixed various read buffer overruns processing ASN.1 strings
3069 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3070 structure which contains a buffer holding the string data and a field
3071 holding the buffer length. This contrasts with normal C strings which
3072 are represented as a buffer for the string data which is terminated
3073 with a NUL (0) byte.
3075 Although not a strict requirement, ASN.1 strings that are parsed using
3076 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3077 well as any string whose value has been set with the ASN1_STRING_set()
3078 function will additionally NUL terminate the byte array in the
3079 ASN1_STRING structure.
3081 However, it is possible for applications to directly construct valid
3082 ASN1_STRING structures which do not NUL terminate the byte array by
3083 directly setting the "data" and "length" fields in the ASN1_STRING
3084 array. This can also happen by using the ASN1_STRING_set0() function.
3086 Numerous OpenSSL functions that print ASN.1 data have been found to
3087 assume that the ASN1_STRING byte array will be NUL terminated, even
3088 though this is not guaranteed for strings that have been directly
3089 constructed. Where an application requests an ASN.1 structure to be
3090 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3091 been directly constructed by the application without NUL terminating
3092 the "data" field, then a read buffer overrun can occur.
3094 The same thing can also occur during name constraints processing
3095 of certificates (for example if a certificate has been directly
3096 constructed by the application instead of loading it via the OpenSSL
3097 parsing functions, and the certificate contains non NUL terminated
3098 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3099 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3101 If a malicious actor can cause an application to directly construct an
3102 ASN1_STRING and then process it through one of the affected OpenSSL
3103 functions then this issue could be hit. This might result in a crash
3104 (causing a Denial of Service attack). It could also result in the
3105 disclosure of private memory contents (such as private keys, or
3106 sensitive plaintext).
3111 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3113 * Fixed a problem with verifying a certificate chain when using the
3114 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3115 the certificates present in a certificate chain. It is not set by default.
3117 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3118 the chain that have explicitly encoded elliptic curve parameters was added
3119 as an additional strict check.
3121 An error in the implementation of this check meant that the result of a
3122 previous check to confirm that certificates in the chain are valid CA
3123 certificates was overwritten. This effectively bypasses the check
3124 that non-CA certificates must not be able to issue other certificates.
3126 If a "purpose" has been configured then there is a subsequent opportunity
3127 for checks that the certificate is a valid CA. All of the named "purpose"
3128 values implemented in libcrypto perform this check. Therefore, where
3129 a purpose is set the certificate chain will still be rejected even when the
3130 strict flag has been used. A purpose is set by default in libssl client and
3131 server certificate verification routines, but it can be overridden or
3132 removed by an application.
3134 In order to be affected, an application must explicitly set the
3135 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3136 for the certificate verification or, in the case of TLS client or server
3137 applications, override the default purpose.
3142 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3143 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3144 renegotiation ClientHello omits the signature_algorithms extension (where it
3145 was present in the initial ClientHello), but includes a
3146 signature_algorithms_cert extension then a NULL pointer dereference will
3147 result, leading to a crash and a denial of service attack.
3149 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3150 (which is the default configuration). OpenSSL TLS clients are not impacted by
3154 *Peter Kästle and Samuel Sapalski*
3156 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3158 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3159 create a unique hash value based on the issuer and serial number data
3160 contained within an X509 certificate. However, it was failing to correctly
3161 handle any errors that may occur while parsing the issuer field (which might
3162 occur if the issuer field is maliciously constructed). This may subsequently
3163 result in a NULL pointer deref and a crash leading to a potential denial of
3169 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3170 padding mode to correctly check for rollback attacks. This is considered a
3171 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3176 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3177 functions. Previously they could overflow the output length argument in some
3178 cases where the input length is close to the maximum permissible length for
3179 an integer on the platform. In such cases the return value from the function
3180 call would be 1 (indicating success), but the output length value would be
3181 negative. This could cause applications to behave incorrectly or crash.
3186 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3187 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3188 could be exploited in a side channel attack to recover the password. Since
3189 the attack is local host only this is outside of the current OpenSSL
3190 threat model and therefore no CVE is assigned.
3192 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3197 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3199 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3200 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3201 If an attacker can control both items being compared then this could lead
3202 to a possible denial of service attack. OpenSSL itself uses the
3203 GENERAL_NAME_cmp function for two purposes:
3204 1) Comparing CRL distribution point names between an available CRL and a
3205 CRL distribution point embedded in an X509 certificate
3206 2) When verifying that a timestamp response token signer matches the
3207 timestamp authority name (exposed via the API functions
3208 TS_RESP_verify_response and TS_RESP_verify_token)
3213 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3215 * Certificates with explicit curve parameters are now disallowed in
3216 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3220 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3221 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3222 conversely, silently ignore DTLS protocol version bounds when configuring
3223 TLS-based contexts. The commands can be repeated to set bounds of both
3224 types. The same applies with the corresponding "min_protocol" and
3225 "max_protocol" command-line switches, in case some application uses both TLS
3228 SSL_CTX instances that are created for a fixed protocol version (e.g.
3229 TLSv1_server_method()) also silently ignore version bounds. Previously
3230 attempts to apply bounds to these protocol versions would result in an
3231 error. Now only the "version-flexible" SSL_CTX instances are subject to
3232 limits in configuration files in command-line options.
3236 * Handshake now fails if Extended Master Secret extension is dropped
3241 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3243 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3245 * Fixed segmentation fault in SSL_check_chain()
3246 Server or client applications that call the SSL_check_chain() function
3247 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3248 dereference as a result of incorrect handling of the
3249 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3250 or unrecognised signature algorithm is received from the peer. This could
3251 be exploited by a malicious peer in a Denial of Service attack.
3256 * Added AES consttime code for no-asm configurations
3257 an optional constant time support for AES was added
3258 when building openssl for no-asm.
3259 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3260 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3261 At this time this feature is by default disabled.
3262 It will be enabled by default in 3.0.
3266 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3268 * Revert the change of EOF detection while reading in libssl to avoid
3269 regressions in applications depending on the current way of reporting
3270 the EOF. As the existing method is not fully accurate the change to
3271 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3272 branch and will be present in the 3.0 release.
3276 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3277 when primes for RSA keys are computed.
3278 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3279 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3280 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3281 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3282 This avoids possible fingerprinting of newly generated RSA modules.
3286 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3288 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3289 while reading in libssl then we would report an error back to the
3290 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3291 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3292 therefore give a hint as to what went wrong.
3296 * Check that ed25519 and ed448 are allowed by the security level. Previously
3297 signature algorithms not using an MD were not being checked that they were
3298 allowed by the security level.
3302 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3303 was not quite right. The behaviour was not consistent between resumption
3304 and normal handshakes, and also not quite consistent with historical
3305 behaviour. The behaviour in various scenarios has been clarified and
3306 it has been updated to make it match historical behaviour as closely as
3311 * *[VMS only]* The header files that the VMS compilers include automatically,
3312 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3313 that the C++ compiler doesn't understand. This is a shortcoming in the
3314 compiler, but can be worked around with `__cplusplus` guards.
3316 C++ applications that use OpenSSL libraries must be compiled using the
3317 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3318 functions. Otherwise, only functions with symbols of less than 31
3319 characters can be used, as the linker will not be able to successfully
3320 resolve symbols with longer names.
3324 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3325 The presence of this system service is determined at run-time.
3329 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3334 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3336 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3337 number generator (RNG). This was intended to include protection in the
3338 event of a fork() system call in order to ensure that the parent and child
3339 processes did not share the same RNG state. However, this protection was not
3340 being used in the default case.
3342 A partial mitigation for this issue is that the output from a high
3343 precision timer is mixed into the RNG state so the likelihood of a parent
3344 and child process sharing state is significantly reduced.
3346 If an application already calls OPENSSL_init_crypto() explicitly using
3347 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3350 *Matthias St. Pierre*
3352 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3353 used even when parsing explicit parameters, when loading a encoded key
3354 or calling `EC_GROUP_new_from_ecpkparameters()`/
3355 `EC_GROUP_new_from_ecparameters()`.
3356 This prevents bypass of security hardening and performance gains,
3357 especially for curves with specialized EC_METHODs.
3358 By default, if a key encoded with explicit parameters is loaded and later
3359 encoded, the output is still encoded with explicit parameters, even if
3360 internally a "named" EC_GROUP is used for computation.
3364 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3365 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3366 NULL. After this change, only the cofactor parameter can be NULL. It also
3367 does some minimal sanity checks on the passed order.
3372 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3373 An attack is simple, if the first CMS_recipientInfo is valid but the
3374 second CMS_recipientInfo is chosen ciphertext. If the second
3375 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3376 encryption key will be replaced by garbage, and the message cannot be
3377 decoded, but if the RSA decryption fails, the correct encryption key is
3378 used and the recipient will not notice the attack.
3379 As a work around for this potential attack the length of the decrypted
3380 key must be equal to the cipher default key length, in case the
3381 certificate is not given and all recipientInfo are tried out.
3382 The old behaviour can be re-enabled in the CMS code by setting the
3383 CMS_DEBUG_DECRYPT flag.
3388 * Early start up entropy quality from the DEVRANDOM seed source has been
3389 improved for older Linux systems. The RAND subsystem will wait for
3390 /dev/random to be producing output before seeding from /dev/urandom.
3391 The seeded state is stored for future library initialisations using
3392 a system global shared memory segment. The shared memory identifier
3393 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3394 the desired value. The default identifier is 114.
3398 * Correct the extended master secret constant on EBCDIC systems. Without this
3399 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3400 negotiate EMS will fail. Unfortunately this also means that TLS connections
3401 between EBCDIC systems with this fix, and EBCDIC systems without this
3402 fix will fail if they negotiate EMS.
3406 * Use Windows installation paths in the mingw builds
3408 Mingw isn't a POSIX environment per se, which means that Windows
3409 paths should be used for installation.
3414 * Changed DH_check to accept parameters with order q and 2q subgroups.
3415 With order 2q subgroups the bit 0 of the private key is not secret
3416 but DH_generate_key works around that by clearing bit 0 of the
3417 private key for those. This avoids leaking bit 0 of the private key.
3421 * Significantly reduce secure memory usage by the randomness pools.
3425 * Revert the DEVRANDOM_WAIT feature for Linux systems
3427 The DEVRANDOM_WAIT feature added a select() call to wait for the
3428 /dev/random device to become readable before reading from the
3429 /dev/urandom device.
3431 It turned out that this change had negative side effects on
3432 performance which were not acceptable. After some discussion it
3433 was decided to revert this feature and leave it up to the OS
3434 resp. the platform maintainer to ensure a proper initialization
3435 during early boot time.
3437 *Matthias St. Pierre*
3439 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3441 * Add build tests for C++. These are generated files that only do one
3442 thing, to include one public OpenSSL head file each. This tests that
3443 the public header files can be usefully included in a C++ application.
3445 This test isn't enabled by default. It can be enabled with the option
3446 'enable-buildtest-c++'.
3450 * Enable SHA3 pre-hashing for ECDSA and DSA.
3454 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3455 This changes the size when using the `genpkey` command when no size is given.
3456 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3457 generation commands to use 2048 bits by default.
3461 * Reorganize the manual pages to consistently have RETURN VALUES,
3462 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3463 util/fix-doc-nits accordingly.
3465 *Paul Yang, Joshua Lock*
3467 * Add the missing accessor EVP_PKEY_get0_engine()
3471 * Have commands like `s_client` and `s_server` output the signature scheme
3472 along with other cipher suite parameters when debugging.
3476 * Make OPENSSL_config() error agnostic again.
3480 * Do the error handling in RSA decryption constant time.
3484 * Prevent over long nonces in ChaCha20-Poly1305.
3486 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3487 for every encryption operation. RFC 7539 specifies that the nonce value
3488 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3489 and front pads the nonce with 0 bytes if it is less than 12
3490 bytes. However it also incorrectly allows a nonce to be set of up to 16
3491 bytes. In this case only the last 12 bytes are significant and any
3492 additional leading bytes are ignored.
3494 It is a requirement of using this cipher that nonce values are
3495 unique. Messages encrypted using a reused nonce value are susceptible to
3496 serious confidentiality and integrity attacks. If an application changes
3497 the default nonce length to be longer than 12 bytes and then makes a
3498 change to the leading bytes of the nonce expecting the new value to be a
3499 new unique nonce then such an application could inadvertently encrypt
3500 messages with a reused nonce.
3502 Additionally the ignored bytes in a long nonce are not covered by the
3503 integrity guarantee of this cipher. Any application that relies on the
3504 integrity of these ignored leading bytes of a long nonce may be further
3505 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3506 is safe because no such use sets such a long nonce value. However user
3507 applications that use this cipher directly and set a non-default nonce
3508 length to be longer than 12 bytes may be vulnerable.
3510 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3516 * Add DEVRANDOM_WAIT feature for Linux systems
3518 On older Linux systems where the getrandom() system call is not available,
3519 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3520 Contrary to getrandom(), the /dev/urandom device will not block during
3521 early boot when the kernel CSPRNG has not been seeded yet.
3523 To mitigate this known weakness, use select() to wait for /dev/random to
3524 become readable before reading from /dev/urandom.
3526 * Ensure that SM2 only uses SM3 as digest algorithm
3530 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3532 * Change the info callback signals for the start and end of a post-handshake
3533 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3534 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3535 confused by this and assume that a TLSv1.2 renegotiation has started. This
3536 can break KeyUpdate handling. Instead we no longer signal the start and end
3537 of a post handshake message exchange (although the messages themselves are
3538 still signalled). This could break some applications that were expecting
3539 the old signals. However without this KeyUpdate is not usable for many
3544 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3546 * Timing vulnerability in DSA signature generation
3548 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3549 timing side channel attack. An attacker could use variations in the signing
3550 algorithm to recover the private key.
3552 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3557 * Timing vulnerability in ECDSA signature generation
3559 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3560 timing side channel attack. An attacker could use variations in the signing
3561 algorithm to recover the private key.
3563 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3568 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3569 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3570 of two gigabytes and the error handling improved.
3572 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3573 categorized as a normal bug, not a security issue, because the DRBG reseeds
3574 automatically and is fully functional even without additional randomness
3575 provided by the application.
3577 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3579 * Add a new ClientHello callback. Provides a callback interface that gives
3580 the application the ability to adjust the nascent SSL object at the
3581 earliest stage of ClientHello processing, immediately after extensions have
3582 been collected but before they have been processed. In particular, this
3583 callback can adjust the supported TLS versions in response to the contents
3588 * Add SM2 base algorithm support.
3592 * s390x assembly pack: add (improved) hardware-support for the following
3593 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3594 aes-cfb/cfb8, aes-ecb.
3598 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3599 parameter is no longer accepted, as it leads to a corrupt table. NULL
3600 pem_str is reserved for alias entries only.
3604 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3605 step for prime curves. The new implementation is based on formulae from
3606 differential addition-and-doubling in homogeneous projective coordinates
3607 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3608 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3609 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3610 to work in projective coordinates.
3612 *Billy Bob Brumley, Nicola Tuveri*
3614 * Change generating and checking of primes so that the error rate of not
3615 being prime depends on the intended use based on the size of the input.
3616 For larger primes this will result in more rounds of Miller-Rabin.
3617 The maximal error rate for primes with more than 1080 bits is lowered
3620 *Kurt Roeckx, Annie Yousar*
3622 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3626 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3627 moving between systems, and to avoid confusion when a Windows build is
3628 done with mingw vs with MSVC. For POSIX installs, there's still a
3629 symlink or copy named 'tsget' to avoid that confusion as well.
3633 * Revert blinding in ECDSA sign and instead make problematic addition
3634 length-invariant. Switch even to fixed-length Montgomery multiplication.
3638 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3639 step for binary curves. The new implementation is based on formulae from
3640 differential addition-and-doubling in mixed Lopez-Dahab projective
3641 coordinates, modified to independently blind the operands.
3643 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3645 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3646 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3647 EC_METHODs to implement their own specialized "ladder step", to take
3648 advantage of more favorable coordinate systems or more efficient
3649 differential addition-and-doubling algorithms.
3651 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3653 * Modified the random device based seed sources to keep the relevant
3654 file descriptors open rather than reopening them on each access.
3655 This allows such sources to operate in a chroot() jail without
3656 the associated device nodes being available. This behaviour can be
3657 controlled using RAND_keep_random_devices_open().
3661 * Numerous side-channel attack mitigations have been applied. This may have
3662 performance impacts for some algorithms for the benefit of improved
3663 security. Specific changes are noted in this change log by their respective
3668 * AIX shared library support overhaul. Switch to AIX "natural" way of
3669 handling shared libraries, which means collecting shared objects of
3670 different versions and bitnesses in one common archive. This allows to
3671 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3672 doesn't affect the way 3rd party applications are linked, only how
3673 multi-version installation is managed.
3677 * Make ec_group_do_inverse_ord() more robust and available to other
3678 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3679 mitigations are applied to the fallback BN_mod_inverse().
3680 When using this function rather than BN_mod_inverse() directly, new
3681 EC cryptosystem implementations are then safer-by-default.
3685 * Add coordinate blinding for EC_POINT and implement projective
3686 coordinate blinding for generic prime curves as a countermeasure to
3687 chosen point SCA attacks.
3689 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3691 * Add blinding to ECDSA and DSA signatures to protect against side channel
3692 attacks discovered by Keegan Ryan (NCC Group).
3696 * Enforce checking in the `pkeyutl` command to ensure that the input
3697 length does not exceed the maximum supported digest length when performing
3698 a sign, verify or verifyrecover operation.
3702 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3703 I/O in combination with something like select() or poll() will hang. This
3704 can be turned off again using SSL_CTX_clear_mode().
3705 Many applications do not properly handle non-application data records, and
3706 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3707 around the problems in those applications, but can also break some.
3708 It's recommended to read the manpages about SSL_read(), SSL_write(),
3709 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3710 SSL_CTX_set_read_ahead() again.
3714 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3715 now allow empty (zero character) pass phrases.
3719 * Apply blinding to binary field modular inversion and remove patent
3720 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3724 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3725 binary and prime elliptic curves.
3729 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3730 constant time fixed point multiplication.
3734 * Revise elliptic curve scalar multiplication with timing attack
3735 defenses: ec_wNAF_mul redirects to a constant time implementation
3736 when computing fixed point and variable point multiplication (which
3737 in OpenSSL are mostly used with secret scalars in keygen, sign,
3738 ECDH derive operations).
3739 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3742 * Updated CONTRIBUTING
3746 * Updated DRBG / RAND to request nonce and additional low entropy
3747 randomness from the system.
3749 *Matthias St. Pierre*
3751 * Updated 'openssl rehash' to use OpenSSL consistent default.
3755 * Moved the load of the ssl_conf module to libcrypto, which helps
3756 loading engines that libssl uses before libssl is initialised.
3760 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3764 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3766 *Ingo Schwarze, Rich Salz*
3768 * Added output of accepting IP address and port for 'openssl s_server'
3772 * Added a new API for TLSv1.3 ciphersuites:
3773 SSL_CTX_set_ciphersuites()
3774 SSL_set_ciphersuites()
3778 * Memory allocation failures consistently add an error to the error
3783 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3784 in libcrypto when run as setuid/setgid.
3788 * Load any config file by default when libssl is used.
3792 * Added new public header file <openssl/rand_drbg.h> and documentation
3793 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3795 *Matthias St. Pierre*
3797 * QNX support removed (cannot find contributors to get their approval
3798 for the license change).
3802 * TLSv1.3 replay protection for early data has been implemented. See the
3803 SSL_read_early_data() man page for further details.
3807 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3808 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3809 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3810 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3811 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3812 configuration has been separated out. See the ciphers man page or the
3813 SSL_CTX_set_ciphersuites() man page for more information.
3817 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3818 in responder mode now supports the new "-multi" option, which
3819 spawns the specified number of child processes to handle OCSP
3820 requests. The "-timeout" option now also limits the OCSP
3821 responder's patience to wait to receive the full client request
3822 on a newly accepted connection. Child processes are respawned
3823 as needed, and the CA index file is automatically reloaded
3824 when changed. This makes it possible to run the "ocsp" responder
3825 as a long-running service, making the OpenSSL CA somewhat more
3826 feature-complete. In this mode, most diagnostic messages logged
3827 after entering the event loop are logged via syslog(3) rather than
3832 * Added support for X448 and Ed448. Heavily based on original work by
3837 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3838 objects loaded. This adds the functions OSSL_STORE_expect() and
3839 OSSL_STORE_find() as well as needed tools to construct searches and
3840 get the search data out of them.
3844 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3845 version of OpenSSL should review their configuration settings to ensure
3846 that they are still appropriate for TLSv1.3. For further information see:
3847 <https://wiki.openssl.org/index.php/TLS1.3>
3851 * Grand redesign of the OpenSSL random generator
3853 The default RAND method now utilizes an AES-CTR DRBG according to
3854 NIST standard SP 800-90Ar1. The new random generator is essentially
3855 a port of the default random generator from the OpenSSL FIPS 2.0
3856 object module. It is a hybrid deterministic random bit generator
3857 using an AES-CTR bit stream and which seeds and reseeds itself
3858 automatically using trusted system entropy sources.
3860 Some of its new features are:
3861 - Support for multiple DRBG instances with seed chaining.
3862 - The default RAND method makes use of a DRBG.
3863 - There is a public and private DRBG instance.
3864 - The DRBG instances are fork-safe.
3865 - Keep all global DRBG instances on the secure heap if it is enabled.
3866 - The public and private DRBG instance are per thread for lock free
3869 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3871 * Changed Configure so it only says what it does and doesn't dump
3872 so much data. Instead, ./configdata.pm should be used as a script
3873 to display all sorts of configuration data.
3877 * Added processing of "make variables" to Configure.
3881 * Added SHA512/224 and SHA512/256 algorithm support.
3885 * The last traces of Netware support, first removed in 1.1.0, have
3890 * Get rid of Makefile.shared, and in the process, make the processing
3891 of certain files (rc.obj, or the .def/.map/.opt files produced from
3892 the ordinal files) more visible and hopefully easier to trace and
3893 debug (or make silent).
3897 * Make it possible to have environment variable assignments as
3898 arguments to config / Configure.
3902 * Add multi-prime RSA (RFC 8017) support.
3906 * Add SM3 implemented according to GB/T 32905-2016
3907 *Jack Lloyd <jack.lloyd@ribose.com>,*
3908 *Ronald Tse <ronald.tse@ribose.com>,*
3909 *Erick Borsboom <erick.borsboom@ribose.com>*
3911 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3912 as documented in RFC6066.
3913 Based on a patch from Tomasz Moń
3915 *Filipe Raimundo da Silva*
3917 * Add SM4 implemented according to GB/T 32907-2016.
3918 *Jack Lloyd <jack.lloyd@ribose.com>,*
3919 *Ronald Tse <ronald.tse@ribose.com>,*
3920 *Erick Borsboom <erick.borsboom@ribose.com>*
3922 * Reimplement -newreq-nodes and ERR_error_string_n; the
3923 original author does not agree with the license change.
3927 * Add ARIA AEAD TLS support.
3931 * Some macro definitions to support VS6 have been removed. Visual
3932 Studio 6 has not worked since 1.1.0
3936 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3937 without clearing the errors.
3941 * Add "atfork" functions. If building on a system that without
3942 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3943 requirements. The RAND facility now uses/requires this.
3951 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3952 not possible to disable entirely. However, it's still possible to
3953 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3956 To disable, configure with 'no-ui-console'. 'no-ui' is still
3957 possible to use as an alias. Check at compile time with the
3958 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3959 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3963 * Add a STORE module, which implements a uniform and URI based reader of
3964 stores that can contain keys, certificates, CRLs and numerous other
3965 objects. The main API is loosely based on a few stdio functions,
3966 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3967 OSSL_STORE_error and OSSL_STORE_close.
3968 The implementation uses backends called "loaders" to implement arbitrary
3969 URI schemes. There is one built in "loader" for the 'file' scheme.
3973 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3974 then adjusted to work on FreeBSD 8.4 as well.
3975 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3976 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3980 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3981 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3982 error code calls like this:
3984 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3986 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3987 that can be encoded in C. For the foreseeable future, this will only
3990 *Richard Levitte and Tim Hudson*
3992 * Removed BSD cryptodev engine.
3996 * Add a build target 'build_all_generated', to build all generated files
3997 and only that. This can be used to prepare everything that requires
3998 things like perl for a system that lacks perl and then move everything
3999 to that system and do the rest of the build there.
4003 * In the UI interface, make it possible to duplicate the user data. This
4004 can be used by engines that need to retain the data for a longer time
4005 than just the call where this user data is passed.
4009 * Ignore the '-named_curve auto' value for compatibility of applications
4012 *Tomáš Mráz <tmraz@fedoraproject.org>*
4014 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
4015 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
4016 alerts across multiple records (some of which could be empty). In practice
4017 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
4018 prohibits this altogether and other libraries (BoringSSL, NSS) do not
4019 support this at all. Supporting it adds significant complexity to the
4020 record layer, and its removal is unlikely to cause interoperability
4025 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
4026 with Z. These are meant to replace LONG and ZLONG and to be size safe.
4027 The use of LONG and ZLONG is discouraged and scheduled for deprecation
4032 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
4033 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
4035 *Richard Levitte, Andy Polyakov*
4037 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
4042 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4043 platform rather than 'mingw'.
4047 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
4048 success if they are asked to add an object which already exists
4049 in the store. This change cascades to other functions which load
4050 certificates and CRLs.
4054 * x86_64 assembly pack: annotate code with DWARF CFI directives to
4055 facilitate stack unwinding even from assembly subroutines.
4059 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
4060 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
4064 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4065 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4066 which is the minimum version we support.
4070 * Certificate time validation (X509_cmp_time) enforces stricter
4071 compliance with RFC 5280. Fractional seconds and timezone offsets
4072 are no longer allowed.
4076 * Add support for ARIA
4080 * s_client will now send the Server Name Indication (SNI) extension by
4081 default unless the new "-noservername" option is used. The server name is
4082 based on the host provided to the "-connect" option unless overridden by
4083 using "-servername".
4087 * Add support for SipHash
4091 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4092 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4093 prevent issues where no progress is being made and the peer continually
4094 sends unrecognised record types, using up resources processing them.
4098 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4099 using the algorithm defined in
4100 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4104 * Heartbeat support has been removed; the ABI is changed for now.
4106 *Richard Levitte, Rich Salz*
4108 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4112 * The RSA "null" method, which was partially supported to avoid patent
4113 issues, has been replaced to always returns NULL.
4120 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4122 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4123 used even when parsing explicit parameters, when loading a encoded key
4124 or calling `EC_GROUP_new_from_ecpkparameters()`/
4125 `EC_GROUP_new_from_ecparameters()`.
4126 This prevents bypass of security hardening and performance gains,
4127 especially for curves with specialized EC_METHODs.
4128 By default, if a key encoded with explicit parameters is loaded and later
4129 encoded, the output is still encoded with explicit parameters, even if
4130 internally a "named" EC_GROUP is used for computation.
4134 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4135 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4136 NULL. After this change, only the cofactor parameter can be NULL. It also
4137 does some minimal sanity checks on the passed order.
4142 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4143 An attack is simple, if the first CMS_recipientInfo is valid but the
4144 second CMS_recipientInfo is chosen ciphertext. If the second
4145 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4146 encryption key will be replaced by garbage, and the message cannot be
4147 decoded, but if the RSA decryption fails, the correct encryption key is
4148 used and the recipient will not notice the attack.
4149 As a work around for this potential attack the length of the decrypted
4150 key must be equal to the cipher default key length, in case the
4151 certificate is not given and all recipientInfo are tried out.
4152 The old behaviour can be re-enabled in the CMS code by setting the
4153 CMS_DEBUG_DECRYPT flag.
4158 * Use Windows installation paths in the mingw builds
4160 Mingw isn't a POSIX environment per se, which means that Windows
4161 paths should be used for installation.
4166 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4168 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4169 This changes the size when using the `genpkey` command when no size is given.
4170 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4171 generation commands to use 2048 bits by default.
4175 * Prevent over long nonces in ChaCha20-Poly1305.
4177 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4178 for every encryption operation. RFC 7539 specifies that the nonce value
4179 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4180 and front pads the nonce with 0 bytes if it is less than 12
4181 bytes. However it also incorrectly allows a nonce to be set of up to 16
4182 bytes. In this case only the last 12 bytes are significant and any
4183 additional leading bytes are ignored.
4185 It is a requirement of using this cipher that nonce values are
4186 unique. Messages encrypted using a reused nonce value are susceptible to
4187 serious confidentiality and integrity attacks. If an application changes
4188 the default nonce length to be longer than 12 bytes and then makes a
4189 change to the leading bytes of the nonce expecting the new value to be a
4190 new unique nonce then such an application could inadvertently encrypt
4191 messages with a reused nonce.
4193 Additionally the ignored bytes in a long nonce are not covered by the
4194 integrity guarantee of this cipher. Any application that relies on the
4195 integrity of these ignored leading bytes of a long nonce may be further
4196 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4197 is safe because no such use sets such a long nonce value. However user
4198 applications that use this cipher directly and set a non-default nonce
4199 length to be longer than 12 bytes may be vulnerable.
4201 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4207 * Added SCA hardening for modular field inversion in EC_GROUP through
4208 a new dedicated field_inv() pointer in EC_METHOD.
4209 This also addresses a leakage affecting conversions from projective
4210 to affine coordinates.
4212 *Billy Bob Brumley, Nicola Tuveri*
4214 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4215 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4219 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4223 * Remove the 'dist' target and add a tarball building script. The
4224 'dist' target has fallen out of use, and it shouldn't be
4225 necessary to configure just to create a source distribution.
4229 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4231 * Timing vulnerability in DSA signature generation
4233 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4234 timing side channel attack. An attacker could use variations in the signing
4235 algorithm to recover the private key.
4237 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4242 * Timing vulnerability in ECDSA signature generation
4244 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4245 timing side channel attack. An attacker could use variations in the signing
4246 algorithm to recover the private key.
4248 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4253 * Add coordinate blinding for EC_POINT and implement projective
4254 coordinate blinding for generic prime curves as a countermeasure to
4255 chosen point SCA attacks.
4257 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4259 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4261 * Client DoS due to large DH parameter
4263 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4264 malicious server can send a very large prime value to the client. This will
4265 cause the client to spend an unreasonably long period of time generating a
4266 key for this prime resulting in a hang until the client has finished. This
4267 could be exploited in a Denial Of Service attack.
4269 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4274 * Cache timing vulnerability in RSA Key Generation
4276 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4277 a cache timing side channel attack. An attacker with sufficient access to
4278 mount cache timing attacks during the RSA key generation process could
4279 recover the private key.
4281 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4282 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4287 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4288 parameter is no longer accepted, as it leads to a corrupt table. NULL
4289 pem_str is reserved for alias entries only.
4293 * Revert blinding in ECDSA sign and instead make problematic addition
4294 length-invariant. Switch even to fixed-length Montgomery multiplication.
4298 * Change generating and checking of primes so that the error rate of not
4299 being prime depends on the intended use based on the size of the input.
4300 For larger primes this will result in more rounds of Miller-Rabin.
4301 The maximal error rate for primes with more than 1080 bits is lowered
4304 *Kurt Roeckx, Annie Yousar*
4306 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4310 * Add blinding to ECDSA and DSA signatures to protect against side channel
4311 attacks discovered by Keegan Ryan (NCC Group).
4315 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4316 now allow empty (zero character) pass phrases.
4320 * Certificate time validation (X509_cmp_time) enforces stricter
4321 compliance with RFC 5280. Fractional seconds and timezone offsets
4322 are no longer allowed.
4326 * Fixed a text canonicalisation bug in CMS
4328 Where a CMS detached signature is used with text content the text goes
4329 through a canonicalisation process first prior to signing or verifying a
4330 signature. This process strips trailing space at the end of lines, converts
4331 line terminators to CRLF and removes additional trailing line terminators
4332 at the end of a file. A bug in the canonicalisation process meant that
4333 some characters, such as form-feed, were incorrectly treated as whitespace
4334 and removed. This is contrary to the specification (RFC5485). This fix
4335 could mean that detached text data signed with an earlier version of
4336 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4337 signed with a fixed OpenSSL may fail to verify with an earlier version of
4338 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4339 and use the "-binary" flag (for the "cms" command line application) or set
4340 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4344 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4346 * Constructed ASN.1 types with a recursive definition could exceed the stack
4348 Constructed ASN.1 types with a recursive definition (such as can be found
4349 in PKCS7) could eventually exceed the stack given malicious input with
4350 excessive recursion. This could result in a Denial Of Service attack. There
4351 are no such structures used within SSL/TLS that come from untrusted sources
4352 so this is considered safe.
4354 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4360 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4362 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4363 effectively reduced to only comparing the least significant bit of each
4364 byte. This allows an attacker to forge messages that would be considered as
4365 authenticated in an amount of tries lower than that guaranteed by the
4366 security claims of the scheme. The module can only be compiled by the
4367 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4369 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4375 * Add a build target 'build_all_generated', to build all generated files
4376 and only that. This can be used to prepare everything that requires
4377 things like perl for a system that lacks perl and then move everything
4378 to that system and do the rest of the build there.
4382 * Backport SSL_OP_NO_RENGOTIATION
4384 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4385 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4386 changes this is no longer possible in 1.1.0. Therefore, the new
4387 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4388 1.1.0 to provide equivalent functionality.
4390 Note that if an application built against 1.1.0h headers (or above) is run
4391 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4392 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4396 * Removed the OS390-Unix config target. It relied on a script that doesn't
4401 * rsaz_1024_mul_avx2 overflow bug on x86_64
4403 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4404 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4405 Analysis suggests that attacks against RSA and DSA as a result of this
4406 defect would be very difficult to perform and are not believed likely.
4407 Attacks against DH1024 are considered just feasible, because most of the
4408 work necessary to deduce information about a private key may be performed
4409 offline. The amount of resources required for such an attack would be
4410 significant. However, for an attack on TLS to be meaningful, the server
4411 would have to share the DH1024 private key among multiple clients, which is
4412 no longer an option since CVE-2016-0701.
4414 This only affects processors that support the AVX2 but not ADX extensions
4415 like Intel Haswell (4th generation).
4417 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4418 was originally found via the OSS-Fuzz project.
4423 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4425 * bn_sqrx8x_internal carry bug on x86_64
4427 There is a carry propagating bug in the x86_64 Montgomery squaring
4428 procedure. No EC algorithms are affected. Analysis suggests that attacks
4429 against RSA and DSA as a result of this defect would be very difficult to
4430 perform and are not believed likely. Attacks against DH are considered just
4431 feasible (although very difficult) because most of the work necessary to
4432 deduce information about a private key may be performed offline. The amount
4433 of resources required for such an attack would be very significant and
4434 likely only accessible to a limited number of attackers. An attacker would
4435 additionally need online access to an unpatched system using the target
4436 private key in a scenario with persistent DH parameters and a private
4437 key that is shared between multiple clients.
4439 This only affects processors that support the BMI1, BMI2 and ADX extensions
4440 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4442 This issue was reported to OpenSSL by the OSS-Fuzz project.
4447 * Malformed X.509 IPAddressFamily could cause OOB read
4449 If an X.509 certificate has a malformed IPAddressFamily extension,
4450 OpenSSL could do a one-byte buffer overread. The most likely result
4451 would be an erroneous display of the certificate in text format.
4453 This issue was reported to OpenSSL by the OSS-Fuzz project.
4458 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4460 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4461 platform rather than 'mingw'.
4465 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4466 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4467 which is the minimum version we support.
4471 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4473 * Encrypt-Then-Mac renegotiation crash
4475 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4476 negotiated where it was not in the original handshake (or vice-versa) then
4477 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4478 and servers are affected.
4480 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4485 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4487 * Truncated packet could crash via OOB read
4489 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4490 cipher is being used, then a truncated packet can cause that host to
4491 perform an out-of-bounds read, usually resulting in a crash.
4493 This issue was reported to OpenSSL by Robert Święcki of Google.
4498 * Bad (EC)DHE parameters cause a client crash
4500 If a malicious server supplies bad parameters for a DHE or ECDHE key
4501 exchange then this can result in the client attempting to dereference a
4502 NULL pointer leading to a client crash. This could be exploited in a Denial
4505 This issue was reported to OpenSSL by Guido Vranken.
4510 * BN_mod_exp may produce incorrect results on x86_64
4512 There is a carry propagating bug in the x86_64 Montgomery squaring
4513 procedure. No EC algorithms are affected. Analysis suggests that attacks
4514 against RSA and DSA as a result of this defect would be very difficult to
4515 perform and are not believed likely. Attacks against DH are considered just
4516 feasible (although very difficult) because most of the work necessary to
4517 deduce information about a private key may be performed offline. The amount
4518 of resources required for such an attack would be very significant and
4519 likely only accessible to a limited number of attackers. An attacker would
4520 additionally need online access to an unpatched system using the target
4521 private key in a scenario with persistent DH parameters and a private
4522 key that is shared between multiple clients. For example this can occur by
4523 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4524 similar to CVE-2015-3193 but must be treated as a separate problem.
4526 This issue was reported to OpenSSL by the OSS-Fuzz project.
4531 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4533 * ChaCha20/Poly1305 heap-buffer-overflow
4535 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4536 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4537 crash. This issue is not considered to be exploitable beyond a DoS.
4539 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4544 * CMS Null dereference
4546 Applications parsing invalid CMS structures can crash with a NULL pointer
4547 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4548 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4549 structure callback if an attempt is made to free certain invalid encodings.
4550 Only CHOICE structures using a callback which do not handle NULL value are
4553 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4558 * Montgomery multiplication may produce incorrect results
4560 There is a carry propagating bug in the Broadwell-specific Montgomery
4561 multiplication procedure that handles input lengths divisible by, but
4562 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4563 and DH private keys are impossible. This is because the subroutine in
4564 question is not used in operations with the private key itself and an input
4565 of the attacker's direct choice. Otherwise the bug can manifest itself as
4566 transient authentication and key negotiation failures or reproducible
4567 erroneous outcome of public-key operations with specially crafted input.
4568 Among EC algorithms only Brainpool P-512 curves are affected and one
4569 presumably can attack ECDH key negotiation. Impact was not analyzed in
4570 detail, because pre-requisites for attack are considered unlikely. Namely
4571 multiple clients have to choose the curve in question and the server has to
4572 share the private key among them, neither of which is default behaviour.
4573 Even then only clients that chose the curve will be affected.
4575 This issue was publicly reported as transient failures and was not
4576 initially recognized as a security issue. Thanks to Richard Morgan for
4577 providing reproducible case.
4582 * Removed automatic addition of RPATH in shared libraries and executables,
4583 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4587 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4589 * Fix Use After Free for large message sizes
4591 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4592 message larger than approx 16k is received then the underlying buffer to
4593 store the incoming message is reallocated and moved. Unfortunately a
4594 dangling pointer to the old location is left which results in an attempt to
4595 write to the previously freed location. This is likely to result in a
4596 crash, however it could potentially lead to execution of arbitrary code.
4598 This issue only affects OpenSSL 1.1.0a.
4600 This issue was reported to OpenSSL by Robert Święcki.
4605 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4607 * OCSP Status Request extension unbounded memory growth
4609 A malicious client can send an excessively large OCSP Status Request
4610 extension. If that client continually requests renegotiation, sending a
4611 large OCSP Status Request extension each time, then there will be unbounded
4612 memory growth on the server. This will eventually lead to a Denial Of
4613 Service attack through memory exhaustion. Servers with a default
4614 configuration are vulnerable even if they do not support OCSP. Builds using
4615 the "no-ocsp" build time option are not affected.
4617 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4622 * SSL_peek() hang on empty record
4624 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4625 sends an empty record. This could be exploited by a malicious peer in a
4626 Denial Of Service attack.
4628 This issue was reported to OpenSSL by Alex Gaynor.
4633 * Excessive allocation of memory in tls_get_message_header() and
4634 dtls1_preprocess_fragment()
4636 A (D)TLS message includes 3 bytes for its length in the header for the
4637 message. This would allow for messages up to 16Mb in length. Messages of
4638 this length are excessive and OpenSSL includes a check to ensure that a
4639 peer is sending reasonably sized messages in order to avoid too much memory
4640 being consumed to service a connection. A flaw in the logic of version
4641 1.1.0 means that memory for the message is allocated too early, prior to
4642 the excessive message length check. Due to way memory is allocated in
4643 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4644 to service a connection. This could lead to a Denial of Service through
4645 memory exhaustion. However, the excessive message length check still takes
4646 place, and this would cause the connection to immediately fail. Assuming
4647 that the application calls SSL_free() on the failed connection in a timely
4648 manner then the 21Mb of allocated memory will then be immediately freed
4649 again. Therefore, the excessive memory allocation will be transitory in
4650 nature. This then means that there is only a security impact if:
4652 1) The application does not call SSL_free() in a timely manner in the event
4653 that the connection fails
4655 2) The application is working in a constrained environment where there is
4656 very little free memory
4658 3) The attacker initiates multiple connection attempts such that there are
4659 multiple connections in a state where memory has been allocated for the
4660 connection; SSL_free() has not yet been called; and there is insufficient
4661 memory to service the multiple requests.
4663 Except in the instance of (1) above any Denial Of Service is likely to be
4664 transitory because as soon as the connection fails the memory is
4665 subsequently freed again in the SSL_free() call. However there is an
4666 increased risk during this period of application crashes due to the lack of
4667 memory - which would then mean a more serious Denial of Service.
4669 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4670 (CVE-2016-6307 and CVE-2016-6308)
4674 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4675 had to be removed. Primary reason is that vendor assembler can't
4676 assemble our modules with -KPIC flag. As result it, assembly
4677 support, was not even available as option. But its lack means
4678 lack of side-channel resistant code, which is incompatible with
4679 security by todays standards. Fortunately gcc is readily available
4680 prepackaged option, which we firmly point at...
4684 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4686 * Windows command-line tool supports UTF-8 opt-in option for arguments
4687 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4688 (to any value) allows Windows user to access PKCS#12 file generated
4689 with Windows CryptoAPI and protected with non-ASCII password, as well
4690 as files generated under UTF-8 locale on Linux also protected with
4695 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4696 have been disabled by default and removed from DEFAULT, just like RC4.
4697 See the RC4 item below to re-enable both.
4701 * The method for finding the storage location for the Windows RAND seed file
4702 has changed. First we check %RANDFILE%. If that is not set then we check
4703 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4704 all else fails we fall back to C:\.
4708 * The EVP_EncryptUpdate() function has had its return type changed from void
4709 to int. A return of 0 indicates and error while a return of 1 indicates
4714 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4715 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4716 off the constant time implementation for RSA, DSA and DH have been made
4717 no-ops and deprecated.
4721 * Windows RAND implementation was simplified to only get entropy by
4722 calling CryptGenRandom(). Various other RAND-related tickets
4725 *Joseph Wylie Yandle, Rich Salz*
4727 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4728 and `OPENSSL_LH_`, respectively. The old names are available
4729 with API compatibility. They new names are now completely documented.
4733 * Unify TYPE_up_ref(obj) methods signature.
4734 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4735 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4736 int (instead of void) like all others TYPE_up_ref() methods.
4737 So now these methods also check the return value of CRYPTO_atomic_add(),
4738 and the validity of object reference counter.
4740 *fdasilvayy@gmail.com*
4742 * With Windows Visual Studio builds, the .pdb files are installed
4743 alongside the installed libraries and executables. For a static
4744 library installation, ossl_static.pdb is the associate compiler
4745 generated .pdb file to be used when linking programs.
4749 * Remove openssl.spec. Packaging files belong with the packagers.
4753 * Automatic Darwin/OSX configuration has had a refresh, it will now
4754 recognise x86_64 architectures automatically. You can still decide
4755 to build for a different bitness with the environment variable
4756 KERNEL_BITS (can be 32 or 64), for example:
4758 KERNEL_BITS=32 ./config
4762 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4763 256 bit AES and HMAC with SHA256.
4767 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4771 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4775 * To enable users to have their own config files and build file templates,
4776 Configure looks in the directory indicated by the environment variable
4777 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4778 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4779 name and is used as is.
4783 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4784 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4785 X509_CERT_FILE_CTX was removed.
4789 * "shared" builds are now the default. To create only static libraries use
4790 the "no-shared" Configure option.
4794 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4795 All of these option have not worked for some while and are fundamental
4800 * Make various cleanup routines no-ops and mark them as deprecated. Most
4801 global cleanup functions are no longer required because they are handled
4802 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4803 Explicitly de-initing can cause problems (e.g. where a library that uses
4804 OpenSSL de-inits, but an application is still using it). The affected
4805 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4806 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4807 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4808 COMP_zlib_cleanup().
4812 * --strict-warnings no longer enables runtime debugging options
4813 such as REF_DEBUG. Instead, debug options are automatically
4814 enabled with '--debug' builds.
4816 *Andy Polyakov, Emilia Käsper*
4818 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4819 have been moved out of the public header files. New functions for managing
4820 these have been added.
4824 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4825 objects have been moved out of the public header files. New
4826 functions for managing these have been added.
4830 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4831 have been moved out of the public header files. New functions for managing
4832 these have been added.
4836 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4837 moved out of the public header files. New functions for managing these
4842 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4846 * Removed the mk1mf build scripts.
4850 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4851 it is always safe to #include a header now.
4855 * Removed the aged BC-32 config and all its supporting scripts
4859 * Removed support for Ultrix, Netware, and OS/2.
4863 * Add support for HKDF.
4865 *Alessandro Ghedini*
4867 * Add support for blake2b and blake2s
4871 * Added support for "pipelining". Ciphers that have the
4872 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4873 encryptions/decryptions simultaneously. There are currently no built-in
4874 ciphers with this property but the expectation is that engines will be able
4875 to offer it to significantly improve throughput. Support has been extended
4876 into libssl so that multiple records for a single connection can be
4877 processed in one go (for >=TLS 1.1).
4881 * Added the AFALG engine. This is an async capable engine which is able to
4882 offload work to the Linux kernel. In this initial version it only supports
4883 AES128-CBC. The kernel must be version 4.1.0 or greater.
4887 * OpenSSL now uses a new threading API. It is no longer necessary to
4888 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4889 are two supported threading models: pthreads and windows threads. It is
4890 also possible to configure OpenSSL at compile time for "no-threads". The
4891 old threading API should no longer be used. The functions have been
4892 replaced with "no-op" compatibility macros.
4894 *Alessandro Ghedini, Matt Caswell*
4896 * Modify behavior of ALPN to invoke callback after SNI/servername
4897 callback, such that updates to the SSL_CTX affect ALPN.
4901 * Add SSL_CIPHER queries for authentication and key-exchange.
4905 * Changes to the DEFAULT cipherlist:
4906 - Prefer (EC)DHE handshakes over plain RSA.
4907 - Prefer AEAD ciphers over legacy ciphers.
4908 - Prefer ECDSA over RSA when both certificates are available.
4909 - Prefer TLSv1.2 ciphers/PRF.
4910 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4915 * Change the ECC default curve list to be this, in order: x25519,
4916 secp256r1, secp521r1, secp384r1.
4920 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4921 disabled by default. They can be re-enabled using the
4922 enable-weak-ssl-ciphers option to Configure.
4926 * If the server has ALPN configured, but supports no protocols that the
4927 client advertises, send a fatal "no_application_protocol" alert.
4928 This behaviour is SHALL in RFC 7301, though it isn't universally
4929 implemented by other servers.
4933 * Add X25519 support.
4934 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4935 for public and private key encoding using the format documented in
4936 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4937 key generation and key derivation.
4939 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4944 * Deprecate SRP_VBASE_get_by_user.
4945 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4946 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4947 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4948 seed, even if the seed is configured.
4950 Users should use SRP_VBASE_get1_by_user instead. Note that in
4951 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4952 also that even though configuring the SRP seed attempts to hide
4953 invalid usernames by continuing the handshake with fake
4954 credentials, this behaviour is not constant time and no strong
4955 guarantees are made that the handshake is indistinguishable from
4956 that of a valid user.
4960 * Configuration change; it's now possible to build dynamic engines
4961 without having to build shared libraries and vice versa. This
4962 only applies to the engines in `engines/`, those in `crypto/engine/`
4963 will always be built into libcrypto (i.e. "static").
4965 Building dynamic engines is enabled by default; to disable, use
4966 the configuration option "disable-dynamic-engine".
4968 The only requirements for building dynamic engines are the
4969 presence of the DSO module and building with position independent
4970 code, so they will also automatically be disabled if configuring
4971 with "disable-dso" or "disable-pic".
4973 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4974 are also taken away from openssl/opensslconf.h, as they are
4979 * Configuration change; if there is a known flag to compile
4980 position independent code, it will always be applied on the
4981 libcrypto and libssl object files, and never on the application
4982 object files. This means other libraries that use routines from
4983 libcrypto / libssl can be made into shared libraries regardless
4984 of how OpenSSL was configured.
4986 If this isn't desirable, the configuration options "disable-pic"
4987 or "no-pic" can be used to disable the use of PIC. This will
4988 also disable building shared libraries and dynamic engines.
4992 * Removed JPAKE code. It was experimental and has no wide use.
4996 * The INSTALL_PREFIX Makefile variable has been renamed to
4997 DESTDIR. That makes for less confusion on what this variable
4998 is for. Also, the configuration option --install_prefix is
5003 * Heartbeat for TLS has been removed and is disabled by default
5004 for DTLS; configure with enable-heartbeats. Code that uses the
5005 old #define's might need to be updated.
5007 *Emilia Käsper, Rich Salz*
5009 * Rename REF_CHECK to REF_DEBUG.
5013 * New "unified" build system
5015 The "unified" build system is aimed to be a common system for all
5016 platforms we support. With it comes new support for VMS.
5018 This system builds supports building in a different directory tree
5019 than the source tree. It produces one Makefile (for unix family
5020 or lookalikes), or one descrip.mms (for VMS).
5022 The source of information to make the Makefile / descrip.mms is
5023 small files called 'build.info', holding the necessary
5024 information for each directory with source to compile, and a
5025 template in Configurations, like unix-Makefile.tmpl or
5028 With this change, the library names were also renamed on Windows
5029 and on VMS. They now have names that are closer to the standard
5030 on Unix, and include the major version number, and in certain
5031 cases, the architecture they are built for. See "Notes on shared
5032 libraries" in INSTALL.
5034 We rely heavily on the perl module Text::Template.
5038 * Added support for auto-initialisation and de-initialisation of the library.
5039 OpenSSL no longer requires explicit init or deinit routines to be called,
5040 except in certain circumstances. See the OPENSSL_init_crypto() and
5041 OPENSSL_init_ssl() man pages for further information.
5045 * The arguments to the DTLSv1_listen function have changed. Specifically the
5046 "peer" argument is now expected to be a BIO_ADDR object.
5048 * Rewrite of BIO networking library. The BIO library lacked consistent
5049 support of IPv6, and adding it required some more extensive
5050 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
5051 which hold all types of addresses and chains of address information.
5052 It also introduces a new API, with functions like BIO_socket,
5053 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
5054 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
5055 have been adapted accordingly.
5059 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
5064 * CRIME protection: disable compression by default, even if OpenSSL is
5065 compiled with zlib enabled. Applications can still enable compression
5066 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5067 using the SSL_CONF library to configure compression.
5071 * The signature of the session callback configured with
5072 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5073 was explicitly marked as `const unsigned char*` instead of
5078 * Always DPURIFY. Remove the use of uninitialized memory in the
5079 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5083 * Removed many obsolete configuration items, including
5084 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5085 MD2_CHAR, MD2_INT, MD2_LONG
5087 IDEA_SHORT, IDEA_LONG
5088 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5090 *Rich Salz, with advice from Andy Polyakov*
5092 * Many BN internals have been moved to an internal header file.
5094 *Rich Salz with help from Andy Polyakov*
5096 * Configuration and writing out the results from it has changed.
5097 Files such as Makefile include/openssl/opensslconf.h and are now
5098 produced through general templates, such as Makefile.in and
5099 crypto/opensslconf.h.in and some help from the perl module
5102 Also, the center of configuration information is no longer
5103 Makefile. Instead, Configure produces a perl module in
5104 configdata.pm which holds most of the config data (in the hash
5105 table %config), the target data that comes from the target
5106 configuration in one of the `Configurations/*.conf` files (in
5111 * To clarify their intended purposes, the Configure options
5112 --prefix and --openssldir change their semantics, and become more
5113 straightforward and less interdependent.
5115 --prefix shall be used exclusively to give the location INSTALLTOP
5116 where programs, scripts, libraries, include files and manuals are
5117 going to be installed. The default is now /usr/local.
5119 --openssldir shall be used exclusively to give the default
5120 location OPENSSLDIR where certificates, private keys, CRLs are
5121 managed. This is also where the default openssl.cnf gets
5123 If the directory given with this option is a relative path, the
5124 values of both the --prefix value and the --openssldir value will
5125 be combined to become OPENSSLDIR.
5126 The default for --openssldir is INSTALLTOP/ssl.
5128 Anyone who uses --openssldir to specify where OpenSSL is to be
5129 installed MUST change to use --prefix instead.
5133 * The GOST engine was out of date and therefore it has been removed. An up
5134 to date GOST engine is now being maintained in an external repository.
5135 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5136 support for GOST ciphersuites (these are only activated if a GOST engine
5141 * EGD is no longer supported by default; use enable-egd when
5144 *Ben Kaduk and Rich Salz*
5146 * The distribution now has Makefile.in files, which are used to
5147 create Makefile's when Configure is run. *Configure must be run
5148 before trying to build now.*
5152 * The return value for SSL_CIPHER_description() for error conditions
5157 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5159 Obtaining and performing DNSSEC validation of TLSA records is
5160 the application's responsibility. The application provides
5161 the TLSA records of its choice to OpenSSL, and these are then
5162 used to authenticate the peer.
5164 The TLSA records need not even come from DNS. They can, for
5165 example, be used to implement local end-entity certificate or
5166 trust-anchor "pinning", where the "pin" data takes the form
5167 of TLSA records, which can augment or replace verification
5168 based on the usual WebPKI public certification authorities.
5172 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5173 continues to support deprecated interfaces in default builds.
5174 However, applications are strongly advised to compile their
5175 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5176 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5177 or the 1.1.0 releases.
5179 In environments in which all applications have been ported to
5180 not use any deprecated interfaces OpenSSL's Configure script
5181 should be used with the --api=1.1.0 option to entirely remove
5182 support for the deprecated features from the library and
5183 unconditionally disable them in the installed headers.
5184 Essentially the same effect can be achieved with the "no-deprecated"
5185 argument to Configure, except that this will always restrict
5186 the build to just the latest API, rather than a fixed API
5189 As applications are ported to future revisions of the API,
5190 they should update their compile-time OPENSSL_API_COMPAT define
5191 accordingly, but in most cases should be able to continue to
5192 compile with later releases.
5194 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5195 0x10000000L and 0x00908000L, respectively. However those
5196 versions did not support the OPENSSL_API_COMPAT feature, and
5197 so applications are not typically tested for explicit support
5198 of just the undeprecated features of either release.
5202 * Add support for setting the minimum and maximum supported protocol.
5203 It can bet set via the SSL_set_min_proto_version() and
5204 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5205 MaxProtocol. It's recommended to use the new APIs to disable
5206 protocols instead of disabling individual protocols using
5207 SSL_set_options() or SSL_CONF's Protocol. This change also
5208 removes support for disabling TLS 1.2 in the OpenSSL TLS
5209 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5213 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5217 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5218 and integrates ECDSA and ECDH functionality into EC. Implementations can
5219 now redirect key generation and no longer need to convert to or from
5222 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5223 include the ec.h header file instead.
5227 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5228 ciphers who are no longer supported and drops support the ephemeral RSA key
5229 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5233 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5234 opaque. For HMAC_CTX, the following constructors and destructors
5237 HMAC_CTX *HMAC_CTX_new(void);
5238 void HMAC_CTX_free(HMAC_CTX *ctx);
5240 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5241 destroy such methods has been added. See EVP_MD_meth_new(3) and
5242 EVP_CIPHER_meth_new(3) for documentation.
5245 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5246 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5247 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5248 an already created structure.
5249 2) For consistency with the majority of our object creators and
5250 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5251 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5252 for deprecated builds.
5256 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5257 cryptographic operations to be performed asynchronously as long as an
5258 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5259 further details. Libssl has also had this capability integrated with the
5260 introduction of the new mode SSL_MODE_ASYNC and associated error
5261 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5262 pages. This work was developed in partnership with Intel Corp.
5266 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5267 always enabled now. If you want to disable the support you should
5268 exclude it using the list of supported ciphers. This also means that the
5269 "-no_ecdhe" option has been removed from s_server.
5273 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5274 SSL_{CTX_}set1_curves() which can set a list.
5278 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5279 curve you want to support using SSL_{CTX_}set1_curves().
5283 * State machine rewrite. The state machine code has been significantly
5284 refactored in order to remove much duplication of code and solve issues
5285 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5286 further details). This change does have some associated API changes.
5287 Notably the SSL_state() function has been removed and replaced by
5288 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5289 SSL_set_state() has been removed altogether. The previous handshake states
5290 defined in ssl.h and ssl3.h have also been removed.
5294 * All instances of the string "ssleay" in the public API were replaced
5295 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5296 Some error codes related to internal RSA_eay API's were renamed.
5300 * The demo files in crypto/threads were moved to demo/threads.
5304 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5307 *Matt Caswell, Rich Salz*
5309 * New ASN.1 embed macro.
5311 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5312 structure is not allocated: it is part of the parent. That is instead of
5320 This reduces memory fragmentation and make it impossible to accidentally
5321 set a mandatory field to NULL.
5323 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5324 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5325 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5330 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5334 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5335 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5336 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5337 DES and RC4 ciphersuites.
5341 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5342 This changes the decoding behaviour for some invalid messages,
5343 though the change is mostly in the more lenient direction, and
5344 legacy behaviour is preserved as much as possible.
5348 * Fix no-stdio build.
5349 *David Woodhouse <David.Woodhouse@intel.com> and also*
5350 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5352 * New testing framework
5353 The testing framework has been largely rewritten and is now using
5354 perl and the perl modules Test::Harness and an extended variant of
5355 Test::More called OpenSSL::Test to do its work. All test scripts in
5356 test/ have been rewritten into test recipes, and all direct calls to
5357 executables in test/Makefile have become individual recipes using the
5358 simplified testing OpenSSL::Test::Simple.
5360 For documentation on our testing modules, do:
5362 perldoc test/testlib/OpenSSL/Test/Simple.pm
5363 perldoc test/testlib/OpenSSL/Test.pm
5367 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5368 are used; the latter aborts on memory leaks (usually checked on exit).
5369 Some undocumented "set malloc, etc., hooks" functions were removed
5370 and others were changed. All are now documented.
5374 * In DSA_generate_parameters_ex, if the provided seed is too short,
5377 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5379 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5380 from RFC4279, RFC4785, RFC5487, RFC5489.
5382 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5383 original RSA_PSK patch.
5387 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5388 era flag was never set throughout the codebase (only read). Also removed
5389 SSL3_FLAGS_POP_BUFFER which was only used if
5390 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5394 * Changed the default name options in the "ca", "crl", "req" and "x509"
5395 to be "oneline" instead of "compat".
5399 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5400 not aware of clients that still exhibit this bug, and the workaround
5401 hasn't been working properly for a while.
5405 * The return type of BIO_number_read() and BIO_number_written() as well as
5406 the corresponding num_read and num_write members in the BIO structure has
5407 changed from unsigned long to uint64_t. On platforms where an unsigned
5408 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5413 * Given the pervasive nature of TLS extensions it is inadvisable to run
5414 OpenSSL without support for them. It also means that maintaining
5415 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5416 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5420 * Removed support for the two export grade static DH ciphersuites
5421 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5422 were newly added (along with a number of other static DH ciphersuites) to
5423 1.0.2. However the two export ones have *never* worked since they were
5424 introduced. It seems strange in any case to be adding new export
5425 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5429 * Version negotiation has been rewritten. In particular SSLv23_method(),
5430 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5431 and turned into macros which simply call the new preferred function names
5432 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5433 should use the new names instead. Also as part of this change the ssl23.h
5434 header file has been removed.
5438 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5439 code and the associated standard is no longer considered fit-for-purpose.
5443 * RT2547 was closed. When generating a private key, try to make the
5444 output file readable only by the owner. This behavior change might
5445 be noticeable when interacting with other software.
5447 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5452 * Added HTTP GET support to the ocsp command.
5456 * Changed default digest for the dgst and enc commands from MD5 to
5461 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5465 * Added support for TLS extended master secret from
5466 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5467 initial patch which was a great help during development.
5471 * All libssl internal structures have been removed from the public header
5472 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5473 now redundant). Users should not attempt to access internal structures
5474 directly. Instead they should use the provided API functions.
5478 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5479 Access to deprecated functions can be re-enabled by running config with
5480 "enable-deprecated". In addition applications wishing to use deprecated
5481 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5482 will, by default, disable some transitive includes that previously existed
5483 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5487 * Added support for OCB mode. OpenSSL has been granted a patent license
5488 compatible with the OpenSSL license for use of OCB. Details are available
5489 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5490 for OCB can be removed by calling config with no-ocb.
5494 * SSLv2 support has been removed. It still supports receiving an SSLv2
5495 compatible client hello.
5499 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5500 done while fixing the error code for the key-too-small case.
5502 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5504 * CA.sh has been removed; use CA.pl instead.
5508 * Removed old DES API.
5512 * Remove various unsupported platforms:
5518 Sinix/ReliantUNIX RM400
5523 16-bit platforms such as WIN16
5527 * Clean up OPENSSL_NO_xxx #define's
5528 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5529 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5530 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5531 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5532 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5533 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5534 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5535 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5536 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5537 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5541 * Cleaned up dead code
5542 Remove all but one '#ifdef undef' which is to be looked at.
5546 * Clean up calling of xxx_free routines.
5547 Just like free(), fix most of the xxx_free routines to accept
5548 NULL. Remove the non-null checks from callers. Save much code.
5552 * Add secure heap for storage of private keys (when possible).
5553 Add BIO_s_secmem(), CBIGNUM, etc.
5554 Contributed by Akamai Technologies under our Corporate CLA.
5558 * Experimental support for a new, fast, unbiased prime candidate generator,
5559 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5561 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5563 * New output format NSS in the sess_id command line tool. This allows
5564 exporting the session id and the master key in NSS keylog format.
5566 *Martin Kaiser <martin@kaiser.cx>*
5568 * Harmonize version and its documentation. -f flag is used to display
5571 *mancha <mancha1@zoho.com>*
5573 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5574 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5576 *mancha <mancha1@zoho.com>*
5578 * Fix some double frees. These are not thought to be exploitable.
5580 *mancha <mancha1@zoho.com>*
5582 * A missing bounds check in the handling of the TLS heartbeat extension
5583 can be used to reveal up to 64k of memory to a connected client or
5586 Thanks for Neel Mehta of Google Security for discovering this bug and to
5587 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5588 preparing the fix ([CVE-2014-0160])
5590 *Adam Langley, Bodo Moeller*
5592 * Fix for the attack described in the paper "Recovering OpenSSL
5593 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5594 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5595 <http://eprint.iacr.org/2014/140>
5597 Thanks to Yuval Yarom and Naomi Benger for discovering this
5598 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5600 *Yuval Yarom and Naomi Benger*
5602 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5603 this fixes a limitation in previous versions of OpenSSL.
5607 * Experimental encrypt-then-mac support.
5609 Experimental support for encrypt then mac from
5610 draft-gutmann-tls-encrypt-then-mac-02.txt
5612 To enable it set the appropriate extension number (0x42 for the test
5613 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5615 For non-compliant peers (i.e. just about everything) this should have no
5618 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5622 * Add EVP support for key wrapping algorithms, to avoid problems with
5623 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5624 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5625 algorithms and include tests cases.
5629 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5634 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5635 MGF1 digest and OAEP label.
5639 * Make openssl verify return errors.
5641 *Chris Palmer <palmer@google.com> and Ben Laurie*
5643 * New function ASN1_TIME_diff to calculate the difference between two
5644 ASN1_TIME structures or one structure and the current time.
5648 * Update fips_test_suite to support multiple command line options. New
5649 test to induce all self test errors in sequence and check expected
5654 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5655 sign or verify all in one operation.
5659 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5660 test programs and fips_test_suite. Includes functionality to parse
5661 the minimal script output of fipsalgest.pl directly.
5665 * Add authorisation parameter to FIPS_module_mode_set().
5669 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5673 * Use separate DRBG fields for internal and external flags. New function
5674 FIPS_drbg_health_check() to perform on demand health checking. Add
5675 generation tests to fips_test_suite with reduced health check interval to
5676 demonstrate periodic health checking. Add "nodh" option to
5677 fips_test_suite to skip very slow DH test.
5681 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5686 * More extensive health check for DRBG checking many more failure modes.
5687 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5688 combination: call this in fips_test_suite.
5692 * Add support for canonical generation of DSA parameter 'g'. See
5695 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5696 POST to handle HMAC cases.
5700 * Add functions FIPS_module_version() and FIPS_module_version_text()
5701 to return numerical and string versions of the FIPS module number.
5705 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5706 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5707 outside the validated module in the FIPS capable OpenSSL.
5711 * Minor change to DRBG entropy callback semantics. In some cases
5712 there is no multiple of the block length between min_len and
5713 max_len. Allow the callback to return more than max_len bytes
5714 of entropy but discard any extra: it is the callback's responsibility
5715 to ensure that the extra data discarded does not impact the
5716 requested amount of entropy.
5720 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5721 information in FIPS186-3, SP800-57 and SP800-131A.
5725 * CCM support via EVP. Interface is very similar to GCM case except we
5726 must supply all data in one chunk (i.e. no update, final) and the
5727 message length must be supplied if AAD is used. Add algorithm test
5732 * Initial version of POST overhaul. Add POST callback to allow the status
5733 of POST to be monitored and/or failures induced. Modify fips_test_suite
5734 to use callback. Always run all selftests even if one fails.
5738 * XTS support including algorithm test driver in the fips_gcmtest program.
5739 Note: this does increase the maximum key length from 32 to 64 bytes but
5740 there should be no binary compatibility issues as existing applications
5741 will never use XTS mode.
5745 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5746 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5747 performs algorithm blocking for unapproved PRNG types. Also do not
5748 set PRNG type in FIPS_mode_set(): leave this to the application.
5749 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5750 the standard OpenSSL PRNG: set additional data to a date time vector.
5754 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5755 This shouldn't present any incompatibility problems because applications
5756 shouldn't be using these directly and any that are will need to rethink
5757 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5761 * Extensive self tests and health checking required by SP800-90 DRBG.
5762 Remove strength parameter from FIPS_drbg_instantiate and always
5763 instantiate at maximum supported strength.
5767 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5771 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5775 * New function DH_compute_key_padded() to compute a DH key and pad with
5776 leading zeroes if needed: this complies with SP800-56A et al.
5780 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5781 anything, incomplete, subject to change and largely untested at present.
5785 * Modify fipscanisteronly build option to only build the necessary object
5786 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5790 * Add experimental option FIPSSYMS to give all symbols in
5791 fipscanister.o and FIPS or fips prefix. This will avoid
5792 conflicts with future versions of OpenSSL. Add perl script
5793 util/fipsas.pl to preprocess assembly language source files
5794 and rename any affected symbols.
5798 * Add selftest checks and algorithm block of non-fips algorithms in
5799 FIPS mode. Remove DES2 from selftests.
5803 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5804 return internal method without any ENGINE dependencies. Add new
5805 tiny fips sign and verify functions.
5809 * New build option no-ec2m to disable characteristic 2 code.
5813 * New build option "fipscanisteronly". This only builds fipscanister.o
5814 and (currently) associated fips utilities. Uses the file Makefile.fips
5815 instead of Makefile.org as the prototype.
5819 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5820 Update fips_gcmtest to use IV generator.
5824 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5825 setting output buffer to NULL. The `*Final` function must be
5826 called although it will not retrieve any additional data. The tag
5827 can be set or retrieved with a ctrl. The IV length is by default 12
5828 bytes (96 bits) but can be set to an alternative value. If the IV
5829 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5834 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5835 underlying do_cipher function handles all cipher semantics itself
5836 including padding and finalisation. This is useful if (for example)
5837 an ENGINE cipher handles block padding itself. The behaviour of
5838 do_cipher is subtly changed if this flag is set: the return value
5839 is the number of characters written to the output buffer (zero is
5840 no longer an error code) or a negative error code. Also if the
5841 input buffer is NULL and length 0 finalisation should be performed.
5845 * If a candidate issuer certificate is already part of the constructed
5846 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5850 * Improve forward-security support: add functions
5852 void SSL_CTX_set_not_resumable_session_callback(
5853 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5854 void SSL_set_not_resumable_session_callback(
5855 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5857 for use by SSL/TLS servers; the callback function will be called whenever a
5858 new session is created, and gets to decide whether the session may be
5859 cached to make it resumable (return 0) or not (return 1). (As by the
5860 SSL/TLS protocol specifications, the session_id sent by the server will be
5861 empty to indicate that the session is not resumable; also, the server will
5862 not generate RFC 4507 (RFC 5077) session tickets.)
5864 A simple reasonable callback implementation is to return is_forward_secure.
5865 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5866 by the SSL/TLS server library, indicating whether it can provide forward
5869 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5871 * New -verify_name option in command line utilities to set verification
5876 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5877 Add CMAC pkey methods.
5881 * Experimental renegotiation in s_server -www mode. If the client
5882 browses /reneg connection is renegotiated. If /renegcert it is
5883 renegotiated requesting a certificate.
5887 * Add an "external" session cache for debugging purposes to s_server. This
5888 should help trace issues which normally are only apparent in deployed
5889 multi-process servers.
5893 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5894 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5895 BIO_set_cipher() and some obscure PEM functions were changed so they
5896 can now return an error. The RAND changes required a change to the
5897 RAND_METHOD structure.
5901 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5902 a gcc attribute to warn if the result of a function is ignored. This
5903 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5904 whose return value is often ignored.
5908 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5909 These allow SCTs (signed certificate timestamps) to be requested and
5910 validated when establishing a connection.
5912 *Rob Percival <robpercival@google.com>*
5917 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5919 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5920 used even when parsing explicit parameters, when loading a encoded key
5921 or calling `EC_GROUP_new_from_ecpkparameters()`/
5922 `EC_GROUP_new_from_ecparameters()`.
5923 This prevents bypass of security hardening and performance gains,
5924 especially for curves with specialized EC_METHODs.
5925 By default, if a key encoded with explicit parameters is loaded and later
5926 encoded, the output is still encoded with explicit parameters, even if
5927 internally a "named" EC_GROUP is used for computation.
5931 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5932 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5933 NULL. After this change, only the cofactor parameter can be NULL. It also
5934 does some minimal sanity checks on the passed order.
5939 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5940 An attack is simple, if the first CMS_recipientInfo is valid but the
5941 second CMS_recipientInfo is chosen ciphertext. If the second
5942 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5943 encryption key will be replaced by garbage, and the message cannot be
5944 decoded, but if the RSA decryption fails, the correct encryption key is
5945 used and the recipient will not notice the attack.
5946 As a work around for this potential attack the length of the decrypted
5947 key must be equal to the cipher default key length, in case the
5948 certificate is not given and all recipientInfo are tried out.
5949 The old behaviour can be re-enabled in the CMS code by setting the
5950 CMS_DEBUG_DECRYPT flag.
5955 * Document issue with installation paths in diverse Windows builds
5957 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5958 binaries and run-time config file.
5963 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5965 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5966 This changes the size when using the `genpkey` command when no size is given.
5967 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5968 generation commands to use 2048 bits by default.
5972 * Add FIPS support for Android Arm 64-bit
5974 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5975 Module in Version 2.0.10. For some reason, the corresponding target
5976 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5977 built with FIPS support on Android Arm 64-bit. This omission has been
5980 *Matthias St. Pierre*
5982 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5984 * 0-byte record padding oracle
5986 If an application encounters a fatal protocol error and then calls
5987 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5988 then OpenSSL can respond differently to the calling application if a 0 byte
5989 record is received with invalid padding compared to if a 0 byte record is
5990 received with an invalid MAC. If the application then behaves differently
5991 based on that in a way that is detectable to the remote peer, then this
5992 amounts to a padding oracle that could be used to decrypt data.
5994 In order for this to be exploitable "non-stitched" ciphersuites must be in
5995 use. Stitched ciphersuites are optimised implementations of certain
5996 commonly used ciphersuites. Also the application must call SSL_shutdown()
5997 twice even if a protocol error has occurred (applications should not do
5998 this but some do anyway).
6000 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
6001 Aviram, with additional investigation by Steven Collison and Andrew
6002 Hourselt. It was reported to OpenSSL on 10th December 2018.
6007 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
6011 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
6013 * Microarchitecture timing vulnerability in ECC scalar multiplication
6015 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
6016 shown to be vulnerable to a microarchitecture timing side channel attack.
6017 An attacker with sufficient access to mount local timing attacks during
6018 ECDSA signature generation could recover the private key.
6020 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
6021 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
6027 * Timing vulnerability in DSA signature generation
6029 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
6030 timing side channel attack. An attacker could use variations in the signing
6031 algorithm to recover the private key.
6033 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
6038 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
6039 Module, accidentally introduced while backporting security fixes from the
6040 development branch and hindering the use of ECC in FIPS mode.
6044 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
6046 * Client DoS due to large DH parameter
6048 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
6049 malicious server can send a very large prime value to the client. This will
6050 cause the client to spend an unreasonably long period of time generating a
6051 key for this prime resulting in a hang until the client has finished. This
6052 could be exploited in a Denial Of Service attack.
6054 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
6059 * Cache timing vulnerability in RSA Key Generation
6061 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
6062 a cache timing side channel attack. An attacker with sufficient access to
6063 mount cache timing attacks during the RSA key generation process could
6064 recover the private key.
6066 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6067 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6072 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6073 parameter is no longer accepted, as it leads to a corrupt table. NULL
6074 pem_str is reserved for alias entries only.
6078 * Revert blinding in ECDSA sign and instead make problematic addition
6079 length-invariant. Switch even to fixed-length Montgomery multiplication.
6083 * Change generating and checking of primes so that the error rate of not
6084 being prime depends on the intended use based on the size of the input.
6085 For larger primes this will result in more rounds of Miller-Rabin.
6086 The maximal error rate for primes with more than 1080 bits is lowered
6089 *Kurt Roeckx, Annie Yousar*
6091 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6095 * Add blinding to ECDSA and DSA signatures to protect against side channel
6096 attacks discovered by Keegan Ryan (NCC Group).
6100 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6101 now allow empty (zero character) pass phrases.
6105 * Certificate time validation (X509_cmp_time) enforces stricter
6106 compliance with RFC 5280. Fractional seconds and timezone offsets
6107 are no longer allowed.
6111 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6113 * Constructed ASN.1 types with a recursive definition could exceed the stack
6115 Constructed ASN.1 types with a recursive definition (such as can be found
6116 in PKCS7) could eventually exceed the stack given malicious input with
6117 excessive recursion. This could result in a Denial Of Service attack. There
6118 are no such structures used within SSL/TLS that come from untrusted sources
6119 so this is considered safe.
6121 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6127 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6129 * Read/write after SSL object in error state
6131 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6132 mechanism. The intent was that if a fatal error occurred during a handshake
6133 then OpenSSL would move into the error state and would immediately fail if
6134 you attempted to continue the handshake. This works as designed for the
6135 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6136 SSL_connect()), however due to a bug it does not work correctly if
6137 SSL_read() or SSL_write() is called directly. In that scenario, if the
6138 handshake fails then a fatal error will be returned in the initial function
6139 call. If SSL_read()/SSL_write() is subsequently called by the application
6140 for the same SSL object then it will succeed and the data is passed without
6141 being decrypted/encrypted directly from the SSL/TLS record layer.
6143 In order to exploit this issue an application bug would have to be present
6144 that resulted in a call to SSL_read()/SSL_write() being issued after having
6145 already received a fatal error.
6147 This issue was reported to OpenSSL by David Benjamin (Google).
6152 * rsaz_1024_mul_avx2 overflow bug on x86_64
6154 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6155 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6156 Analysis suggests that attacks against RSA and DSA as a result of this
6157 defect would be very difficult to perform and are not believed likely.
6158 Attacks against DH1024 are considered just feasible, because most of the
6159 work necessary to deduce information about a private key may be performed
6160 offline. The amount of resources required for such an attack would be
6161 significant. However, for an attack on TLS to be meaningful, the server
6162 would have to share the DH1024 private key among multiple clients, which is
6163 no longer an option since CVE-2016-0701.
6165 This only affects processors that support the AVX2 but not ADX extensions
6166 like Intel Haswell (4th generation).
6168 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6169 was originally found via the OSS-Fuzz project.
6174 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6176 * bn_sqrx8x_internal carry bug on x86_64
6178 There is a carry propagating bug in the x86_64 Montgomery squaring
6179 procedure. No EC algorithms are affected. Analysis suggests that attacks
6180 against RSA and DSA as a result of this defect would be very difficult to
6181 perform and are not believed likely. Attacks against DH are considered just
6182 feasible (although very difficult) because most of the work necessary to
6183 deduce information about a private key may be performed offline. The amount
6184 of resources required for such an attack would be very significant and
6185 likely only accessible to a limited number of attackers. An attacker would
6186 additionally need online access to an unpatched system using the target
6187 private key in a scenario with persistent DH parameters and a private
6188 key that is shared between multiple clients.
6190 This only affects processors that support the BMI1, BMI2 and ADX extensions
6191 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6193 This issue was reported to OpenSSL by the OSS-Fuzz project.
6198 * Malformed X.509 IPAddressFamily could cause OOB read
6200 If an X.509 certificate has a malformed IPAddressFamily extension,
6201 OpenSSL could do a one-byte buffer overread. The most likely result
6202 would be an erroneous display of the certificate in text format.
6204 This issue was reported to OpenSSL by the OSS-Fuzz project.
6208 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6210 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6211 platform rather than 'mingw'.
6215 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6217 * Truncated packet could crash via OOB read
6219 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6220 cipher is being used, then a truncated packet can cause that host to
6221 perform an out-of-bounds read, usually resulting in a crash.
6223 This issue was reported to OpenSSL by Robert Święcki of Google.
6228 * BN_mod_exp may produce incorrect results on x86_64
6230 There is a carry propagating bug in the x86_64 Montgomery squaring
6231 procedure. No EC algorithms are affected. Analysis suggests that attacks
6232 against RSA and DSA as a result of this defect would be very difficult to
6233 perform and are not believed likely. Attacks against DH are considered just
6234 feasible (although very difficult) because most of the work necessary to
6235 deduce information about a private key may be performed offline. The amount
6236 of resources required for such an attack would be very significant and
6237 likely only accessible to a limited number of attackers. An attacker would
6238 additionally need online access to an unpatched system using the target
6239 private key in a scenario with persistent DH parameters and a private
6240 key that is shared between multiple clients. For example this can occur by
6241 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6242 similar to CVE-2015-3193 but must be treated as a separate problem.
6244 This issue was reported to OpenSSL by the OSS-Fuzz project.
6249 * Montgomery multiplication may produce incorrect results
6251 There is a carry propagating bug in the Broadwell-specific Montgomery
6252 multiplication procedure that handles input lengths divisible by, but
6253 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6254 and DH private keys are impossible. This is because the subroutine in
6255 question is not used in operations with the private key itself and an input
6256 of the attacker's direct choice. Otherwise the bug can manifest itself as
6257 transient authentication and key negotiation failures or reproducible
6258 erroneous outcome of public-key operations with specially crafted input.
6259 Among EC algorithms only Brainpool P-512 curves are affected and one
6260 presumably can attack ECDH key negotiation. Impact was not analyzed in
6261 detail, because pre-requisites for attack are considered unlikely. Namely
6262 multiple clients have to choose the curve in question and the server has to
6263 share the private key among them, neither of which is default behaviour.
6264 Even then only clients that chose the curve will be affected.
6266 This issue was publicly reported as transient failures and was not
6267 initially recognized as a security issue. Thanks to Richard Morgan for
6268 providing reproducible case.
6273 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6274 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6275 prevent issues where no progress is being made and the peer continually
6276 sends unrecognised record types, using up resources processing them.
6280 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6282 * Missing CRL sanity check
6284 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6285 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6286 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6288 This issue only affects the OpenSSL 1.0.2i
6293 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6295 * OCSP Status Request extension unbounded memory growth
6297 A malicious client can send an excessively large OCSP Status Request
6298 extension. If that client continually requests renegotiation, sending a
6299 large OCSP Status Request extension each time, then there will be unbounded
6300 memory growth on the server. This will eventually lead to a Denial Of
6301 Service attack through memory exhaustion. Servers with a default
6302 configuration are vulnerable even if they do not support OCSP. Builds using
6303 the "no-ocsp" build time option are not affected.
6305 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6310 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6313 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6319 * OOB write in MDC2_Update()
6321 An overflow can occur in MDC2_Update() either if called directly or
6322 through the EVP_DigestUpdate() function using MDC2. If an attacker
6323 is able to supply very large amounts of input data after a previous
6324 call to EVP_EncryptUpdate() with a partial block then a length check
6325 can overflow resulting in a heap corruption.
6327 The amount of data needed is comparable to SIZE_MAX which is impractical
6330 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6335 * Malformed SHA512 ticket DoS
6337 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6338 DoS attack where a malformed ticket will result in an OOB read which will
6341 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6342 a custom server callback and ticket lookup mechanism.
6344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6349 * OOB write in BN_bn2dec()
6351 The function BN_bn2dec() does not check the return value of BN_div_word().
6352 This can cause an OOB write if an application uses this function with an
6353 overly large BIGNUM. This could be a problem if an overly large certificate
6354 or CRL is printed out from an untrusted source. TLS is not affected because
6355 record limits will reject an oversized certificate before it is parsed.
6357 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6362 * OOB read in TS_OBJ_print_bio()
6364 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6365 the total length the OID text representation would use and not the amount
6366 of data written. This will result in OOB reads when large OIDs are
6369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6374 * Pointer arithmetic undefined behaviour
6376 Avoid some undefined pointer arithmetic
6378 A common idiom in the codebase is to check limits in the following manner:
6381 Where "p" points to some malloc'd data of SIZE bytes and
6384 "len" here could be from some externally supplied data (e.g. from a TLS
6387 The rules of C pointer arithmetic are such that "p + len" is only well
6388 defined where len <= SIZE. Therefore the above idiom is actually
6389 undefined behaviour.
6391 For example this could cause problems if some malloc implementation
6392 provides an address for "p" such that "p + len" actually overflows for
6393 values of len that are too big and therefore p + len < limit.
6395 This issue was reported to OpenSSL by Guido Vranken
6400 * Constant time flag not preserved in DSA signing
6402 Operations in the DSA signing algorithm should run in constant time in
6403 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6404 implementation means that a non-constant time codepath is followed for
6405 certain operations. This has been demonstrated through a cache-timing
6406 attack to be sufficient for an attacker to recover the private DSA key.
6408 This issue was reported by César Pereida (Aalto University), Billy Brumley
6409 (Tampere University of Technology), and Yuval Yarom (The University of
6410 Adelaide and NICTA).
6415 * DTLS buffered message DoS
6417 In a DTLS connection where handshake messages are delivered out-of-order
6418 those messages that OpenSSL is not yet ready to process will be buffered
6419 for later use. Under certain circumstances, a flaw in the logic means that
6420 those messages do not get removed from the buffer even though the handshake
6421 has been completed. An attacker could force up to approx. 15 messages to
6422 remain in the buffer when they are no longer required. These messages will
6423 be cleared when the DTLS connection is closed. The default maximum size for
6424 a message is 100k. Therefore, the attacker could force an additional 1500k
6425 to be consumed per connection. By opening many simultaneous connections an
6426 attacker could cause a DoS attack through memory exhaustion.
6428 This issue was reported to OpenSSL by Quan Luo.
6433 * DTLS replay protection DoS
6435 A flaw in the DTLS replay attack protection mechanism means that records
6436 that arrive for future epochs update the replay protection "window" before
6437 the MAC for the record has been validated. This could be exploited by an
6438 attacker by sending a record for the next epoch (which does not have to
6439 decrypt or have a valid MAC), with a very large sequence number. This means
6440 that all subsequent legitimate packets are dropped causing a denial of
6441 service for a specific DTLS connection.
6443 This issue was reported to OpenSSL by the OCAP audit team.
6448 * Certificate message OOB reads
6450 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6451 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6452 theoretical DoS risk but this has not been observed in practice on common
6455 The messages affected are client certificate, client certificate request
6456 and server certificate. As a result the attack can only be performed
6457 against a client or a server which enables client authentication.
6459 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6464 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6466 * Prevent padding oracle in AES-NI CBC MAC check
6468 A MITM attacker can use a padding oracle attack to decrypt traffic
6469 when the connection uses an AES CBC cipher and the server support
6472 This issue was introduced as part of the fix for Lucky 13 padding
6473 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6474 constant time by making sure that always the same bytes are read and
6475 compared against either the MAC or padding bytes. But it no longer
6476 checked that there was enough data to have both the MAC and padding
6479 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6483 * Fix EVP_EncodeUpdate overflow
6485 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6486 Base64 encoding of binary data. If an attacker is able to supply very large
6487 amounts of input data then a length check can overflow resulting in a heap
6490 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6491 the `PEM_write_bio*` family of functions. These are mainly used within the
6492 OpenSSL command line applications, so any application which processes data
6493 from an untrusted source and outputs it as a PEM file should be considered
6494 vulnerable to this issue. User applications that call these APIs directly
6495 with large amounts of untrusted data may also be vulnerable.
6497 This issue was reported by Guido Vranken.
6502 * Fix EVP_EncryptUpdate overflow
6504 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6505 is able to supply very large amounts of input data after a previous call to
6506 EVP_EncryptUpdate() with a partial block then a length check can overflow
6507 resulting in a heap corruption. Following an analysis of all OpenSSL
6508 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6509 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6510 the first called function after an EVP_EncryptInit(), and therefore that
6511 specific call must be safe. The second form is where the length passed to
6512 EVP_EncryptUpdate() can be seen from the code to be some small value and
6513 therefore there is no possibility of an overflow. Since all instances are
6514 one of these two forms, it is believed that there can be no overflows in
6515 internal code due to this problem. It should be noted that
6516 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6517 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6518 of these calls have also been analysed too and it is believed there are no
6519 instances in internal usage where an overflow could occur.
6521 This issue was reported by Guido Vranken.
6526 * Prevent ASN.1 BIO excessive memory allocation
6528 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6529 a short invalid encoding can cause allocation of large amounts of memory
6530 potentially consuming excessive resources or exhausting memory.
6532 Any application parsing untrusted data through d2i BIO functions is
6533 affected. The memory based functions such as d2i_X509() are *not* affected.
6534 Since the memory based functions are used by the TLS library, TLS
6535 applications are not affected.
6537 This issue was reported by Brian Carpenter.
6544 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6545 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6546 in arbitrary stack data being returned in the buffer.
6548 This issue was reported by Guido Vranken.
6553 * Modify behavior of ALPN to invoke callback after SNI/servername
6554 callback, such that updates to the SSL_CTX affect ALPN.
6558 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6563 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6564 methods are enabled and ssl2 is disabled the methods return NULL.
6568 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6570 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6571 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6572 provide any "EXPORT" or "LOW" strength ciphers.
6576 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6577 is by default disabled at build-time. Builds that are not configured with
6578 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6579 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6580 will need to explicitly call either of:
6582 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6584 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6586 as appropriate. Even if either of those is used, or the application
6587 explicitly uses the version-specific SSLv2_method() or its client and
6588 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6589 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6590 ciphers, and SSLv2 56-bit DES are no longer available.
6595 * Fix a double-free in DSA code
6597 A double free bug was discovered when OpenSSL parses malformed DSA private
6598 keys and could lead to a DoS attack or memory corruption for applications
6599 that receive DSA private keys from untrusted sources. This scenario is
6602 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6608 * Disable SRP fake user seed to address a server memory leak.
6610 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6612 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6613 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6614 was changed to ignore the "fake user" SRP seed, even if the seed
6617 Users should use SRP_VBASE_get1_by_user instead. Note that in
6618 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6619 also that even though configuring the SRP seed attempts to hide
6620 invalid usernames by continuing the handshake with fake
6621 credentials, this behaviour is not constant time and no strong
6622 guarantees are made that the handshake is indistinguishable from
6623 that of a valid user.
6628 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6630 In the BN_hex2bn function the number of hex digits is calculated using an
6631 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6632 large values of `i` this can result in `bn_expand` not allocating any
6633 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6634 field as NULL leading to a subsequent NULL ptr deref. For very large values
6635 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6636 In this case memory is allocated to the internal BIGNUM data field, but it
6637 is insufficiently sized leading to heap corruption. A similar issue exists
6638 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6639 is ever called by user applications with very large untrusted hex/dec data.
6640 This is anticipated to be a rare occurrence.
6642 All OpenSSL internal usage of these functions use data that is not expected
6643 to be untrusted, e.g. config file data or application command line
6644 arguments. If user developed applications generate config file data based
6645 on untrusted data then it is possible that this could also lead to security
6646 consequences. This is also anticipated to be rare.
6648 This issue was reported to OpenSSL by Guido Vranken.
6653 * Fix memory issues in `BIO_*printf` functions
6655 The internal `fmtstr` function used in processing a "%s" format string in
6656 the `BIO_*printf` functions could overflow while calculating the length of a
6657 string and cause an OOB read when printing very long strings.
6659 Additionally the internal `doapr_outch` function can attempt to write to an
6660 OOB memory location (at an offset from the NULL pointer) in the event of a
6661 memory allocation failure. In 1.0.2 and below this could be caused where
6662 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6663 could be in processing a very long "%s" format string. Memory leaks can
6666 The first issue may mask the second issue dependent on compiler behaviour.
6667 These problems could enable attacks where large amounts of untrusted data
6668 is passed to the `BIO_*printf` functions. If applications use these functions
6669 in this way then they could be vulnerable. OpenSSL itself uses these
6670 functions when printing out human-readable dumps of ASN.1 data. Therefore
6671 applications that print this data could be vulnerable if the data is from
6672 untrusted sources. OpenSSL command line applications could also be
6673 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6674 as command line arguments.
6676 Libssl is not considered directly vulnerable. Additionally certificates etc
6677 received via remote connections via libssl are also unlikely to be able to
6678 trigger these issues because of message size limits enforced within libssl.
6680 This issue was reported to OpenSSL Guido Vranken.
6685 * Side channel attack on modular exponentiation
6687 A side-channel attack was found which makes use of cache-bank conflicts on
6688 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6689 of RSA keys. The ability to exploit this issue is limited as it relies on
6690 an attacker who has control of code in a thread running on the same
6691 hyper-threaded core as the victim thread which is performing decryptions.
6693 This issue was reported to OpenSSL by Yuval Yarom, The University of
6694 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6695 Nadia Heninger, University of Pennsylvania with more information at
6696 <http://cachebleed.info>.
6701 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6702 if no keysize is specified with default_bits. This fixes an
6703 omission in an earlier change that changed all RSA/DSA key generation
6704 commands to use 2048 bits by default.
6708 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6710 * DH small subgroups
6712 Historically OpenSSL only ever generated DH parameters based on "safe"
6713 primes. More recently (in version 1.0.2) support was provided for
6714 generating X9.42 style parameter files such as those required for RFC 5114
6715 support. The primes used in such files may not be "safe". Where an
6716 application is using DH configured with parameters based on primes that are
6717 not "safe" then an attacker could use this fact to find a peer's private
6718 DH exponent. This attack requires that the attacker complete multiple
6719 handshakes in which the peer uses the same private DH exponent. For example
6720 this could be used to discover a TLS server's private DH exponent if it's
6721 reusing the private DH exponent or it's using a static DH ciphersuite.
6723 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6724 TLS. It is not on by default. If the option is not set then the server
6725 reuses the same private DH exponent for the life of the server process and
6726 would be vulnerable to this attack. It is believed that many popular
6727 applications do set this option and would therefore not be at risk.
6729 The fix for this issue adds an additional check where a "q" parameter is
6730 available (as is the case in X9.42 based parameters). This detects the
6731 only known attack, and is the only possible defense for static DH
6732 ciphersuites. This could have some performance impact.
6734 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6735 default and cannot be disabled. This could have some performance impact.
6737 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6742 * SSLv2 doesn't block disabled ciphers
6744 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6745 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6746 been disabled, provided that the SSLv2 protocol was not also disabled via
6749 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6750 and Sebastian Schinzel.
6755 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6757 * BN_mod_exp may produce incorrect results on x86_64
6759 There is a carry propagating bug in the x86_64 Montgomery squaring
6760 procedure. No EC algorithms are affected. Analysis suggests that attacks
6761 against RSA and DSA as a result of this defect would be very difficult to
6762 perform and are not believed likely. Attacks against DH are considered just
6763 feasible (although very difficult) because most of the work necessary to
6764 deduce information about a private key may be performed offline. The amount
6765 of resources required for such an attack would be very significant and
6766 likely only accessible to a limited number of attackers. An attacker would
6767 additionally need online access to an unpatched system using the target
6768 private key in a scenario with persistent DH parameters and a private
6769 key that is shared between multiple clients. For example this can occur by
6770 default in OpenSSL DHE based SSL/TLS ciphersuites.
6772 This issue was reported to OpenSSL by Hanno Böck.
6777 * Certificate verify crash with missing PSS parameter
6779 The signature verification routines will crash with a NULL pointer
6780 dereference if presented with an ASN.1 signature using the RSA PSS
6781 algorithm and absent mask generation function parameter. Since these
6782 routines are used to verify certificate signature algorithms this can be
6783 used to crash any certificate verification operation and exploited in a
6784 DoS attack. Any application which performs certificate verification is
6785 vulnerable including OpenSSL clients and servers which enable client
6788 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6793 * X509_ATTRIBUTE memory leak
6795 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6796 memory. This structure is used by the PKCS#7 and CMS routines so any
6797 application which reads PKCS#7 or CMS data from untrusted sources is
6798 affected. SSL/TLS is not affected.
6800 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6807 This changes the decoding behaviour for some invalid messages,
6808 though the change is mostly in the more lenient direction, and
6809 legacy behaviour is preserved as much as possible.
6813 * In DSA_generate_parameters_ex, if the provided seed is too short,
6816 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6818 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6820 * Alternate chains certificate forgery
6822 During certificate verification, OpenSSL will attempt to find an
6823 alternative certificate chain if the first attempt to build such a chain
6824 fails. An error in the implementation of this logic can mean that an
6825 attacker could cause certain checks on untrusted certificates to be
6826 bypassed, such as the CA flag, enabling them to use a valid leaf
6827 certificate to act as a CA and "issue" an invalid certificate.
6829 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6834 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6836 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6837 incompatibility in the handling of HMAC. The previous ABI has now been
6842 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6844 * Malformed ECParameters causes infinite loop
6846 When processing an ECParameters structure OpenSSL enters an infinite loop
6847 if the curve specified is over a specially malformed binary polynomial
6850 This can be used to perform denial of service against any
6851 system which processes public keys, certificate requests or
6852 certificates. This includes TLS clients and TLS servers with
6853 client authentication enabled.
6855 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6860 * Exploitable out-of-bounds read in X509_cmp_time
6862 X509_cmp_time does not properly check the length of the ASN1_TIME
6863 string and can read a few bytes out of bounds. In addition,
6864 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6867 An attacker can use this to craft malformed certificates and CRLs of
6868 various sizes and potentially cause a segmentation fault, resulting in
6869 a DoS on applications that verify certificates or CRLs. TLS clients
6870 that verify CRLs are affected. TLS clients and servers with client
6871 authentication enabled may be affected if they use custom verification
6874 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6875 independently by Hanno Böck.
6880 * PKCS7 crash with missing EnvelopedContent
6882 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6883 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6884 with missing content and trigger a NULL pointer dereference on parsing.
6886 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6887 structures from untrusted sources are affected. OpenSSL clients and
6888 servers are not affected.
6890 This issue was reported to OpenSSL by Michal Zalewski (Google).
6895 * CMS verify infinite loop with unknown hash function
6897 When verifying a signedData message the CMS code can enter an infinite loop
6898 if presented with an unknown hash function OID. This can be used to perform
6899 denial of service against any system which verifies signedData messages using
6901 This issue was reported to OpenSSL by Johannes Bauer.
6906 * Race condition handling NewSessionTicket
6908 If a NewSessionTicket is received by a multi-threaded client when attempting to
6909 reuse a previous ticket then a race condition can occur potentially leading to
6910 a double free of the ticket data.
6915 * Only support 256-bit or stronger elliptic curves with the
6916 'ecdh_auto' setting (server) or by default (client). Of supported
6917 curves, prefer P-256 (both).
6921 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6923 * ClientHello sigalgs DoS fix
6925 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6926 invalid signature algorithms extension a NULL pointer dereference will
6927 occur. This can be exploited in a DoS attack against the server.
6929 This issue was was reported to OpenSSL by David Ramos of Stanford
6933 *Stephen Henson and Matt Caswell*
6935 * Multiblock corrupted pointer fix
6937 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6938 feature only applies on 64 bit x86 architecture platforms that support AES
6939 NI instructions. A defect in the implementation of "multiblock" can cause
6940 OpenSSL's internal write buffer to become incorrectly set to NULL when
6941 using non-blocking IO. Typically, when the user application is using a
6942 socket BIO for writing, this will only result in a failed connection.
6943 However if some other BIO is used then it is likely that a segmentation
6944 fault will be triggered, thus enabling a potential DoS attack.
6946 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6951 * Segmentation fault in DTLSv1_listen fix
6953 The DTLSv1_listen function is intended to be stateless and processes the
6954 initial ClientHello from many peers. It is common for user code to loop
6955 over the call to DTLSv1_listen until a valid ClientHello is received with
6956 an associated cookie. A defect in the implementation of DTLSv1_listen means
6957 that state is preserved in the SSL object from one invocation to the next
6958 that can lead to a segmentation fault. Errors processing the initial
6959 ClientHello can trigger this scenario. An example of such an error could be
6960 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6963 This issue was reported to OpenSSL by Per Allansson.
6968 * Segmentation fault in ASN1_TYPE_cmp fix
6970 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6971 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6972 certificate signature algorithm consistency this can be used to crash any
6973 certificate verification operation and exploited in a DoS attack. Any
6974 application which performs certificate verification is vulnerable including
6975 OpenSSL clients and servers which enable client authentication.
6980 * Segmentation fault for invalid PSS parameters fix
6982 The signature verification routines will crash with a NULL pointer
6983 dereference if presented with an ASN.1 signature using the RSA PSS
6984 algorithm and invalid parameters. Since these routines are used to verify
6985 certificate signature algorithms this can be used to crash any
6986 certificate verification operation and exploited in a DoS attack. Any
6987 application which performs certificate verification is vulnerable including
6988 OpenSSL clients and servers which enable client authentication.
6990 This issue was was reported to OpenSSL by Brian Carpenter.
6995 * ASN.1 structure reuse memory corruption fix
6997 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6998 memory corruption via an invalid write. Such reuse is and has been
6999 strongly discouraged and is believed to be rare.
7001 Applications that parse structures containing CHOICE or ANY DEFINED BY
7002 components may be affected. Certificate parsing (d2i_X509 and related
7003 functions) are however not affected. OpenSSL clients and servers are
7009 * PKCS7 NULL pointer dereferences fix
7011 The PKCS#7 parsing code does not handle missing outer ContentInfo
7012 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7013 missing content and trigger a NULL pointer dereference on parsing.
7015 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7016 otherwise parse PKCS#7 structures from untrusted sources are
7017 affected. OpenSSL clients and servers are not affected.
7019 This issue was reported to OpenSSL by Michal Zalewski (Google).
7024 * DoS via reachable assert in SSLv2 servers fix
7026 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7027 servers that both support SSLv2 and enable export cipher suites by sending
7028 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7030 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7031 (OpenSSL development team).
7036 * Empty CKE with client auth and DHE fix
7038 If client auth is used then a server can seg fault in the event of a DHE
7039 ciphersuite being selected and a zero length ClientKeyExchange message
7040 being sent by the client. This could be exploited in a DoS attack.
7045 * Handshake with unseeded PRNG fix
7047 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
7048 with an unseeded PRNG. The conditions are:
7049 - The client is on a platform where the PRNG has not been seeded
7050 automatically, and the user has not seeded manually
7051 - A protocol specific client method version has been used (i.e. not
7052 SSL_client_methodv23)
7053 - A ciphersuite is used that does not require additional random data from
7054 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
7056 If the handshake succeeds then the client random that has been used will
7057 have been generated from a PRNG with insufficient entropy and therefore the
7058 output may be predictable.
7060 For example using the following command with an unseeded openssl will
7061 succeed on an unpatched platform:
7063 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7068 * Use After Free following d2i_ECPrivatekey error fix
7070 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7071 could cause a use after free condition. This, in turn, could cause a double
7072 free in several private key parsing functions (such as d2i_PrivateKey
7073 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7074 for applications that receive EC private keys from untrusted
7075 sources. This scenario is considered rare.
7077 This issue was discovered by the BoringSSL project and fixed in their
7083 * X509_to_X509_REQ NULL pointer deref fix
7085 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7086 the certificate key is invalid. This function is rarely used in practice.
7088 This issue was discovered by Brian Carpenter.
7093 * Removed the export ciphers from the DEFAULT ciphers
7097 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7099 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7100 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7101 So far those who have to target multiple platforms would compromise
7102 and argue that binary targeting say ARMv5 would still execute on
7103 ARMv8. "Universal" build resolves this compromise by providing
7104 near-optimal performance even on newer platforms.
7108 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7109 (other platforms pending).
7111 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7113 * Add support for the SignedCertificateTimestampList certificate and
7114 OCSP response extensions from RFC6962.
7118 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7119 for corner cases. (Certain input points at infinity could lead to
7120 bogus results, with non-infinity inputs mapped to infinity too.)
7124 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7125 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7126 common cases are optimized and there still is room for further
7127 improvements. Vector Permutation AES for Altivec is also added.
7131 * Add support for little-endian ppc64 Linux target.
7133 *Marcelo Cerri (IBM)*
7135 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7136 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7137 are optimized and there still is room for further improvements.
7138 Both 32- and 64-bit modes are supported.
7140 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7142 * Improved ARMv7 NEON support.
7146 * Support for SPARC Architecture 2011 crypto extensions, first
7147 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7148 SHA256/512, MD5, GHASH and modular exponentiation.
7150 *Andy Polyakov, David Miller*
7152 * Accelerated modular exponentiation for Intel processors, a.k.a.
7155 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7157 * Support for new and upcoming Intel processors, including AVX2,
7158 BMI and SHA ISA extensions. This includes additional "stitched"
7159 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7162 This work was sponsored by Intel Corp.
7166 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7167 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7168 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7172 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7173 this fixes a limitation in previous versions of OpenSSL.
7177 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7178 MGF1 digest and OAEP label.
7182 * Add EVP support for key wrapping algorithms, to avoid problems with
7183 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7184 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7185 algorithms and include tests cases.
7189 * Add functions to allocate and set the fields of an ECDSA_METHOD
7192 *Douglas E. Engert, Steve Henson*
7194 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7195 difference in days and seconds between two tm or ASN1_TIME structures.
7199 * Add -rev test option to s_server to just reverse order of characters
7200 received by client and send back to server. Also prints an abbreviated
7201 summary of the connection parameters.
7205 * New option -brief for s_client and s_server to print out a brief summary
7206 of connection parameters.
7210 * Add callbacks for arbitrary TLS extensions.
7212 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7214 * New option -crl_download in several openssl utilities to download CRLs
7215 from CRLDP extension in certificates.
7219 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7223 * New function X509_CRL_diff to generate a delta CRL from the difference
7224 of two full CRLs. Add support to "crl" utility.
7228 * New functions to set lookup_crls function and to retrieve
7229 X509_STORE from X509_STORE_CTX.
7233 * Print out deprecated issuer and subject unique ID fields in
7238 * Extend OCSP I/O functions so they can be used for simple general purpose
7239 HTTP as well as OCSP. New wrapper function which can be used to download
7240 CRLs using the OCSP API.
7244 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7248 * `SSL_CONF*` functions. These provide a common framework for application
7249 configuration using configuration files or command lines.
7253 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7254 message callback and prints the results. Needs compile time option
7255 "enable-ssl-trace". New options to s_client and s_server to enable
7260 * New ctrl and macro to retrieve supported points extensions.
7261 Print out extension in s_server and s_client.
7265 * New functions to retrieve certificate signature and signature
7270 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7275 * New Suite B modes for TLS code. These use and enforce the requirements
7276 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7277 only use Suite B curves. The Suite B modes can be set by using the
7278 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7282 * New chain verification flags for Suite B levels of security. Check
7283 algorithms are acceptable when flags are set in X509_verify_cert.
7287 * Make tls1_check_chain return a set of flags indicating checks passed
7288 by a certificate chain. Add additional tests to handle client
7289 certificates: checks for matching certificate type and issuer name
7294 * If an attempt is made to use a signature algorithm not in the peer
7295 preference list abort the handshake. If client has no suitable
7296 signature algorithms in response to a certificate request do not
7297 use the certificate.
7301 * If server EC tmp key is not in client preference list abort handshake.
7305 * Add support for certificate stores in CERT structure. This makes it
7306 possible to have different stores per SSL structure or one store in
7307 the parent SSL_CTX. Include distinct stores for certificate chain
7308 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7309 to build and store a certificate chain in CERT structure: returning
7310 an error if the chain cannot be built: this will allow applications
7311 to test if a chain is correctly configured.
7313 Note: if the CERT based stores are not set then the parent SSL_CTX
7314 store is used to retain compatibility with existing behaviour.
7318 * New function ssl_set_client_disabled to set a ciphersuite disabled
7319 mask based on the current session, check mask when sending client
7320 hello and checking the requested ciphersuite.
7324 * New ctrls to retrieve and set certificate types in a certificate
7325 request message. Print out received values in s_client. If certificate
7326 types is not set with custom values set sensible values based on
7327 supported signature algorithms.
7331 * Support for distinct client and server supported signature algorithms.
7335 * Add certificate callback. If set this is called whenever a certificate
7336 is required by client or server. An application can decide which
7337 certificate chain to present based on arbitrary criteria: for example
7338 supported signature algorithms. Add very simple example to s_server.
7339 This fixes many of the problems and restrictions of the existing client
7340 certificate callback: for example you can now clear an existing
7341 certificate and specify the whole chain.
7345 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7346 the certificate can be used for (if anything). Set valid_flags field
7347 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7348 to have similar checks in it.
7350 Add new "cert_flags" field to CERT structure and include a "strict mode".
7351 This enforces some TLS certificate requirements (such as only permitting
7352 certificate signature algorithms contained in the supported algorithms
7353 extension) which some implementations ignore: this option should be used
7354 with caution as it could cause interoperability issues.
7358 * Update and tidy signature algorithm extension processing. Work out
7359 shared signature algorithms based on preferences and peer algorithms
7360 and print them out in s_client and s_server. Abort handshake if no
7361 shared signature algorithms.
7365 * Add new functions to allow customised supported signature algorithms
7366 for SSL and SSL_CTX structures. Add options to s_client and s_server
7371 * New function SSL_certs_clear() to delete all references to certificates
7372 from an SSL structure. Before this once a certificate had been added
7373 it couldn't be removed.
7377 * Integrate hostname, email address and IP address checking with certificate
7378 verification. New verify options supporting checking in openssl utility.
7382 * Fixes and wildcard matching support to hostname and email checking
7383 functions. Add manual page.
7385 *Florian Weimer (Red Hat Product Security Team)*
7387 * New functions to check a hostname email or IP address against a
7388 certificate. Add options x509 utility to print results of checks against
7393 * Fix OCSP checking.
7395 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7397 * Initial experimental support for explicitly trusted non-root CAs.
7398 OpenSSL still tries to build a complete chain to a root but if an
7399 intermediate CA has a trust setting included that is used. The first
7400 setting is used: whether to trust (e.g., -addtrust option to the x509
7405 * Add -trusted_first option which attempts to find certificates in the
7406 trusted store even if an untrusted chain is also supplied.
7410 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7411 platform support for Linux and Android.
7415 * Support for linux-x32, ILP32 environment in x86_64 framework.
7419 * Experimental multi-implementation support for FIPS capable OpenSSL.
7420 When in FIPS mode the approved implementations are used as normal,
7421 when not in FIPS mode the internal unapproved versions are used instead.
7422 This means that the FIPS capable OpenSSL isn't forced to use the
7423 (often lower performance) FIPS implementations outside FIPS mode.
7427 * Transparently support X9.42 DH parameters when calling
7428 PEM_read_bio_DHparameters. This means existing applications can handle
7429 the new parameter format automatically.
7433 * Initial experimental support for X9.42 DH parameter format: mainly
7434 to support use of 'q' parameter for RFC5114 parameters.
7438 * Add DH parameters from RFC5114 including test data to dhtest.
7442 * Support for automatic EC temporary key parameter selection. If enabled
7443 the most preferred EC parameters are automatically used instead of
7444 hardcoded fixed parameters. Now a server just has to call:
7445 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7446 support ECDH and use the most appropriate parameters.
7450 * Enhance and tidy EC curve and point format TLS extension code. Use
7451 static structures instead of allocation if default values are used.
7452 New ctrls to set curves we wish to support and to retrieve shared curves.
7453 Print out shared curves in s_server. New options to s_server and s_client
7454 to set list of supported curves.
7458 * New ctrls to retrieve supported signature algorithms and
7459 supported curve values as an array of NIDs. Extend openssl utility
7460 to print out received values.
7464 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7465 between NIDs and the more common NIST names such as "P-256". Enhance
7466 ecparam utility and ECC method to recognise the NIST names for curves.
7470 * Enhance SSL/TLS certificate chain handling to support different
7471 chains for each certificate instead of one chain in the parent SSL_CTX.
7475 * Support for fixed DH ciphersuite client authentication: where both
7476 server and client use DH certificates with common parameters.
7480 * Support for fixed DH ciphersuites: those requiring DH server
7485 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7487 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7488 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7489 X509_CINF_get_signature were reverted post internal team review.
7494 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7496 * OCSP Status Request extension unbounded memory growth
7498 A malicious client can send an excessively large OCSP Status Request
7499 extension. If that client continually requests renegotiation, sending a
7500 large OCSP Status Request extension each time, then there will be unbounded
7501 memory growth on the server. This will eventually lead to a Denial Of
7502 Service attack through memory exhaustion. Servers with a default
7503 configuration are vulnerable even if they do not support OCSP. Builds using
7504 the "no-ocsp" build time option are not affected.
7506 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7511 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7514 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7520 * OOB write in MDC2_Update()
7522 An overflow can occur in MDC2_Update() either if called directly or
7523 through the EVP_DigestUpdate() function using MDC2. If an attacker
7524 is able to supply very large amounts of input data after a previous
7525 call to EVP_EncryptUpdate() with a partial block then a length check
7526 can overflow resulting in a heap corruption.
7528 The amount of data needed is comparable to SIZE_MAX which is impractical
7531 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7536 * Malformed SHA512 ticket DoS
7538 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7539 DoS attack where a malformed ticket will result in an OOB read which will
7542 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7543 a custom server callback and ticket lookup mechanism.
7545 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7550 * OOB write in BN_bn2dec()
7552 The function BN_bn2dec() does not check the return value of BN_div_word().
7553 This can cause an OOB write if an application uses this function with an
7554 overly large BIGNUM. This could be a problem if an overly large certificate
7555 or CRL is printed out from an untrusted source. TLS is not affected because
7556 record limits will reject an oversized certificate before it is parsed.
7558 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7563 * OOB read in TS_OBJ_print_bio()
7565 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7566 the total length the OID text representation would use and not the amount
7567 of data written. This will result in OOB reads when large OIDs are
7570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7575 * Pointer arithmetic undefined behaviour
7577 Avoid some undefined pointer arithmetic
7579 A common idiom in the codebase is to check limits in the following manner:
7582 Where "p" points to some malloc'd data of SIZE bytes and
7585 "len" here could be from some externally supplied data (e.g. from a TLS
7588 The rules of C pointer arithmetic are such that "p + len" is only well
7589 defined where len <= SIZE. Therefore, the above idiom is actually
7590 undefined behaviour.
7592 For example this could cause problems if some malloc implementation
7593 provides an address for "p" such that "p + len" actually overflows for
7594 values of len that are too big and therefore p + len < limit.
7596 This issue was reported to OpenSSL by Guido Vranken
7601 * Constant time flag not preserved in DSA signing
7603 Operations in the DSA signing algorithm should run in constant time in
7604 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7605 implementation means that a non-constant time codepath is followed for
7606 certain operations. This has been demonstrated through a cache-timing
7607 attack to be sufficient for an attacker to recover the private DSA key.
7609 This issue was reported by César Pereida (Aalto University), Billy Brumley
7610 (Tampere University of Technology), and Yuval Yarom (The University of
7611 Adelaide and NICTA).
7616 * DTLS buffered message DoS
7618 In a DTLS connection where handshake messages are delivered out-of-order
7619 those messages that OpenSSL is not yet ready to process will be buffered
7620 for later use. Under certain circumstances, a flaw in the logic means that
7621 those messages do not get removed from the buffer even though the handshake
7622 has been completed. An attacker could force up to approx. 15 messages to
7623 remain in the buffer when they are no longer required. These messages will
7624 be cleared when the DTLS connection is closed. The default maximum size for
7625 a message is 100k. Therefore, the attacker could force an additional 1500k
7626 to be consumed per connection. By opening many simultaneous connections an
7627 attacker could cause a DoS attack through memory exhaustion.
7629 This issue was reported to OpenSSL by Quan Luo.
7634 * DTLS replay protection DoS
7636 A flaw in the DTLS replay attack protection mechanism means that records
7637 that arrive for future epochs update the replay protection "window" before
7638 the MAC for the record has been validated. This could be exploited by an
7639 attacker by sending a record for the next epoch (which does not have to
7640 decrypt or have a valid MAC), with a very large sequence number. This means
7641 that all subsequent legitimate packets are dropped causing a denial of
7642 service for a specific DTLS connection.
7644 This issue was reported to OpenSSL by the OCAP audit team.
7649 * Certificate message OOB reads
7651 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7652 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7653 theoretical DoS risk but this has not been observed in practice on common
7656 The messages affected are client certificate, client certificate request
7657 and server certificate. As a result the attack can only be performed
7658 against a client or a server which enables client authentication.
7660 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7665 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7667 * Prevent padding oracle in AES-NI CBC MAC check
7669 A MITM attacker can use a padding oracle attack to decrypt traffic
7670 when the connection uses an AES CBC cipher and the server support
7673 This issue was introduced as part of the fix for Lucky 13 padding
7674 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7675 constant time by making sure that always the same bytes are read and
7676 compared against either the MAC or padding bytes. But it no longer
7677 checked that there was enough data to have both the MAC and padding
7680 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7685 * Fix EVP_EncodeUpdate overflow
7687 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7688 Base64 encoding of binary data. If an attacker is able to supply very large
7689 amounts of input data then a length check can overflow resulting in a heap
7692 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7693 the `PEM_write_bio*` family of functions. These are mainly used within the
7694 OpenSSL command line applications, so any application which processes data
7695 from an untrusted source and outputs it as a PEM file should be considered
7696 vulnerable to this issue. User applications that call these APIs directly
7697 with large amounts of untrusted data may also be vulnerable.
7699 This issue was reported by Guido Vranken.
7704 * Fix EVP_EncryptUpdate overflow
7706 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7707 is able to supply very large amounts of input data after a previous call to
7708 EVP_EncryptUpdate() with a partial block then a length check can overflow
7709 resulting in a heap corruption. Following an analysis of all OpenSSL
7710 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7711 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7712 the first called function after an EVP_EncryptInit(), and therefore that
7713 specific call must be safe. The second form is where the length passed to
7714 EVP_EncryptUpdate() can be seen from the code to be some small value and
7715 therefore there is no possibility of an overflow. Since all instances are
7716 one of these two forms, it is believed that there can be no overflows in
7717 internal code due to this problem. It should be noted that
7718 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7719 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7720 of these calls have also been analysed too and it is believed there are no
7721 instances in internal usage where an overflow could occur.
7723 This issue was reported by Guido Vranken.
7728 * Prevent ASN.1 BIO excessive memory allocation
7730 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7731 a short invalid encoding can casuse allocation of large amounts of memory
7732 potentially consuming excessive resources or exhausting memory.
7734 Any application parsing untrusted data through d2i BIO functions is
7735 affected. The memory based functions such as d2i_X509() are *not* affected.
7736 Since the memory based functions are used by the TLS library, TLS
7737 applications are not affected.
7739 This issue was reported by Brian Carpenter.
7746 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7747 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7748 in arbitrary stack data being returned in the buffer.
7750 This issue was reported by Guido Vranken.
7755 * Modify behavior of ALPN to invoke callback after SNI/servername
7756 callback, such that updates to the SSL_CTX affect ALPN.
7760 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7765 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7766 methods are enabled and ssl2 is disabled the methods return NULL.
7770 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7772 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7773 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7774 provide any "EXPORT" or "LOW" strength ciphers.
7778 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7779 is by default disabled at build-time. Builds that are not configured with
7780 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7781 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7782 will need to explicitly call either of:
7784 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7786 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7788 as appropriate. Even if either of those is used, or the application
7789 explicitly uses the version-specific SSLv2_method() or its client and
7790 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7791 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7792 ciphers, and SSLv2 56-bit DES are no longer available.
7797 * Fix a double-free in DSA code
7799 A double free bug was discovered when OpenSSL parses malformed DSA private
7800 keys and could lead to a DoS attack or memory corruption for applications
7801 that receive DSA private keys from untrusted sources. This scenario is
7804 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7810 * Disable SRP fake user seed to address a server memory leak.
7812 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7814 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7815 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7816 was changed to ignore the "fake user" SRP seed, even if the seed
7819 Users should use SRP_VBASE_get1_by_user instead. Note that in
7820 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7821 also that even though configuring the SRP seed attempts to hide
7822 invalid usernames by continuing the handshake with fake
7823 credentials, this behaviour is not constant time and no strong
7824 guarantees are made that the handshake is indistinguishable from
7825 that of a valid user.
7830 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7832 In the BN_hex2bn function the number of hex digits is calculated using an
7833 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7834 large values of `i` this can result in `bn_expand` not allocating any
7835 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7836 field as NULL leading to a subsequent NULL ptr deref. For very large values
7837 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7838 In this case memory is allocated to the internal BIGNUM data field, but it
7839 is insufficiently sized leading to heap corruption. A similar issue exists
7840 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7841 is ever called by user applications with very large untrusted hex/dec data.
7842 This is anticipated to be a rare occurrence.
7844 All OpenSSL internal usage of these functions use data that is not expected
7845 to be untrusted, e.g. config file data or application command line
7846 arguments. If user developed applications generate config file data based
7847 on untrusted data then it is possible that this could also lead to security
7848 consequences. This is also anticipated to be rare.
7850 This issue was reported to OpenSSL by Guido Vranken.
7855 * Fix memory issues in `BIO_*printf` functions
7857 The internal `fmtstr` function used in processing a "%s" format string in
7858 the `BIO_*printf` functions could overflow while calculating the length of a
7859 string and cause an OOB read when printing very long strings.
7861 Additionally the internal `doapr_outch` function can attempt to write to an
7862 OOB memory location (at an offset from the NULL pointer) in the event of a
7863 memory allocation failure. In 1.0.2 and below this could be caused where
7864 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7865 could be in processing a very long "%s" format string. Memory leaks can
7868 The first issue may mask the second issue dependent on compiler behaviour.
7869 These problems could enable attacks where large amounts of untrusted data
7870 is passed to the `BIO_*printf` functions. If applications use these functions
7871 in this way then they could be vulnerable. OpenSSL itself uses these
7872 functions when printing out human-readable dumps of ASN.1 data. Therefore
7873 applications that print this data could be vulnerable if the data is from
7874 untrusted sources. OpenSSL command line applications could also be
7875 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7876 as command line arguments.
7878 Libssl is not considered directly vulnerable. Additionally certificates etc
7879 received via remote connections via libssl are also unlikely to be able to
7880 trigger these issues because of message size limits enforced within libssl.
7882 This issue was reported to OpenSSL Guido Vranken.
7887 * Side channel attack on modular exponentiation
7889 A side-channel attack was found which makes use of cache-bank conflicts on
7890 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7891 of RSA keys. The ability to exploit this issue is limited as it relies on
7892 an attacker who has control of code in a thread running on the same
7893 hyper-threaded core as the victim thread which is performing decryptions.
7895 This issue was reported to OpenSSL by Yuval Yarom, The University of
7896 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7897 Nadia Heninger, University of Pennsylvania with more information at
7898 <http://cachebleed.info>.
7903 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7904 if no keysize is specified with default_bits. This fixes an
7905 omission in an earlier change that changed all RSA/DSA key generation
7906 commands to use 2048 bits by default.
7910 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7912 * Protection for DH small subgroup attacks
7914 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7915 switched on by default and cannot be disabled. This could have some
7920 * SSLv2 doesn't block disabled ciphers
7922 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7923 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7924 been disabled, provided that the SSLv2 protocol was not also disabled via
7927 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7928 and Sebastian Schinzel.
7933 * Reject DH handshakes with parameters shorter than 1024 bits.
7937 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7939 * Certificate verify crash with missing PSS parameter
7941 The signature verification routines will crash with a NULL pointer
7942 dereference if presented with an ASN.1 signature using the RSA PSS
7943 algorithm and absent mask generation function parameter. Since these
7944 routines are used to verify certificate signature algorithms this can be
7945 used to crash any certificate verification operation and exploited in a
7946 DoS attack. Any application which performs certificate verification is
7947 vulnerable including OpenSSL clients and servers which enable client
7950 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7955 * X509_ATTRIBUTE memory leak
7957 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7958 memory. This structure is used by the PKCS#7 and CMS routines so any
7959 application which reads PKCS#7 or CMS data from untrusted sources is
7960 affected. SSL/TLS is not affected.
7962 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7968 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7969 This changes the decoding behaviour for some invalid messages,
7970 though the change is mostly in the more lenient direction, and
7971 legacy behaviour is preserved as much as possible.
7975 * In DSA_generate_parameters_ex, if the provided seed is too short,
7976 use a random seed, as already documented.
7978 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7980 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7982 * Alternate chains certificate forgery
7984 During certificate verification, OpenSSL will attempt to find an
7985 alternative certificate chain if the first attempt to build such a chain
7986 fails. An error in the implementation of this logic can mean that an
7987 attacker could cause certain checks on untrusted certificates to be
7988 bypassed, such as the CA flag, enabling them to use a valid leaf
7989 certificate to act as a CA and "issue" an invalid certificate.
7991 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7997 * Race condition handling PSK identify hint
7999 If PSK identity hints are received by a multi-threaded client then
8000 the values are wrongly updated in the parent SSL_CTX structure. This can
8001 result in a race condition potentially leading to a double free of the
8007 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
8009 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
8010 incompatibility in the handling of HMAC. The previous ABI has now been
8013 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
8015 * Malformed ECParameters causes infinite loop
8017 When processing an ECParameters structure OpenSSL enters an infinite loop
8018 if the curve specified is over a specially malformed binary polynomial
8021 This can be used to perform denial of service against any
8022 system which processes public keys, certificate requests or
8023 certificates. This includes TLS clients and TLS servers with
8024 client authentication enabled.
8026 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8031 * Exploitable out-of-bounds read in X509_cmp_time
8033 X509_cmp_time does not properly check the length of the ASN1_TIME
8034 string and can read a few bytes out of bounds. In addition,
8035 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8038 An attacker can use this to craft malformed certificates and CRLs of
8039 various sizes and potentially cause a segmentation fault, resulting in
8040 a DoS on applications that verify certificates or CRLs. TLS clients
8041 that verify CRLs are affected. TLS clients and servers with client
8042 authentication enabled may be affected if they use custom verification
8045 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8046 independently by Hanno Böck.
8051 * PKCS7 crash with missing EnvelopedContent
8053 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8054 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8055 with missing content and trigger a NULL pointer dereference on parsing.
8057 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8058 structures from untrusted sources are affected. OpenSSL clients and
8059 servers are not affected.
8061 This issue was reported to OpenSSL by Michal Zalewski (Google).
8066 * CMS verify infinite loop with unknown hash function
8068 When verifying a signedData message the CMS code can enter an infinite loop
8069 if presented with an unknown hash function OID. This can be used to perform
8070 denial of service against any system which verifies signedData messages using
8072 This issue was reported to OpenSSL by Johannes Bauer.
8077 * Race condition handling NewSessionTicket
8079 If a NewSessionTicket is received by a multi-threaded client when attempting to
8080 reuse a previous ticket then a race condition can occur potentially leading to
8081 a double free of the ticket data.
8086 * Reject DH handshakes with parameters shorter than 768 bits.
8088 *Kurt Roeckx and Emilia Kasper*
8090 * dhparam: generate 2048-bit parameters by default.
8092 *Kurt Roeckx and Emilia Kasper*
8094 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8096 * Segmentation fault in ASN1_TYPE_cmp fix
8098 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8099 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8100 certificate signature algorithm consistency this can be used to crash any
8101 certificate verification operation and exploited in a DoS attack. Any
8102 application which performs certificate verification is vulnerable including
8103 OpenSSL clients and servers which enable client authentication.
8108 * ASN.1 structure reuse memory corruption fix
8110 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8111 memory corruption via an invalid write. Such reuse is and has been
8112 strongly discouraged and is believed to be rare.
8114 Applications that parse structures containing CHOICE or ANY DEFINED BY
8115 components may be affected. Certificate parsing (d2i_X509 and related
8116 functions) are however not affected. OpenSSL clients and servers are
8122 * PKCS7 NULL pointer dereferences fix
8124 The PKCS#7 parsing code does not handle missing outer ContentInfo
8125 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8126 missing content and trigger a NULL pointer dereference on parsing.
8128 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8129 otherwise parse PKCS#7 structures from untrusted sources are
8130 affected. OpenSSL clients and servers are not affected.
8132 This issue was reported to OpenSSL by Michal Zalewski (Google).
8137 * DoS via reachable assert in SSLv2 servers fix
8139 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8140 servers that both support SSLv2 and enable export cipher suites by sending
8141 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8143 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8144 (OpenSSL development team).
8149 * Use After Free following d2i_ECPrivatekey error fix
8151 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8152 could cause a use after free condition. This, in turn, could cause a double
8153 free in several private key parsing functions (such as d2i_PrivateKey
8154 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8155 for applications that receive EC private keys from untrusted
8156 sources. This scenario is considered rare.
8158 This issue was discovered by the BoringSSL project and fixed in their
8164 * X509_to_X509_REQ NULL pointer deref fix
8166 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8167 the certificate key is invalid. This function is rarely used in practice.
8169 This issue was discovered by Brian Carpenter.
8174 * Removed the export ciphers from the DEFAULT ciphers
8178 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8180 * Build fixes for the Windows and OpenVMS platforms
8182 *Matt Caswell and Richard Levitte*
8184 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8186 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8187 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8188 dereference. This could lead to a Denial Of Service attack. Thanks to
8189 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8194 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8195 dtls1_buffer_record function under certain conditions. In particular this
8196 could occur if an attacker sent repeated DTLS records with the same
8197 sequence number but for the next epoch. The memory leak could be exploited
8198 by an attacker in a Denial of Service attack through memory exhaustion.
8199 Thanks to Chris Mueller for reporting this issue.
8204 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8205 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8206 method would be set to NULL which could later result in a NULL pointer
8207 dereference. Thanks to Frank Schmirler for reporting this issue.
8212 * Abort handshake if server key exchange message is omitted for ephemeral
8215 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8216 reporting this issue.
8221 * Remove non-export ephemeral RSA code on client and server. This code
8222 violated the TLS standard by allowing the use of temporary RSA keys in
8223 non-export ciphersuites and could be used by a server to effectively
8224 downgrade the RSA key length used to a value smaller than the server
8225 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8226 INRIA or reporting this issue.
8231 * Fixed issue where DH client certificates are accepted without verification.
8232 An OpenSSL server will accept a DH certificate for client authentication
8233 without the certificate verify message. This effectively allows a client to
8234 authenticate without the use of a private key. This only affects servers
8235 which trust a client certificate authority which issues certificates
8236 containing DH keys: these are extremely rare and hardly ever encountered.
8237 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8243 * Ensure that the session ID context of an SSL is updated when its
8244 SSL_CTX is updated via SSL_set_SSL_CTX.
8246 The session ID context is typically set from the parent SSL_CTX,
8247 and can vary with the CTX.
8251 * Fix various certificate fingerprint issues.
8253 By using non-DER or invalid encodings outside the signed portion of a
8254 certificate the fingerprint can be changed without breaking the signature.
8255 Although no details of the signed portion of the certificate can be changed
8256 this can cause problems with some applications: e.g. those using the
8257 certificate fingerprint for blacklists.
8259 1. Reject signatures with non zero unused bits.
8261 If the BIT STRING containing the signature has non zero unused bits reject
8262 the signature. All current signature algorithms require zero unused bits.
8264 2. Check certificate algorithm consistency.
8266 Check the AlgorithmIdentifier inside TBS matches the one in the
8267 certificate signature. NB: this will result in signature failure
8268 errors for some broken certificates.
8270 Thanks to Konrad Kraszewski from Google for reporting this issue.
8272 3. Check DSA/ECDSA signatures use DER.
8274 Re-encode DSA/ECDSA signatures and compare with the original received
8275 signature. Return an error if there is a mismatch.
8277 This will reject various cases including garbage after signature
8278 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8279 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8280 (negative or with leading zeroes).
8282 Further analysis was conducted and fixes were developed by Stephen Henson
8283 of the OpenSSL core team.
8289 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8290 results on some platforms, including x86_64. This bug occurs at random
8291 with a very low probability, and is not known to be exploitable in any
8292 way, though its exact impact is difficult to determine. Thanks to Pieter
8293 Wuille (Blockstream) who reported this issue and also suggested an initial
8294 fix. Further analysis was conducted by the OpenSSL development team and
8295 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8296 the OpenSSL core team.
8301 * Do not resume sessions on the server if the negotiated protocol
8302 version does not match the session's version. Resuming with a different
8303 version, while not strictly forbidden by the RFC, is of questionable
8304 sanity and breaks all known clients.
8306 *David Benjamin, Emilia Käsper*
8308 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8309 early CCS messages during renegotiation. (Note that because
8310 renegotiation is encrypted, this early CCS was not exploitable.)
8314 * Tighten client-side session ticket handling during renegotiation:
8315 ensure that the client only accepts a session ticket if the server sends
8316 the extension anew in the ServerHello. Previously, a TLS client would
8317 reuse the old extension state and thus accept a session ticket if one was
8318 announced in the initial ServerHello.
8320 Similarly, ensure that the client requires a session ticket if one
8321 was advertised in the ServerHello. Previously, a TLS client would
8322 ignore a missing NewSessionTicket message.
8326 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8330 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8331 sends a carefully crafted handshake message, to cause OpenSSL to fail
8332 to free up to 64k of memory causing a memory leak. This could be
8333 exploited in a Denial Of Service attack. This issue affects OpenSSL
8334 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8335 whether SRTP is used or configured. Implementations of OpenSSL that
8336 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8338 The fix was developed by the OpenSSL team.
8343 * Session Ticket Memory Leak.
8345 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8346 integrity of that ticket is first verified. In the event of a session
8347 ticket integrity check failing, OpenSSL will fail to free memory
8348 causing a memory leak. By sending a large number of invalid session
8349 tickets an attacker could exploit this issue in a Denial Of Service
8355 * Build option no-ssl3 is incomplete.
8357 When OpenSSL is configured with "no-ssl3" as a build option, servers
8358 could accept and complete an SSL 3.0 handshake, and clients could be
8359 configured to send them.
8362 *Akamai and the OpenSSL team*
8364 * Add support for TLS_FALLBACK_SCSV.
8365 Client applications doing fallback retries should call
8366 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8369 *Adam Langley, Bodo Moeller*
8371 * Add additional DigestInfo checks.
8373 Re-encode DigestInto in DER and check against the original when
8374 verifying RSA signature: this will reject any improperly encoded
8375 DigestInfo structures.
8377 Note: this is a precautionary measure and no attacks are currently known.
8381 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8383 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8384 SRP code can be overrun an internal buffer. Add sanity check that
8385 g, A, B < N to SRP code.
8387 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8388 Group for discovering this issue.
8393 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8394 TLS 1.0 instead of higher protocol versions when the ClientHello message
8395 is badly fragmented. This allows a man-in-the-middle attacker to force a
8396 downgrade to TLS 1.0 even if both the server and the client support a
8397 higher protocol version, by modifying the client's TLS records.
8399 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8400 researching this issue.
8405 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8406 to a denial of service attack. A malicious server can crash the client
8407 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8408 ciphersuite and sending carefully crafted handshake messages.
8410 Thanks to Felix Gröbert (Google) for discovering and researching this
8416 * By sending carefully crafted DTLS packets an attacker could cause openssl
8417 to leak memory. This can be exploited through a Denial of Service attack.
8418 Thanks to Adam Langley for discovering and researching this issue.
8423 * An attacker can force openssl to consume large amounts of memory whilst
8424 processing DTLS handshake messages. This can be exploited through a
8425 Denial of Service attack.
8426 Thanks to Adam Langley for discovering and researching this issue.
8431 * An attacker can force an error condition which causes openssl to crash
8432 whilst processing DTLS packets due to memory being freed twice. This
8433 can be exploited through a Denial of Service attack.
8434 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8440 * If a multithreaded client connects to a malicious server using a resumed
8441 session and the server sends an ec point format extension it could write
8442 up to 255 bytes to freed memory.
8444 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8450 * A malicious server can crash an OpenSSL client with a null pointer
8451 dereference (read) by specifying an SRP ciphersuite even though it was not
8452 properly negotiated with the client. This can be exploited through a
8453 Denial of Service attack.
8455 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8456 discovering and researching this issue.
8461 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8462 X509_name_oneline, X509_name_print_ex et al. to leak some information
8463 from the stack. Applications may be affected if they echo pretty printing
8464 output to the attacker.
8466 Thanks to Ivan Fratric (Google) for discovering this issue.
8469 *Emilia Käsper, and Steve Henson*
8471 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8472 for corner cases. (Certain input points at infinity could lead to
8473 bogus results, with non-infinity inputs mapped to infinity too.)
8477 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8479 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8480 handshake can force the use of weak keying material in OpenSSL
8481 SSL/TLS clients and servers.
8483 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8484 researching this issue. ([CVE-2014-0224])
8486 *KIKUCHI Masashi, Steve Henson*
8488 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8489 OpenSSL DTLS client the code can be made to recurse eventually crashing
8492 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8495 *Imre Rad, Steve Henson*
8497 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8498 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8499 client or server. This is potentially exploitable to run arbitrary
8500 code on a vulnerable client or server.
8502 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8504 *Jüri Aedla, Steve Henson*
8506 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8507 are subject to a denial of service attack.
8509 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8510 this issue. ([CVE-2014-3470])
8512 *Felix Gröbert, Ivan Fratric, Steve Henson*
8514 * Harmonize version and its documentation. -f flag is used to display
8517 *mancha <mancha1@zoho.com>*
8519 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8520 in i2d_ECPrivateKey.
8522 *mancha <mancha1@zoho.com>*
8524 * Fix some double frees. These are not thought to be exploitable.
8526 *mancha <mancha1@zoho.com>*
8528 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8530 * A missing bounds check in the handling of the TLS heartbeat extension
8531 can be used to reveal up to 64k of memory to a connected client or
8534 Thanks for Neel Mehta of Google Security for discovering this bug and to
8535 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8536 preparing the fix ([CVE-2014-0160])
8538 *Adam Langley, Bodo Moeller*
8540 * Fix for the attack described in the paper "Recovering OpenSSL
8541 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8542 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8543 <http://eprint.iacr.org/2014/140>
8545 Thanks to Yuval Yarom and Naomi Benger for discovering this
8546 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8548 *Yuval Yarom and Naomi Benger*
8550 * TLS pad extension: draft-agl-tls-padding-03
8552 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8553 TLS client Hello record length value would otherwise be > 255 and
8554 less that 512 pad with a dummy extension containing zeroes so it
8555 is at least 512 bytes long.
8557 *Adam Langley, Steve Henson*
8559 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8561 * Fix for TLS record tampering bug. A carefully crafted invalid
8562 handshake could crash OpenSSL with a NULL pointer exception.
8563 Thanks to Anton Johansson for reporting this issues.
8566 * Keep original DTLS digest and encryption contexts in retransmission
8567 structures so we can use the previous session parameters if they need
8568 to be resent. ([CVE-2013-6450])
8572 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8573 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8574 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8575 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8576 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8577 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8579 *Rob Stradling, Adam Langley*
8581 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8583 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8584 supporting platforms or when small records were transferred.
8586 *Andy Polyakov, Steve Henson*
8588 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8590 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8592 This addresses the flaw in CBC record processing discovered by
8593 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8594 at: <http://www.isg.rhul.ac.uk/tls/>
8596 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8597 Security Group at Royal Holloway, University of London
8598 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8599 Emilia Käsper for the initial patch.
8602 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8604 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8605 ciphersuites which can be exploited in a denial of service attack.
8606 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8607 and detecting this bug and to Wolfgang Ettlinger
8608 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8613 * Return an error when checking OCSP signatures when key is NULL.
8614 This fixes a DoS attack. ([CVE-2013-0166])
8618 * Make openssl verify return errors.
8620 *Chris Palmer <palmer@google.com> and Ben Laurie*
8622 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8623 the right response is stapled. Also change SSL_get_certificate()
8624 so it returns the certificate actually sent.
8625 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8627 *Rob Stradling <rob.stradling@comodo.com>*
8629 * Fix possible deadlock when decoding public keys.
8633 * Don't use TLS 1.0 record version number in initial client hello
8638 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8640 * Sanity check record length before skipping explicit IV in TLS
8641 1.2, 1.1 and DTLS to fix DoS attack.
8643 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8644 fuzzing as a service testing platform.
8649 * Initialise tkeylen properly when encrypting CMS messages.
8650 Thanks to Solar Designer of Openwall for reporting this issue.
8654 * In FIPS mode don't try to use composite ciphers as they are not
8659 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8661 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8662 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8663 mean any application compiled against OpenSSL 1.0.0 headers setting
8664 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8665 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8666 0x10000000L Any application which was previously compiled against
8667 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8668 will need to be recompiled as a result. Letting be results in
8669 inability to disable specifically TLS 1.1 and in client context,
8670 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8674 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8675 disable just protocol X, but all protocols above X *if* there are
8676 protocols *below* X still enabled. In more practical terms it means
8677 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8678 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8679 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8684 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8686 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8687 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8688 in CRYPTO_realloc_clean.
8690 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8691 issue and to Adam Langley <agl@chromium.org> for fixing it.
8694 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8696 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8700 * Workarounds for some broken servers that "hang" if a client hello
8701 record length exceeds 255 bytes.
8703 1. Do not use record version number > TLS 1.0 in initial client
8704 hello: some (but not all) hanging servers will now work.
8705 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8706 the number of ciphers sent in the client hello. This should be
8707 set to an even number, such as 50, for example by passing:
8708 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8709 Most broken servers should now work.
8710 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8711 TLS 1.2 client support entirely.
8715 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8719 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8721 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8722 STRING form instead of a DigestInfo.
8726 * The format used for MDC2 RSA signatures is inconsistent between EVP
8727 and the RSA_sign/RSA_verify functions. This was made more apparent when
8728 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8729 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8730 the correct format in RSA_verify so both forms transparently work.
8734 * Some servers which support TLS 1.0 can choke if we initially indicate
8735 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8736 encrypted premaster secret. As a workaround use the maximum permitted
8737 client version in client hello, this should keep such servers happy
8738 and still work with previous versions of OpenSSL.
8742 * Add support for TLS/DTLS heartbeats.
8744 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8746 * Add support for SCTP.
8748 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8750 * Improved PRNG seeding for VOS.
8752 *Paul Green <Paul.Green@stratus.com>*
8754 * Extensive assembler packs updates, most notably:
8756 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8757 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8758 - x86_64: bit-sliced AES implementation;
8759 - ARM: NEON support, contemporary platforms optimizations;
8760 - s390x: z196 support;
8761 - `*`: GHASH and GF(2^m) multiplication implementations;
8765 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8766 (removal of unnecessary code)
8768 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8770 * Add TLS key material exporter from RFC 5705.
8774 * Add DTLS-SRTP negotiation from RFC 5764.
8778 * Add Next Protocol Negotiation,
8779 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8780 disabled with a no-npn flag to config or Configure. Code donated
8783 *Adam Langley <agl@google.com> and Ben Laurie*
8785 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8786 NIST-P256, NIST-P521, with constant-time single point multiplication on
8787 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8788 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8789 Code made available under Apache License version 2.0.
8791 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8792 line to include this in your build of OpenSSL, and run "make depend" (or
8793 "make update"). This enables the following EC_METHODs:
8795 EC_GFp_nistp224_method()
8796 EC_GFp_nistp256_method()
8797 EC_GFp_nistp521_method()
8799 EC_GROUP_new_by_curve_name() will automatically use these (while
8800 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8803 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8805 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8806 all platforms. Move ssize_t definition from e_os.h to the public
8807 header file e_os2.h as it now appears in public header file cms.h
8811 * New -sigopt option to the ca, req and x509 utilities. Additional
8812 signature parameters can be passed using this option and in
8817 * Add RSA PSS signing function. This will generate and set the
8818 appropriate AlgorithmIdentifiers for PSS based on those in the
8819 corresponding EVP_MD_CTX structure. No application support yet.
8823 * Support for companion algorithm specific ASN1 signing routines.
8824 New function ASN1_item_sign_ctx() signs a pre-initialised
8825 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8826 the appropriate parameters.
8830 * Add new algorithm specific ASN1 verification initialisation function
8831 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8832 handling will be the same no matter what EVP_PKEY_METHOD is used.
8833 Add a PSS handler to support verification of PSS signatures: checked
8834 against a number of sample certificates.
8838 * Add signature printing for PSS. Add PSS OIDs.
8840 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8842 * Add algorithm specific signature printing. An individual ASN1 method
8843 can now print out signatures instead of the standard hex dump.
8845 More complex signatures (e.g. PSS) can print out more meaningful
8846 information. Include DSA version that prints out the signature
8851 * Password based recipient info support for CMS library: implementing
8856 * Split password based encryption into PBES2 and PBKDF2 functions. This
8857 neatly separates the code into cipher and PBE sections and is required
8858 for some algorithms that split PBES2 into separate pieces (such as
8859 password based CMS).
8863 * Session-handling fixes:
8864 - Fix handling of connections that are resuming with a session ID,
8865 but also support Session Tickets.
8866 - Fix a bug that suppressed issuing of a new ticket if the client
8867 presented a ticket with an expired session.
8868 - Try to set the ticket lifetime hint to something reasonable.
8869 - Make tickets shorter by excluding irrelevant information.
8870 - On the client side, don't ignore renewed tickets.
8872 *Adam Langley, Bodo Moeller (Google)*
8874 * Fix PSK session representation.
8878 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8880 This work was sponsored by Intel.
8884 * Add GCM support to TLS library. Some custom code is needed to split
8885 the IV between the fixed (from PRF) and explicit (from TLS record)
8886 portions. This adds all GCM ciphersuites supported by RFC5288 and
8887 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8888 add a special AESGCM string for GCM only.
8892 * Expand range of ctrls for AES GCM. Permit setting invocation
8893 field on decrypt and retrieval of invocation field only on encrypt.
8897 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8898 As required by RFC5289 these ciphersuites cannot be used if for
8899 versions of TLS earlier than 1.2.
8903 * For FIPS capable OpenSSL interpret a NULL default public key method
8904 as unset and return the appropriate default but do *not* set the default.
8905 This means we can return the appropriate method in applications that
8906 switch between FIPS and non-FIPS modes.
8910 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8911 ENGINE is used then we cannot handle that in the FIPS module so we
8912 keep original code iff non-FIPS operations are allowed.
8916 * Add -attime option to openssl utilities.
8918 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8920 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8924 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8925 FIPS EC methods unconditionally for now.
8929 * New build option no-ec2m to disable characteristic 2 code.
8933 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8934 all cases can be covered as some introduce binary incompatibilities.
8938 * Redirect RSA operations to FIPS module including keygen,
8939 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8943 * Add similar low-level API blocking to ciphers.
8947 * low-level digest APIs are not approved in FIPS mode: any attempt
8948 to use these will cause a fatal error. Applications that *really* want
8949 to use them can use the `private_*` version instead.
8953 * Redirect cipher operations to FIPS module for FIPS builds.
8957 * Redirect digest operations to FIPS module for FIPS builds.
8961 * Update build system to add "fips" flag which will link in fipscanister.o
8962 for static and shared library builds embedding a signature if needed.
8966 * Output TLS supported curves in preference order instead of numerical
8967 order. This is currently hardcoded for the highest order curves first.
8968 This should be configurable so applications can judge speed vs strength.
8972 * Add TLS v1.2 server support for client authentication.
8976 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8981 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8982 FIPS modules versions.
8986 * Add TLS v1.2 client side support for client authentication. Keep cache
8987 of handshake records longer as we don't know the hash algorithm to use
8988 until after the certificate request message is received.
8992 * Initial TLS v1.2 client support. Add a default signature algorithms
8993 extension including all the algorithms we support. Parse new signature
8994 format in client key exchange. Relax some ECC signing restrictions for
8995 TLS v1.2 as indicated in RFC5246.
8999 * Add server support for TLS v1.2 signature algorithms extension. Switch
9000 to new signature format when needed using client digest preference.
9001 All server ciphersuites should now work correctly in TLS v1.2. No client
9002 support yet and no support for client certificates.
9006 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
9007 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
9008 ciphersuites. At present only RSA key exchange ciphersuites work with
9009 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
9010 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
9011 and version checking.
9015 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
9016 with this defined it will not be affected by any changes to ssl internal
9017 structures. Add several utility functions to allow openssl application
9018 to work with OPENSSL_NO_SSL_INTERN defined.
9022 * A long standing patch to add support for SRP from EdelWeb (Peter
9023 Sylvester and Christophe Renou) was integrated.
9024 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
9025 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
9028 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
9032 * Permit abbreviated handshakes when renegotiating using the function
9033 SSL_renegotiate_abbreviated().
9035 *Robin Seggelmann <seggelmann@fh-muenster.de>*
9037 * Add call to ENGINE_register_all_complete() to
9038 ENGINE_load_builtin_engines(), so some implementations get used
9039 automatically instead of needing explicit application support.
9043 * Add support for TLS key exporter as described in RFC5705.
9045 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
9047 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
9048 a few changes are required:
9050 Add SSL_OP_NO_TLSv1_1 flag.
9051 Add TLSv1_1 methods.
9052 Update version checking logic to handle version 1.1.
9053 Add explicit IV handling (ported from DTLS code).
9054 Add command line options to s_client/s_server.
9061 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
9063 * X509_ATTRIBUTE memory leak
9065 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
9066 memory. This structure is used by the PKCS#7 and CMS routines so any
9067 application which reads PKCS#7 or CMS data from untrusted sources is
9068 affected. SSL/TLS is not affected.
9070 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9076 * Race condition handling PSK identify hint
9078 If PSK identity hints are received by a multi-threaded client then
9079 the values are wrongly updated in the parent SSL_CTX structure. This can
9080 result in a race condition potentially leading to a double free of the
9086 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9088 * Malformed ECParameters causes infinite loop
9090 When processing an ECParameters structure OpenSSL enters an infinite loop
9091 if the curve specified is over a specially malformed binary polynomial
9094 This can be used to perform denial of service against any
9095 system which processes public keys, certificate requests or
9096 certificates. This includes TLS clients and TLS servers with
9097 client authentication enabled.
9099 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9104 * Exploitable out-of-bounds read in X509_cmp_time
9106 X509_cmp_time does not properly check the length of the ASN1_TIME
9107 string and can read a few bytes out of bounds. In addition,
9108 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9111 An attacker can use this to craft malformed certificates and CRLs of
9112 various sizes and potentially cause a segmentation fault, resulting in
9113 a DoS on applications that verify certificates or CRLs. TLS clients
9114 that verify CRLs are affected. TLS clients and servers with client
9115 authentication enabled may be affected if they use custom verification
9118 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9119 independently by Hanno Böck.
9124 * PKCS7 crash with missing EnvelopedContent
9126 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9127 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9128 with missing content and trigger a NULL pointer dereference on parsing.
9130 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9131 structures from untrusted sources are affected. OpenSSL clients and
9132 servers are not affected.
9134 This issue was reported to OpenSSL by Michal Zalewski (Google).
9139 * CMS verify infinite loop with unknown hash function
9141 When verifying a signedData message the CMS code can enter an infinite loop
9142 if presented with an unknown hash function OID. This can be used to perform
9143 denial of service against any system which verifies signedData messages using
9145 This issue was reported to OpenSSL by Johannes Bauer.
9150 * Race condition handling NewSessionTicket
9152 If a NewSessionTicket is received by a multi-threaded client when attempting to
9153 reuse a previous ticket then a race condition can occur potentially leading to
9154 a double free of the ticket data.
9159 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9161 * Segmentation fault in ASN1_TYPE_cmp fix
9163 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9164 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9165 certificate signature algorithm consistency this can be used to crash any
9166 certificate verification operation and exploited in a DoS attack. Any
9167 application which performs certificate verification is vulnerable including
9168 OpenSSL clients and servers which enable client authentication.
9173 * ASN.1 structure reuse memory corruption fix
9175 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9176 memory corruption via an invalid write. Such reuse is and has been
9177 strongly discouraged and is believed to be rare.
9179 Applications that parse structures containing CHOICE or ANY DEFINED BY
9180 components may be affected. Certificate parsing (d2i_X509 and related
9181 functions) are however not affected. OpenSSL clients and servers are
9187 * PKCS7 NULL pointer dereferences fix
9189 The PKCS#7 parsing code does not handle missing outer ContentInfo
9190 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9191 missing content and trigger a NULL pointer dereference on parsing.
9193 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9194 otherwise parse PKCS#7 structures from untrusted sources are
9195 affected. OpenSSL clients and servers are not affected.
9197 This issue was reported to OpenSSL by Michal Zalewski (Google).
9202 * DoS via reachable assert in SSLv2 servers fix
9204 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9205 servers that both support SSLv2 and enable export cipher suites by sending
9206 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9208 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9209 (OpenSSL development team).
9214 * Use After Free following d2i_ECPrivatekey error fix
9216 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9217 could cause a use after free condition. This, in turn, could cause a double
9218 free in several private key parsing functions (such as d2i_PrivateKey
9219 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9220 for applications that receive EC private keys from untrusted
9221 sources. This scenario is considered rare.
9223 This issue was discovered by the BoringSSL project and fixed in their
9229 * X509_to_X509_REQ NULL pointer deref fix
9231 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9232 the certificate key is invalid. This function is rarely used in practice.
9234 This issue was discovered by Brian Carpenter.
9239 * Removed the export ciphers from the DEFAULT ciphers
9243 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9245 * Build fixes for the Windows and OpenVMS platforms
9247 *Matt Caswell and Richard Levitte*
9249 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9251 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9252 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9253 dereference. This could lead to a Denial Of Service attack. Thanks to
9254 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9259 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9260 dtls1_buffer_record function under certain conditions. In particular this
9261 could occur if an attacker sent repeated DTLS records with the same
9262 sequence number but for the next epoch. The memory leak could be exploited
9263 by an attacker in a Denial of Service attack through memory exhaustion.
9264 Thanks to Chris Mueller for reporting this issue.
9269 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9270 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9271 method would be set to NULL which could later result in a NULL pointer
9272 dereference. Thanks to Frank Schmirler for reporting this issue.
9277 * Abort handshake if server key exchange message is omitted for ephemeral
9280 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9281 reporting this issue.
9286 * Remove non-export ephemeral RSA code on client and server. This code
9287 violated the TLS standard by allowing the use of temporary RSA keys in
9288 non-export ciphersuites and could be used by a server to effectively
9289 downgrade the RSA key length used to a value smaller than the server
9290 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9291 INRIA or reporting this issue.
9296 * Fixed issue where DH client certificates are accepted without verification.
9297 An OpenSSL server will accept a DH certificate for client authentication
9298 without the certificate verify message. This effectively allows a client to
9299 authenticate without the use of a private key. This only affects servers
9300 which trust a client certificate authority which issues certificates
9301 containing DH keys: these are extremely rare and hardly ever encountered.
9302 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9308 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9309 results on some platforms, including x86_64. This bug occurs at random
9310 with a very low probability, and is not known to be exploitable in any
9311 way, though its exact impact is difficult to determine. Thanks to Pieter
9312 Wuille (Blockstream) who reported this issue and also suggested an initial
9313 fix. Further analysis was conducted by the OpenSSL development team and
9314 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9315 the OpenSSL core team.
9320 * Fix various certificate fingerprint issues.
9322 By using non-DER or invalid encodings outside the signed portion of a
9323 certificate the fingerprint can be changed without breaking the signature.
9324 Although no details of the signed portion of the certificate can be changed
9325 this can cause problems with some applications: e.g. those using the
9326 certificate fingerprint for blacklists.
9328 1. Reject signatures with non zero unused bits.
9330 If the BIT STRING containing the signature has non zero unused bits reject
9331 the signature. All current signature algorithms require zero unused bits.
9333 2. Check certificate algorithm consistency.
9335 Check the AlgorithmIdentifier inside TBS matches the one in the
9336 certificate signature. NB: this will result in signature failure
9337 errors for some broken certificates.
9339 Thanks to Konrad Kraszewski from Google for reporting this issue.
9341 3. Check DSA/ECDSA signatures use DER.
9343 Re-encode DSA/ECDSA signatures and compare with the original received
9344 signature. Return an error if there is a mismatch.
9346 This will reject various cases including garbage after signature
9347 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9348 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9349 (negative or with leading zeroes).
9351 Further analysis was conducted and fixes were developed by Stephen Henson
9352 of the OpenSSL core team.
9358 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9360 * Session Ticket Memory Leak.
9362 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9363 integrity of that ticket is first verified. In the event of a session
9364 ticket integrity check failing, OpenSSL will fail to free memory
9365 causing a memory leak. By sending a large number of invalid session
9366 tickets an attacker could exploit this issue in a Denial Of Service
9372 * Build option no-ssl3 is incomplete.
9374 When OpenSSL is configured with "no-ssl3" as a build option, servers
9375 could accept and complete an SSL 3.0 handshake, and clients could be
9376 configured to send them.
9379 *Akamai and the OpenSSL team*
9381 * Add support for TLS_FALLBACK_SCSV.
9382 Client applications doing fallback retries should call
9383 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9386 *Adam Langley, Bodo Moeller*
9388 * Add additional DigestInfo checks.
9390 Re-encode DigestInto in DER and check against the original when
9391 verifying RSA signature: this will reject any improperly encoded
9392 DigestInfo structures.
9394 Note: this is a precautionary measure and no attacks are currently known.
9398 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9400 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9401 to a denial of service attack. A malicious server can crash the client
9402 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9403 ciphersuite and sending carefully crafted handshake messages.
9405 Thanks to Felix Gröbert (Google) for discovering and researching this
9411 * By sending carefully crafted DTLS packets an attacker could cause openssl
9412 to leak memory. This can be exploited through a Denial of Service attack.
9413 Thanks to Adam Langley for discovering and researching this issue.
9418 * An attacker can force openssl to consume large amounts of memory whilst
9419 processing DTLS handshake messages. This can be exploited through a
9420 Denial of Service attack.
9421 Thanks to Adam Langley for discovering and researching this issue.
9426 * An attacker can force an error condition which causes openssl to crash
9427 whilst processing DTLS packets due to memory being freed twice. This
9428 can be exploited through a Denial of Service attack.
9429 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9435 * If a multithreaded client connects to a malicious server using a resumed
9436 session and the server sends an ec point format extension it could write
9437 up to 255 bytes to freed memory.
9439 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9445 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9446 X509_name_oneline, X509_name_print_ex et al. to leak some information
9447 from the stack. Applications may be affected if they echo pretty printing
9448 output to the attacker.
9450 Thanks to Ivan Fratric (Google) for discovering this issue.
9453 *Emilia Käsper, and Steve Henson*
9455 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9456 for corner cases. (Certain input points at infinity could lead to
9457 bogus results, with non-infinity inputs mapped to infinity too.)
9461 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9463 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9464 handshake can force the use of weak keying material in OpenSSL
9465 SSL/TLS clients and servers.
9467 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9468 researching this issue. ([CVE-2014-0224])
9470 *KIKUCHI Masashi, Steve Henson*
9472 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9473 OpenSSL DTLS client the code can be made to recurse eventually crashing
9476 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9479 *Imre Rad, Steve Henson*
9481 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9482 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9483 client or server. This is potentially exploitable to run arbitrary
9484 code on a vulnerable client or server.
9486 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9488 *Jüri Aedla, Steve Henson*
9490 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9491 are subject to a denial of service attack.
9493 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9494 this issue. ([CVE-2014-3470])
9496 *Felix Gröbert, Ivan Fratric, Steve Henson*
9498 * Harmonize version and its documentation. -f flag is used to display
9501 *mancha <mancha1@zoho.com>*
9503 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9504 in i2d_ECPrivateKey.
9506 *mancha <mancha1@zoho.com>*
9508 * Fix some double frees. These are not thought to be exploitable.
9510 *mancha <mancha1@zoho.com>*
9512 * Fix for the attack described in the paper "Recovering OpenSSL
9513 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9514 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9515 <http://eprint.iacr.org/2014/140>
9517 Thanks to Yuval Yarom and Naomi Benger for discovering this
9518 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9520 *Yuval Yarom and Naomi Benger*
9522 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9524 * Keep original DTLS digest and encryption contexts in retransmission
9525 structures so we can use the previous session parameters if they need
9526 to be resent. ([CVE-2013-6450])
9530 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9531 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9532 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9533 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9534 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9535 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9537 *Rob Stradling, Adam Langley*
9539 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9541 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9543 This addresses the flaw in CBC record processing discovered by
9544 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9545 at: <http://www.isg.rhul.ac.uk/tls/>
9547 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9548 Security Group at Royal Holloway, University of London
9549 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9550 Emilia Käsper for the initial patch.
9553 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9555 * Return an error when checking OCSP signatures when key is NULL.
9556 This fixes a DoS attack. ([CVE-2013-0166])
9560 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9561 the right response is stapled. Also change SSL_get_certificate()
9562 so it returns the certificate actually sent.
9563 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9564 (This is a backport)
9566 *Rob Stradling <rob.stradling@comodo.com>*
9568 * Fix possible deadlock when decoding public keys.
9572 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9574 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9577 * Sanity check record length before skipping explicit IV in DTLS
9580 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9581 fuzzing as a service testing platform.
9586 * Initialise tkeylen properly when encrypting CMS messages.
9587 Thanks to Solar Designer of Openwall for reporting this issue.
9591 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9593 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9594 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9595 in CRYPTO_realloc_clean.
9597 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9598 issue and to Adam Langley <agl@chromium.org> for fixing it.
9601 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9603 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9605 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9606 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9607 content decryption and always return the same error. Note: this attack
9608 needs on average 2^20 messages so it only affects automated senders. The
9609 old behaviour can be re-enabled in the CMS code by setting the
9610 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9611 an MMA defence is not necessary.
9612 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9613 this issue. ([CVE-2012-0884])
9617 * Fix CVE-2011-4619: make sure we really are receiving a
9618 client hello before rejecting multiple SGC restarts. Thanks to
9619 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9623 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9625 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9626 Thanks to Antonio Martin, Enterprise Secure Access Research and
9627 Development, Cisco Systems, Inc. for discovering this bug and
9628 preparing a fix. ([CVE-2012-0050])
9632 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9634 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9635 of the Vaudenay padding oracle attack on CBC mode encryption
9636 which enables an efficient plaintext recovery attack against
9637 the OpenSSL implementation of DTLS. Their attack exploits timing
9638 differences arising during decryption processing. A research
9639 paper describing this attack can be found at:
9640 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9641 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9642 Security Group at Royal Holloway, University of London
9643 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9644 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9645 for preparing the fix. ([CVE-2011-4108])
9647 *Robin Seggelmann, Michael Tuexen*
9649 * Clear bytes used for block padding of SSL 3.0 records.
9652 *Adam Langley (Google)*
9654 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9655 Kadianakis <desnacked@gmail.com> for discovering this issue and
9656 Adam Langley for preparing the fix. ([CVE-2011-4619])
9658 *Adam Langley (Google)*
9660 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9662 *Andrey Kulikov <amdeich@gmail.com>*
9664 * Prevent malformed RFC3779 data triggering an assertion failure.
9665 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9666 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9668 *Rob Austein <sra@hactrn.net>*
9670 * Improved PRNG seeding for VOS.
9672 *Paul Green <Paul.Green@stratus.com>*
9674 * Fix ssl_ciph.c set-up race.
9676 *Adam Langley (Google)*
9678 * Fix spurious failures in ecdsatest.c.
9680 *Emilia Käsper (Google)*
9682 * Fix the BIO_f_buffer() implementation (which was mixing different
9683 interpretations of the `..._len` fields).
9685 *Adam Langley (Google)*
9687 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9688 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9689 threads won't reuse the same blinding coefficients.
9691 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9692 lock to call BN_BLINDING_invert_ex, and avoids one use of
9693 BN_BLINDING_update for each BN_BLINDING structure (previously,
9694 the last update always remained unused).
9696 *Emilia Käsper (Google)*
9698 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9700 *Bob Buckholz (Google)*
9702 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9704 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9705 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9707 *Kaspar Brand <ossl@velox.ch>*
9709 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9710 for multi-threaded use of ECDH. ([CVE-2011-3210])
9712 *Adam Langley (Google)*
9714 * Fix x509_name_ex_d2i memory leak on bad inputs.
9718 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9719 signature public key algorithm by using OID xref utilities instead.
9720 Before this you could only use some ECC ciphersuites with SHA1 only.
9724 * Add protection against ECDSA timing attacks as mentioned in the paper
9725 by Billy Bob Brumley and Nicola Tuveri, see:
9726 <http://eprint.iacr.org/2011/232.pdf>
9728 *Billy Bob Brumley and Nicola Tuveri*
9730 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9732 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9734 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9736 * Fix bug in string printing code: if *any* escaping is enabled we must
9737 escape the escape character (backslash) or the resulting string is
9742 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9744 * Disable code workaround for ancient and obsolete Netscape browsers
9745 and servers: an attacker can use it in a ciphersuite downgrade attack.
9746 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9750 * Fixed J-PAKE implementation error, originally discovered by
9751 Sebastien Martini, further info and confirmation from Stefan
9752 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9756 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9758 * Fix extension code to avoid race conditions which can result in a buffer
9759 overrun vulnerability: resumed sessions must not be modified as they can
9760 be shared by multiple threads. CVE-2010-3864
9764 * Fix WIN32 build system to correctly link an ENGINE directory into
9769 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9771 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9774 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9776 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9778 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9779 context. The operation can be customised via the ctrl mechanism in
9780 case ENGINEs want to include additional functionality.
9784 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9788 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9789 output hashes compatible with older versions of OpenSSL.
9791 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9793 * Fix compression algorithm handling: if resuming a session use the
9794 compression algorithm of the resumed session instead of determining
9795 it from client hello again. Don't allow server to change algorithm.
9799 * Add load_crls() function to commands tidying load_certs() too. Add option
9800 to verify utility to allow additional CRLs to be included.
9804 * Update OCSP request code to permit adding custom headers to the request:
9805 some responders need this.
9809 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9812 *Julia Lawall <julia@diku.dk>*
9814 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9815 needlessly dereferenced structures, used obsolete functions and
9816 didn't handle all updated verify codes correctly.
9820 * Disable MD2 in the default configuration.
9824 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9825 indicate the initial BIO being pushed or popped. This makes it possible
9826 to determine whether the BIO is the one explicitly called or as a result
9827 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9828 it handles reference counts correctly and doesn't zero out the I/O bio
9829 when it is not being explicitly popped. WARNING: applications which
9830 included workarounds for the old buggy behaviour will need to be modified
9831 or they could free up already freed BIOs.
9835 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9836 renaming to all platforms (within the 0.9.8 branch, this was
9837 done conditionally on Netware platforms to avoid a name clash).
9839 *Guenter <lists@gknw.net>*
9841 * Add ECDHE and PSK support to DTLS.
9843 *Michael Tuexen <tuexen@fh-muenster.de>*
9845 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9850 * Add "missing" function EVP_MD_flags() (without this the only way to
9851 retrieve a digest flags is by accessing the structure directly. Update
9852 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9853 or cipher is registered as in the "from" argument. Print out all
9854 registered digests in the dgst usage message instead of manually
9855 attempting to work them out.
9859 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9860 this allows the use of compression and extensions. Change default cipher
9861 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9862 by default unless an application cipher string requests it.
9866 * Alter match criteria in PKCS12_parse(). It used to try to use local
9867 key ids to find matching certificates and keys but some PKCS#12 files
9868 don't follow the (somewhat unwritten) rules and this strategy fails.
9869 Now just gather all certificates together and the first private key
9870 then look for the first certificate that matches the key.
9874 * Support use of registered digest and cipher names for dgst and cipher
9875 commands instead of having to add each one as a special case. So now
9882 openssl dgst -sha256 foo
9884 and this works for ENGINE based algorithms too.
9888 * Update Gost ENGINE to support parameter files.
9890 *Victor B. Wagner <vitus@cryptocom.ru>*
9892 * Support GeneralizedTime in ca utility.
9894 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9896 * Enhance the hash format used for certificate directory links. The new
9897 form uses the canonical encoding (meaning equivalent names will work
9898 even if they aren't identical) and uses SHA1 instead of MD5. This form
9899 is incompatible with the older format and as a result c_rehash should
9900 be used to rebuild symbolic links.
9904 * Make PKCS#8 the default write format for private keys, replacing the
9905 traditional format. This form is standardised, more secure and doesn't
9906 include an implicit MD5 dependency.
9910 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9911 committed to OpenSSL should pass this lot as a minimum.
9915 * Add session ticket override functionality for use by EAP-FAST.
9917 *Jouni Malinen <j@w1.fi>*
9919 * Modify HMAC functions to return a value. Since these can be implemented
9920 in an ENGINE errors can occur.
9924 * Type-checked OBJ_bsearch_ex.
9928 * Type-checked OBJ_bsearch. Also some constification necessitated
9929 by type-checking. Still to come: TXT_DB, bsearch(?),
9930 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9935 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9936 seconds to a tm structure directly, instead of going through OS
9937 specific date routines. This avoids any issues with OS routines such
9938 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9939 and X509_time_adj_ex() to cover the extended range. The existing
9940 X509_time_adj() is still usable and will no longer have any date issues.
9944 * Delta CRL support. New use deltas option which will attempt to locate
9945 and search any appropriate delta CRLs available.
9947 This work was sponsored by Google.
9951 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9952 code and add additional score elements. Validate alternate CRL paths
9953 as part of the CRL checking and indicate a new error "CRL path validation
9954 error" in this case. Applications wanting additional details can use
9955 the verify callback and check the new "parent" field. If this is not
9956 NULL CRL path validation is taking place. Existing applications won't
9957 see this because it requires extended CRL support which is off by
9960 This work was sponsored by Google.
9964 * Support for freshest CRL extension.
9966 This work was sponsored by Google.
9970 * Initial indirect CRL support. Currently only supported in the CRLs
9971 passed directly and not via lookup. Process certificate issuer
9972 CRL entry extension and lookup CRL entries by bother issuer name
9973 and serial number. Check and process CRL issuer entry in IDP extension.
9975 This work was sponsored by Google.
9979 * Add support for distinct certificate and CRL paths. The CRL issuer
9980 certificate is validated separately in this case. Only enabled if
9981 an extended CRL support flag is set: this flag will enable additional
9982 CRL functionality in future.
9984 This work was sponsored by Google.
9988 * Add support for policy mappings extension.
9990 This work was sponsored by Google.
9994 * Fixes to pathlength constraint, self issued certificate handling,
9995 policy processing to align with RFC3280 and PKITS tests.
9997 This work was sponsored by Google.
10001 * Support for name constraints certificate extension. DN, email, DNS
10002 and URI types are currently supported.
10004 This work was sponsored by Google.
10008 * To cater for systems that provide a pointer-based thread ID rather
10009 than numeric, deprecate the current numeric thread ID mechanism and
10010 replace it with a structure and associated callback type. This
10011 mechanism allows a numeric "hash" to be extracted from a thread ID in
10012 either case, and on platforms where pointers are larger than 'long',
10013 mixing is done to help ensure the numeric 'hash' is usable even if it
10014 can't be guaranteed unique. The default mechanism is to use "&errno"
10015 as a pointer-based thread ID to distinguish between threads.
10017 Applications that want to provide their own thread IDs should now use
10018 CRYPTO_THREADID_set_callback() to register a callback that will call
10019 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
10021 Note that ERR_remove_state() is now deprecated, because it is tied
10022 to the assumption that thread IDs are numeric. ERR_remove_state(0)
10023 to free the current thread's error state should be replaced by
10024 ERR_remove_thread_state(NULL).
10026 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
10027 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
10028 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
10029 application was previously providing a numeric thread callback that
10030 was inappropriate for distinguishing threads, then uniqueness might
10031 have been obtained with &errno that happened immediately in the
10032 intermediate development versions of OpenSSL; this is no longer the
10033 case, the numeric thread callback will now override the automatic use
10036 *Geoff Thorpe, with help from Bodo Moeller*
10038 * Initial support for different CRL issuing certificates. This covers a
10039 simple case where the self issued certificates in the chain exist and
10040 the real CRL issuer is higher in the existing chain.
10042 This work was sponsored by Google.
10046 * Removed effectively defunct crypto/store from the build.
10050 * Revamp of STACK to provide stronger type-checking. Still to come:
10051 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
10052 ASN1_STRING, CONF_VALUE.
10056 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
10057 RAM on SSL connections. This option can save about 34k per idle SSL.
10061 * Revamp of LHASH to provide stronger type-checking. Still to come:
10062 STACK, TXT_DB, bsearch, qsort.
10066 * Initial support for Cryptographic Message Syntax (aka CMS) based
10067 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10068 support for data, signedData, compressedData, digestedData and
10069 encryptedData, envelopedData types included. Scripts to check against
10070 RFC4134 examples draft and interop and consistency checks of many
10071 content types and variants.
10075 * Add options to enc utility to support use of zlib compression BIO.
10079 * Extend mk1mf to support importing of options and assembly language
10080 files from Configure script, currently only included in VC-WIN32.
10081 The assembly language rules can now optionally generate the source
10082 files from the associated perl scripts.
10086 * Implement remaining functionality needed to support GOST ciphersuites.
10087 Interop testing has been performed using CryptoPro implementations.
10089 *Victor B. Wagner <vitus@cryptocom.ru>*
10091 * s390x assembler pack.
10095 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10100 * Implement Opaque PRF Input TLS extension as specified in
10101 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10102 official specification yet and no extension type assignment by
10103 IANA exists, this extension (for now) will have to be explicitly
10104 enabled when building OpenSSL by providing the extension number
10105 to use. For example, specify an option
10107 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10109 to the "config" or "Configure" script to enable the extension,
10110 assuming extension number 0x9527 (which is a completely arbitrary
10111 and unofficial assignment based on the MD5 hash of the Internet
10112 Draft). Note that by doing so, you potentially lose
10113 interoperability with other TLS implementations since these might
10114 be using the same extension number for other purposes.
10116 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10117 opaque PRF input value to use in the handshake. This will create
10118 an internal copy of the length-'len' string at 'src', and will
10119 return non-zero for success.
10121 To get more control and flexibility, provide a callback function
10124 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10125 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10129 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10132 Callback function 'cb' will be called in handshakes, and is
10133 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10134 Argument 'arg' is for application purposes (the value as given to
10135 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10136 be provided to the callback function). The callback function
10137 has to return non-zero to report success: usually 1 to use opaque
10138 PRF input just if possible, or 2 to enforce use of the opaque PRF
10139 input. In the latter case, the library will abort the handshake
10140 if opaque PRF input is not successfully negotiated.
10142 Arguments 'peerinput' and 'len' given to the callback function
10143 will always be NULL and 0 in the case of a client. A server will
10144 see the client's opaque PRF input through these variables if
10145 available (NULL and 0 otherwise). Note that if the server
10146 provides an opaque PRF input, the length must be the same as the
10147 length of the client's opaque PRF input.
10149 Note that the callback function will only be called when creating
10150 a new session (session resumption can resume whatever was
10151 previously negotiated), and will not be called in SSL 2.0
10152 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10153 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10154 for applications that need to enforce opaque PRF input.
10158 * Update ssl code to support digests other than SHA1+MD5 for handshake
10161 *Victor B. Wagner <vitus@cryptocom.ru>*
10163 * Add RFC4507 support to OpenSSL. This includes the corrections in
10164 RFC4507bis. The encrypted ticket format is an encrypted encoded
10165 SSL_SESSION structure, that way new session features are automatically
10168 If a client application caches session in an SSL_SESSION structure
10169 support is transparent because tickets are now stored in the encoded
10172 The SSL_CTX structure automatically generates keys for ticket
10173 protection in servers so again support should be possible
10174 with no application modification.
10176 If a client or server wishes to disable RFC4507 support then the option
10177 SSL_OP_NO_TICKET can be set.
10179 Add a TLS extension debugging callback to allow the contents of any client
10180 or server extensions to be examined.
10182 This work was sponsored by Google.
10186 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10187 OpenSSL should now compile cleanly on gcc 4.2
10189 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10191 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10192 support including streaming MAC support: this is required for GOST
10193 ciphersuite support.
10195 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10197 * Add option -stream to use PKCS#7 streaming in smime utility. New
10198 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10199 to output in BER and PEM format.
10203 * Experimental support for use of HMAC via EVP_PKEY interface. This
10204 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10205 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10206 ENGINE support for HMAC keys which are unextractable. New -mac and
10207 -macopt options to dgst utility.
10211 * New option -sigopt to dgst utility. Update dgst to use
10212 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10213 alternative signing parameters such as X9.31 or PSS in the dgst
10218 * Change ssl_cipher_apply_rule(), the internal function that does
10219 the work each time a ciphersuite string requests enabling
10220 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10221 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10222 the order of disabled ciphersuites such that those ciphersuites
10223 that most recently went from enabled to disabled not only stay
10224 in order with respect to each other, but also have higher priority
10225 than other disabled ciphersuites the next time ciphersuites are
10228 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10229 the same ciphersuites as with "HIGH" alone, but in a specific
10230 order where the PSK ciphersuites come first (since they are the
10231 most recently disabled ciphersuites when "HIGH" is parsed).
10233 Also, change ssl_create_cipher_list() (using this new
10234 functionality) such that between otherwise identical
10235 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10240 * Change ssl_create_cipher_list() so that it automatically
10241 arranges the ciphersuites in reasonable order before starting
10242 to process the rule string. Thus, the definition for "DEFAULT"
10243 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10244 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10245 This makes it much easier to arrive at a reasonable default order
10246 in applications for which anonymous ciphers are OK (meaning
10247 that you can't actually use DEFAULT).
10249 *Bodo Moeller; suggested by Victor Duchovni*
10251 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10252 processing) into multiple integers instead of setting
10253 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10254 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10255 (These masks as well as the individual bit definitions are hidden
10256 away into the non-exported interface ssl/ssl_locl.h, so this
10257 change to the definition of the SSL_CIPHER structure shouldn't
10258 affect applications.) This give us more bits for each of these
10259 categories, so there is no longer a need to coagulate AES128 and
10260 AES256 into a single algorithm bit, and to coagulate Camellia128
10261 and Camellia256 into a single algorithm bit, which has led to all
10264 Thus, among other things, the kludge introduced in 0.9.7m and
10265 0.9.8e for masking out AES256 independently of AES128 or masking
10266 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10268 With the change, we also introduce new ciphersuite aliases that
10269 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10274 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10275 Use the leftmost N bytes of the signature input if the input is
10276 larger than the prime q (with N being the size in bytes of q).
10280 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10281 it yet and it is largely untested.
10285 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10289 * Initial incomplete changes to avoid need for function casts in OpenSSL
10290 some compilers (gcc 4.2 and later) reject their use. Safestack is
10291 reimplemented. Update ASN1 to avoid use of legacy functions.
10295 * Win32/64 targets are linked with Winsock2.
10299 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10300 to external functions. This can be used to increase CRL handling
10301 efficiency especially when CRLs are very large by (for example) storing
10302 the CRL revoked certificates in a database.
10306 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10307 new CRLs added to a directory can be used. New command line option
10308 -verify_return_error to s_client and s_server. This causes real errors
10309 to be returned by the verify callback instead of carrying on no matter
10310 what. This reflects the way a "real world" verify callback would behave.
10314 * GOST engine, supporting several GOST algorithms and public key formats.
10315 Kindly donated by Cryptocom.
10319 * Partial support for Issuing Distribution Point CRL extension. CRLs
10320 partitioned by DP are handled but no indirect CRL or reason partitioning
10321 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10322 selected via a scoring technique which handles IDP and AKID in CRLs.
10326 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10327 will ultimately be used for all verify operations: this will remove the
10328 X509_STORE dependency on certificate verification and allow alternative
10329 lookup methods. X509_STORE based implementations of these two callbacks.
10333 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10334 Modify get_crl() to find a valid (unexpired) CRL if possible.
10338 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10339 this would be called X509_CRL_cmp() but that name is already used by
10340 a function that just compares CRL issuer names. Cache several CRL
10341 extensions in X509_CRL structure and cache CRLDP in X509.
10345 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10346 this maps equivalent X509_NAME structures into a consistent structure.
10347 Name comparison can then be performed rapidly using memcmp().
10351 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10356 * Allow digests to supply their own micalg string for S/MIME type using
10357 the ctrl EVP_MD_CTRL_MICALG.
10361 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10362 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10363 ctrl. It can then customise the structure before and/or after signing
10368 * New function OBJ_add_sigid() to allow application defined signature OIDs
10369 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10370 to free up any added signature OIDs.
10374 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10375 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10376 digest and cipher tables. New options added to openssl utility:
10377 list-message-digest-algorithms and list-cipher-algorithms.
10381 * Change the array representation of binary polynomials: the list
10382 of degrees of non-zero coefficients is now terminated with -1.
10383 Previously it was terminated with 0, which was also part of the
10384 value; thus, the array representation was not applicable to
10385 polynomials where t^0 has coefficient zero. This change makes
10386 the array representation useful in a more general context.
10390 * Various modifications and fixes to SSL/TLS cipher string
10391 handling. For ECC, the code now distinguishes between fixed ECDH
10392 with RSA certificates on the one hand and with ECDSA certificates
10393 on the other hand, since these are separate ciphersuites. The
10394 unused code for Fortezza ciphersuites has been removed.
10396 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10397 (not "ECDHE"). For consistency with the code for DH
10398 certificates, use of ECDH certificates is now considered ECDH
10399 authentication, not RSA or ECDSA authentication (the latter is
10400 merely the CA's signing algorithm and not actively used in the
10403 The temporary ciphersuite alias "ECCdraft" is no longer
10404 available, and ECC ciphersuites are no longer excluded from "ALL"
10405 and "DEFAULT". The following aliases now exist for RFC 4492
10406 ciphersuites, most of these by analogy with the DH case:
10408 kECDHr - ECDH cert, signed with RSA
10409 kECDHe - ECDH cert, signed with ECDSA
10410 kECDH - ECDH cert (signed with either RSA or ECDSA)
10411 kEECDH - ephemeral ECDH
10412 ECDH - ECDH cert or ephemeral ECDH
10415 aECDSA - ECDSA cert
10418 AECDH - anonymous ECDH
10419 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10423 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10424 Use correct micalg parameters depending on digest(s) in signed message.
10428 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10429 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10433 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10434 an engine to register a method. Add ENGINE lookups for methods and
10435 functional reference processing.
10439 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10440 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10445 * New -resign option to smime utility. This adds one or more signers
10446 to an existing PKCS#7 signedData structure. Also -md option to use an
10447 alternative message digest algorithm for signing.
10451 * Tidy up PKCS#7 routines and add new functions to make it easier to
10452 create PKCS7 structures containing multiple signers. Update smime
10453 application to support multiple signers.
10457 * New -macalg option to pkcs12 utility to allow setting of an alternative
10462 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10463 Reorganize PBE internals to lookup from a static table using NIDs,
10464 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10465 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10466 PRF which will be automatically used with PBES2.
10470 * Replace the algorithm specific calls to generate keys in "req" with the
10475 * Update PKCS#7 enveloped data routines to use new API. This is now
10476 supported by any public key method supporting the encrypt operation. A
10477 ctrl is added to allow the public key algorithm to examine or modify
10478 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10483 * Add a ctrl to asn1 method to allow a public key algorithm to express
10484 a default digest type to use. In most cases this will be SHA1 but some
10485 algorithms (such as GOST) need to specify an alternative digest. The
10486 return value indicates how strong the preference is 1 means optional and
10487 2 is mandatory (that is it is the only supported type). Modify
10488 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10489 use the default md. Update openssl utilities to use the default digest
10490 type for signing if it is not explicitly indicated.
10494 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10495 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10496 signing method from the key type. This effectively removes the link
10497 between digests and public key types.
10501 * Add an OID cross reference table and utility functions. Its purpose is to
10502 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10503 rsaEncryption. This will allow some of the algorithm specific hackery
10504 needed to use the correct OID to be removed.
10508 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10509 structures for PKCS7_sign(). They are now set up by the relevant public
10514 * Add provisional EC pkey method with support for ECDSA and ECDH.
10518 * Add support for key derivation (agreement) in the API, DH method and
10523 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10524 public and private key formats. As a side effect these add additional
10525 command line functionality not previously available: DSA signatures can be
10526 generated and verified using pkeyutl and DH key support and generation in
10533 *Oliver Tappe <zooey@hirschkaefer.de>*
10535 * New make target "install_html_docs" installs HTML renditions of the
10538 *Oliver Tappe <zooey@hirschkaefer.de>*
10540 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10541 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10542 support key and parameter generation and add initial key generation
10543 functionality for RSA.
10547 * Add functions for main EVP_PKEY_method operations. The undocumented
10548 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10549 `EVP_PKEY_{encrypt,decrypt}_old`.
10553 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10554 key API, doesn't do much yet.
10558 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10559 public key algorithms. New option to openssl utility:
10560 "list-public-key-algorithms" to print out info.
10564 * Implement the Supported Elliptic Curves Extension for
10565 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10569 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10570 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10574 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10575 utilities such as rsa, dsa, dsaparam etc except they process any key
10580 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10581 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10582 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10587 * Initial support for pluggable public key ASN1.
10588 De-spaghettify the public key ASN1 handling. Move public and private
10589 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10590 algorithm specific handling to a single module within the relevant
10591 algorithm directory. Add functions to allow (near) opaque processing
10592 of public and private key structures.
10596 * Implement the Supported Point Formats Extension for
10597 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10601 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10602 for the psk identity [hint] and the psk callback functions to the
10603 SSL_SESSION, SSL and SSL_CTX structure.
10606 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10610 SSL_CTX_use_psk_identity_hint
10611 SSL_get_psk_identity_hint
10612 SSL_get_psk_identity
10613 SSL_use_psk_identity_hint
10615 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10617 * Add RFC 3161 compliant time stamp request creation, response generation
10618 and response verification functionality.
10620 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10622 * Add initial support for TLS extensions, specifically for the server_name
10623 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10624 have new members for a hostname. The SSL data structure has an
10625 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10626 stored in that context to allow for session resumption, even after the
10627 SSL has been switched to a new SSL_CTX in reaction to a client's
10628 server_name extension.
10630 New functions (subject to change):
10632 SSL_get_servername()
10633 SSL_get_servername_type()
10636 New CTRL codes and macros (subject to change):
10638 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10639 - SSL_CTX_set_tlsext_servername_callback()
10640 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10641 - SSL_CTX_set_tlsext_servername_arg()
10642 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10644 openssl s_client has a new '-servername ...' option.
10646 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10647 '-key2 ...', '-servername_fatal' (subject to change). This allows
10648 testing the HostName extension for a specific single hostname ('-cert'
10649 and '-key' remain fallbacks for handshakes without HostName
10650 negotiation). If the unrecognized_name alert has to be sent, this by
10651 default is a warning; it becomes fatal with the '-servername_fatal'
10654 *Peter Sylvester, Remy Allais, Christophe Renou*
10656 * Whirlpool hash implementation is added.
10660 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10661 bn(64,32). Because of instruction set limitations it doesn't have
10662 any negative impact on performance. This was done mostly in order
10663 to make it possible to share assembler modules, such as bn_mul_mont
10664 implementations, between 32- and 64-bit builds without hassle.
10668 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10669 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10674 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10675 dedicated Montgomery multiplication procedure, is introduced.
10676 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10677 "64-bit" performance on certain 32-bit targets.
10681 * New option SSL_OP_NO_COMP to disable use of compression selectively
10682 in SSL structures. New SSL ctrl to set maximum send fragment size.
10683 Save memory by setting the I/O buffer sizes dynamically instead of
10684 using the maximum available value.
10688 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10689 in addition to the text details.
10693 * Very, very preliminary EXPERIMENTAL support for printing of general
10694 ASN1 structures. This currently produces rather ugly output and doesn't
10695 handle several customised structures at all.
10699 * Integrated support for PVK file format and some related formats such
10700 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10701 these in the 'rsa' and 'dsa' utilities.
10705 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10709 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10710 place for the (very old) "NETSCAPE" format certificates which are now
10711 handled using new ASN1 code equivalents.
10715 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10716 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10717 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10721 * Modify CRL distribution points extension code to print out previously
10722 unsupported fields. Enhance extension setting code to allow setting of
10727 * Add print and set support for Issuing Distribution Point CRL extension.
10731 * Change 'Configure' script to enable Camellia by default.
10738 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10740 * When rejecting SSL/TLS records due to an incorrect version number, never
10741 update s->server with a new major version number. As of
10742 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10743 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10744 the previous behavior could result in a read attempt at NULL when
10745 receiving specific incorrect SSL/TLS records once record payload
10746 protection is active. ([CVE-2010-0740])
10748 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10750 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10751 could be crashed if the relevant tables were not present (e.g. chrooted).
10753 *Tomas Hoger <thoger@redhat.com>*
10755 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10757 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10759 *Martin Olsson, Neel Mehta*
10761 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10762 accommodate for stack sorting, always a write lock!).
10766 * On some versions of WIN32 Heap32Next is very slow. This can cause
10767 excessive delays in the RAND_poll(): over a minute. As a workaround
10768 include a time check in the inner Heap32Next loop too.
10772 * The code that handled flushing of data in SSL/TLS originally used the
10773 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10774 the problem outlined in PR#1949. The fix suggested there however can
10775 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10776 of Apache). So instead simplify the code to flush unconditionally.
10777 This should be fine since flushing with no data to flush is a no op.
10781 * Handle TLS versions 2.0 and later properly and correctly use the
10782 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10783 off ancient servers have a habit of sticking around for a while...
10787 * Modify compression code so it frees up structures without using the
10788 ex_data callbacks. This works around a problem where some applications
10789 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10790 restarting) then use compression (e.g. SSL with compression) later.
10791 This results in significant per-connection memory leaks and
10792 has caused some security issues including CVE-2008-1678 and
10797 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10798 change when encrypting or decrypting.
10802 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10803 connect and renegotiate with servers which do not support RI.
10804 Until RI is more widely deployed this option is enabled by default.
10808 * Add "missing" ssl ctrls to clear options and mode.
10812 * If client attempts to renegotiate and doesn't support RI respond with
10813 a no_renegotiation alert as required by RFC5746. Some renegotiating
10814 TLS clients will continue a connection gracefully when they receive
10815 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10816 waiting for a server hello which it will never receive. Now we treat a
10817 received no_renegotiation alert as a fatal error. This is because
10818 applications requesting a renegotiation might well expect it to succeed
10819 and would have no code in place to handle the server denying it so the
10820 only safe thing to do is to terminate the connection.
10824 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10825 peer supports secure renegotiation and 0 otherwise. Print out peer
10826 renegotiation support in s_client/s_server.
10830 * Replace the highly broken and deprecated SPKAC certification method with
10831 the updated NID creation version. This should correctly handle UTF8.
10835 * Implement RFC5746. Re-enable renegotiation but require the extension
10836 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10837 turns out to be a bad idea. It has been replaced by
10838 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10839 SSL_CTX_set_options(). This is really not recommended unless you
10840 know what you are doing.
10842 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10844 * Fixes to stateless session resumption handling. Use initial_ctx when
10845 issuing and attempting to decrypt tickets in case it has changed during
10846 servername handling. Use a non-zero length session ID when attempting
10847 stateless session resumption: this makes it possible to determine if
10848 a resumption has occurred immediately after receiving server hello
10849 (several places in OpenSSL subtly assume this) instead of later in
10854 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10855 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10856 fixes for a few places where the return code is not checked
10859 *Julia Lawall <julia@diku.dk>*
10861 * Add --strict-warnings option to Configure script to include devteam
10862 warnings in other configurations.
10866 * Add support for --libdir option and LIBDIR variable in makefiles. This
10867 makes it possible to install openssl libraries in locations which
10868 have names other than "lib", for example "/usr/lib64" which some
10871 *Steve Henson, based on patch from Jeremy Utley*
10873 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10874 X690 8.9.12 and can produce some misleading textual output of OIDs.
10876 *Steve Henson, reported by Dan Kaminsky*
10878 * Delete MD2 from algorithm tables. This follows the recommendation in
10879 several standards that it is not used in new applications due to
10880 several cryptographic weaknesses. For binary compatibility reasons
10881 the MD2 API is still compiled in by default.
10885 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10890 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10891 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10894 *Guenter <lists@gknw.net>*
10896 * Fix the server certificate chain building code to use X509_verify_cert(),
10897 it used to have an ad-hoc builder which was unable to cope with anything
10898 other than a simple chain.
10900 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10902 * Don't check self signed certificate signatures in X509_verify_cert()
10903 by default (a flag can override this): it just wastes time without
10904 adding any security. As a useful side effect self signed root CAs
10905 with non-FIPS digests are now usable in FIPS mode.
10909 * In dtls1_process_out_of_seq_message() the check if the current message
10910 is already buffered was missing. For every new message was memory
10911 allocated, allowing an attacker to perform an denial of service attack
10912 with sending out of seq handshake messages until there is no memory
10913 left. Additionally every future message was buffered, even if the
10914 sequence number made no sense and would be part of another handshake.
10915 So only messages with sequence numbers less than 10 in advance will be
10916 buffered. ([CVE-2009-1378])
10918 *Robin Seggelmann, discovered by Daniel Mentz*
10920 * Records are buffered if they arrive with a future epoch to be
10921 processed after finishing the corresponding handshake. There is
10922 currently no limitation to this buffer allowing an attacker to perform
10923 a DOS attack with sending records with future epochs until there is no
10924 memory left. This patch adds the pqueue_size() function to determine
10925 the size of a buffer and limits the record buffer to 100 entries.
10928 *Robin Seggelmann, discovered by Daniel Mentz*
10930 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10931 parent structure is freed. ([CVE-2009-1379])
10935 * Handle non-blocking I/O properly in SSL_shutdown() call.
10937 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10939 * Add `2.5.4.*` OIDs
10941 *Ilya O. <vrghost@gmail.com>*
10943 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10945 * Disable renegotiation completely - this fixes a severe security
10946 problem ([CVE-2009-3555]) at the cost of breaking all
10947 renegotiation. Renegotiation can be re-enabled by setting
10948 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10949 run-time. This is really not recommended unless you know what
10954 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10956 * Don't set val to NULL when freeing up structures, it is freed up by
10957 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10958 zeroing past the valid field. ([CVE-2009-0789])
10960 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10962 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10963 checked correctly. This would allow some invalid signed attributes to
10964 appear to verify correctly. ([CVE-2009-0591])
10966 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10968 * Reject UniversalString and BMPString types with invalid lengths. This
10969 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10970 a legal length. ([CVE-2009-0590])
10974 * Set S/MIME signing as the default purpose rather than setting it
10975 unconditionally. This allows applications to override it at the store
10980 * Permit restricted recursion of ASN1 strings. This is needed in practice
10981 to handle some structures.
10985 * Improve efficiency of mem_gets: don't search whole buffer each time
10988 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10990 * New -hex option for openssl rand.
10994 * Print out UTF8String and NumericString when parsing ASN1.
10998 * Support NumericString type for name components.
11002 * Allow CC in the environment to override the automatically chosen
11003 compiler. Note that nothing is done to ensure flags work with the
11008 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
11010 * Properly check EVP_VerifyFinal() and similar return values
11013 *Ben Laurie, Bodo Moeller, Google Security Team*
11015 * Enable TLS extensions by default.
11019 * Allow the CHIL engine to be loaded, whether the application is
11020 multithreaded or not. (This does not release the developer from the
11021 obligation to set up the dynamic locking callbacks.)
11023 *Sander Temme <sander@temme.net>*
11025 * Use correct exit code if there is an error in dgst command.
11027 *Steve Henson; problem pointed out by Roland Dirlewanger*
11029 * Tweak Configure so that you need to say "experimental-jpake" to enable
11030 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
11034 * Add experimental JPAKE support, including demo authentication in
11035 s_client and s_server.
11039 * Set the comparison function in v3_addr_canonize().
11041 *Rob Austein <sra@hactrn.net>*
11043 * Add support for XMPP STARTTLS in s_client.
11045 *Philip Paeps <philip@freebsd.org>*
11047 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
11048 to ensure that even with this option, only ciphersuites in the
11049 server's preference list will be accepted. (Note that the option
11050 applies only when resuming a session, so the earlier behavior was
11051 just about the algorithm choice for symmetric cryptography.)
11055 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
11057 * Fix NULL pointer dereference if a DTLS server received
11058 ChangeCipherSpec as first record ([CVE-2009-1386]).
11062 * Fix a state transition in s3_srvr.c and d1_srvr.c
11063 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
11065 *Nagendra Modadugu*
11067 * The fix in 0.9.8c that supposedly got rid of unsafe
11068 double-checked locking was incomplete for RSA blinding,
11069 addressing just one layer of what turns out to have been
11070 doubly unsafe triple-checked locking.
11072 So now fix this for real by retiring the MONT_HELPER macro
11073 in crypto/rsa/rsa_eay.c.
11075 *Bodo Moeller; problem pointed out by Marius Schilder*
11077 * Various precautionary measures:
11079 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11081 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11082 (NB: This would require knowledge of the secret session ticket key
11083 to exploit, in which case you'd be SOL either way.)
11085 - Change bn_nist.c so that it will properly handle input BIGNUMs
11086 outside the expected range.
11088 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11091 *Neel Mehta, Bodo Moeller*
11093 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11094 the load fails. Useful for distros.
11096 *Ben Laurie and the FreeBSD team*
11098 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11102 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11106 * Expand ENGINE to support engine supplied SSL client certificate functions.
11108 This work was sponsored by Logica.
11112 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11113 keystores. Support for SSL/TLS client authentication too.
11114 Not compiled unless enable-capieng specified to Configure.
11116 This work was sponsored by Logica.
11120 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11121 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11122 attribute creation routines such as certificate requests and PKCS#12
11127 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11129 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11130 handshake which could lead to a client crash as found using the
11131 Codenomicon TLS test suite ([CVE-2008-1672])
11133 *Steve Henson, Mark Cox*
11135 * Fix double free in TLS server name extensions which could lead to
11136 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11140 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11142 Clear the error queue to ensure that error entries left from
11143 older function calls do not interfere with the correct operation.
11145 *Lutz Jaenicke, Erik de Castro Lopo*
11147 * Remove root CA certificates of commercial CAs:
11149 The OpenSSL project does not recommend any specific CA and does not
11150 have any policy with respect to including or excluding any CA.
11151 Therefore, it does not make any sense to ship an arbitrary selection
11152 of root CA certificates with the OpenSSL software.
11156 * RSA OAEP patches to fix two separate invalid memory reads.
11157 The first one involves inputs when 'lzero' is greater than
11158 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11159 before the beginning of from). The second one involves inputs where
11160 the 'db' section contains nothing but zeroes (there is a one-byte
11161 invalid read after the end of 'db').
11163 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11165 * Partial backport from 0.9.9-dev:
11167 Introduce bn_mul_mont (dedicated Montgomery multiplication
11168 procedure) as a candidate for BIGNUM assembler implementation.
11169 While 0.9.9-dev uses assembler for various architectures, only
11170 x86_64 is available by default here in the 0.9.8 branch, and
11171 32-bit x86 is available through a compile-time setting.
11173 To try the 32-bit x86 assembler implementation, use Configure
11174 option "enable-montasm" (which exists only for this backport).
11176 As "enable-montasm" for 32-bit x86 disclaims code stability
11177 anyway, in this constellation we activate additional code
11178 backported from 0.9.9-dev for further performance improvements,
11179 namely BN_from_montgomery_word. (To enable this otherwise,
11180 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11182 *Andy Polyakov (backport partially by Bodo Moeller)*
11184 * Add TLS session ticket callback. This allows an application to set
11185 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11186 values. This is useful for key rollover for example where several key
11187 sets may exist with different names.
11191 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11192 This was broken until now in 0.9.8 releases, such that the only way
11193 a registered ENGINE could be used (assuming it initialises
11194 successfully on the host) was to explicitly set it as the default
11195 for the relevant algorithms. This is in contradiction with 0.9.7
11196 behaviour and the documentation. With this fix, when an ENGINE is
11197 registered into a given algorithm's table of implementations, the
11198 'uptodate' flag is reset so that auto-discovery will be used next
11199 time a new context for that algorithm attempts to select an
11202 *Ian Lister (tweaked by Geoff Thorpe)*
11204 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11205 implementation in the following ways:
11207 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11210 Lack of BER streaming support means one pass streaming processing is
11211 only supported if data is detached: setting the streaming flag is
11212 ignored for embedded content.
11214 CMS support is disabled by default and must be explicitly enabled
11215 with the enable-cms configuration option.
11219 * Update the GMP engine glue to do direct copies between BIGNUM and
11220 mpz_t when openssl and GMP use the same limb size. Otherwise the
11221 existing "conversion via a text string export" trick is still used.
11223 *Paul Sheer <paulsheer@gmail.com>*
11225 * Zlib compression BIO. This is a filter BIO which compressed and
11226 uncompresses any data passed through it.
11230 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11231 RFC3394 compatible AES key wrapping.
11235 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11236 sets string data without copying. X509_ALGOR_set0() and
11237 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11238 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11239 from an X509_ATTRIBUTE structure optionally checking it occurs only
11240 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11245 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11246 to get the expected BN_FLG_CONSTTIME behavior.
11248 *Bodo Moeller (Google)*
11252 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11253 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11254 - added some more tests to do_tests.pl
11255 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11256 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11257 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11258 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11259 - various changes to netware.pl to enable gcc-cross builds on Win32
11261 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11262 - various changes to fix missing prototype warnings
11263 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11264 - added AES, WHIRLPOOL and CPUID assembler code to build files
11265 - added missing AES assembler make rules to mk1mf.pl
11266 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11268 *Guenter Knauf <eflash@gmx.net>*
11270 * Implement certificate status request TLS extension defined in RFC3546.
11271 A client can set the appropriate parameters and receive the encoded
11272 OCSP response via a callback. A server can query the supplied parameters
11273 and set the encoded OCSP response in the callback. Add simplified examples
11274 to s_client and s_server.
11278 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11280 * Fix various bugs:
11281 + Binary incompatibility of ssl_ctx_st structure
11282 + DTLS interoperation with non-compliant servers
11283 + Don't call get_session_cb() without proposed session
11284 + Fix ia64 assembler code
11286 *Andy Polyakov, Steve Henson*
11288 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11290 * DTLS Handshake overhaul. There were longstanding issues with
11291 OpenSSL DTLS implementation, which were making it impossible for
11292 RFC 4347 compliant client to communicate with OpenSSL server.
11293 Unfortunately just fixing these incompatibilities would "cut off"
11294 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11295 server keeps tolerating non RFC compliant syntax. The opposite is
11296 not true, 0.9.8f client can not communicate with earlier server.
11297 This update even addresses CVE-2007-4995.
11301 * Changes to avoid need for function casts in OpenSSL: some compilers
11302 (gcc 4.2 and later) reject their use.
11303 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11306 * Add RFC4507 support to OpenSSL. This includes the corrections in
11307 RFC4507bis. The encrypted ticket format is an encrypted encoded
11308 SSL_SESSION structure, that way new session features are automatically
11311 If a client application caches session in an SSL_SESSION structure
11312 support is transparent because tickets are now stored in the encoded
11315 The SSL_CTX structure automatically generates keys for ticket
11316 protection in servers so again support should be possible
11317 with no application modification.
11319 If a client or server wishes to disable RFC4507 support then the option
11320 SSL_OP_NO_TICKET can be set.
11322 Add a TLS extension debugging callback to allow the contents of any client
11323 or server extensions to be examined.
11325 This work was sponsored by Google.
11329 * Add initial support for TLS extensions, specifically for the server_name
11330 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11331 have new members for a hostname. The SSL data structure has an
11332 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11333 stored in that context to allow for session resumption, even after the
11334 SSL has been switched to a new SSL_CTX in reaction to a client's
11335 server_name extension.
11337 New functions (subject to change):
11339 SSL_get_servername()
11340 SSL_get_servername_type()
11343 New CTRL codes and macros (subject to change):
11345 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11346 - SSL_CTX_set_tlsext_servername_callback()
11347 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11348 - SSL_CTX_set_tlsext_servername_arg()
11349 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11351 openssl s_client has a new '-servername ...' option.
11353 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11354 '-key2 ...', '-servername_fatal' (subject to change). This allows
11355 testing the HostName extension for a specific single hostname ('-cert'
11356 and '-key' remain fallbacks for handshakes without HostName
11357 negotiation). If the unrecognized_name alert has to be sent, this by
11358 default is a warning; it becomes fatal with the '-servername_fatal'
11361 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11363 * Add AES and SSE2 assembly language support to VC++ build.
11367 * Mitigate attack on final subtraction in Montgomery reduction.
11371 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11372 (which previously caused an internal error).
11376 * Squeeze another 10% out of IGE mode when in != out.
11380 * AES IGE mode speedup.
11382 *Dean Gaudet (Google)*
11384 * Add the Korean symmetric 128-bit cipher SEED (see
11385 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11386 add SEED ciphersuites from RFC 4162:
11388 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11389 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11390 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11391 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11393 To minimize changes between patchlevels in the OpenSSL 0.9.8
11394 series, SEED remains excluded from compilation unless OpenSSL
11395 is configured with 'enable-seed'.
11397 *KISA, Bodo Moeller*
11399 * Mitigate branch prediction attacks, which can be practical if a
11400 single processor is shared, allowing a spy process to extract
11401 information. For detailed background information, see
11402 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11403 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11404 and Necessary Software Countermeasures"). The core of the change
11405 are new versions BN_div_no_branch() and
11406 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11407 respectively, which are slower, but avoid the security-relevant
11408 conditional branches. These are automatically called by BN_div()
11409 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11410 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11411 remove a conditional branch.
11413 BN_FLG_CONSTTIME is the new name for the previous
11414 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11415 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11416 in the exponent causes BN_mod_exp_mont() to use the alternative
11417 implementation in BN_mod_exp_mont_consttime().) The old name
11418 remains as a deprecated alias.
11420 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11421 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11422 constant-time implementations for more than just exponentiation.
11423 Here too the old name is kept as a deprecated alias.
11425 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11426 the BN_BLINDING structure gets an independent copy of the
11427 modulus. This means that the previous `BIGNUM *m` argument to
11428 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11429 essentially becomes `const BIGNUM *m`, although we can't actually
11430 change this in the header file before 0.9.9. It allows
11431 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11432 enable BN_FLG_CONSTTIME.
11434 *Matthew D Wood (Intel Corp)*
11436 * In the SSL/TLS server implementation, be strict about session ID
11437 context matching (which matters if an application uses a single
11438 external cache for different purposes). Previously,
11439 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11440 set. This did ensure strict client verification, but meant that,
11441 with applications using a single external cache for quite
11442 different requirements, clients could circumvent ciphersuite
11443 restrictions for a given session ID context by starting a session
11444 in a different context.
11448 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11449 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11450 authentication-only ciphersuites.
11454 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11455 not complete and could lead to a possible single byte overflow
11456 ([CVE-2007-5135]) [Ben Laurie]
11458 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11460 * Since AES128 and AES256 (and similarly Camellia128 and
11461 Camellia256) share a single mask bit in the logic of
11462 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11463 kludge to work properly if AES128 is available and AES256 isn't
11464 (or if Camellia128 is available and Camellia256 isn't).
11468 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11469 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11470 When a point or a seed is encoded in a BIT STRING, we need to
11471 prevent the removal of trailing zero bits to get the proper DER
11472 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11473 of a NamedBitList, for which trailing 0 bits need to be removed.)
11477 * Have SSL/TLS server implementation tolerate "mismatched" record
11478 protocol version while receiving ClientHello even if the
11479 ClientHello is fragmented. (The server can't insist on the
11480 particular protocol version it has chosen before the ServerHello
11481 message has informed the client about his choice.)
11485 * Add RFC 3779 support.
11487 *Rob Austein for ARIN, Ben Laurie*
11489 * Load error codes if they are not already present instead of using a
11490 static variable. This allows them to be cleanly unloaded and reloaded.
11491 Improve header file function name parsing.
11495 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11496 or CAPABILITY handshake as required by RFCs.
11498 *Goetz Babin-Ebell*
11500 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11502 * Introduce limits to prevent malicious keys being able to
11503 cause a denial of service. ([CVE-2006-2940])
11505 *Steve Henson, Bodo Moeller*
11507 * Fix ASN.1 parsing of certain invalid structures that can result
11508 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11510 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11511 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11513 * Fix SSL client code which could crash if connecting to a
11514 malicious SSLv2 server. ([CVE-2006-4343])
11516 *Tavis Ormandy and Will Drewry, Google Security Team*
11518 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11519 match only those. Before that, "AES256-SHA" would be interpreted
11520 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11521 the same strength classification in 0.9.7h) as we currently only
11522 have a single AES bit in the ciphersuite description bitmap.
11523 That change, however, also applied to ciphersuite strings such as
11524 "RC4-MD5" that intentionally matched multiple ciphersuites --
11525 namely, SSL 2.0 ciphersuites in addition to the more common ones
11526 from SSL 3.0/TLS 1.0.
11528 So we change the selection algorithm again: Naming an explicit
11529 ciphersuite selects this one ciphersuite, and any other similar
11530 ciphersuite (same bitmap) from *other* protocol versions.
11531 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11532 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11534 Since SSL 2.0 does not have any ciphersuites for which the
11535 128/256 bit distinction would be relevant, this works for now.
11536 The proper fix will be to use different bits for AES128 and
11537 AES256, which would have avoided the problems from the beginning;
11538 however, bits are scarce, so we can only do this in a new release
11539 (not just a patchlevel) when we can change the SSL_CIPHER
11540 definition to split the single 'unsigned long mask' bitmap into
11541 multiple values to extend the available space.
11545 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11547 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11548 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11550 * Add AES IGE and biIGE modes.
11554 * Change the Unix randomness entropy gathering to use poll() when
11555 possible instead of select(), since the latter has some
11556 undesirable limitations.
11558 *Darryl Miles via Richard Levitte and Bodo Moeller*
11560 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11561 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11562 cannot be implicitly activated as part of, e.g., the "AES" alias.
11563 However, please upgrade to OpenSSL 0.9.9[-dev] for
11564 non-experimental use of the ECC ciphersuites to get TLS extension
11565 support, which is required for curve and point format negotiation
11566 to avoid potential handshake problems.
11570 * Disable rogue ciphersuites:
11572 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11573 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11574 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11576 The latter two were purportedly from
11577 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11580 Also deactivate the remaining ciphersuites from
11581 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11582 unofficial, and the ID has long expired.
11586 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11587 dual-core machines) and other potential thread-safety issues.
11591 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11592 versions), which is now available for royalty-free use
11593 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11594 Also, add Camellia TLS ciphersuites from RFC 4132.
11596 To minimize changes between patchlevels in the OpenSSL 0.9.8
11597 series, Camellia remains excluded from compilation unless OpenSSL
11598 is configured with 'enable-camellia'.
11602 * Disable the padding bug check when compression is in use. The padding
11603 bug check assumes the first packet is of even length, this is not
11604 necessarily true if compression is enabled and can result in false
11605 positives causing handshake failure. The actual bug test is ancient
11606 code so it is hoped that implementations will either have fixed it by
11607 now or any which still have the bug do not support compression.
11611 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11613 * When applying a cipher rule check to see if string match is an explicit
11614 cipher suite and only match that one cipher suite if it is.
11618 * Link in manifests for VC++ if needed.
11620 *Austin Ziegler <halostatue@gmail.com>*
11622 * Update support for ECC-based TLS ciphersuites according to
11623 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11624 TLS extensions, which are supported starting with the 0.9.9
11625 branch, not in the OpenSSL 0.9.8 branch).
11629 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11630 opaque EVP_CIPHER_CTX handling.
11634 * Fixes and enhancements to zlib compression code. We now only use
11635 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11636 to conform with the standards mentioned here:
11637 <http://www.zlib.net/DLL_FAQ.txt>
11638 Static zlib linking now works on Windows and the new --with-zlib-include
11639 --with-zlib-lib options to Configure can be used to supply the location
11640 of the headers and library. Gracefully handle case where zlib library
11645 * Several fixes and enhancements to the OID generation code. The old code
11646 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11647 handle numbers larger than ULONG_MAX, truncated printing and had a
11648 non standard OBJ_obj2txt() behaviour.
11652 * Add support for building of engines under engine/ as shared libraries
11653 under VC++ build system.
11657 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11658 Hopefully, we will not see any false combination of paths any more.
11662 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11664 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11665 (part of SSL_OP_ALL). This option used to disable the
11666 countermeasure against man-in-the-middle protocol-version
11667 rollback in the SSL 2.0 server implementation, which is a bad
11668 idea. ([CVE-2005-2969])
11670 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11671 for Information Security, National Institute of Advanced Industrial
11672 Science and Technology [AIST], Japan)*
11674 * Add two function to clear and return the verify parameter flags.
11678 * Keep cipherlists sorted in the source instead of sorting them at
11679 runtime, thus removing the need for a lock.
11683 * Avoid some small subgroup attacks in Diffie-Hellman.
11685 *Nick Mathewson and Ben Laurie*
11687 * Add functions for well-known primes.
11691 * Extended Windows CE support.
11693 *Satoshi Nakamura and Andy Polyakov*
11695 * Initialize SSL_METHOD structures at compile time instead of during
11696 runtime, thus removing the need for a lock.
11700 * Make PKCS7_decrypt() work even if no certificate is supplied by
11701 attempting to decrypt each encrypted key in turn. Add support to
11706 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11708 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11711 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11715 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11716 key into the same file any more.
11720 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11724 * Add -utf8 command line and config file option to 'ca'.
11726 *Stefan <stf@udoma.org*
11728 * Removed the macro des_crypt(), as it seems to conflict with some
11729 libraries. Use DES_crypt().
11733 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11734 involves renaming the source and generated shared-libs for
11735 both. The engines will accept the corrected or legacy ids
11736 ('ncipher' and '4758_cca' respectively) when binding. NB,
11737 this only applies when building 'shared'.
11739 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11741 * Add attribute functions to EVP_PKEY structure. Modify
11742 PKCS12_create() to recognize a CSP name attribute and
11743 use it. Make -CSP option work again in pkcs12 utility.
11747 * Add new functionality to the bn blinding code:
11748 - automatic re-creation of the BN_BLINDING parameters after
11749 a fixed number of uses (currently 32)
11750 - add new function for parameter creation
11751 - introduce flags to control the update behaviour of the
11752 BN_BLINDING parameters
11753 - hide BN_BLINDING structure
11754 Add a second BN_BLINDING slot to the RSA structure to improve
11755 performance when a single RSA object is shared among several
11760 * Add support for DTLS.
11762 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11764 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11765 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11769 * Remove buggy and incomplete DH cert support from
11770 ssl/ssl_rsa.c and ssl/s3_both.c
11774 * Use SHA-1 instead of MD5 as the default digest algorithm for
11775 the `apps/openssl` commands.
11779 * Compile clean with "-Wall -Wmissing-prototypes
11780 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11781 DEBUG_SAFESTACK must also be set.
11785 * Change ./Configure so that certain algorithms can be disabled by default.
11786 The new counterpiece to "no-xxx" is "enable-xxx".
11788 The patented RC5 and MDC2 algorithms will now be disabled unless
11789 "enable-rc5" and "enable-mdc2", respectively, are specified.
11791 (IDEA remains enabled despite being patented. This is because IDEA
11792 is frequently required for interoperability, and there is no license
11793 fee for non-commercial use. As before, "no-idea" can be used to
11794 avoid this algorithm.)
11798 * Add processing of proxy certificates (see RFC 3820). This work was
11799 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11800 EGEE (Enabling Grids for E-science in Europe).
11804 * RC4 performance overhaul on modern architectures/implementations, such
11805 as Intel P4, IA-64 and AMD64.
11809 * New utility extract-section.pl. This can be used specify an alternative
11810 section number in a pod file instead of having to treat each file as
11811 a separate case in Makefile. This can be done by adding two lines to the
11814 =for comment openssl_section:XXX
11816 The blank line is mandatory.
11820 * New arguments -certform, -keyform and -pass for s_client and s_server
11821 to allow alternative format key and certificate files and passphrase
11826 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11827 update associated structures and add various utility functions.
11829 Add new policy related verify parameters, include policy checking in
11830 standard verify code. Enhance 'smime' application with extra parameters
11831 to support policy checking and print out.
11835 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11836 Nehemiah processors. These extensions support AES encryption in hardware
11837 as well as RNG (though RNG support is currently disabled).
11839 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11841 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11845 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11847 *Andy Polyakov and a number of other people*
11849 * Improved PowerPC platform support. Most notably BIGNUM assembler
11850 implementation contributed by IBM.
11852 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11854 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11855 exponent rather than 'unsigned long'. There is a corresponding change to
11856 the new 'rsa_keygen' element of the RSA_METHOD structure.
11858 *Jelte Jansen, Geoff Thorpe*
11860 * Functionality for creating the initial serial number file is now
11861 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11863 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11864 number file to 1, which is bound to cause problems. To avoid
11865 the problems while respecting compatibility between different 0.9.7
11866 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11867 CA.pl for serial number initialization. With the new release 0.9.8,
11868 we can fix the problem directly in the 'ca' utility.)
11872 * Reduced header interdependencies by declaring more opaque objects in
11873 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11874 give fewer recursive includes, which could break lazy source code - so
11875 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11876 developers should define this symbol when building and using openssl to
11877 ensure they track the recommended behaviour, interfaces, [etc], but
11878 backwards-compatible behaviour prevails when this isn't defined.
11882 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11886 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11887 This will generate a random key of the appropriate length based on the
11888 cipher context. The EVP_CIPHER can provide its own random key generation
11889 routine to support keys of a specific form. This is used in the des and
11890 3des routines to generate a key of the correct parity. Update S/MIME
11891 code to use new functions and hence generate correct parity DES keys.
11892 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11893 valid (weak or incorrect parity).
11897 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11898 as looking them up. This is useful when the verified structure may contain
11899 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11900 present unless the new PKCS7_NO_CRL flag is asserted.
11904 * Extend ASN1 oid configuration module. It now additionally accepts the
11907 shortName = some long name, 1.2.3.4
11911 * Reimplemented the BN_CTX implementation. There is now no more static
11912 limitation on the number of variables it can handle nor the depth of the
11913 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11914 information can now expand as required, and rather than having a single
11915 static array of bignums, BN_CTX now uses a linked-list of such arrays
11916 allowing it to expand on demand whilst maintaining the usefulness of
11917 BN_CTX's "bundling".
11921 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11922 to allow all RSA operations to function using a single BN_CTX.
11926 * Preliminary support for certificate policy evaluation and checking. This
11927 is initially intended to pass the tests outlined in "Conformance Testing
11928 of Relying Party Client Certificate Path Processing Logic" v1.07.
11932 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11933 remained unused and not that useful. A variety of other little bignum
11934 tweaks and fixes have also been made continuing on from the audit (see
11939 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11940 associated ASN1, EVP and SSL functions and old ASN1 macros.
11944 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11945 and this should never fail. So the return value from the use of
11946 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11947 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11951 * BN_CTX_get() should return zero-valued bignums, providing the same
11952 initialised value as BN_new().
11954 *Geoff Thorpe, suggested by Ulf Möller*
11956 * Support for inhibitAnyPolicy certificate extension.
11960 * An audit of the BIGNUM code is underway, for which debugging code is
11961 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11962 is considered valid when processing BIGNUMs, and causes execution to
11963 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11964 further steps are taken to deliberately pollute unused data in BIGNUM
11965 structures to try and expose faulty code further on. For now, openssl will
11966 (in its default mode of operation) continue to tolerate the inconsistent
11967 forms that it has tolerated in the past, but authors and packagers should
11968 consider trying openssl and their own applications when compiled with
11969 these debugging symbols defined. It will help highlight potential bugs in
11970 their own code, and will improve the test coverage for OpenSSL itself. At
11971 some point, these tighter rules will become openssl's default to improve
11972 maintainability, though the assert()s and other overheads will remain only
11973 in debugging configurations. See bn.h for more details.
11975 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11977 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11978 that can only be obtained through BN_CTX_new() (which implicitly
11979 initialises it). The presence of this function only made it possible
11980 to overwrite an existing structure (and cause memory leaks).
11984 * Because of the callback-based approach for implementing LHASH as a
11985 template type, lh_insert() adds opaque objects to hash-tables and
11986 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11987 to clean up those corresponding objects before destroying the hash table
11988 (and losing the object pointers). So some over-zealous constifications in
11989 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11990 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11991 prototyped to have "const" restrictions on the object pointers they are
11992 given (and so aren't required to cast them away any more).
11996 * The tmdiff.h API was so ugly and minimal that our own timing utility
11997 (speed) prefers to use its own implementation. The two implementations
11998 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11999 its object type properly exposed (MS_TM) instead of casting to/from
12000 `char *`. This may still change yet if someone realises MS_TM and
12002 aren't necessarily the greatest nomenclatures - but this is what was used
12003 internally to the implementation so I've used that for now.
12007 * Ensure that deprecated functions do not get compiled when
12008 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
12009 the self-tests were still using deprecated key-generation functions so
12010 these have been updated also.
12014 * Reorganise PKCS#7 code to separate the digest location functionality
12015 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
12016 New function PKCS7_set_digest() to set the digest type for PKCS#7
12017 digestedData type. Add additional code to correctly generate the
12018 digestedData type and add support for this type in PKCS7 initialization
12023 * New function PKCS7_set0_type_other() this initializes a PKCS7
12024 structure of type "other".
12028 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
12029 sure the loop does correctly stop and breaking ("division by zero")
12030 modulus operations are not performed. The (pre-generated) prime
12031 table crypto/bn/bn_prime.h was already correct, but it could not be
12032 re-generated on some platforms because of the "division by zero"
12033 situation in the script.
12035 *Ralf S. Engelschall*
12037 * Update support for ECC-based TLS ciphersuites according to
12038 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
12039 SHA-1 now is only used for "small" curves (where the
12040 representation of a field element takes up to 24 bytes); for
12041 larger curves, the field element resulting from ECDH is directly
12042 used as premaster secret.
12044 *Douglas Stebila (Sun Microsystems Laboratories)*
12046 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
12047 curve secp160r1 to the tests.
12049 *Douglas Stebila (Sun Microsystems Laboratories)*
12051 * Add the possibility to load symbols globally with DSO.
12053 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
12055 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
12056 control of the error stack.
12060 * Add support for STORE in ENGINE.
12064 * Add the STORE type. The intention is to provide a common interface
12065 to certificate and key stores, be they simple file-based stores, or
12066 HSM-type store, or LDAP stores, or...
12067 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12071 * Add a generic structure called OPENSSL_ITEM. This can be used to
12072 pass a list of arguments to any function as well as provide a way
12073 for a function to pass data back to the caller.
12077 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12078 works like BUF_strdup() but can be used to duplicate a portion of
12079 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12084 * Add the function sk_find_ex() which works like sk_find(), but will
12085 return an index to an element even if an exact match couldn't be
12086 found. The index is guaranteed to point at the element where the
12087 searched-for key would be inserted to preserve sorting order.
12091 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12092 takes an extra flags argument for optional functionality. Currently,
12093 the following flags are defined:
12095 OBJ_BSEARCH_VALUE_ON_NOMATCH
12096 This one gets OBJ_bsearch_ex() to return a pointer to the first
12097 element where the comparing function returns a negative or zero
12100 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12101 This one gets OBJ_bsearch_ex() to return a pointer to the first
12102 element where the comparing function returns zero. This is useful
12103 if there are more than one element where the comparing function
12108 * Make it possible to create self-signed certificates with 'openssl ca'
12109 in such a way that the self-signed certificate becomes part of the
12110 CA database and uses the same mechanisms for serial number generation
12111 as all other certificate signing. The new flag '-selfsign' enables
12112 this functionality. Adapt CA.sh and CA.pl.in.
12116 * Add functionality to check the public key of a certificate request
12117 against a given private. This is useful to check that a certificate
12118 request can be signed by that key (self-signing).
12122 * Make it possible to have multiple active certificates with the same
12123 subject in the CA index file. This is done only if the keyword
12124 'unique_subject' is set to 'no' in the main CA section (default
12125 if 'CA_default') of the configuration file. The value is saved
12126 with the database itself in a separate index attribute file,
12127 named like the index file with '.attr' appended to the name.
12131 * Generate multi-valued AVAs using '+' notation in config files for
12136 * Support for nameConstraints certificate extension.
12140 * Support for policyConstraints certificate extension.
12144 * Support for policyMappings certificate extension.
12148 * Make sure the default DSA_METHOD implementation only uses its
12149 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12150 and change its own handlers to be NULL so as to remove unnecessary
12151 indirection. This lets alternative implementations fallback to the
12152 default implementation more easily.
12156 * Support for directoryName in GeneralName related extensions
12161 * Make it possible to link applications using Makefile.shared.
12162 Make that possible even when linking against static libraries!
12166 * Support for single pass processing for S/MIME signing. This now
12167 means that S/MIME signing can be done from a pipe, in addition
12168 cleartext signing (multipart/signed type) is effectively streaming
12169 and the signed data does not need to be all held in memory.
12171 This is done with a new flag PKCS7_STREAM. When this flag is set
12172 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12173 is done after the data is output (and digests calculated) in
12174 SMIME_write_PKCS7().
12178 * Add full support for -rpath/-R, both in shared libraries and
12179 applications, at least on the platforms where it's known how
12184 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12185 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12186 will now compute a table of multiples of the generator that
12187 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12188 faster (notably in the case of a single point multiplication,
12189 scalar * generator).
12191 *Nils Larsch, Bodo Moeller*
12193 * IPv6 support for certificate extensions. The various extensions
12194 which use the IP:a.b.c.d can now take IPv6 addresses using the
12195 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12200 * Added an ENGINE that implements RSA by performing private key
12201 exponentiations with the GMP library. The conversions to and from
12202 GMP's mpz_t format aren't optimised nor are any montgomery forms
12203 cached, and on x86 it appears OpenSSL's own performance has caught up.
12204 However there are likely to be other architectures where GMP could
12205 provide a boost. This ENGINE is not built in by default, but it can be
12206 specified at Configure time and should be accompanied by the necessary
12207 linker additions, eg;
12208 ./config -DOPENSSL_USE_GMP -lgmp
12212 * "openssl engine" will not display ENGINE/DSO load failure errors when
12213 testing availability of engines with "-t" - the old behaviour is
12214 produced by increasing the feature's verbosity with "-tt".
12218 * ECDSA routines: under certain error conditions uninitialized BN objects
12219 could be freed. Solution: make sure initialization is performed early
12220 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12225 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12226 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12227 software implementations. For DSA and DH, parameter generation can
12228 also be overridden by providing the appropriate method callbacks.
12232 * Change the "progress" mechanism used in key-generation and
12233 primality testing to functions that take a new BN_GENCB pointer in
12234 place of callback/argument pairs. The new API functions have `_ex`
12235 postfixes and the older functions are reimplemented as wrappers for
12236 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12237 declarations of the old functions to help (graceful) attempts to
12238 migrate to the new functions. Also, the new key-generation API
12239 functions operate on a caller-supplied key-structure and return
12240 success/failure rather than returning a key or NULL - this is to
12241 help make "keygen" another member function of RSA_METHOD etc.
12243 Example for using the new callback interface:
12245 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12246 void *my_arg = ...;
12249 BN_GENCB_set(&my_cb, my_callback, my_arg);
12251 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12252 /* For the meaning of a, b in calls to my_callback(), see the
12253 * documentation of the function that calls the callback.
12254 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12255 * my_callback should return 1 if it wants BN_is_prime_ex()
12256 * to continue, or 0 to stop.
12261 * Change the ZLIB compression method to be stateful, and make it
12262 available to TLS with the number defined in
12263 draft-ietf-tls-compression-04.txt.
12267 * Add the ASN.1 structures and functions for CertificatePair, which
12268 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12270 CertificatePair ::= SEQUENCE {
12271 forward [0] Certificate OPTIONAL,
12272 reverse [1] Certificate OPTIONAL,
12273 -- at least one of the pair shall be present -- }
12275 Also implement the PEM functions to read and write certificate
12276 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12278 This needed to be defined, mostly for the sake of the LDAP
12279 attribute crossCertificatePair, but may prove useful elsewhere as
12284 * Make it possible to inhibit symlinking of shared libraries in
12285 Makefile.shared, for Cygwin's sake.
12289 * Extend the BIGNUM API by creating a function
12290 void BN_set_negative(BIGNUM *a, int neg);
12291 and a macro that behave like
12292 int BN_is_negative(const BIGNUM *a);
12294 to avoid the need to access 'a->neg' directly in applications.
12298 * Implement fast modular reduction for pseudo-Mersenne primes
12299 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12300 EC_GROUP_new_curve_GFp() will now automatically use this
12303 *Nils Larsch <nla@trustcenter.de>*
12305 * Add new lock type (CRYPTO_LOCK_BN).
12309 * Change the ENGINE framework to automatically load engines
12310 dynamically from specific directories unless they could be
12311 found to already be built in or loaded. Move all the
12312 current engines except for the cryptodev one to a new
12313 directory engines/.
12314 The engines in engines/ are built as shared libraries if
12315 the "shared" options was given to ./Configure or ./config.
12316 Otherwise, they are inserted in libcrypto.a.
12317 /usr/local/ssl/engines is the default directory for dynamic
12318 engines, but that can be overridden at configure time through
12319 the usual use of --prefix and/or --openssldir, and at run
12320 time with the environment variable OPENSSL_ENGINES.
12322 *Geoff Thorpe and Richard Levitte*
12324 * Add Makefile.shared, a helper makefile to build shared
12325 libraries. Adapt Makefile.org.
12329 * Add version info to Win32 DLLs.
12331 *Peter 'Luna' Runestig" <peter@runestig.com>*
12333 * Add new 'medium level' PKCS#12 API. Certificates and keys
12334 can be added using this API to created arbitrary PKCS#12
12335 files while avoiding the low-level API.
12337 New options to PKCS12_create(), key or cert can be NULL and
12338 will then be omitted from the output file. The encryption
12339 algorithm NIDs can be set to -1 for no encryption, the mac
12340 iteration count can be set to 0 to omit the mac.
12342 Enhance pkcs12 utility by making the -nokeys and -nocerts
12343 options work when creating a PKCS#12 file. New option -nomac
12344 to omit the mac, NONE can be set for an encryption algorithm.
12345 New code is modified to use the enhanced PKCS12_create()
12346 instead of the low-level API.
12350 * Extend ASN1 encoder to support indefinite length constructed
12351 encoding. This can output sequences tags and octet strings in
12352 this form. Modify pk7_asn1.c to support indefinite length
12353 encoding. This is experimental and needs additional code to
12354 be useful, such as an ASN1 bio and some enhanced streaming
12357 Extend template encode functionality so that tagging is passed
12358 down to the template encoder.
12362 * Let 'openssl req' fail if an argument to '-newkey' is not
12363 recognized instead of using RSA as a default.
12367 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12368 As these are not official, they are not included in "ALL";
12369 the "ECCdraft" ciphersuite group alias can be used to select them.
12371 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12373 * Add ECDH engine support.
12375 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12377 * Add ECDH in new directory crypto/ecdh/.
12379 *Douglas Stebila (Sun Microsystems Laboratories)*
12381 * Let BN_rand_range() abort with an error after 100 iterations
12382 without success (which indicates a broken PRNG).
12386 * Change BN_mod_sqrt() so that it verifies that the input value
12387 is really the square of the return value. (Previously,
12388 BN_mod_sqrt would show GIGO behaviour.)
12392 * Add named elliptic curves over binary fields from X9.62, SECG,
12393 and WAP/WTLS; add OIDs that were still missing.
12395 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12397 * Extend the EC library for elliptic curves over binary fields
12398 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12401 EC_GF2m_simple_method
12405 EC_GROUP_new_curve_GF2m
12406 EC_GROUP_set_curve_GF2m
12407 EC_GROUP_get_curve_GF2m
12408 EC_POINT_set_affine_coordinates_GF2m
12409 EC_POINT_get_affine_coordinates_GF2m
12410 EC_POINT_set_compressed_coordinates_GF2m
12412 Point compression for binary fields is disabled by default for
12413 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12416 As binary polynomials are represented as BIGNUMs, various members
12417 of the EC_GROUP and EC_POINT data structures can be shared
12418 between the implementations for prime fields and binary fields;
12419 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12420 are essentially identical to their `..._GFp` counterparts.
12421 (For simplicity, the `..._GFp` prefix has been dropped from
12422 various internal method names.)
12424 An internal 'field_div' method (similar to 'field_mul' and
12425 'field_sqr') has been added; this is used only for binary fields.
12427 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12429 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12430 through methods ('mul', 'precompute_mult').
12432 The generic implementations (now internally called 'ec_wNAF_mul'
12433 and 'ec_wNAF_precomputed_mult') remain the default if these
12434 methods are undefined.
12436 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12438 * New function EC_GROUP_get_degree, which is defined through
12439 EC_METHOD. For curves over prime fields, this returns the bit
12440 length of the modulus.
12442 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12444 * New functions EC_GROUP_dup, EC_POINT_dup.
12445 (These simply call ..._new and ..._copy).
12447 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12449 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12450 Polynomials are represented as BIGNUMs (where the sign bit is not
12451 used) in the following functions [macros]:
12454 BN_GF2m_sub [= BN_GF2m_add]
12455 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12456 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12457 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12459 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12460 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12461 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12462 BN_GF2m_cmp [= BN_ucmp]
12464 (Note that only the 'mod' functions are actually for fields GF(2^m).
12465 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12467 For some functions, an the irreducible polynomial defining a
12468 field can be given as an 'unsigned int[]' with strictly
12469 decreasing elements giving the indices of those bits that are set;
12470 i.e., p[] represents the polynomial
12471 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12473 p[0] > p[1] > ... > p[k] = 0.
12474 This applies to the following functions:
12477 BN_GF2m_mod_mul_arr
12478 BN_GF2m_mod_sqr_arr
12479 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12480 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12481 BN_GF2m_mod_exp_arr
12482 BN_GF2m_mod_sqrt_arr
12483 BN_GF2m_mod_solve_quad_arr
12487 Conversion can be performed by the following functions:
12492 bntest.c has additional tests for binary polynomial arithmetic.
12494 Two implementations for BN_GF2m_mod_div() are available.
12495 The default algorithm simply uses BN_GF2m_mod_inv() and
12496 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12497 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12498 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12500 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12502 * Add new error code 'ERR_R_DISABLED' that can be used when some
12503 functionality is disabled at compile-time.
12505 *Douglas Stebila <douglas.stebila@sun.com>*
12507 * Change default behaviour of 'openssl asn1parse' so that more
12508 information is visible when viewing, e.g., a certificate:
12510 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12511 mode the content of non-printable OCTET STRINGs is output in a
12512 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12513 avoid the appearance of a printable string.
12515 *Nils Larsch <nla@trustcenter.de>*
12517 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12519 EC_GROUP_set_asn1_flag()
12520 EC_GROUP_get_asn1_flag()
12521 EC_GROUP_set_point_conversion_form()
12522 EC_GROUP_get_point_conversion_form()
12523 These control ASN1 encoding details:
12524 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12525 has been set to OPENSSL_EC_NAMED_CURVE.
12526 - Points are encoded in uncompressed form by default; options for
12527 asn1_for are as for point2oct, namely
12528 POINT_CONVERSION_COMPRESSED
12529 POINT_CONVERSION_UNCOMPRESSED
12530 POINT_CONVERSION_HYBRID
12532 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12534 EC_GROUP_set_seed()
12535 EC_GROUP_get0_seed()
12536 EC_GROUP_get_seed_len()
12537 This is used only for ASN1 purposes (so far).
12539 *Nils Larsch <nla@trustcenter.de>*
12541 * Add 'field_type' member to EC_METHOD, which holds the NID
12542 of the appropriate field type OID. The new function
12543 EC_METHOD_get_field_type() returns this value.
12545 *Nils Larsch <nla@trustcenter.de>*
12548 EC_POINT_point2bn()
12549 EC_POINT_bn2point()
12550 EC_POINT_point2hex()
12551 EC_POINT_hex2point()
12552 providing useful interfaces to EC_POINT_point2oct() and
12553 EC_POINT_oct2point().
12555 *Nils Larsch <nla@trustcenter.de>*
12557 * Change internals of the EC library so that the functions
12558 EC_GROUP_set_generator()
12559 EC_GROUP_get_generator()
12560 EC_GROUP_get_order()
12561 EC_GROUP_get_cofactor()
12562 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12563 to methods, which would lead to unnecessary code duplication when
12564 adding different types of curves.
12566 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12568 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12569 arithmetic, and such that modified wNAFs are generated
12570 (which avoid length expansion in many cases).
12574 * Add a function EC_GROUP_check_discriminant() (defined via
12575 EC_METHOD) that verifies that the curve discriminant is non-zero.
12577 Add a function EC_GROUP_check() that makes some sanity tests
12578 on a EC_GROUP, its generator and order. This includes
12579 EC_GROUP_check_discriminant().
12581 *Nils Larsch <nla@trustcenter.de>*
12583 * Add ECDSA in new directory crypto/ecdsa/.
12585 Add applications 'openssl ecparam' and 'openssl ecdsa'
12586 (these are based on 'openssl dsaparam' and 'openssl dsa').
12588 ECDSA support is also included in various other files across the
12589 library. Most notably,
12590 - 'openssl req' now has a '-newkey ecdsa:file' option;
12591 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12592 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12593 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12594 them suitable for ECDSA where domain parameters must be
12595 extracted before the specific public key;
12596 - ECDSA engine support has been added.
12598 *Nils Larsch <nla@trustcenter.de>*
12600 * Include some named elliptic curves, and add OIDs from X9.62,
12601 SECG, and WAP/WTLS. Each curve can be obtained from the new
12603 EC_GROUP_new_by_curve_name(),
12604 and the list of available named curves can be obtained with
12605 EC_get_builtin_curves().
12606 Also add a 'curve_name' member to EC_GROUP objects, which can be
12608 EC_GROUP_set_curve_name()
12609 EC_GROUP_get_curve_name()
12611 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12613 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12614 was actually never needed) and in BN_mul(). The removal in BN_mul()
12615 required a small change in bn_mul_part_recursive() and the addition
12616 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12617 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12618 bn_sub_words() and bn_add_words() except they take arrays with
12623 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12625 * Cleanse PEM buffers before freeing them since they may contain
12628 *Benjamin Bennett <ben@psc.edu>*
12630 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12631 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12632 authentication-only ciphersuites.
12636 * Since AES128 and AES256 share a single mask bit in the logic of
12637 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12638 kludge to work properly if AES128 is available and AES256 isn't.
12642 * Expand security boundary to match 1.1.1 module.
12646 * Remove redundant features: hash file source, editing of test vectors
12647 modify fipsld to use external fips_premain.c signature.
12651 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12652 run algorithm test programs.
12656 * Make algorithm test programs more tolerant of whitespace.
12660 * Have SSL/TLS server implementation tolerate "mismatched" record
12661 protocol version while receiving ClientHello even if the
12662 ClientHello is fragmented. (The server can't insist on the
12663 particular protocol version it has chosen before the ServerHello
12664 message has informed the client about his choice.)
12668 * Load error codes if they are not already present instead of using a
12669 static variable. This allows them to be cleanly unloaded and reloaded.
12673 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12675 * Introduce limits to prevent malicious keys being able to
12676 cause a denial of service. ([CVE-2006-2940])
12678 *Steve Henson, Bodo Moeller*
12680 * Fix ASN.1 parsing of certain invalid structures that can result
12681 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12683 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12684 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12686 * Fix SSL client code which could crash if connecting to a
12687 malicious SSLv2 server. ([CVE-2006-4343])
12689 *Tavis Ormandy and Will Drewry, Google Security Team*
12691 * Change ciphersuite string processing so that an explicit
12692 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12693 will no longer include "AES128-SHA"), and any other similar
12694 ciphersuite (same bitmap) from *other* protocol versions (so that
12695 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12696 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12697 changes from 0.9.8b and 0.9.8d.
12701 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12703 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12704 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12706 * Change the Unix randomness entropy gathering to use poll() when
12707 possible instead of select(), since the latter has some
12708 undesirable limitations.
12710 *Darryl Miles via Richard Levitte and Bodo Moeller*
12712 * Disable rogue ciphersuites:
12714 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12715 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12716 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12718 The latter two were purportedly from
12719 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12722 Also deactivate the remaining ciphersuites from
12723 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12724 unofficial, and the ID has long expired.
12728 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12729 dual-core machines) and other potential thread-safety issues.
12733 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12735 * Adapt fipsld and the build system to link against the validated FIPS
12736 module in FIPS mode.
12740 * Fixes for VC++ 2005 build under Windows.
12744 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12745 from a Windows bash shell such as MSYS. It is autodetected from the
12746 "config" script when run from a VC++ environment. Modify standard VC++
12747 build to use fipscanister.o from the GNU make build.
12751 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12753 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12754 The value now differs depending on if you build for FIPS or not.
12755 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12756 safely run with a non-FIPSed libcrypto, as it may crash because of
12757 the difference induced by this change.
12761 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12763 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12764 (part of SSL_OP_ALL). This option used to disable the
12765 countermeasure against man-in-the-middle protocol-version
12766 rollback in the SSL 2.0 server implementation, which is a bad
12767 idea. ([CVE-2005-2969])
12769 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12770 for Information Security, National Institute of Advanced Industrial
12771 Science and Technology [AIST, Japan)]*
12773 * Minimal support for X9.31 signatures and PSS padding modes. This is
12774 mainly for FIPS compliance and not fully integrated at this stage.
12778 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12779 the exponentiation using a fixed-length exponent. (Otherwise,
12780 the information leaked through timing could expose the secret key
12781 after many signatures; cf. Bleichenbacher's attack on DSA with
12786 * Make a new fixed-window mod_exp implementation the default for
12787 RSA, DSA, and DH private-key operations so that the sequence of
12788 squares and multiplies and the memory access pattern are
12789 independent of the particular secret key. This will mitigate
12790 cache-timing and potential related attacks.
12792 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12793 and this is automatically used by BN_mod_exp_mont() if the new flag
12794 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12795 will use this BN flag for private exponents unless the flag
12796 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12797 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12799 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12801 * Change the client implementation for SSLv23_method() and
12802 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12803 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12804 (Previously, the SSL 2.0 backwards compatible Client Hello
12805 message format would be used even with SSL_OP_NO_SSLv2.)
12809 * Add support for smime-type MIME parameter in S/MIME messages which some
12814 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12815 a threadsafe manner. Modify rsa code to use new function and add calls
12816 to dsa and dh code (which had race conditions before).
12820 * Include the fixed error library code in the C error file definitions
12821 instead of fixing them up at runtime. This keeps the error code
12822 structures constant.
12826 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12828 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12831 * Fixes for newer kerberos headers. NB: the casts are needed because
12832 the 'length' field is signed on one version and unsigned on another
12833 with no (?) obvious way to tell the difference, without these VC++
12834 complains. Also the "definition" of FAR (blank) is no longer included
12835 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12836 some needed definitions.
12840 * Undo Cygwin change.
12844 * Added support for proxy certificates according to RFC 3820.
12845 Because they may be a security thread to unaware applications,
12846 they must be explicitly allowed in run-time. See
12847 docs/HOWTO/proxy_certificates.txt for further information.
12851 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12853 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12854 server and client random values. Previously
12855 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12856 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12858 This change has negligible security impact because:
12860 1. Server and client random values still have 24 bytes of pseudo random
12863 2. Server and client random values are sent in the clear in the initial
12866 3. The master secret is derived using the premaster secret (48 bytes in
12867 size for static RSA ciphersuites) as well as client server and random
12870 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12873 *Stephen Henson, reported by UK NISCC*
12875 * Use Windows randomness collection on Cygwin.
12879 * Fix hang in EGD/PRNGD query when communication socket is closed
12880 prematurely by EGD/PRNGD.
12882 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12884 * Prompt for pass phrases when appropriate for PKCS12 input format.
12888 * Back-port of selected performance improvements from development
12889 branch, as well as improved support for PowerPC platforms.
12893 * Add lots of checks for memory allocation failure, error codes to indicate
12894 failure and freeing up memory if a failure occurs.
12896 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12898 * Add new -passin argument to dgst.
12902 * Perform some character comparisons of different types in X509_NAME_cmp:
12903 this is needed for some certificates that re-encode DNs into UTF8Strings
12904 (in violation of RFC3280) and can't or won't issue name rollover
12909 * Make an explicit check during certificate validation to see that
12910 the CA setting in each certificate on the chain is correct. As a
12911 side effect always do the following basic checks on extensions,
12912 not just when there's an associated purpose to the check:
12914 - if there is an unhandled critical extension (unless the user
12915 has chosen to ignore this fault)
12916 - if the path length has been exceeded (if one is set at all)
12917 - that certain extensions fit the associated purpose (if one has
12922 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12924 * Avoid a race condition when CRLs are checked in a multi threaded
12925 environment. This would happen due to the reordering of the revoked
12926 entries during signature checking and serial number lookup. Now the
12927 encoding is cached and the serial number sort performed under a lock.
12928 Add new STACK function sk_is_sorted().
12932 * Add Delta CRL to the extension code.
12936 * Various fixes to s3_pkt.c so alerts are sent properly.
12938 *David Holmes <d.holmes@f5.com>*
12940 * Reduce the chances of duplicate issuer name and serial numbers (in
12941 violation of RFC3280) using the OpenSSL certificate creation utilities.
12942 This is done by creating a random 64 bit value for the initial serial
12943 number when a serial number file is created or when a self signed
12944 certificate is created using 'openssl req -x509'. The initial serial
12945 number file is created using 'openssl x509 -next_serial' in CA.pl
12946 rather than being initialized to 1.
12950 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12952 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12953 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12955 *Joe Orton, Steve Henson*
12957 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12960 *Joe Orton, Steve Henson*
12962 * Make it possible to have multiple active certificates with the same
12963 subject in the CA index file. This is done only if the keyword
12964 'unique_subject' is set to 'no' in the main CA section (default
12965 if 'CA_default') of the configuration file. The value is saved
12966 with the database itself in a separate index attribute file,
12967 named like the index file with '.attr' appended to the name.
12971 * X509 verify fixes. Disable broken certificate workarounds when
12972 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12973 keyUsage extension present. Don't accept CRLs with unhandled critical
12974 extensions: since verify currently doesn't process CRL extensions this
12975 rejects a CRL with *any* critical extensions. Add new verify error codes
12980 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12981 A clarification of RFC2560 will require the use of OCTET STRINGs and
12982 some implementations cannot handle the current raw format. Since OpenSSL
12983 copies and compares OCSP nonces as opaque blobs without any attempt at
12984 parsing them this should not create any compatibility issues.
12988 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12989 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12990 this HMAC (and other) operations are several times slower than OpenSSL
12995 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12997 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12999 * Use the correct content when signing type "other".
13003 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
13005 * Fix various bugs revealed by running the NISCC test suite:
13007 Stop out of bounds reads in the ASN1 code when presented with
13008 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13010 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
13012 If verify callback ignores invalid public key errors don't try to check
13013 certificate signature with the NULL public key.
13017 * New -ignore_err option in ocsp application to stop the server
13018 exiting on the first error in a request.
13022 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13023 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13028 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13029 extra data after the compression methods not only for TLS 1.0
13030 but also for SSL 3.0 (as required by the specification).
13032 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13034 * Change X509_certificate_type() to mark the key as exported/exportable
13035 when it's 512 *bits* long, not 512 bytes.
13039 * Change AES_cbc_encrypt() so it outputs exact multiple of
13040 blocks during encryption.
13044 * Various fixes to base64 BIO and non blocking I/O. On write
13045 flushes were not handled properly if the BIO retried. On read
13046 data was not being buffered properly and had various logic bugs.
13047 This also affects blocking I/O when the data being decoded is a
13052 * Various S/MIME bugfixes and compatibility changes:
13053 output correct application/pkcs7 MIME type if
13054 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
13055 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
13056 of files as .eml work). Correctly handle very long lines in MIME
13061 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
13063 * Countermeasure against the Klima-Pokorny-Rosa extension of
13064 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13065 a protocol version number mismatch like a decryption error
13066 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13070 * Turn on RSA blinding by default in the default implementation
13071 to avoid a timing attack. Applications that don't want it can call
13072 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13073 They would be ill-advised to do so in most cases.
13075 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13077 * Change RSA blinding code so that it works when the PRNG is not
13078 seeded (in this case, the secret RSA exponent is abused as
13079 an unpredictable seed -- if it is not unpredictable, there
13080 is no point in blinding anyway). Make RSA blinding thread-safe
13081 by remembering the creator's thread ID in rsa->blinding and
13082 having all other threads use local one-time blinding factors
13083 (this requires more computation than sharing rsa->blinding, but
13084 avoids excessive locking; and if an RSA object is not shared
13085 between threads, blinding will still be very fast).
13089 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13090 ENGINE as defaults for all supported algorithms irrespective of
13091 the 'flags' parameter. 'flags' is now honoured, so applications
13092 should make sure they are passing it correctly.
13096 * Target "mingw" now allows native Windows code to be generated in
13097 the Cygwin environment as well as with the MinGW compiler.
13101 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13103 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13104 via timing by performing a MAC computation even if incorrect
13105 block cipher padding has been found. This is a countermeasure
13106 against active attacks where the attacker has to distinguish
13107 between bad padding and a MAC verification error. ([CVE-2003-0078])
13109 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13110 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13111 Martin Vuagnoux (EPFL, Ilion)*
13113 * Make the no-err option work as intended. The intention with no-err
13114 is not to have the whole error stack handling routines removed from
13115 libcrypto, it's only intended to remove all the function name and
13116 reason texts, thereby removing some of the footprint that may not
13117 be interesting if those errors aren't displayed anyway.
13119 NOTE: it's still possible for any application or module to have its
13120 own set of error texts inserted. The routines are there, just not
13121 used by default when no-err is given.
13125 * Add support for FreeBSD on IA64.
13127 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13129 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13130 Kerberos function mit_des_cbc_cksum(). Before this change,
13131 the value returned by DES_cbc_cksum() was like the one from
13132 mit_des_cbc_cksum(), except the bytes were swapped.
13134 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13136 * Allow an application to disable the automatic SSL chain building.
13137 Before this a rather primitive chain build was always performed in
13138 ssl3_output_cert_chain(): an application had no way to send the
13139 correct chain if the automatic operation produced an incorrect result.
13141 Now the chain builder is disabled if either:
13143 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13145 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13147 The reasoning behind this is that an application would not want the
13148 auto chain building to take place if extra chain certificates are
13149 present and it might also want a means of sending no additional
13150 certificates (for example the chain has two certificates and the
13155 * Add the possibility to build without the ENGINE framework.
13157 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13159 * Under Win32 gmtime() can return NULL: check return value in
13160 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13164 * DSA routines: under certain error conditions uninitialized BN objects
13165 could be freed. Solution: make sure initialization is performed early
13166 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13167 Nils Larsch <nla@trustcenter.de> via PR#459)
13171 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13172 checked on reconnect on the client side, therefore session resumption
13173 could still fail with a "ssl session id is different" error. This
13174 behaviour is masked when SSL_OP_ALL is used due to
13175 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13176 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13177 followup to PR #377.
13181 * IA-32 assembler support enhancements: unified ELF targets, support
13182 for SCO/Caldera platforms, fix for Cygwin shared build.
13186 * Add support for FreeBSD on sparc64. As a consequence, support for
13187 FreeBSD on non-x86 processors is separate from x86 processors on
13188 the config script, much like the NetBSD support.
13190 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13192 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13194 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13197 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13198 code (06) was taken as the first octet of the session ID and the last
13199 octet was ignored consequently. As a result SSLv2 client side session
13200 caching could not have worked due to the session ID mismatch between
13202 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13207 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13208 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13213 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13214 seems that in spite of existing for more than a year, many application
13215 author have done nothing to provide the necessary callbacks, which
13216 means that this particular engine will not work properly anywhere.
13217 This is a very unfortunate situation which forces us, in the name
13218 of usability, to give the hw_ncipher.c a static lock, which is part
13220 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13221 appear in 0.9.8 or later. We EXPECT application authors to have
13222 dealt properly with this when 0.9.8 is released (unless we actually
13223 make such changes in the libcrypto locking code that changes will
13224 have to be made anyway).
13228 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13229 octets have been read, EOF or an error occurs. Without this change
13230 some truncated ASN1 structures will not produce an error.
13234 * Disable Heimdal support, since it hasn't been fully implemented.
13235 Still give the possibility to force the use of Heimdal, but with
13236 warnings and a request that patches get sent to openssl-dev.
13240 * Add the VC-CE target, introduce the WINCE sysname, and add
13241 INSTALL.WCE and appropriate conditionals to make it build.
13243 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13245 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13246 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13247 edit numbers of the version.
13249 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13251 * Introduce safe string copy and catenation functions
13252 (BUF_strlcpy() and BUF_strlcat()).
13254 *Ben Laurie (CHATS) and Richard Levitte*
13256 * Avoid using fixed-size buffers for one-line DNs.
13258 *Ben Laurie (CHATS)*
13260 * Add BUF_MEM_grow_clean() to avoid information leakage when
13261 resizing buffers containing secrets, and use where appropriate.
13263 *Ben Laurie (CHATS)*
13265 * Avoid using fixed size buffers for configuration file location.
13267 *Ben Laurie (CHATS)*
13269 * Avoid filename truncation for various CA files.
13271 *Ben Laurie (CHATS)*
13273 * Use sizeof in preference to magic numbers.
13275 *Ben Laurie (CHATS)*
13277 * Avoid filename truncation in cert requests.
13279 *Ben Laurie (CHATS)*
13281 * Add assertions to check for (supposedly impossible) buffer
13284 *Ben Laurie (CHATS)*
13286 * Don't cache truncated DNS entries in the local cache (this could
13287 potentially lead to a spoofing attack).
13289 *Ben Laurie (CHATS)*
13291 * Fix various buffers to be large enough for hex/decimal
13292 representations in a platform independent manner.
13294 *Ben Laurie (CHATS)*
13296 * Add CRYPTO_realloc_clean() to avoid information leakage when
13297 resizing buffers containing secrets, and use where appropriate.
13299 *Ben Laurie (CHATS)*
13301 * Add BIO_indent() to avoid much slightly worrying code to do
13304 *Ben Laurie (CHATS)*
13306 * Convert sprintf()/BIO_puts() to BIO_printf().
13308 *Ben Laurie (CHATS)*
13310 * buffer_gets() could terminate with the buffer only half
13313 *Ben Laurie (CHATS)*
13315 * Add assertions to prevent user-supplied crypto functions from
13316 overflowing internal buffers by having large block sizes, etc.
13318 *Ben Laurie (CHATS)*
13320 * New OPENSSL_assert() macro (similar to assert(), but enabled
13323 *Ben Laurie (CHATS)*
13325 * Eliminate unused copy of key in RC4.
13327 *Ben Laurie (CHATS)*
13329 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13331 *Ben Laurie (CHATS)*
13333 * Fix off-by-one error in EGD path.
13335 *Ben Laurie (CHATS)*
13337 * If RANDFILE path is too long, ignore instead of truncating.
13339 *Ben Laurie (CHATS)*
13341 * Eliminate unused and incorrectly sized X.509 structure
13344 *Ben Laurie (CHATS)*
13346 * Eliminate unused and dangerous function knumber().
13348 *Ben Laurie (CHATS)*
13350 * Eliminate unused and dangerous structure, KSSL_ERR.
13352 *Ben Laurie (CHATS)*
13354 * Protect against overlong session ID context length in an encoded
13355 session object. Since these are local, this does not appear to be
13358 *Ben Laurie (CHATS)*
13360 * Change from security patch (see 0.9.6e below) that did not affect
13361 the 0.9.6 release series:
13363 Remote buffer overflow in SSL3 protocol - an attacker could
13364 supply an oversized master key in Kerberos-enabled versions.
13367 *Ben Laurie (CHATS)*
13369 * Change the SSL kerb5 codes to match RFC 2712.
13373 * Make -nameopt work fully for req and add -reqopt switch.
13375 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13377 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13379 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13381 * Make sure tests can be performed even if the corresponding algorithms
13382 have been removed entirely. This was also the last step to make
13383 OpenSSL compilable with DJGPP under all reasonable conditions.
13385 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13387 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13388 to allow version independent disabling of normally unselected ciphers,
13389 which may be activated as a side-effect of selecting a single cipher.
13391 (E.g., cipher list string "RSA" enables ciphersuites that are left
13392 out of "ALL" because they do not provide symmetric encryption.
13393 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13395 *Lutz Jaenicke, Bodo Moeller*
13397 * Add appropriate support for separate platform-dependent build
13398 directories. The recommended way to make a platform-dependent
13399 build directory is the following (tested on Linux), maybe with
13402 # Place yourself outside of the OpenSSL source tree. In
13403 # this example, the environment variable OPENSSL_SOURCE
13404 # is assumed to contain the absolute OpenSSL source directory.
13405 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13406 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13407 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13408 mkdir -p `dirname $F`
13409 ln -s $OPENSSL_SOURCE/$F $F
13412 To be absolutely sure not to disturb the source tree, a "make clean"
13413 is a good thing. If it isn't successful, don't worry about it,
13414 it probably means the source directory is very clean.
13418 * Make sure any ENGINE control commands make local copies of string
13419 pointers passed to them whenever necessary. Otherwise it is possible
13420 the caller may have overwritten (or deallocated) the original string
13421 data when a later ENGINE operation tries to use the stored values.
13423 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13425 * Improve diagnostics in file reading and command-line digests.
13427 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13429 * Add AES modes CFB and OFB to the object database. Correct an
13430 error in AES-CFB decryption.
13434 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13435 allows existing EVP_CIPHER_CTX structures to be reused after
13436 calling `EVP_*Final()`. This behaviour is used by encryption
13437 BIOs and some applications. This has the side effect that
13438 applications must explicitly clean up cipher contexts with
13439 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13443 * Check the values of dna and dnb in bn_mul_recursive before calling
13444 bn_mul_comba (a non zero value means the a or b arrays do not contain
13445 n2 elements) and fallback to bn_mul_normal if either is not zero.
13449 * Fix escaping of non-ASCII characters when using the -subj option
13450 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13454 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13455 form for "surname", serialNumber has no short form.
13456 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13457 therefore remove "mail" short name for "internet 7".
13458 The OID for unique identifiers in X509 certificates is
13459 x500UniqueIdentifier, not uniqueIdentifier.
13460 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13464 * Add an "init" command to the ENGINE config module and auto initialize
13465 ENGINEs. Without any "init" command the ENGINE will be initialized
13466 after all ctrl commands have been executed on it. If init=1 the
13467 ENGINE is initialized at that point (ctrls before that point are run
13468 on the uninitialized ENGINE and after on the initialized one). If
13469 init=0 then the ENGINE will not be initialized at all.
13473 * Fix the 'app_verify_callback' interface so that the user-defined
13474 argument is actually passed to the callback: In the
13475 SSL_CTX_set_cert_verify_callback() prototype, the callback
13476 declaration has been changed from
13479 int (*cb)(X509_STORE_CTX *,void *);
13480 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13481 i=s->ctx->app_verify_callback(&ctx)
13482 has been changed into
13483 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13485 To update applications using SSL_CTX_set_cert_verify_callback(),
13486 a dummy argument can be added to their callback functions.
13488 *D. K. Smetters <smetters@parc.xerox.com>*
13490 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13492 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13494 * Add and OPENSSL_LOAD_CONF define which will cause
13495 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13496 This allows older applications to transparently support certain
13497 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13498 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13499 load the config file and OPENSSL_add_all_algorithms_conf() which will
13500 always load it have also been added.
13504 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13505 Adjust NIDs and EVP layer.
13507 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13509 * Config modules support in openssl utility.
13511 Most commands now load modules from the config file,
13512 though in a few (such as version) this isn't done
13513 because it couldn't be used for anything.
13515 In the case of ca and req the config file used is
13516 the same as the utility itself: that is the -config
13517 command line option can be used to specify an
13522 * Move default behaviour from OPENSSL_config(). If appname is NULL
13523 use "openssl_conf" if filename is NULL use default openssl config file.
13527 * Add an argument to OPENSSL_config() to allow the use of an alternative
13528 config section name. Add a new flag to tolerate a missing config file
13529 and move code to CONF_modules_load_file().
13533 * Support for crypto accelerator cards from Accelerated Encryption
13534 Processing, www.aep.ie. (Use engine 'aep')
13535 The support was copied from 0.9.6c [engine] and adapted/corrected
13536 to work with the new engine framework.
13538 *AEP Inc. and Richard Levitte*
13540 * Support for SureWare crypto accelerator cards from Baltimore
13541 Technologies. (Use engine 'sureware')
13542 The support was copied from 0.9.6c [engine] and adapted
13543 to work with the new engine framework.
13547 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13548 make the newer ENGINE framework commands for the CHIL engine work.
13550 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13552 * Make it possible to produce shared libraries on ReliantUNIX.
13554 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13556 * Add the configuration target debug-linux-ppro.
13557 Make 'openssl rsa' use the general key loading routines
13558 implemented in `apps.c`, and make those routines able to
13559 handle the key format FORMAT_NETSCAPE and the variant
13562 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13564 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13566 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13568 * Add -keyform to rsautl, and document -engine.
13570 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13572 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13573 BIO_R_NO_SUCH_FILE error code rather than the generic
13574 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13578 * Add new functions
13579 ERR_peek_last_error
13580 ERR_peek_last_error_line
13581 ERR_peek_last_error_line_data.
13582 These are similar to
13584 ERR_peek_error_line
13585 ERR_peek_error_line_data,
13586 but report on the latest error recorded rather than the first one
13587 still in the error queue.
13589 *Ben Laurie, Bodo Moeller*
13591 * default_algorithms option in ENGINE config module. This allows things
13593 default_algorithms = ALL
13594 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13598 * Preliminary ENGINE config module.
13602 * New experimental application configuration code.
13606 * Change the AES code to follow the same name structure as all other
13607 symmetric ciphers, and behave the same way. Move everything to
13608 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13610 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13612 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13614 *Ben Laurie and Theo de Raadt*
13616 * Add option to output public keys in req command.
13618 *Massimiliano Pala madwolf@openca.org*
13620 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13621 (up to about 10% better than before for P-192 and P-224).
13625 * New functions/macros
13627 SSL_CTX_set_msg_callback(ctx, cb)
13628 SSL_CTX_set_msg_callback_arg(ctx, arg)
13629 SSL_set_msg_callback(ssl, cb)
13630 SSL_set_msg_callback_arg(ssl, arg)
13632 to request calling a callback function
13634 void cb(int write_p, int version, int content_type,
13635 const void *buf, size_t len, SSL *ssl, void *arg)
13637 whenever a protocol message has been completely received
13638 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13639 protocol version according to which the SSL library interprets
13640 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13641 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13642 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13643 specification (change_cipher_spec(20), alert(21), handshake(22)).
13644 'buf' and 'len' point to the actual message, 'ssl' to the
13645 SSL object, and 'arg' is the application-defined value set by
13646 SSL[_CTX]_set_msg_callback_arg().
13648 'openssl s_client' and 'openssl s_server' have new '-msg' options
13649 to enable a callback that displays all protocol messages.
13653 * Change the shared library support so shared libraries are built as
13654 soon as the corresponding static library is finished, and thereby get
13655 openssl and the test programs linked against the shared library.
13656 This still only happens when the keyword "shard" has been given to
13657 the configuration scripts.
13659 NOTE: shared library support is still an experimental thing, and
13660 backward binary compatibility is still not guaranteed.
13662 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13664 * Add support for Subject Information Access extension.
13666 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13668 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13669 additional bytes when new memory had to be allocated, not just
13670 when reusing an existing buffer.
13674 * New command line and configuration option 'utf8' for the req command.
13675 This allows field values to be specified as UTF8 strings.
13679 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13680 runs for the former and machine-readable output for the latter.
13684 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13685 of the e-mail address in the DN (i.e., it will go into a certificate
13686 extension only). The new configuration file option 'email_in_dn = no'
13687 has the same effect.
13689 *Massimiliano Pala madwolf@openca.org*
13691 * Change all functions with names starting with `des_` to be starting
13692 with `DES_` instead. Add wrappers that are compatible with libdes,
13693 but are named `_ossl_old_des_*`. Finally, add macros that map the
13694 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13695 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13696 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13699 Since we provide two compatibility mappings, the user needs to
13700 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13701 compatibility is desired. The default (i.e., when that macro
13702 isn't defined) is OpenSSL 0.9.6c compatibility.
13704 There are also macros that enable and disable the support of old
13705 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13706 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13707 are defined, the default will apply: to support the old des routines.
13709 In either case, one must include openssl/des.h to get the correct
13710 definitions. Do not try to just include openssl/des_old.h, that
13713 NOTE: This is a major break of an old API into a new one. Software
13714 authors are encouraged to switch to the `DES_` style functions. Some
13715 time in the future, des_old.h and the libdes compatibility functions
13716 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13717 default), and then completely removed.
13721 * Test for certificates which contain unsupported critical extensions.
13722 If such a certificate is found during a verify operation it is
13723 rejected by default: this behaviour can be overridden by either
13724 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13725 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13726 X509_supported_extension() has also been added which returns 1 if a
13727 particular extension is supported.
13731 * Modify the behaviour of EVP cipher functions in similar way to digests
13732 to retain compatibility with existing code.
13736 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13737 compatibility with existing code. In particular the 'ctx' parameter does
13738 not have to be to be initialized before the call to EVP_DigestInit() and
13739 it is tidied up after a call to EVP_DigestFinal(). New function
13740 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13741 EVP_MD_CTX_copy() changed to not require the destination to be
13742 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13743 requires the destination to be valid.
13745 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13746 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13750 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13751 so that complete 'Handshake' protocol structures are kept in memory
13752 instead of overwriting 'msg_type' and 'length' with 'body' data.
13756 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13758 *Massimo Santin via Richard Levitte*
13760 * Major restructuring to the underlying ENGINE code. This includes
13761 reduction of linker bloat, separation of pure "ENGINE" manipulation
13762 (initialisation, etc) from functionality dealing with implementations
13763 of specific crypto interfaces. This change also introduces integrated
13764 support for symmetric ciphers and digest implementations - so ENGINEs
13765 can now accelerate these by providing EVP_CIPHER and EVP_MD
13766 implementations of their own. This is detailed in
13767 [crypto/engine/README.md](crypto/engine/README.md)
13768 as it couldn't be adequately described here. However, there are a few
13769 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13770 were changed in the original introduction of ENGINE code have now
13771 reverted back - the hooking from this code to ENGINE is now a good
13772 deal more passive and at run-time, operations deal directly with
13773 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13774 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13775 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13776 they were not being used by the framework as there is no concept of a
13777 BIGNUM_METHOD and they could not be generalised to the new
13778 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13779 ENGINE_cpy() has been removed as it cannot be consistently defined in
13784 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13788 * Change mkdef.pl to sort symbols that get the same entry number,
13789 and make sure the automatically generated functions `ERR_load_*`
13790 become part of libeay.num as well.
13794 * New function SSL_renegotiate_pending(). This returns true once
13795 renegotiation has been requested (either SSL_renegotiate() call
13796 or HelloRequest/ClientHello received from the peer) and becomes
13797 false once a handshake has been completed.
13798 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13799 sends a HelloRequest, but does not ensure that a handshake takes
13800 place. SSL_renegotiate_pending() is useful for checking if the
13801 client has followed the request.)
13805 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13806 By default, clients may request session resumption even during
13807 renegotiation (if session ID contexts permit); with this option,
13808 session resumption is possible only in the first handshake.
13810 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13811 more bits available for options that should not be part of
13812 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13816 * Add some demos for certificate and certificate request creation.
13820 * Make maximum certificate chain size accepted from the peer application
13821 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13822 "Douglas E. Engert" <deengert@anl.gov>.
13826 * Add support for shared libraries for Unixware-7
13827 (Boyd Lynn Gerber <gerberb@zenez.com>).
13831 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13832 be done prior to destruction. Use this to unload error strings from
13833 ENGINEs that load their own error strings. NB: This adds two new API
13834 functions to "get" and "set" this destroy handler in an ENGINE.
13838 * Alter all existing ENGINE implementations (except "openssl" and
13839 "openbsd") to dynamically instantiate their own error strings. This
13840 makes them more flexible to be built both as statically-linked ENGINEs
13841 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13842 Also, add stub code to each that makes building them as self-contained
13843 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13847 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13848 implementations into applications that are completely implemented in
13849 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13850 commands that can be used to configure what shared-library to load and
13851 to control aspects of the way it is handled. Also, made an update to
13852 the [README-Engine.md](README-Engine.md) file
13853 that brings its information up-to-date and
13854 provides some information and instructions on the "dynamic" ENGINE
13855 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13859 * Make it possible to unload ranges of ERR strings with a new
13860 "ERR_unload_strings" function.
13864 * Add a copy() function to EVP_MD.
13868 * Make EVP_MD routines take a context pointer instead of just the
13869 md_data void pointer.
13873 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13874 that the digest can only process a single chunk of data
13875 (typically because it is provided by a piece of
13876 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13877 is only going to provide a single chunk of data, and hence the
13878 framework needn't accumulate the data for oneshot drivers.
13882 * As with "ERR", make it possible to replace the underlying "ex_data"
13883 functions. This change also alters the storage and management of global
13884 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13885 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13886 index counters. The API functions that use this state have been changed
13887 to take a "class_index" rather than pointers to the class's local STACK
13888 and counter, and there is now an API function to dynamically create new
13889 classes. This centralisation allows us to (a) plug a lot of the
13890 thread-safety problems that existed, and (b) makes it possible to clean
13891 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13892 such data would previously have always leaked in application code and
13893 workarounds were in place to make the memory debugging turn a blind eye
13894 to it. Application code that doesn't use this new function will still
13895 leak as before, but their memory debugging output will announce it now
13896 rather than letting it slide.
13898 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13899 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13900 has a return value to indicate success or failure.
13904 * Make it possible to replace the underlying "ERR" functions such that the
13905 global state (2 LHASH tables and 2 locks) is only used by the "default"
13906 implementation. This change also adds two functions to "get" and "set"
13907 the implementation prior to it being automatically set the first time
13908 any other ERR function takes place. Ie. an application can call "get",
13909 pass the return value to a module it has just loaded, and that module
13910 can call its own "set" function using that value. This means the
13911 module's "ERR" operations will use (and modify) the error state in the
13912 application and not in its own statically linked copy of OpenSSL code.
13916 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13917 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13918 the operation, and provides a more encapsulated way for external code
13919 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13920 to use these functions rather than manually incrementing the counts.
13922 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13926 * Add EVP test program.
13930 * Add symmetric cipher support to ENGINE. Expect the API to change!
13934 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13935 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13936 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13937 These allow a CRL to be built without having to access X509_CRL fields
13938 directly. Modify 'ca' application to use new functions.
13942 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13943 bug workarounds. Rollback attack detection is a security feature.
13944 The problem will only arise on OpenSSL servers when TLSv1 is not
13945 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13946 Software authors not wanting to support TLSv1 will have special reasons
13947 for their choice and can explicitly enable this option.
13949 *Bodo Moeller, Lutz Jaenicke*
13951 * Rationalise EVP so it can be extended: don't include a union of
13952 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13953 (similar to those existing for EVP_CIPHER_CTX).
13958 EVP_MD_CTX_init(&md); /* new function call */
13959 EVP_DigestInit(&md, EVP_sha1());
13960 EVP_DigestUpdate(&md, in, len);
13961 EVP_DigestFinal(&md, out, NULL);
13962 EVP_MD_CTX_cleanup(&md); /* new function call */
13966 * Make DES key schedule conform to the usual scheme, as well as
13967 correcting its structure. This means that calls to DES functions
13968 now have to pass a pointer to a des_key_schedule instead of a
13969 plain des_key_schedule (which was actually always a pointer
13972 des_key_schedule ks;
13974 des_set_key_checked(..., &ks);
13975 des_ncbc_encrypt(..., &ks, ...);
13977 (Note that a later change renames 'des_...' into 'DES_...'.)
13981 * Initial reduction of linker bloat: the use of some functions, such as
13982 PEM causes large amounts of unused functions to be linked in due to
13983 poor organisation. For example pem_all.c contains every PEM function
13984 which has a knock on effect of linking in large amounts of (unused)
13985 ASN1 code. Grouping together similar functions and splitting unrelated
13986 functions prevents this.
13990 * Cleanup of EVP macros.
13994 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13995 correct `_ecb suffix`.
13999 * Add initial OCSP responder support to ocsp application. The
14000 revocation information is handled using the text based index
14001 use by the ca application. The responder can either handle
14002 requests generated internally, supplied in files (for example
14003 via a CGI script) or using an internal minimal server.
14007 * Add configuration choices to get zlib compression for TLS.
14011 * Changes to Kerberos SSL for RFC 2712 compliance:
14012 1. Implemented real KerberosWrapper, instead of just using
14013 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
14014 2. Implemented optional authenticator field of KerberosWrapper.
14016 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
14017 and authenticator structs; see crypto/krb5/.
14019 Generalized Kerberos calls to support multiple Kerberos libraries.
14020 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
14021 via Richard Levitte*
14023 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
14024 already does with RSA. testdsa.h now has 'priv_key/pub_key'
14025 values for each of the key sizes rather than having just
14026 parameters (and 'speed' generating keys each time).
14030 * Speed up EVP routines.
14033 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
14034 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
14035 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
14036 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
14038 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
14039 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
14040 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
14043 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
14045 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
14049 * Added the OS2-EMX target.
14051 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
14053 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
14054 New functions to support `NCONF` routines in extension code.
14055 New function `CONF_set_nconf()`
14056 to allow functions which take an `NCONF` to also handle the old `LHASH`
14057 structure: this means that the old `CONF` compatible routines can be
14058 retained (in particular w.rt. extensions) without having to duplicate the
14059 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
14063 * Enhance the general user interface with mechanisms for inner control
14064 and with possibilities to have yes/no kind of prompts.
14068 * Change all calls to low-level digest routines in the library and
14069 applications to use EVP. Add missing calls to HMAC_cleanup() and
14070 don't assume HMAC_CTX can be copied using memcpy().
14072 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14074 * Add the possibility to control engines through control names but with
14075 arbitrary arguments instead of just a string.
14076 Change the key loaders to take a UI_METHOD instead of a callback
14077 function pointer. NOTE: this breaks binary compatibility with earlier
14078 versions of OpenSSL [engine].
14079 Adapt the nCipher code for these new conditions and add a card insertion
14084 * Enhance the general user interface with mechanisms to better support
14085 dialog box interfaces, application-defined prompts, the possibility
14086 to use defaults (for example default passwords from somewhere else)
14087 and interrupts/cancellations.
14091 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14092 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14096 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14097 tidy up some unnecessarily weird code in 'sk_new()').
14099 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14101 * Change the key loading routines for ENGINEs to use the same kind
14102 callback (pem_password_cb) as all other routines that need this
14107 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14108 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14109 than this minimum value is recommended.
14113 * New random seeder for OpenVMS, using the system process statistics
14114 that are easily reachable.
14118 * Windows apparently can't transparently handle global
14119 variables defined in DLLs. Initialisations such as:
14121 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14123 won't compile. This is used by the any applications that need to
14124 declare their own ASN1 modules. This was fixed by adding the option
14125 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14126 needed for static libraries under Win32.
14130 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14131 setting of purpose and trust fields. New X509_STORE trust and
14132 purpose functions and tidy up setting in other SSL functions.
14136 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14137 structure. These are inherited by X509_STORE_CTX when it is
14138 initialised. This allows various defaults to be set in the
14139 X509_STORE structure (such as flags for CRL checking and custom
14140 purpose or trust settings) for functions which only use X509_STORE_CTX
14141 internally such as S/MIME.
14143 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14144 trust settings if they are not set in X509_STORE. This allows X509_STORE
14145 purposes and trust (in S/MIME for example) to override any set by default.
14147 Add command line options for CRL checking to smime, s_client and s_server
14152 * Initial CRL based revocation checking. If the CRL checking flag(s)
14153 are set then the CRL is looked up in the X509_STORE structure and
14154 its validity and signature checked, then if the certificate is found
14155 in the CRL the verify fails with a revoked error.
14157 Various new CRL related callbacks added to X509_STORE_CTX structure.
14159 Command line options added to 'verify' application to support this.
14161 This needs some additional work, such as being able to handle multiple
14162 CRLs with different times, extension based lookup (rather than just
14163 by subject name) and ultimately more complete V2 CRL extension
14168 * Add a general user interface API (crypto/ui/). This is designed
14169 to replace things like des_read_password and friends (backward
14170 compatibility functions using this new API are provided).
14171 The purpose is to remove prompting functions from the DES code
14172 section as well as provide for prompting through dialog boxes in
14173 a window system and the like.
14177 * Add "ex_data" support to ENGINE so implementations can add state at a
14178 per-structure level rather than having to store it globally.
14182 * Make it possible for ENGINE structures to be copied when retrieved by
14183 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14184 This causes the "original" ENGINE structure to act like a template,
14185 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14186 operational state can be localised to each ENGINE structure, despite the
14187 fact they all share the same "methods". New ENGINE structures returned in
14188 this case have no functional references and the return value is the single
14189 structural reference. This matches the single structural reference returned
14190 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14195 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14196 needs to match any other type at all we need to manually clear the
14201 * Changes to the "openssl engine" utility to include;
14202 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14203 about an ENGINE's available control commands.
14204 - executing control commands from command line arguments using the
14205 '-pre' and '-post' switches. '-post' is only used if '-t' is
14206 specified and the ENGINE is successfully initialised. The syntax for
14207 the individual commands are colon-separated, for example;
14208 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14212 * New dynamic control command support for ENGINEs. ENGINEs can now
14213 declare their own commands (numbers), names (strings), descriptions,
14214 and input types for run-time discovery by calling applications. A
14215 subset of these commands are implicitly classed as "executable"
14216 depending on their input type, and only these can be invoked through
14217 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14218 can be based on user input, config files, etc). The distinction is
14219 that "executable" commands cannot return anything other than a boolean
14220 result and can only support numeric or string input, whereas some
14221 discoverable commands may only be for direct use through
14222 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14223 pointers, or other custom uses. The "executable" commands are to
14224 support parameterisations of ENGINE behaviour that can be
14225 unambiguously defined by ENGINEs and used consistently across any
14226 OpenSSL-based application. Commands have been added to all the
14227 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14228 control over shared-library paths without source code alterations.
14232 * Changed all ENGINE implementations to dynamically allocate their
14233 ENGINEs rather than declaring them statically. Apart from this being
14234 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14235 this also allows the implementations to compile without using the
14236 internal engine_int.h header.
14240 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14241 'const' value. Any code that should be able to modify a RAND_METHOD
14242 should already have non-const pointers to it (ie. they should only
14243 modify their own ones).
14247 * Made a variety of little tweaks to the ENGINE code.
14248 - "atalla" and "ubsec" string definitions were moved from header files
14249 to C code. "nuron" string definitions were placed in variables
14250 rather than hard-coded - allowing parameterisation of these values
14251 later on via ctrl() commands.
14252 - Removed unused "#if 0"'d code.
14253 - Fixed engine list iteration code so it uses ENGINE_free() to release
14254 structural references.
14255 - Constified the RAND_METHOD element of ENGINE structures.
14256 - Constified various get/set functions as appropriate and added
14257 missing functions (including a catch-all ENGINE_cpy that duplicates
14258 all ENGINE values onto a new ENGINE except reference counts/state).
14259 - Removed NULL parameter checks in get/set functions. Setting a method
14260 or function to NULL is a way of cancelling out a previously set
14261 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14262 and doesn't justify the extra error symbols and code.
14263 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14264 flags from engine_int.h to engine.h.
14265 - Changed prototypes for ENGINE handler functions (init(), finish(),
14266 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14270 * Implement binary inversion algorithm for BN_mod_inverse in addition
14271 to the algorithm using long division. The binary algorithm can be
14272 used only if the modulus is odd. On 32-bit systems, it is faster
14273 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14274 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14275 up to 450 bits. In 64-bit environments, the binary algorithm
14276 appears to be advantageous for much longer moduli; here we use it
14277 for moduli up to 2048 bits.
14281 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14282 could not support the combine flag in choice fields.
14286 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14287 extensions from a certificate request to the certificate.
14291 * Allow multiple 'certopt' and 'nameopt' options to be separated
14292 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14293 file: this allows the display of the certificate about to be
14294 signed to be customised, to allow certain fields to be included
14295 or excluded and extension details. The old system didn't display
14296 multicharacter strings properly, omitted fields not in the policy
14297 and couldn't display additional details such as extensions.
14301 * Function EC_POINTs_mul for multiple scalar multiplication
14302 of an arbitrary number of elliptic curve points
14303 \sum scalars[i]*points[i],
14304 optionally including the generator defined for the EC_GROUP:
14305 scalar*generator + \sum scalars[i]*points[i].
14307 EC_POINT_mul is a simple wrapper function for the typical case
14308 that the point list has just one item (besides the optional
14313 * First EC_METHODs for curves over GF(p):
14315 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14316 operations and provides various method functions that can also
14317 operate with faster implementations of modular arithmetic.
14319 EC_GFp_mont_method() reuses most functions that are part of
14320 EC_GFp_simple_method, but uses Montgomery arithmetic.
14322 *Bodo Moeller; point addition and point doubling
14323 implementation directly derived from source code provided by
14324 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14326 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14327 crypto/ec/ec_lib.c):
14329 Curves are EC_GROUP objects (with an optional group generator)
14330 based on EC_METHODs that are built into the library.
14332 Points are EC_POINT objects based on EC_GROUP objects.
14334 Most of the framework would be able to handle curves over arbitrary
14335 finite fields, but as there are no obvious types for fields other
14336 than GF(p), some functions are limited to that for now.
14340 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14341 that the file contains a complete HTTP response.
14345 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14346 change the def and num file printf format specifier from "%-40sXXX"
14347 to "%-39s XXX". The latter will always guarantee a space after the
14348 field while the former will cause them to run together if the field
14349 is 40 of more characters long.
14353 * Constify the cipher and digest 'method' functions and structures
14354 and modify related functions to take constant EVP_MD and EVP_CIPHER
14359 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14360 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14364 * Modify `EVP_Digest*()` routines so they now return values. Although the
14365 internal software routines can never fail additional hardware versions
14370 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14372 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14373 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14376 ERR_R_NESTED_ASN1_ERROR
14378 ERR_R_MISSING_ASN1_EOS
14379 were 4 .. 9, conflicting with
14380 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14382 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14383 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14385 Add new error code 'ERR_R_INTERNAL_ERROR'.
14389 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14394 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14395 sets the subject name for a new request or supersedes the
14396 subject name in a given request. Formats that can be parsed are
14397 'CN=Some Name, OU=myOU, C=IT'
14399 'CN=Some Name/OU=myOU/C=IT'.
14401 Add options '-batch' and '-verbose' to 'openssl req'.
14403 *Massimiliano Pala <madwolf@hackmasters.net>*
14405 * Introduce the possibility to access global variables through
14406 functions on platform were that's the best way to handle exporting
14407 global variables in shared libraries. To enable this functionality,
14408 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14409 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14410 is normally done by Configure or something similar).
14412 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14413 in the source file (foo.c) like this:
14415 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14416 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14418 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14419 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14421 OPENSSL_DECLARE_GLOBAL(int,foo);
14422 #define foo OPENSSL_GLOBAL_REF(foo)
14423 OPENSSL_DECLARE_GLOBAL(double,bar);
14424 #define bar OPENSSL_GLOBAL_REF(bar)
14426 The #defines are very important, and therefore so is including the
14427 header file everywhere where the defined globals are used.
14429 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14430 of ASN.1 items, but that structure is a bit different.
14432 The largest change is in util/mkdef.pl which has been enhanced with
14433 better and easier to understand logic to choose which symbols should
14434 go into the Windows .def files as well as a number of fixes and code
14435 cleanup (among others, algorithm keywords are now sorted
14436 lexicographically to avoid constant rewrites).
14440 * In BN_div() keep a copy of the sign of 'num' before writing the
14441 result to 'rm' because if rm==num the value will be overwritten
14442 and produce the wrong result if 'num' is negative: this caused
14443 problems with BN_mod() and BN_nnmod().
14447 * Function OCSP_request_verify(). This checks the signature on an
14448 OCSP request and verifies the signer certificate. The signer
14449 certificate is just checked for a generic purpose and OCSP request
14454 * Add OCSP_check_validity() function to check the validity of OCSP
14455 responses. OCSP responses are prepared in real time and may only
14456 be a few seconds old. Simply checking that the current time lies
14457 between thisUpdate and nextUpdate max reject otherwise valid responses
14458 caused by either OCSP responder or client clock inaccuracy. Instead
14459 we allow thisUpdate and nextUpdate to fall within a certain period of
14460 the current time. The age of the response can also optionally be
14461 checked. Two new options -validity_period and -status_age added to
14466 * If signature or public key algorithm is unrecognized print out its
14467 OID rather that just UNKNOWN.
14471 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14472 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14473 ID to be generated from the issuer certificate alone which can then be
14474 passed to OCSP_id_issuer_cmp().
14478 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14479 ASN1 modules to export functions returning ASN1_ITEM pointers
14480 instead of the ASN1_ITEM structures themselves. This adds several
14481 new macros which allow the underlying ASN1 function/structure to
14482 be accessed transparently. As a result code should not use ASN1_ITEM
14483 references directly (such as &X509_it) but instead use the relevant
14484 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14485 use of the new ASN1 code on platforms where exporting structures
14486 is problematical (for example in shared libraries) but exporting
14487 functions returning pointers to structures is not.
14491 * Add support for overriding the generation of SSL/TLS session IDs.
14492 These callbacks can be registered either in an SSL_CTX or per SSL.
14493 The purpose of this is to allow applications to control, if they wish,
14494 the arbitrary values chosen for use as session IDs, particularly as it
14495 can be useful for session caching in multiple-server environments. A
14496 command-line switch for testing this (and any client code that wishes
14497 to use such a feature) has been added to "s_server".
14499 *Geoff Thorpe, Lutz Jaenicke*
14501 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14502 of the form `#if defined(...) || defined(...) || ...` and
14503 `#if !defined(...) && !defined(...) && ...`. This also avoids
14504 the growing number of special cases it was previously handling.
14508 * Make all configuration macros available for application by making
14509 sure they are available in opensslconf.h, by giving them names starting
14510 with `OPENSSL_` to avoid conflicts with other packages and by making
14511 sure e_os2.h will cover all platform-specific cases together with
14513 Additionally, it is now possible to define configuration/platform-
14514 specific names (called "system identities"). In the C code, these
14515 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14516 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14517 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14522 * New option -set_serial to 'req' and 'x509' this allows the serial
14523 number to use to be specified on the command line. Previously self
14524 signed certificates were hard coded with serial number 0 and the
14525 CA options of 'x509' had to use a serial number in a file which was
14530 * New options to 'ca' utility to support V2 CRL entry extensions.
14531 Currently CRL reason, invalidity date and hold instruction are
14532 supported. Add new CRL extensions to V3 code and some new objects.
14536 * New function EVP_CIPHER_CTX_set_padding() this is used to
14537 disable standard block padding (aka PKCS#5 padding) in the EVP
14538 API, which was previously mandatory. This means that the data is
14539 not padded in any way and so the total length much be a multiple
14540 of the block size, otherwise an error occurs.
14544 * Initial (incomplete) OCSP SSL support.
14548 * New function OCSP_parse_url(). This splits up a URL into its host,
14549 port and path components: primarily to parse OCSP URLs. New -url
14550 option to ocsp utility.
14554 * New nonce behavior. The return value of OCSP_check_nonce() now
14555 reflects the various checks performed. Applications can decide
14556 whether to tolerate certain situations such as an absent nonce
14557 in a response when one was present in a request: the ocsp application
14558 just prints out a warning. New function OCSP_add1_basic_nonce()
14559 this is to allow responders to include a nonce in a response even if
14560 the request is nonce-less.
14564 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14565 skipped when using openssl x509 multiple times on a single input file,
14566 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14570 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14571 set string type: to handle setting ASN1_TIME structures. Fix ca
14572 utility to correctly initialize revocation date of CRLs.
14576 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14577 the clients preferred ciphersuites and rather use its own preferences.
14578 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14579 Internet Explorer by ensuring unchanged hash method during stepup.
14580 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14584 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14585 to aes and add a new 'exist' option to print out symbols that don't
14590 * Additional options to ocsp utility to allow flags to be set and
14591 additional certificates supplied.
14595 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14596 OCSP client a number of certificate to only verify the response
14601 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14602 handle the new API. Currently only ECB, CBC modes supported. Add new
14605 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14606 Encryption Standard (AES) Ciphersuites for Transport Layer
14607 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14608 not enabled by default and were not part of the "ALL" ciphersuite
14609 alias because they were not yet official; they could be
14610 explicitly requested by specifying the "AESdraft" ciphersuite
14611 group alias. In the final release of OpenSSL 0.9.7, the group
14612 alias is called "AES" and is part of "ALL".)
14614 *Ben Laurie, Steve Henson, Bodo Moeller*
14616 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14617 request to response.
14621 * Functions for OCSP responders. OCSP_request_onereq_count(),
14622 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14623 extract information from a certificate request. OCSP_response_create()
14624 creates a response and optionally adds a basic response structure.
14625 OCSP_basic_add1_status() adds a complete single response to a basic
14626 response and returns the OCSP_SINGLERESP structure just added (to allow
14627 extensions to be included for example). OCSP_basic_add1_cert() adds a
14628 certificate to a basic response and OCSP_basic_sign() signs a basic
14629 response with various flags. New helper functions ASN1_TIME_check()
14630 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14631 (converts ASN1_TIME to GeneralizedTime).
14635 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14636 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14637 structure from a certificate. X509_pubkey_digest() digests the public_key
14638 contents: this is used in various key identifiers.
14642 * Make sk_sort() tolerate a NULL argument.
14644 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14646 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14647 passed by the function are trusted implicitly. If any of them signed the
14648 response then it is assumed to be valid and is not verified.
14652 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14653 to data. This was previously part of the PKCS7 ASN1 code. This
14654 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14655 *Steve Henson, reported by Kenneth R. Robinette
14656 <support@securenetterm.com>*
14658 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14659 routines: without these tracing memory leaks is very painful.
14660 Fix leaks in PKCS12 and PKCS7 routines.
14664 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14665 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14666 effectively meant GeneralizedTime would never be used. Now it
14667 is initialised to -1 but X509_time_adj() now has to check the value
14668 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14669 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14670 *Steve Henson, reported by Kenneth R. Robinette
14671 <support@securenetterm.com>*
14673 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14674 result in a zero length in the ASN1_INTEGER structure which was
14675 not consistent with the structure when d2i_ASN1_INTEGER() was used
14676 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14677 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14678 where it did not print out a minus for negative ASN1_INTEGER.
14682 * Add summary printout to ocsp utility. The various functions which
14683 convert status values to strings have been renamed to:
14684 OCSP_response_status_str(), OCSP_cert_status_str() and
14685 OCSP_crl_reason_str() and are no longer static. New options
14686 to verify nonce values and to disable verification. OCSP response
14687 printout format cleaned up.
14691 * Add additional OCSP certificate checks. These are those specified
14692 in RFC2560. This consists of two separate checks: the CA of the
14693 certificate being checked must either be the OCSP signer certificate
14694 or the issuer of the OCSP signer certificate. In the latter case the
14695 OCSP signer certificate must contain the OCSP signing extended key
14696 usage. This check is performed by attempting to match the OCSP
14697 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14698 in the OCSP_CERTID structures of the response.
14702 * Initial OCSP certificate verification added to OCSP_basic_verify()
14703 and related routines. This uses the standard OpenSSL certificate
14704 verify routines to perform initial checks (just CA validity) and
14705 to obtain the certificate chain. Then additional checks will be
14706 performed on the chain. Currently the root CA is checked to see
14707 if it is explicitly trusted for OCSP signing. This is used to set
14708 a root CA as a global signing root: that is any certificate that
14709 chains to that CA is an acceptable OCSP signing certificate.
14713 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14714 extensions from a separate configuration file.
14715 As when reading extensions from the main configuration file,
14716 the '-extensions ...' option may be used for specifying the
14719 *Massimiliano Pala <madwolf@comune.modena.it>*
14721 * New OCSP utility. Allows OCSP requests to be generated or
14722 read. The request can be sent to a responder and the output
14723 parsed, outputted or printed in text form. Not complete yet:
14724 still needs to check the OCSP response validity.
14728 * New subcommands for 'openssl ca':
14729 `openssl ca -status <serial>` prints the status of the cert with
14730 the given serial number (according to the index file).
14731 `openssl ca -updatedb` updates the expiry status of certificates
14734 *Massimiliano Pala <madwolf@comune.modena.it>*
14736 * New '-newreq-nodes' command option to CA.pl. This is like
14737 '-newreq', but calls 'openssl req' with the '-nodes' option
14738 so that the resulting key is not encrypted.
14740 *Damien Miller <djm@mindrot.org>*
14742 * New configuration for the GNU Hurd.
14744 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14746 * Initial code to implement OCSP basic response verify. This
14747 is currently incomplete. Currently just finds the signer's
14748 certificate and verifies the signature on the response.
14752 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14753 value of OPENSSLDIR. This is available via the new '-d' option
14754 to 'openssl version', and is also included in 'openssl version -a'.
14758 * Allowing defining memory allocation callbacks that will be given
14759 file name and line number information in additional arguments
14760 (a `const char*` and an int). The basic functionality remains, as
14761 well as the original possibility to just replace malloc(),
14762 realloc() and free() by functions that do not know about these
14763 additional arguments. To register and find out the current
14764 settings for extended allocation functions, the following
14765 functions are provided:
14767 CRYPTO_set_mem_ex_functions
14768 CRYPTO_set_locked_mem_ex_functions
14769 CRYPTO_get_mem_ex_functions
14770 CRYPTO_get_locked_mem_ex_functions
14772 These work the same way as CRYPTO_set_mem_functions and friends.
14773 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14774 extended allocation function is enabled.
14775 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14776 a conventional allocation function is enabled.
14778 *Richard Levitte, Bodo Moeller*
14780 * Finish off removing the remaining LHASH function pointer casts.
14781 There should no longer be any prototype-casting required when using
14782 the LHASH abstraction, and any casts that remain are "bugs". See
14783 the callback types and macros at the head of lhash.h for details
14784 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14788 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14789 If /dev/[u]random devices are not available or do not return enough
14790 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14792 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14793 /etc/entropy will be queried once each in this sequence, querying stops
14794 when enough entropy was collected without querying more sockets.
14798 * Change the Unix RAND_poll() variant to be able to poll several
14799 random devices, as specified by DEVRANDOM, until a sufficient amount
14800 of data has been collected. We spend at most 10 ms on each file
14801 (select timeout) and read in non-blocking mode. DEVRANDOM now
14802 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14803 (previously it was just the string "/dev/urandom"), so on typical
14804 platforms the 10 ms delay will never occur.
14805 Also separate out the Unix variant to its own file, rand_unix.c.
14806 For VMS, there's a currently-empty rand_vms.c.
14810 * Move OCSP client related routines to ocsp_cl.c. These
14811 provide utility functions which an application needing
14812 to issue a request to an OCSP responder and analyse the
14813 response will typically need: as opposed to those which an
14814 OCSP responder itself would need which will be added later.
14816 OCSP_request_sign() signs an OCSP request with an API similar
14817 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14818 response. OCSP_response_get1_basic() extracts basic response
14819 from response. OCSP_resp_find_status(): finds and extracts status
14820 information from an OCSP_CERTID structure (which will be created
14821 when the request structure is built). These are built from lower
14822 level functions which work on OCSP_SINGLERESP structures but
14823 won't normally be used unless the application wishes to examine
14824 extensions in the OCSP response for example.
14826 Replace nonce routines with a pair of functions.
14827 OCSP_request_add1_nonce() adds a nonce value and optionally
14828 generates a random value. OCSP_check_nonce() checks the
14829 validity of the nonce in an OCSP response.
14833 * Change function OCSP_request_add() to OCSP_request_add0_id().
14834 This doesn't copy the supplied OCSP_CERTID and avoids the
14835 need to free up the newly created id. Change return type
14836 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14837 This can then be used to add extensions to the request.
14838 Deleted OCSP_request_new(), since most of its functionality
14839 is now in OCSP_REQUEST_new() (and the case insensitive name
14840 clash) apart from the ability to set the request name which
14841 will be added elsewhere.
14845 * Update OCSP API. Remove obsolete extensions argument from
14846 various functions. Extensions are now handled using the new
14847 OCSP extension code. New simple OCSP HTTP function which
14848 can be used to send requests and parse the response.
14852 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14853 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14854 uses the special reorder version of SET OF to sort the attributes
14855 and reorder them to match the encoded order. This resolves a long
14856 standing problem: a verify on a PKCS7 structure just after signing
14857 it used to fail because the attribute order did not match the
14858 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14859 it uses the received order. This is necessary to tolerate some broken
14860 software that does not order SET OF. This is handled by encoding
14861 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14862 to produce the required SET OF.
14866 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14867 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14868 files to get correct declarations of the ASN.1 item variables.
14872 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14873 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14874 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14875 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14876 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14877 ASN1_ITEM and no wrapper functions.
14881 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14882 replace the old function pointer based I/O routines. Change most of
14883 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14887 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14888 lines, recognize more "algorithms" that can be deselected, and make
14889 it complain about algorithm deselection that isn't recognised.
14893 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14894 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14895 to use new functions. Add NO_ASN1_OLD which can be set to remove
14896 some old style ASN1 functions: this can be used to determine if old
14897 code will still work when these eventually go away.
14901 * New extension functions for OCSP structures, these follow the
14902 same conventions as certificates and CRLs.
14906 * New function X509V3_add1_i2d(). This automatically encodes and
14907 adds an extension. Its behaviour can be customised with various
14908 flags to append, replace or delete. Various wrappers added for
14909 certificates and CRLs.
14913 * Fix to avoid calling the underlying ASN1 print routine when
14914 an extension cannot be parsed. Correct a typo in the
14915 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14919 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14920 entries for variables.
14924 * Add functionality to `apps/openssl.c` for detecting locking
14925 problems: As the program is single-threaded, all we have
14926 to do is register a locking callback using an array for
14927 storing which locks are currently held by the program.
14931 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14932 SSL_get_ex_data_X509_STORE_idx(), which is used in
14933 ssl_verify_cert_chain() and thus can be called at any time
14934 during TLS/SSL handshakes so that thread-safety is essential.
14935 Unfortunately, the ex_data design is not at all suited
14936 for multi-threaded use, so it probably should be abolished.
14940 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14942 *Broadcom, tweaked and integrated by Geoff Thorpe*
14944 * Move common extension printing code to new function
14945 X509V3_print_extensions(). Reorganise OCSP print routines and
14946 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14950 * New function X509_signature_print() to remove duplication in some
14955 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14956 set (this was treated exactly the same as SET OF previously). This
14957 is used to reorder the STACK representing the structure to match the
14958 encoding. This will be used to get round a problem where a PKCS7
14959 structure which was signed could not be verified because the STACK
14960 order did not reflect the encoded order.
14964 * Reimplement the OCSP ASN1 module using the new code.
14968 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14969 for its ASN1 operations. The old style function pointers still exist
14970 for now but they will eventually go away.
14974 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14975 completely replaces the old ASN1 functionality with a table driven
14976 encoder and decoder which interprets an ASN1_ITEM structure describing
14977 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14978 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14979 has also been converted to the new form.
14983 * Change BN_mod_exp_recp so that negative moduli are tolerated
14984 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14985 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14986 for negative moduli.
14990 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14991 of not touching the result's sign bit.
14995 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
15000 * Changed the LHASH code to use prototypes for callbacks, and created
15001 macros to declare and implement thin (optionally static) functions
15002 that provide type-safety and avoid function pointer casting for the
15003 type-specific callbacks.
15007 * Added Kerberos Cipher Suites to be used with TLS, as written in
15009 *Veers Staats <staatsvr@asc.hpc.mil>,
15010 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
15012 * Reformat the FAQ so the different questions and answers can be divided
15013 in sections depending on the subject.
15017 * Have the zlib compression code load ZLIB.DLL dynamically under
15022 * New function BN_mod_sqrt for computing square roots modulo a prime
15023 (using the probabilistic Tonelli-Shanks algorithm unless
15024 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
15025 be handled deterministically).
15027 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15029 * Make BN_mod_inverse faster by explicitly handling small quotients
15030 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
15031 512 bits], about 30% for larger ones [1024 or 2048 bits].)
15035 * New function BN_kronecker.
15039 * Fix BN_gcd so that it works on negative inputs; the result is
15040 positive unless both parameters are zero.
15041 Previously something reasonably close to an infinite loop was
15042 possible because numbers could be growing instead of shrinking
15043 in the implementation of Euclid's algorithm.
15047 * Fix BN_is_word() and BN_is_one() macros to take into account the
15048 sign of the number in question.
15050 Fix BN_is_word(a,w) to work correctly for w == 0.
15052 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
15053 because its test if the absolute value of 'a' equals 'w'.
15054 Note that BN_abs_is_word does *not* handle w == 0 reliably;
15055 it exists mostly for use in the implementations of BN_is_zero(),
15056 BN_is_one(), and BN_is_word().
15060 * New function BN_swap.
15064 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
15065 the exponentiation functions are more likely to produce reasonable
15066 results on negative inputs.
15070 * Change BN_mod_mul so that the result is always non-negative.
15071 Previously, it could be negative if one of the factors was negative;
15072 I don't think anyone really wanted that behaviour.
15076 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15077 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15078 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15079 and add new functions:
15088 BN_mod_lshift1_quick
15090 BN_mod_lshift_quick
15092 These functions always generate non-negative results.
15094 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15095 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15097 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15098 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15099 be reduced modulo `m`.
15101 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15104 The following entry accidentally appeared in the CHANGES file
15105 distributed with OpenSSL 0.9.7. The modifications described in
15106 it do *not* apply to OpenSSL 0.9.7.
15108 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15109 was actually never needed) and in BN_mul(). The removal in BN_mul()
15110 required a small change in bn_mul_part_recursive() and the addition
15111 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15112 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15113 bn_sub_words() and bn_add_words() except they take arrays with
15119 * In 'openssl passwd', verify passwords read from the terminal
15120 unless the '-salt' option is used (which usually means that
15121 verification would just waste user's time since the resulting
15122 hash is going to be compared with some given password hash)
15123 or the new '-noverify' option is used.
15125 This is an incompatible change, but it does not affect
15126 non-interactive use of 'openssl passwd' (passwords on the command
15127 line, '-stdin' option, '-in ...' option) and thus should not
15128 cause any problems.
15132 * Remove all references to RSAref, since there's no more need for it.
15136 * Make DSO load along a path given through an environment variable
15137 (SHLIB_PATH) with shl_load().
15141 * Constify the ENGINE code as a result of BIGNUM constification.
15142 Also constify the RSA code and most things related to it. In a
15143 few places, most notable in the depth of the ASN.1 code, ugly
15144 casts back to non-const were required (to be solved at a later
15149 * Make it so the openssl application has all engines loaded by default.
15153 * Constify the BIGNUM routines a little more.
15157 * Add the following functions:
15159 ENGINE_load_cswift()
15161 ENGINE_load_atalla()
15162 ENGINE_load_nuron()
15163 ENGINE_load_builtin_engines()
15165 That way, an application can itself choose if external engines that
15166 are built-in in OpenSSL shall ever be used or not. The benefit is
15167 that applications won't have to be linked with libdl or other dso
15168 libraries unless it's really needed.
15170 Changed 'openssl engine' to load all engines on demand.
15171 Changed the engine header files to avoid the duplication of some
15172 declarations (they differed!).
15176 * 'openssl engine' can now list capabilities.
15180 * Better error reporting in 'openssl engine'.
15184 * Never call load_dh_param(NULL) in s_server.
15188 * Add engine application. It can currently list engines by name and
15189 identity, and test if they are actually available.
15193 * Improve RPM specification file by forcing symbolic linking and making
15194 sure the installed documentation is also owned by root.root.
15196 *Damien Miller <djm@mindrot.org>*
15198 * Give the OpenSSL applications more possibilities to make use of
15199 keys (public as well as private) handled by engines.
15203 * Add OCSP code that comes from CertCo.
15207 * Add VMS support for the Rijndael code.
15211 * Added untested support for Nuron crypto accelerator.
15215 * Add support for external cryptographic devices. This code was
15216 previously distributed separately as the "engine" branch.
15218 *Geoff Thorpe, Richard Levitte*
15220 * Rework the filename-translation in the DSO code. It is now possible to
15221 have far greater control over how a "name" is turned into a filename
15222 depending on the operating environment and any oddities about the
15223 different shared library filenames on each system.
15227 * Support threads on FreeBSD-elf in Configure.
15231 * Fix for SHA1 assembly problem with MASM: it produces
15232 warnings about corrupt line number information when assembling
15233 with debugging information. This is caused by the overlapping
15236 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15239 NCONF_get_number() has no error checking at all. As a replacement,
15240 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15241 promoted strongly. The old NCONF_get_number is kept around for
15242 binary backward compatibility.
15243 Make it possible for methods to load from something other than a BIO,
15244 by providing a function pointer that is given a name instead of a BIO.
15245 For example, this could be used to load configuration data from an
15250 * Fix for non blocking accept BIOs. Added new I/O special reason
15251 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15252 with non blocking I/O was not possible because no retry code was
15253 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15258 * Added the beginnings of Rijndael support.
15262 * Fix for bug in DirectoryString mask setting. Add support for
15263 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15264 to allow certificate printing to more controllable, additional
15265 'certopt' option to 'x509' to allow new printing options to be
15270 * Clean old EAY MD5 hack from e_os.h.
15274 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15276 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15277 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15279 *Joe Orton, Steve Henson*
15281 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15283 * Fix additional bug revealed by the NISCC test suite:
15285 Stop bug triggering large recursion when presented with
15286 certain ASN.1 tags ([CVE-2003-0851])
15290 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15292 * Fix various bugs revealed by running the NISCC test suite:
15294 Stop out of bounds reads in the ASN1 code when presented with
15295 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15297 If verify callback ignores invalid public key errors don't try to check
15298 certificate signature with the NULL public key.
15302 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15303 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15308 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15309 extra data after the compression methods not only for TLS 1.0
15310 but also for SSL 3.0 (as required by the specification).
15312 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15314 * Change X509_certificate_type() to mark the key as exported/exportable
15315 when it's 512 *bits* long, not 512 bytes.
15319 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15321 * Countermeasure against the Klima-Pokorny-Rosa extension of
15322 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15323 a protocol version number mismatch like a decryption error
15324 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15328 * Turn on RSA blinding by default in the default implementation
15329 to avoid a timing attack. Applications that don't want it can call
15330 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15331 They would be ill-advised to do so in most cases.
15333 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15335 * Change RSA blinding code so that it works when the PRNG is not
15336 seeded (in this case, the secret RSA exponent is abused as
15337 an unpredictable seed -- if it is not unpredictable, there
15338 is no point in blinding anyway). Make RSA blinding thread-safe
15339 by remembering the creator's thread ID in rsa->blinding and
15340 having all other threads use local one-time blinding factors
15341 (this requires more computation than sharing rsa->blinding, but
15342 avoids excessive locking; and if an RSA object is not shared
15343 between threads, blinding will still be very fast).
15347 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15349 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15350 via timing by performing a MAC computation even if incorrect
15351 block cipher padding has been found. This is a countermeasure
15352 against active attacks where the attacker has to distinguish
15353 between bad padding and a MAC verification error. ([CVE-2003-0078])
15355 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15356 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15357 Martin Vuagnoux (EPFL, Ilion)*
15359 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15361 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15362 memory from its contents. This is done with a counter that will
15363 place alternating values in each byte. This can be used to solve
15364 two issues: 1) the removal of calls to memset() by highly optimizing
15365 compilers, and 2) cleansing with other values than 0, since those can
15366 be read through on certain media, for example a swap space on disk.
15370 * Bugfix: client side session caching did not work with external caching,
15371 because the session->cipher setting was not restored when reloading
15372 from the external cache. This problem was masked, when
15373 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15374 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15378 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15379 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15381 *Zeev Lieber <zeev-l@yahoo.com>*
15383 * Undo an undocumented change introduced in 0.9.6e which caused
15384 repeated calls to OpenSSL_add_all_ciphers() and
15385 OpenSSL_add_all_digests() to be ignored, even after calling
15390 * Change the default configuration reader to deal with last line not
15391 being properly terminated.
15395 * Change X509_NAME_cmp() so it applies the special rules on handling
15396 DN values that are of type PrintableString, as well as RDNs of type
15397 emailAddress where the value has the type ia5String.
15399 *stefank@valicert.com via Richard Levitte*
15401 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15402 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15403 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15404 the bitwise-OR of the two for use by the majority of applications
15405 wanting this behaviour, and update the docs. The documented
15406 behaviour and actual behaviour were inconsistent and had been
15407 changing anyway, so this is more a bug-fix than a behavioural
15410 *Geoff Thorpe, diagnosed by Nadav Har'El*
15412 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15413 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15417 * Fix initialization code race conditions in
15418 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15419 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15420 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15421 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15422 ssl2_get_cipher_by_char(),
15423 ssl3_get_cipher_by_char().
15425 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15427 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15428 the cached sessions are flushed, as the remove_cb() might use ex_data
15429 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15430 (see [openssl.org #212]).
15432 *Geoff Thorpe, Lutz Jaenicke*
15434 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15435 length, instead of the encoding length to d2i_ASN1_OBJECT.
15439 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15441 * [In 0.9.6g-engine release:]
15442 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15444 *Lynn Gazis <lgazis@rainbow.com>*
15446 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15448 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15449 and get fix the header length calculation.
15450 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15451 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15453 * Use proper error handling instead of 'assertions' in buffer
15454 overflow checks added in 0.9.6e. This prevents DoS (the
15455 assertions could call abort()).
15457 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15459 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15461 * Add various sanity checks to asn1_get_length() to reject
15462 the ASN1 length bytes if they exceed sizeof(long), will appear
15463 negative or the content length exceeds the length of the
15466 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15468 * Fix cipher selection routines: ciphers without encryption had no flags
15469 for the cipher strength set and where therefore not handled correctly
15470 by the selection routines (PR #130).
15474 * Fix EVP_dsa_sha macro.
15479 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15480 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15481 that was added in OpenSSL 0.9.6d.
15483 As the countermeasure turned out to be incompatible with some
15484 broken SSL implementations, the new option is part of SSL_OP_ALL.
15485 SSL_OP_ALL is usually employed when compatibility with weird SSL
15486 implementations is desired (e.g. '-bugs' option to 's_client' and
15487 's_server'), so the new option is automatically set in many
15492 * Changes in security patch:
15494 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15495 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15496 Air Force Materiel Command, USAF, under agreement number
15499 * Add various sanity checks to asn1_get_length() to reject
15500 the ASN1 length bytes if they exceed sizeof(long), will appear
15501 negative or the content length exceeds the length of the
15502 supplied buffer. ([CVE-2002-0659])
15504 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15506 * Assertions for various potential buffer overflows, not known to
15507 happen in practice.
15509 *Ben Laurie (CHATS)*
15511 * Various temporary buffers to hold ASCII versions of integers were
15512 too small for 64 bit platforms. ([CVE-2002-0655])
15513 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15515 * Remote buffer overflow in SSL3 protocol - an attacker could
15516 supply an oversized session ID to a client. ([CVE-2002-0656])
15518 *Ben Laurie (CHATS)*
15520 * Remote buffer overflow in SSL2 protocol - an attacker could
15521 supply an oversized client master key. ([CVE-2002-0656])
15523 *Ben Laurie (CHATS)*
15525 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15527 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15528 encoded as NULL) with id-dsa-with-sha1.
15530 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15532 * Check various `X509_...()` return values in `apps/req.c`.
15534 *Nils Larsch <nla@trustcenter.de>*
15536 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15537 an end-of-file condition would erroneously be flagged, when the CRLF
15538 was just at the end of a processed block. The bug was discovered when
15539 processing data through a buffering memory BIO handing the data to a
15540 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15541 <ptsekov@syntrex.com> and Nedelcho Stanev.
15545 * Implement a countermeasure against a vulnerability recently found
15546 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15547 before application data chunks to avoid the use of known IVs
15548 with data potentially chosen by the attacker.
15552 * Fix length checks in ssl3_get_client_hello().
15556 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15557 to prevent ssl3_read_internal() from incorrectly assuming that
15558 ssl3_read_bytes() found application data while handshake
15559 processing was enabled when in fact s->s3->in_read_app_data was
15560 merely automatically cleared during the initial handshake.
15562 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15564 * Fix object definitions for Private and Enterprise: they were not
15565 recognized in their shortname (=lowercase) representation. Extend
15566 obj_dat.pl to issue an error when using undefined keywords instead
15567 of silently ignoring the problem (Svenning Sorensen
15568 <sss@sss.dnsalias.net>).
15572 * Fix DH_generate_parameters() so that it works for 'non-standard'
15573 generators, i.e. generators other than 2 and 5. (Previously, the
15574 code did not properly initialise the 'add' and 'rem' values to
15575 BN_generate_prime().)
15577 In the new general case, we do not insist that 'generator' is
15578 actually a primitive root: This requirement is rather pointless;
15579 a generator of the order-q subgroup is just as good, if not
15584 * Map new X509 verification errors to alerts. Discovered and submitted by
15585 Tom Wu <tom@arcot.com>.
15589 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15590 returning non-zero before the data has been completely received
15591 when using non-blocking I/O.
15593 *Bodo Moeller; problem pointed out by John Hughes*
15595 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15597 *Ben Laurie, Lutz Jaenicke*
15599 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15600 Yoram Zahavi <YoramZ@gilian.com>).
15604 * Add information about CygWin 1.3 and on, and preserve proper
15605 configuration for the versions before that.
15607 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15609 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15610 check whether we deal with a copy of a session and do not delete from
15611 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15612 <izhar@checkpoint.com>.
15616 * Do not store session data into the internal session cache, if it
15617 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15618 flag is set). Proposed by Aslam <aslam@funk.com>.
15622 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15627 * [In 0.9.6d-engine release:]
15628 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15630 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15632 * Add the configuration target linux-s390x.
15634 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15636 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15637 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15638 variable as an indication that a ClientHello message has been
15639 received. As the flag value will be lost between multiple
15640 invocations of ssl3_accept when using non-blocking I/O, the
15641 function may not be aware that a handshake has actually taken
15642 place, thus preventing a new session from being added to the
15645 To avoid this problem, we now set s->new_session to 2 instead of
15646 using a local variable.
15648 *Lutz Jaenicke, Bodo Moeller*
15650 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15651 if the SSL_R_LENGTH_MISMATCH error is detected.
15653 *Geoff Thorpe, Bodo Moeller*
15655 * New 'shared_ldflag' column in Configure platform table.
15659 * Fix EVP_CIPHER_mode macro.
15661 *"Dan S. Camper" <dan@bti.net>*
15663 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15664 type, we must throw them away by setting rr->length to 0.
15666 *D P Chang <dpc@qualys.com>*
15668 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15670 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15671 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15672 worked incorrectly for those cases where range = `10..._2` and
15673 `3*range` is two bits longer than range.)
15677 * Only add signing time to PKCS7 structures if it is not already
15682 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15683 OBJ_ld_ce should be OBJ_id_ce.
15684 Also some ip-pda OIDs in crypto/objects/objects.txt were
15685 incorrect (cf. RFC 3039).
15687 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15689 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15690 returns early because it has nothing to do.
15692 *Andy Schneider <andy.schneider@bjss.co.uk>*
15694 * [In 0.9.6c-engine release:]
15695 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15697 *Andy Schneider <andy.schneider@bjss.co.uk>*
15699 * [In 0.9.6c-engine release:]
15700 Add support for Cryptographic Appliance's keyserver technology.
15701 (Use engine 'keyclient')
15703 *Cryptographic Appliances and Geoff Thorpe*
15705 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15706 is called via tools/c89.sh because arguments have to be
15707 rearranged (all '-L' options must appear before the first object
15710 *Richard Shapiro <rshapiro@abinitio.com>*
15712 * [In 0.9.6c-engine release:]
15713 Add support for Broadcom crypto accelerator cards, backported
15716 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15718 * [In 0.9.6c-engine release:]
15719 Add support for SureWare crypto accelerator cards from
15720 Baltimore Technologies. (Use engine 'sureware')
15722 *Baltimore Technologies and Mark Cox*
15724 * [In 0.9.6c-engine release:]
15725 Add support for crypto accelerator cards from Accelerated
15726 Encryption Processing, www.aep.ie. (Use engine 'aep')
15728 *AEP Inc. and Mark Cox*
15730 * Add a configuration entry for gcc on UnixWare.
15732 *Gary Benson <gbenson@redhat.com>*
15734 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15735 messages are stored in a single piece (fixed-length part and
15736 variable-length part combined) and fix various bugs found on the way.
15740 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15741 instead. BIO_gethostbyname() does not know what timeouts are
15742 appropriate, so entries would stay in cache even when they have
15744 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15746 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15747 faced with a pathologically small ClientHello fragment that does
15748 not contain client_version: Instead of aborting with an error,
15749 simply choose the highest available protocol version (i.e.,
15750 TLS 1.0 unless it is disabled). In practice, ClientHello
15751 messages are never sent like this, but this change gives us
15752 strictly correct behaviour at least for TLS.
15756 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15757 never resets s->method to s->ctx->method when called from within
15758 one of the SSL handshake functions.
15760 *Bodo Moeller; problem pointed out by Niko Baric*
15762 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15763 (sent using the client's version number) if client_version is
15764 smaller than the protocol version in use. Also change
15765 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15766 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15767 the client will at least see that alert.
15771 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15776 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15777 client receives HelloRequest while in a handshake.
15779 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15781 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15782 should end in 'break', not 'goto end' which circumvents various
15783 cleanups done in state SSL_ST_OK. But session related stuff
15784 must be disabled for SSL_ST_OK in the case that we just sent a
15787 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15788 before just sending a HelloRequest.
15790 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15792 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15793 reveal whether illegal block cipher padding was found or a MAC
15794 verification error occurred. (Neither SSLerr() codes nor alerts
15795 are directly visible to potential attackers, but the information
15796 may leak via logfiles.)
15798 Similar changes are not required for the SSL 2.0 implementation
15799 because the number of padding bytes is sent in clear for SSL 2.0,
15800 and the extra bytes are just ignored. However ssl/s2_pkt.c
15801 failed to verify that the purported number of padding bytes is in
15806 * Add OpenUNIX-8 support including shared libraries
15807 (Boyd Lynn Gerber <gerberb@zenez.com>).
15811 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15812 'wristwatch attack' using huge encoding parameters (cf.
15813 James H. Manger's CRYPTO 2001 paper). Note that the
15814 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15815 encoding parameters and hence was not vulnerable.
15819 * BN_sqr() bug fix.
15821 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15823 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15824 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15825 followed by modular reduction.
15827 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15829 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15830 equivalent based on BN_pseudo_rand() instead of BN_rand().
15834 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15835 This function was broken, as the check for a new client hello message
15836 to handle SGC did not allow these large messages.
15837 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15841 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15845 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15846 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15850 * Rework the configuration and shared library support for Tru64 Unix.
15851 The configuration part makes use of modern compiler features and
15852 still retains old compiler behavior for those that run older versions
15853 of the OS. The shared library support part includes a variant that
15854 uses the RPATH feature, and is available through the special
15855 configuration target "alpha-cc-rpath", which will never be selected
15858 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15860 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15861 with the same message size as in ssl3_get_certificate_request().
15862 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15863 messages might inadvertently be reject as too long.
15865 *Petr Lampa <lampa@fee.vutbr.cz>*
15867 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15871 * Modified SSL library such that the verify_callback that has been set
15872 specifically for an SSL object with SSL_set_verify() is actually being
15873 used. Before the change, a verify_callback set with this function was
15874 ignored and the verify_callback() set in the SSL_CTX at the time of
15875 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15876 to allow the necessary settings.
15880 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15881 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15882 done automatically (in contradiction to the requirements of the C
15883 standard). This made problems when used from OpenSSH.
15887 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15888 dh->length and always used
15890 BN_rand_range(priv_key, dh->p).
15892 BN_rand_range() is not necessary for Diffie-Hellman, and this
15893 specific range makes Diffie-Hellman unnecessarily inefficient if
15894 dh->length (recommended exponent length) is much smaller than the
15895 length of dh->p. We could use BN_rand_range() if the order of
15896 the subgroup was stored in the DH structure, but we only have
15901 BN_rand(priv_key, l, ...)
15903 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15910 RSA_eay_public_encrypt
15911 RSA_eay_private_decrypt
15912 RSA_eay_private_encrypt (signing)
15913 RSA_eay_public_decrypt (signature verification)
15915 (default implementations for RSA_public_encrypt,
15916 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15917 always reject numbers >= n.
15921 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15922 to synchronize access to 'locking_thread'. This is necessary on
15923 systems where access to 'locking_thread' (an 'unsigned long'
15924 variable) is not atomic.
15928 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15929 *before* setting the 'crypto_lock_rand' flag. The previous code had
15930 a race condition if 0 is a valid thread ID.
15932 *Travis Vitek <vitek@roguewave.com>*
15934 * Add support for shared libraries under Irix.
15936 *Albert Chin-A-Young <china@thewrittenword.com>*
15938 * Add configuration option to build on Linux on both big-endian and
15939 little-endian MIPS.
15941 *Ralf Baechle <ralf@uni-koblenz.de>*
15943 * Add the possibility to create shared libraries on HP-UX.
15947 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15949 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15950 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15951 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15952 PRNG state recovery was possible based on the output of
15953 one PRNG request appropriately sized to gain knowledge on
15954 'md' followed by enough consecutive 1-byte PRNG requests
15955 to traverse all of 'state'.
15957 1. When updating 'md_local' (the current thread's copy of 'md')
15958 during PRNG output generation, hash all of the previous
15959 'md_local' value, not just the half used for PRNG output.
15961 2. Make the number of bytes from 'state' included into the hash
15962 independent from the number of PRNG bytes requested.
15964 The first measure alone would be sufficient to avoid
15965 Markku-Juhani's attack. (Actually it had never occurred
15966 to me that the half of 'md_local' used for chaining was the
15967 half from which PRNG output bytes were taken -- I had always
15968 assumed that the secret half would be used.) The second
15969 measure makes sure that additional data from 'state' is never
15970 mixed into 'md_local' in small portions; this heuristically
15971 further strengthens the PRNG.
15975 * Fix crypto/bn/asm/mips3.s.
15979 * When only the key is given to "enc", the IV is undefined. Print out
15980 an error message in this case.
15984 * Handle special case when X509_NAME is empty in X509 printing routines.
15988 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15989 positive and less than q.
15993 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15994 used: it isn't thread safe and the add_lock_callback should handle
15997 *Paul Rose <Paul.Rose@bridge.com>*
15999 * Verify that incoming data obeys the block size in
16000 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
16006 *Ulf Möller, Bodo Möller*
16008 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
16009 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
16010 when fixing the server behaviour for backwards-compatible 'client
16011 hello' messages. (Note that the attack is impractical against
16012 SSL 3.0 and TLS 1.0 anyway because length and version checking
16013 means that the probability of guessing a valid ciphertext is
16014 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
16017 Before 0.9.5, the countermeasure (hide the error by generating a
16018 random 'decryption result') did not work properly because
16019 ERR_clear_error() was missing, meaning that SSL_get_error() would
16020 detect the supposedly ignored error.
16022 Both problems are now fixed.
16026 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
16027 (previously it was 1024).
16031 * Fix for compatibility mode trust settings: ignore trust settings
16032 unless some valid trust or reject settings are present.
16036 * Fix for blowfish EVP: its a variable length cipher.
16040 * Fix various bugs related to DSA S/MIME verification. Handle missing
16041 parameters in DSA public key structures and return an error in the
16042 DSA routines if parameters are absent.
16046 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
16047 in the current directory if neither $RANDFILE nor $HOME was set.
16048 RAND_file_name() in 0.9.6a returned NULL in this case. This has
16049 caused some confusion to Windows users who haven't defined $HOME.
16050 Thus RAND_file_name() is changed again: e_os.h can define a
16051 DEFAULT_HOME, which will be used if $HOME is not set.
16052 For Windows, we use "C:"; on other platforms, we still require
16053 environment variables.
16055 * Move 'if (!initialized) RAND_poll()' into regions protected by
16056 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
16057 having multiple threads call RAND_poll() concurrently.
16061 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
16062 combination of a flag and a thread ID variable.
16063 Otherwise while one thread is in ssleay_rand_bytes (which sets the
16064 flag), *other* threads can enter ssleay_add_bytes without obeying
16065 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
16066 that they do not hold after the first thread unsets add_do_not_lock).
16070 * Change bctest again: '-x' expressions are not available in all
16071 versions of 'test'.
16075 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16077 * Fix a couple of memory leaks in PKCS7_dataDecode()
16079 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16081 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16082 the default extension for executables, if any. Also, make the perl
16083 scripts that use symlink() to test if it really exists and use "cp"
16084 if it doesn't. All this made OpenSSL compilable and installable in
16089 * Fix for asn1_GetSequence() for indefinite length constructed data.
16090 If SEQUENCE is length is indefinite just set c->slen to the total
16091 amount of data available.
16093 *Steve Henson, reported by shige@FreeBSD.org*
16095 *This change does not apply to 0.9.7.*
16097 * Change bctest to avoid here-documents inside command substitution
16098 (workaround for FreeBSD /bin/sh bug).
16099 For compatibility with Ultrix, avoid shell functions (introduced
16100 in the bctest version that searches along $PATH).
16104 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16105 with des_encrypt() defined on some operating systems, like Solaris
16110 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16111 On the Importance of Eliminating Errors in Cryptographic
16112 Computations, J. Cryptology 14 (2001) 2, 101-119,
16113 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16117 * MIPS assembler BIGNUM division bug fix.
16121 * Disabled incorrect Alpha assembler code.
16125 * Fix PKCS#7 decode routines so they correctly update the length
16126 after reading an EOC for the EXPLICIT tag.
16130 *This change does not apply to 0.9.7.*
16132 * Fix bug in PKCS#12 key generation routines. This was triggered
16133 if a 3DES key was generated with a 0 initial byte. Include
16134 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16135 (but broken) behaviour.
16139 * Enhance bctest to search for a working bc along $PATH and print
16142 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16144 * Fix memory leaks in err.c: free err_data string if necessary;
16145 don't write to the wrong index in ERR_set_error_data.
16149 * Implement ssl23_peek (analogous to ssl23_read), which previously
16154 * Replace rdtsc with `_emit` statements for VC++ version 5.
16156 *Jeremy Cooper <jeremy@baymoo.org>*
16158 * Make it possible to reuse SSLv2 sessions.
16162 * In copy_email() check for >= 0 as a return value for
16163 X509_NAME_get_index_by_NID() since 0 is a valid index.
16165 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16167 * Avoid coredump with unsupported or invalid public keys by checking if
16168 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16169 PKCS7_verify() fails with non detached data.
16173 * Don't use getenv in library functions when run as setuid/setgid.
16174 New function OPENSSL_issetugid().
16178 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16179 due to incorrect handling of multi-threading:
16181 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16183 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16185 3. Count how many times MemCheck_off() has been called so that
16186 nested use can be treated correctly. This also avoids
16187 inband-signalling in the previous code (which relied on the
16188 assumption that thread ID 0 is impossible).
16192 * Add "-rand" option also to s_client and s_server.
16196 * Fix CPU detection on Irix 6.x.
16197 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16198 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16200 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16205 *This change does not apply to 0.9.7.*
16207 * Use the cached encoding of an X509_NAME structure rather than
16208 copying it. This is apparently the reason for the libsafe "errors"
16209 but the code is actually correct.
16213 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16214 Bleichenbacher's DSA attack.
16215 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16216 to be set and top=0 forces the highest bit to be set; top=-1 is new
16217 and leaves the highest bit random.
16219 *Ulf Moeller, Bodo Moeller*
16221 * In the `NCONF_...`-based implementations for `CONF_...` queries
16222 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16223 a temporary CONF structure with the data component set to NULL
16224 (which gives segmentation faults in lh_retrieve).
16225 Instead, use NULL for the CONF pointer in CONF_get_string and
16226 CONF_get_number (which may use environment variables) and directly
16227 return NULL from CONF_get_section.
16231 * Fix potential buffer overrun for EBCDIC.
16235 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16236 keyUsage if basicConstraints absent for a CA.
16240 * Make SMIME_write_PKCS7() write mail header values with a format that
16241 is more generally accepted (no spaces before the semicolon), since
16242 some programs can't parse those values properly otherwise. Also make
16243 sure BIO's that break lines after each write do not create invalid
16248 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16249 macros previously used would not encode an empty SEQUENCE OF
16250 and break the signature.
16254 *This change does not apply to 0.9.7.*
16256 * Zero the premaster secret after deriving the master secret in
16261 * Add some EVP_add_digest_alias registrations (as found in
16262 OpenSSL_add_all_digests()) to SSL_library_init()
16263 aka OpenSSL_add_ssl_algorithms(). This provides improved
16264 compatibility with peers using X.509 certificates
16265 with unconventional AlgorithmIdentifier OIDs.
16269 * Fix for Irix with NO_ASM.
16271 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16273 * ./config script fixes.
16275 *Ulf Moeller, Richard Levitte*
16277 * Fix 'openssl passwd -1'.
16281 * Change PKCS12_key_gen_asc() so it can cope with non null
16282 terminated strings whose length is passed in the passlen
16283 parameter, for example from PEM callbacks. This was done
16284 by adding an extra length parameter to asc2uni().
16286 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16288 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16289 call failed, free the DSA structure.
16293 * Fix to uni2asc() to cope with zero length Unicode strings.
16294 These are present in some PKCS#12 files.
16298 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16299 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16300 when writing a 32767 byte record.
16302 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16304 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16305 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16307 (RSA objects have a reference count access to which is protected
16308 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16309 so they are meant to be shared between threads.)
16310 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16311 "Reddie, Steven" <Steven.Reddie@ca.com>*
16313 * Fix a deadlock in CRYPTO_mem_leaks().
16317 * Use better test patterns in bntest.
16321 * rand_win.c fix for Borland C.
16325 * BN_rshift bugfix for n == 0.
16329 * Add a 'bctest' script that checks for some known 'bc' bugs
16330 so that 'make test' does not abort just because 'bc' is broken.
16334 * Store verify_result within SSL_SESSION also for client side to
16335 avoid potential security hole. (Re-used sessions on the client side
16336 always resulted in verify_result==X509_V_OK, not using the original
16337 result of the server certificate verification.)
16341 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16342 SSL3_RT_APPLICATION_DATA, return 0.
16343 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16348 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16349 releases, have been re-implemented by renaming the previous
16350 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16351 and ssl3_read_internal, respectively, and adding 'peek' parameters
16352 to them. The new ssl[23]_{read,peek} functions are calls to
16353 ssl[23]_read_internal with the 'peek' flag set appropriately.
16354 A 'peek' parameter has also been added to ssl3_read_bytes, which
16355 does the actual work for ssl3_read_internal.
16359 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16360 the method-specific "init()" handler. Also clean up ex_data after
16361 calling the method-specific "finish()" handler. Previously, this was
16362 happening the other way round.
16366 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16367 The previous value, 12, was not always sufficient for BN_mod_exp().
16371 * Make sure that shared libraries get the internal name engine with
16372 the full version number and not just 0. This should mark the
16373 shared libraries as not backward compatible. Of course, this should
16374 be changed again when we can guarantee backward binary compatibility.
16378 * Fix typo in get_cert_by_subject() in by_dir.c
16380 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16382 * Rework the system to generate shared libraries:
16384 - Make note of the expected extension for the shared libraries and
16385 if there is a need for symbolic links from for example libcrypto.so.0
16386 to libcrypto.so.0.9.7. There is extended info in Configure for
16389 - Make as few rebuilds of the shared libraries as possible.
16391 - Still avoid linking the OpenSSL programs with the shared libraries.
16393 - When installing, install the shared libraries separately from the
16398 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16400 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16401 and not in SSL_clear because the latter is also used by the
16402 accept/connect functions; previously, the settings made by
16403 SSL_set_read_ahead would be lost during the handshake.
16405 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16407 * Correct util/mkdef.pl to be selective about disabled algorithms.
16408 Previously, it would create entries for disabled algorithms no
16413 * Added several new manual pages for SSL_* function.
16417 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16419 * In ssl23_get_client_hello, generate an error message when faced
16420 with an initial SSL 3.0/TLS record that is too small to contain the
16421 first two bytes of the ClientHello message, i.e. client_version.
16422 (Note that this is a pathologic case that probably has never happened
16423 in real life.) The previous approach was to use the version number
16424 from the record header as a substitute; but our protocol choice
16425 should not depend on that one because it is not authenticated
16426 by the Finished messages.
16430 * More robust randomness gathering functions for Windows.
16432 *Jeffrey Altman <jaltman@columbia.edu>*
16434 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16435 not set then we don't setup the error code for issuer check errors
16436 to avoid possibly overwriting other errors which the callback does
16437 handle. If an application does set the flag then we assume it knows
16438 what it is doing and can handle the new informational codes
16443 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16444 a general "ANY" type, as such it should be able to decode anything
16445 including tagged types. However it didn't check the class so it would
16446 wrongly interpret tagged types in the same way as their universal
16447 counterpart and unknown types were just rejected. Changed so that the
16448 tagged and unknown types are handled in the same way as a SEQUENCE:
16449 that is the encoding is stored intact. There is also a new type
16450 "V_ASN1_OTHER" which is used when the class is not universal, in this
16451 case we have no idea what the actual type is so we just lump them all
16456 * On VMS, stdout may very well lead to a file that is written to
16457 in a record-oriented fashion. That means that every write() will
16458 write a separate record, which will be read separately by the
16459 programs trying to read from it. This can be very confusing.
16461 The solution is to put a BIO filter in the way that will buffer
16462 text until a linefeed is reached, and then write everything a
16463 line at a time, so every record written will be an actual line,
16464 not chunks of lines and not (usually doesn't happen, but I've
16465 seen it once) several lines in one record. BIO_f_linebuffer() is
16468 Currently, it's a VMS-only method, because that's where it has
16469 been tested well enough.
16473 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16474 it can return incorrect results.
16475 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16476 but it was in 0.9.6-beta[12].)
16480 * Disable the check for content being present when verifying detached
16481 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16482 include zero length content when signing messages.
16486 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16487 BIO_ctrl (for BIO pairs).
16491 * Add DSO method for VMS.
16495 * Bug fix: Montgomery multiplication could produce results with the
16500 * Add RPM specification openssl.spec and modify it to build three
16501 packages. The default package contains applications, application
16502 documentation and run-time libraries. The devel package contains
16503 include files, static libraries and function documentation. The
16504 doc package contains the contents of the doc directory. The original
16505 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16509 * Add a large number of documentation files for many SSL routines.
16511 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16513 * Add a configuration entry for Sony News 4.
16515 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16517 * Don't set the two most significant bits to one when generating a
16518 random number < q in the DSA library.
16522 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16523 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16524 the underlying transport is blocking) if a handshake took place.
16525 (The default behaviour is needed by applications such as s_client
16526 and s_server that use select() to determine when to use SSL_read;
16527 but for applications that know in advance when to expect data, it
16528 just makes things more complicated.)
16532 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16537 * Add a few more EBCDIC conditionals that make `req` and `x509`
16538 work better on such systems.
16540 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16542 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16543 Update PKCS12_parse() so it copies the friendlyName and the
16544 keyid to the certificates aux info.
16548 * Fix bug in PKCS7_verify() which caused an infinite loop
16549 if there was more than one signature.
16551 *Sven Uszpelkat <su@celocom.de>*
16553 * Major change in util/mkdef.pl to include extra information
16554 about each symbol, as well as presenting variables as well
16555 as functions. This change means that there's n more need
16556 to rebuild the .num files when some algorithms are excluded.
16560 * Allow the verify time to be set by an application,
16561 rather than always using the current time.
16565 * Phase 2 verify code reorganisation. The certificate
16566 verify code now looks up an issuer certificate by a
16567 number of criteria: subject name, authority key id
16568 and key usage. It also verifies self signed certificates
16569 by the same criteria. The main comparison function is
16570 X509_check_issued() which performs these checks.
16572 Lot of changes were necessary in order to support this
16573 without completely rewriting the lookup code.
16575 Authority and subject key identifier are now cached.
16577 The LHASH 'certs' is X509_STORE has now been replaced
16578 by a STACK_OF(X509_OBJECT). This is mainly because an
16579 LHASH can't store or retrieve multiple objects with
16580 the same hash value.
16582 As a result various functions (which were all internal
16583 use only) have changed to handle the new X509_STORE
16584 structure. This will break anything that messed round
16585 with X509_STORE internally.
16587 The functions X509_STORE_add_cert() now checks for an
16588 exact match, rather than just subject name.
16590 The X509_STORE API doesn't directly support the retrieval
16591 of multiple certificates matching a given criteria, however
16592 this can be worked round by performing a lookup first
16593 (which will fill the cache with candidate certificates)
16594 and then examining the cache for matches. This is probably
16595 the best we can do without throwing out X509_LOOKUP
16596 entirely (maybe later...).
16598 The X509_VERIFY_CTX structure has been enhanced considerably.
16600 All certificate lookup operations now go via a get_issuer()
16601 callback. Although this currently uses an X509_STORE it
16602 can be replaced by custom lookups. This is a simple way
16603 to bypass the X509_STORE hackery necessary to make this
16604 work and makes it possible to use more efficient techniques
16605 in future. A very simple version which uses a simple
16606 STACK for its trusted certificate store is also provided
16607 using X509_STORE_CTX_trusted_stack().
16609 The verify_cb() and verify() callbacks now have equivalents
16610 in the X509_STORE_CTX structure.
16612 X509_STORE_CTX also has a 'flags' field which can be used
16613 to customise the verify behaviour.
16617 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16618 excludes S/MIME capabilities.
16622 * When a certificate request is read in keep a copy of the
16623 original encoding of the signed data and use it when outputting
16624 again. Signatures then use the original encoding rather than
16625 a decoded, encoded version which may cause problems if the
16626 request is improperly encoded.
16630 * For consistency with other BIO_puts implementations, call
16631 buffer_write(b, ...) directly in buffer_puts instead of calling
16634 In BIO_puts, increment b->num_write as in BIO_write.
16636 *Peter.Sylvester@EdelWeb.fr*
16638 * Fix BN_mul_word for the case where the word is 0. (We have to use
16639 BN_zero, we may not return a BIGNUM with an array consisting of
16640 words set to zero.)
16644 * Avoid calling abort() from within the library when problems are
16645 detected, except if preprocessor symbols have been defined
16646 (such as REF_CHECK, BN_DEBUG etc.).
16650 * New openssl application 'rsautl'. This utility can be
16651 used for low-level RSA operations. DER public key
16652 BIO/fp routines also added.
16656 * New Configure entry and patches for compiling on QNX 4.
16658 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16660 * A demo state-machine implementation was sponsored by
16661 Nuron (<http://www.nuron.com/>) and is now available in
16662 demos/state_machine.
16666 * New options added to the 'dgst' utility for signature
16667 generation and verification.
16671 * Unrecognized PKCS#7 content types are now handled via a
16672 catch all ASN1_TYPE structure. This allows unsupported
16673 types to be stored as a "blob" and an application can
16674 encode and decode it manually.
16678 * Fix various signed/unsigned issues to make a_strex.c
16679 compile under VC++.
16681 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16683 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16684 length if passed a buffer. ASN1_INTEGER_to_BN failed
16685 if passed a NULL BN and its argument was negative.
16687 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16689 * Modification to PKCS#7 encoding routines to output definite
16690 length encoding. Since currently the whole structures are in
16691 memory there's not real point in using indefinite length
16692 constructed encoding. However if OpenSSL is compiled with
16693 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16697 * Added BIO_vprintf() and BIO_vsnprintf().
16701 * Added more prefixes to parse for in the strings written
16702 through a logging bio, to cover all the levels that are available
16703 through syslog. The prefixes are now:
16705 PANIC, EMERG, EMR => LOG_EMERG
16706 ALERT, ALR => LOG_ALERT
16707 CRIT, CRI => LOG_CRIT
16708 ERROR, ERR => LOG_ERR
16709 WARNING, WARN, WAR => LOG_WARNING
16710 NOTICE, NOTE, NOT => LOG_NOTICE
16711 INFO, INF => LOG_INFO
16712 DEBUG, DBG => LOG_DEBUG
16714 and as before, if none of those prefixes are present at the
16715 beginning of the string, LOG_ERR is chosen.
16717 On Win32, the `LOG_*` levels are mapped according to this:
16719 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16720 LOG_WARNING => EVENTLOG_WARNING_TYPE
16721 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16725 * Made it possible to reconfigure with just the configuration
16726 argument "reconf" or "reconfigure". The command line arguments
16727 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16728 and are retrieved from there when reconfiguring.
16734 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16736 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16740 * The obj_dat.pl script was messing up the sorting of object
16741 names. The reason was that it compared the quoted version
16742 of strings as a result "OCSP" > "OCSP Signing" because
16743 " > SPACE. Changed script to store unquoted versions of
16744 names and add quotes on output. It was also omitting some
16745 names from the lookup table if they were given a default
16746 value (that is if SN is missing it is given the same
16747 value as LN and vice versa), these are now added on the
16748 grounds that if an object has a name we should be able to
16749 look it up. Finally added warning output when duplicate
16750 short or long names are found.
16754 * Changes needed for Tandem NSK.
16756 *Scott Uroff <scott@xypro.com>*
16758 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16759 RSA_padding_check_SSLv23(), special padding was never detected
16760 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16761 version rollback attacks was not effective.
16763 In s23_clnt.c, don't use special rollback-attack detection padding
16764 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16765 client; similarly, in s23_srvr.c, don't do the rollback check if
16766 SSL 2.0 is the only protocol enabled in the server.
16770 * Make it possible to get hexdumps of unprintable data with 'openssl
16771 asn1parse'. By implication, the functions ASN1_parse_dump() and
16772 BIO_dump_indent() are added.
16776 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16777 these print out strings and name structures based on various
16778 flags including RFC2253 support and proper handling of
16779 multibyte characters. Added options to the 'x509' utility
16780 to allow the various flags to be set.
16784 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16785 Also change the functions X509_cmp_current_time() and
16786 X509_gmtime_adj() work with an ASN1_TIME structure,
16787 this will enable certificates using GeneralizedTime in validity
16788 dates to be checked.
16792 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16793 negative public key encodings) on by default,
16794 NO_NEG_PUBKEY_BUG can be set to disable it.
16798 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16799 content octets. An i2c_ASN1_OBJECT is unnecessary because
16800 the encoding can be trivially obtained from the structure.
16804 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16805 not read locks (`CRYPTO_r_[un]lock`).
16809 * A first attempt at creating official support for shared
16810 libraries through configuration. I've kept it so the
16811 default is static libraries only, and the OpenSSL programs
16812 are always statically linked for now, but there are
16813 preparations for dynamic linking in place.
16814 This has been tested on Linux and Tru64.
16818 * Randomness polling function for Win9x, as described in:
16819 Peter Gutmann, Software Generation of Practically Strong
16824 * Fix so PRNG is seeded in req if using an already existing
16829 * New options to smime application. -inform and -outform
16830 allow alternative formats for the S/MIME message including
16831 PEM and DER. The -content option allows the content to be
16832 specified separately. This should allow things like Netscape
16833 form signing output easier to verify.
16837 * Fix the ASN1 encoding of tags using the 'long form'.
16841 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16842 STRING types. These convert content octets to and from the
16843 underlying type. The actual tag and length octets are
16844 already assumed to have been read in and checked. These
16845 are needed because all other string types have virtually
16846 identical handling apart from the tag. By having versions
16847 of the ASN1 functions that just operate on content octets
16848 IMPLICIT tagging can be handled properly. It also allows
16849 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16850 and ASN1_INTEGER are identical apart from the tag.
16854 * Change the handling of OID objects as follows:
16856 - New object identifiers are inserted in objects.txt, following
16857 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16858 - objects.pl is used to process obj_mac.num and create a new
16860 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16863 This is currently kind of a hack, and the perl code in objects.pl
16864 isn't very elegant, but it works as I intended. The simplest way
16865 to check that it worked correctly is to look in obj_dat.h and
16866 check the array nid_objs and make sure the objects haven't moved
16867 around (this is important!). Additions are OK, as well as
16868 consistent name changes.
16872 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16876 * Addition of the command line parameter '-rand file' to 'openssl req'.
16877 The given file adds to whatever has already been seeded into the
16878 random pool through the RANDFILE configuration file option or
16879 environment variable, or the default random state file.
16883 * mkstack.pl now sorts each macro group into lexical order.
16884 Previously the output order depended on the order the files
16885 appeared in the directory, resulting in needless rewriting
16890 * Patches to make OpenSSL compile under Win32 again. Mostly
16891 work arounds for the VC++ problem that it treats func() as
16892 func(void). Also stripped out the parts of mkdef.pl that
16893 added extra typesafe functions: these no longer exist.
16897 * Reorganisation of the stack code. The macros are now all
16898 collected in safestack.h . Each macro is defined in terms of
16899 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16900 DEBUG_SAFESTACK is now handled in terms of function casts,
16901 this has the advantage of retaining type safety without the
16902 use of additional functions. If DEBUG_SAFESTACK is not defined
16903 then the non typesafe macros are used instead. Also modified the
16904 mkstack.pl script to handle the new form. Needs testing to see
16905 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16906 the default if no major problems. Similar behaviour for ASN1_SET_OF
16907 and PKCS12_STACK_OF.
16911 * When some versions of IIS use the 'NET' form of private key the
16912 key derivation algorithm is different. Normally MD5(password) is
16913 used as a 128 bit RC4 key. In the modified case
16914 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16915 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16916 as the old Netscape_RSA functions except they have an additional
16917 'sgckey' parameter which uses the modified algorithm. Also added
16918 an -sgckey command line option to the rsa utility. Thanks to
16919 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16920 algorithm to openssl-dev.
16924 * The evp_local.h macros were using 'c.##kname' which resulted in
16925 invalid expansion on some systems (SCO 5.0.5 for example).
16926 Corrected to 'c.kname'.
16928 *Phillip Porch <root@theporch.com>*
16930 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16931 a STACK of email addresses from a certificate or request, these look
16932 in the subject name and the subject alternative name extensions and
16933 omit any duplicate addresses.
16937 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16938 This makes DSA verification about 2 % faster.
16942 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16943 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16944 plus overhead for 1024 bit moduli).
16945 This makes exponentiations about 0.5 % faster for 1024 bit
16946 exponents (as measured by "openssl speed rsa2048").
16950 * Rename memory handling macros to avoid conflicts with other
16952 Malloc => OPENSSL_malloc
16953 Malloc_locked => OPENSSL_malloc_locked
16954 Realloc => OPENSSL_realloc
16955 Free => OPENSSL_free
16959 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16960 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16964 * CygWin32 support.
16966 *John Jarvie <jjarvie@newsguy.com>*
16968 * The type-safe stack code has been rejigged. It is now only compiled
16969 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16970 by default all type-specific stack functions are "#define"d back to
16971 standard stack functions. This results in more streamlined output
16972 but retains the type-safety checking possibilities of the original
16977 * The STACK code has been cleaned up, and certain type declarations
16978 that didn't make a lot of sense have been brought in line. This has
16979 also involved a cleanup of sorts in safestack.h to more correctly
16980 map type-safe stack functions onto their plain stack counterparts.
16981 This work has also resulted in a variety of "const"ifications of
16982 lots of the code, especially `_cmp` operations which should normally
16983 be prototyped with "const" parameters anyway.
16987 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16988 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16989 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16990 where all of 'md' is used each time the PRNG is used, but 'state'
16991 is used only indexed by a cyclic counter. As entropy may not be
16992 well distributed from the beginning, 'md' is important as a
16993 chaining variable. However, the output function chains only half
16994 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16995 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16996 in all of 'state' being rewritten, with the new values depending
16997 on virtually all of 'md'. This overcomes the 80 bit limitation.)
17001 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
17002 the handshake is continued after ssl_verify_cert_chain();
17003 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
17004 can lead to 'unexplainable' connection aborts later.
17006 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
17008 * Major EVP API cipher revision.
17009 Add hooks for extra EVP features. This allows various cipher
17010 parameters to be set in the EVP interface. Support added for variable
17011 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
17012 setting of RC2 and RC5 parameters.
17014 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
17017 Remove lots of duplicated code from the EVP library. For example *every*
17018 cipher init() function handles the 'iv' in the same way according to the
17019 cipher mode. They also all do nothing if the 'key' parameter is NULL and
17020 for CFB and OFB modes they zero ctx->num.
17022 New functionality allows removal of S/MIME code RC2 hack.
17024 Most of the routines have the same form and so can be declared in terms
17027 By shifting this to the top level EVP_CipherInit() it can be removed from
17028 all individual ciphers. If the cipher wants to handle IVs or keys
17029 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
17032 Change lots of functions like EVP_EncryptUpdate() to now return a
17033 value: although software versions of the algorithms cannot fail
17034 any installed hardware versions can.
17038 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
17039 this option is set, tolerate broken clients that send the negotiated
17040 protocol version number instead of the requested protocol version
17045 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
17046 i.e. non-zero for export ciphersuites, zero otherwise.
17047 Previous versions had this flag inverted, inconsistent with
17048 rsa_tmp_cb (..._TMP_RSA_CB).
17050 *Bodo Moeller; problem reported by Amit Chopra*
17052 * Add missing DSA library text string. Work around for some IIS
17053 key files with invalid SEQUENCE encoding.
17057 * Add a document (doc/standards.txt) that list all kinds of standards
17058 and so on that are implemented in OpenSSL.
17062 * Enhance c_rehash script. Old version would mishandle certificates
17063 with the same subject name hash and wouldn't handle CRLs at all.
17064 Added -fingerprint option to crl utility, to support new c_rehash
17069 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17073 * Fix for SSL server purpose checking. Server checking was
17074 rejecting certificates which had extended key usage present
17075 but no ssl client purpose.
17077 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17079 * Make PKCS#12 code work with no password. The PKCS#12 spec
17080 is a little unclear about how a blank password is handled.
17081 Since the password in encoded as a BMPString with terminating
17082 double NULL a zero length password would end up as just the
17083 double NULL. However no password at all is different and is
17084 handled differently in the PKCS#12 key generation code. NS
17085 treats a blank password as zero length. MSIE treats it as no
17086 password on export: but it will try both on import. We now do
17087 the same: PKCS12_parse() tries zero length and no password if
17088 the password is set to "" or NULL (NULL is now a valid password:
17089 it wasn't before) as does the pkcs12 application.
17093 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17094 perror when PEM_read_bio_X509_REQ fails, the error message must
17095 be obtained from the error queue.
17099 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17100 it in ERR_remove_state if appropriate, and change ERR_get_state
17101 accordingly to avoid race conditions (this is necessary because
17102 thread_hash is no longer constant once set).
17106 * Bugfix for linux-elf makefile.one.
17110 * RSA_get_default_method() will now cause a default
17111 RSA_METHOD to be chosen if one doesn't exist already.
17112 Previously this was only set during a call to RSA_new()
17113 or RSA_new_method(NULL) meaning it was possible for
17114 RSA_get_default_method() to return NULL.
17118 * Added native name translation to the existing DSO code
17119 that will convert (if the flag to do so is set) filenames
17120 that are sufficiently small and have no path information
17121 into a canonical native form. Eg. "blah" converted to
17122 "libblah.so" or "blah.dll" etc.
17126 * New function ERR_error_string_n(e, buf, len) which is like
17127 ERR_error_string(e, buf), but writes at most 'len' bytes
17128 including the 0 terminator. For ERR_error_string_n, 'buf'
17131 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17133 * CONF library reworked to become more general. A new CONF
17134 configuration file reader "class" is implemented as well as a
17135 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17136 old `CONF_*` functions are still there, but are reimplemented to
17137 work in terms of the new functions. Also, a set of functions
17138 to handle the internal storage of the configuration data is
17139 provided to make it easier to write new configuration file
17140 reader "classes" (I can definitely see something reading a
17141 configuration file in XML format, for example), called `_CONF_*`,
17142 or "the configuration storage API"...
17144 The new configuration file reading functions are:
17146 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17147 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17149 NCONF_default, NCONF_WIN32
17151 NCONF_dump_fp, NCONF_dump_bio
17153 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17154 NCONF_new creates a new CONF object. This works in the same way
17155 as other interfaces in OpenSSL, like the BIO interface.
17156 `NCONF_dump_*` dump the internal storage of the configuration file,
17157 which is useful for debugging. All other functions take the same
17158 arguments as the old `CONF_*` functions with the exception of the
17159 first that must be a `CONF *` instead of a `LHASH *`.
17161 To make it easier to use the new classes with the old `CONF_*` functions,
17162 the function CONF_set_default_method is provided.
17166 * Add '-tls1' option to 'openssl ciphers', which was already
17167 mentioned in the documentation but had not been implemented.
17168 (This option is not yet really useful because even the additional
17169 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17173 * Initial DSO code added into libcrypto for letting OpenSSL (and
17174 OpenSSL-based applications) load shared libraries and bind to
17175 them in a portable way.
17177 *Geoff Thorpe, with contributions from Richard Levitte*
17179 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17181 * Make sure _lrotl and _lrotr are only used with MSVC.
17183 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17184 (the default implementation of RAND_status).
17186 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17187 to '-clrext' (= clear extensions), as intended and documented.
17188 *Bodo Moeller; inconsistency pointed out by Michael Attili
17189 <attili@amaxo.com>*
17191 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17192 was larger than the MD block size.
17194 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17196 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17197 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17198 using the passed key: if the passed key was a private key the result
17199 of X509_print(), for example, would be to print out all the private key
17204 * des_quad_cksum() byte order bug fix.
17205 *Ulf Möller, using the problem description in krb4-0.9.7, where
17206 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17208 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17211 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17213 * For easily testing in shell scripts whether some command
17214 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17215 returns with exit code 0 iff no command of the given name is available.
17216 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17217 the output goes to stdout and nothing is printed to stderr.
17218 Additional arguments are always ignored.
17220 Since for each cipher there is a command of the same name,
17221 the 'no-cipher' compilation switches can be tested this way.
17223 ('openssl no-XXX' is not able to detect pseudo-commands such
17224 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17228 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17232 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17233 is set; it will be thrown away anyway because each handshake creates
17235 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17236 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17237 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17238 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17242 * New s_client option -ign_eof: EOF at stdin is ignored, and
17243 'Q' and 'R' lose their special meanings (quit/renegotiate).
17244 This is part of what -quiet does; unlike -quiet, -ign_eof
17245 does not suppress any output.
17249 * Add compatibility options to the purpose and trust code. The
17250 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17251 accepts a certificate or CA, this was the previous behaviour,
17252 with all the associated security issues.
17254 X509_TRUST_COMPAT is the old trust behaviour: only and
17255 automatically trust self signed roots in certificate store. A
17256 new trust setting X509_TRUST_DEFAULT is used to specify that
17257 a purpose has no associated trust setting and it should instead
17258 use the value in the default purpose.
17262 * Fix the PKCS#8 DSA private key code so it decodes keys again
17263 and fix a memory leak.
17267 * In util/mkerr.pl (which implements 'make errors'), preserve
17268 reason strings from the previous version of the .c file, as
17269 the default to have only downcase letters (and digits) in
17270 automatically generated reasons codes is not always appropriate.
17274 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17275 using strerror. Previously, ERR_reason_error_string() returned
17276 library names as reason strings for SYSerr; but SYSerr is a special
17277 case where small numbers are errno values, not library numbers.
17281 * Add '-dsaparam' option to 'openssl dhparam' application. This
17282 converts DSA parameters into DH parameters. (When creating parameters,
17283 DSA_generate_parameters is used.)
17287 * Include 'length' (recommended exponent length) in C code generated
17288 by 'openssl dhparam -C'.
17292 * The second argument to set_label in perlasm was already being used
17293 so couldn't be used as a "file scope" flag. Moved to third argument
17298 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17299 instead of RAND_bytes for encryption IVs and salts.
17303 * Include RAND_status() into RAND_METHOD instead of implementing
17304 it only for md_rand.c Otherwise replacing the PRNG by calling
17305 RAND_set_rand_method would be impossible.
17309 * Don't let DSA_generate_key() enter an infinite loop if the random
17310 number generation fails.
17314 * New 'rand' application for creating pseudo-random output.
17318 * Added configuration support for Linux/IA64
17320 *Rolf Haberrecker <rolf@suse.de>*
17322 * Assembler module support for Mingw32.
17326 * Shared library support for HPUX (in shlib/).
17328 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17330 * Shared library support for Solaris gcc.
17332 *Lutz Behnke <behnke@trustcenter.de>*
17334 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17336 * PKCS7_encrypt() was adding text MIME headers twice because they
17337 were added manually and by SMIME_crlf_copy().
17341 * In bntest.c don't call BN_rand with zero bits argument.
17343 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17345 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17346 case was implemented. This caused BN_div_recp() to fail occasionally.
17350 * Add an optional second argument to the set_label() in the perl
17351 assembly language builder. If this argument exists and is set
17352 to 1 it signals that the assembler should use a symbol whose
17353 scope is the entire file, not just the current function. This
17354 is needed with MASM which uses the format label:: for this scope.
17356 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17358 * Change the ASN1 types so they are typedefs by default. Before
17359 almost all types were #define'd to ASN1_STRING which was causing
17360 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17365 * Change names of new functions to the new get1/get0 naming
17366 convention: After 'get1', the caller owns a reference count
17367 and has to call `..._free`; 'get0' returns a pointer to some
17368 data structure without incrementing reference counters.
17369 (Some of the existing 'get' functions increment a reference
17370 counter, some don't.)
17371 Similarly, 'set1' and 'add1' functions increase reference
17372 counters or duplicate objects.
17376 * Allow for the possibility of temp RSA key generation failure:
17377 the code used to assume it always worked and crashed on failure.
17381 * Fix potential buffer overrun problem in BIO_printf().
17382 *Ulf Möller, using public domain code by Patrick Powell; problem
17383 pointed out by David Sacerdote <das33@cornell.edu>*
17385 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17386 RAND_egd() and RAND_status(). In the command line application,
17387 the EGD socket can be specified like a seed file using RANDFILE
17392 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17393 Some CAs (e.g. Verisign) distribute certificates in this form.
17397 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17398 list to exclude them. This means that no special compilation option
17399 is needed to use anonymous DH: it just needs to be included in the
17404 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17405 EVP_MD_type. The old functionality is available in a new macro called
17406 EVP_MD_md(). Change code that uses it and update docs.
17410 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17411 where the `void *` argument is replaced by a function pointer argument.
17412 Previously `void *` was abused to point to functions, which works on
17413 many platforms, but is not correct. As these functions are usually
17414 called by macros defined in OpenSSL header files, most source code
17415 should work without changes.
17419 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17420 sections with information on -D... compiler switches used for
17421 compiling the library so that applications can see them. To enable
17422 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17423 must be defined. E.g.,
17424 #define OPENSSL_ALGORITHM_DEFINES
17425 #include <openssl/opensslconf.h>
17426 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17428 *Richard Levitte, Ulf and Bodo Möller*
17430 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17435 * Change the 'other' type in certificate aux info to a STACK_OF
17436 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17437 the required ASN1 format: arbitrary types determined by an OID.
17441 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17442 argument to 'req'. This is not because the function is newer or
17443 better than others it just uses the work 'NEW' in the certificate
17444 request header lines. Some software needs this.
17448 * Reorganise password command line arguments: now passwords can be
17449 obtained from various sources. Delete the PEM_cb function and make
17450 it the default behaviour: i.e. if the callback is NULL and the
17451 usrdata argument is not NULL interpret it as a null terminated pass
17452 phrase. If usrdata and the callback are NULL then the pass phrase
17453 is prompted for as usual.
17457 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17458 the support is automatically enabled. The resulting binaries will
17459 autodetect the card and use it if present.
17461 *Ben Laurie and Compaq Inc.*
17463 * Work around for Netscape hang bug. This sends certificate request
17464 and server done in one record. Since this is perfectly legal in the
17465 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17466 the bugs/SSLv3 entry for more info.
17470 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17474 * Add -rand argument to smime and pkcs12 applications and read/write
17479 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17483 * Add command line password options to the remaining applications.
17487 * Bug fix for BN_div_recp() for numerators with an even number of
17492 * More tests in bntest.c, and changed test_bn output.
17496 * ./config recognizes MacOS X now.
17500 * Bug fix for BN_div() when the first words of num and divisor are
17501 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17505 * Add support for various broken PKCS#8 formats, and command line
17506 options to produce them.
17510 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17511 get temporary BIGNUMs from a BN_CTX.
17515 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17520 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17521 include a #define from the old name to the new. The original intent
17522 was that statically linked binaries could for example just call
17523 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17524 link with digests. This never worked because SSLeay_add_all_digests()
17525 and SSLeay_add_all_ciphers() were in the same source file so calling
17526 one would link with the other. They are now in separate source files.
17530 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17534 * Use a less unusual form of the Miller-Rabin primality test (it used
17535 a binary algorithm for exponentiation integrated into the Miller-Rabin
17536 loop, our standard modexp algorithms are faster).
17540 * Support for the EBCDIC character set completed.
17542 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17544 * Source code cleanups: use const where appropriate, eliminate casts,
17545 use `void *` instead of `char *` in lhash.
17549 * Bugfix: ssl3_send_server_key_exchange was not restartable
17550 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17551 this the server could overwrite ephemeral keys that the client
17556 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17557 using 50 iterations of the Rabin-Miller test.
17559 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17560 iterations of the Rabin-Miller test as required by the appendix
17561 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17562 As BN_is_prime_fasttest includes trial division, DSA parameter
17563 generation becomes much faster.
17565 This implies a change for the callback functions in DSA_is_prime
17566 and DSA_generate_parameters: The callback function is called once
17567 for each positive witness in the Rabin-Miller test, not just
17568 occasionally in the inner loop; and the parameters to the
17569 callback function now provide an iteration count for the outer
17570 loop rather than for the current invocation of the inner loop.
17571 DSA_generate_parameters additionally can call the callback
17572 function with an 'iteration count' of -1, meaning that a
17573 candidate has passed the trial division test (when q is generated
17574 from an application-provided seed, trial division is skipped).
17578 * New function BN_is_prime_fasttest that optionally does trial
17579 division before starting the Rabin-Miller test and has
17580 an additional BN_CTX * argument (whereas BN_is_prime always
17581 has to allocate at least one BN_CTX).
17582 'callback(1, -1, cb_arg)' is called when a number has passed the
17583 trial division stage.
17587 * Fix for bug in CRL encoding. The validity dates weren't being handled
17592 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17596 * New function BN_pseudo_rand().
17600 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17601 bignum version of BN_from_montgomery() with the working code from
17602 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17607 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17608 made it impossible to use the same SSL_SESSION data structure in
17609 SSL2 clients in multiple threads.
17613 * The return value of RAND_load_file() no longer counts bytes obtained
17614 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17615 to seed the PRNG (previously an explicit byte count was required).
17617 *Ulf Möller, Bodo Möller*
17619 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17620 used `char *` instead of `void *` and had casts all over the place.
17624 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17628 * Retain source code compatibility for BN_prime_checks macro:
17629 BN_is_prime(..., BN_prime_checks, ...) now uses
17630 BN_prime_checks_for_size to determine the appropriate number of
17631 Rabin-Miller iterations.
17635 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17636 DH_CHECK_P_NOT_SAFE_PRIME.
17637 (Check if this is true? OpenPGP calls them "strong".)
17641 * Merge the functionality of "dh" and "gendh" programs into a new program
17642 "dhparam". The old programs are retained for now but will handle DH keys
17643 (instead of parameters) in future.
17647 * Make the ciphers, s_server and s_client programs check the return values
17648 when a new cipher list is set.
17652 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17653 ciphers. Before when the 56bit ciphers were enabled the sorting was
17656 The syntax for the cipher sorting has been extended to support sorting by
17657 cipher-strength (using the strength_bits hard coded in the tables).
17658 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17660 Fix a bug in the cipher-command parser: when supplying a cipher command
17661 string with an "undefined" symbol (neither command nor alphanumeric
17662 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17663 an error is flagged.
17665 Due to the strength-sorting extension, the code of the
17666 ssl_create_cipher_list() function was completely rearranged. I hope that
17667 the readability was also increased :-)
17669 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17671 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17672 for the first serial number and places 2 in the serial number file. This
17673 avoids problems when the root CA is created with serial number zero and
17674 the first user certificate has the same issuer name and serial number
17679 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17680 the new code. Add documentation for this stuff.
17684 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17685 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17686 structures and behave in an analogous way to the X509v3 functions:
17687 they shouldn't be called directly but wrapper functions should be used
17690 So we also now have some wrapper functions that call the X509at functions
17691 when passed certificate requests. (TO DO: similar things can be done with
17692 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17693 things. Some of these need some d2i or i2d and print functionality
17694 because they handle more complex structures.)
17698 * Add missing #ifndefs that caused missing symbols when building libssl
17699 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17700 NO_RSA in `ssl/s2*.c`.
17702 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17704 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17705 has a return value which indicates the quality of the random data
17706 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17707 error queue. New function RAND_pseudo_bytes() generates output that is
17708 guaranteed to be unique but not unpredictable. RAND_add is like
17709 RAND_seed, but takes an extra argument for an entropy estimate
17710 (RAND_seed always assumes full entropy).
17714 * Do more iterations of Rabin-Miller probable prime test (specifically,
17715 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17716 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17717 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17718 false-positive rate of at most 2^-80 for random input.
17722 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17726 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17727 in the 0.9.5 release), this returns the chain
17728 from an X509_CTX structure with a dup of the stack and all
17729 the X509 reference counts upped: so the stack will exist
17730 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17733 Also make SSL_SESSION_print() print out the verify return
17738 * Add manpage for the pkcs12 command. Also change the default
17739 behaviour so MAC iteration counts are used unless the new
17740 -nomaciter option is used. This improves file security and
17741 only older versions of MSIE (4.0 for example) need it.
17745 * Honor the no-xxx Configure options when creating .DEF files.
17749 * Add PKCS#10 attributes to field table: challengePassword,
17750 unstructuredName and unstructuredAddress. These are taken from
17751 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17752 international characters are used.
17754 More changes to X509_ATTRIBUTE code: allow the setting of types
17755 based on strings. Remove the 'loc' parameter when adding
17756 attributes because these will be a SET OF encoding which is sorted
17761 * Initial changes to the 'req' utility to allow request generation
17762 automation. This will allow an application to just generate a template
17763 file containing all the field values and have req construct the
17766 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17767 used all over the place including certificate requests and PKCS#7
17768 structures. They are currently handled manually where necessary with
17769 some primitive wrappers for PKCS#7. The new functions behave in a
17770 manner analogous to the X509 extension functions: they allow
17771 attributes to be looked up by NID and added.
17773 Later something similar to the X509V3 code would be desirable to
17774 automatically handle the encoding, decoding and printing of the
17775 more complex types. The string types like challengePassword can
17776 be handled by the string table functions.
17778 Also modified the multi byte string table handling. Now there is
17779 a 'global mask' which masks out certain types. The table itself
17780 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17781 is useful when for example there is only one permissible type
17782 (as in countryName) and using the mask might result in no valid
17787 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17788 SSL_get_peer_finished to allow applications to obtain the latest
17789 Finished messages sent to the peer or expected from the peer,
17790 respectively. (SSL_get_peer_finished is usually the Finished message
17791 actually received from the peer, otherwise the protocol will be aborted.)
17793 As the Finished message are message digests of the complete handshake
17794 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17795 be used for external authentication procedures when the authentication
17796 provided by SSL/TLS is not desired or is not enough.
17800 * Enhanced support for Alpha Linux is added. Now ./config checks if
17801 the host supports BWX extension and if Compaq C is present on the
17802 $PATH. Just exploiting of the BWX extension results in 20-30%
17803 performance kick for some algorithms, e.g. DES and RC4 to mention
17804 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17809 * Add support for MS "fast SGC". This is arguably a violation of the
17810 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17811 weak crypto and after checking the certificate is SGC a second one
17812 with strong crypto. MS SGC stops the first handshake after receiving
17813 the server certificate message and sends a second client hello. Since
17814 a server will typically do all the time consuming operations before
17815 expecting any further messages from the client (server key exchange
17816 is the most expensive) there is little difference between the two.
17818 To get OpenSSL to support MS SGC we have to permit a second client
17819 hello message after we have sent server done. In addition we have to
17820 reset the MAC if we do get this second client hello.
17824 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17825 if a DER encoded private key is RSA or DSA traditional format. Changed
17826 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17827 format DER encoded private key. Newer code should use PKCS#8 format which
17828 has the key type encoded in the ASN1 structure. Added DER private key
17829 support to pkcs8 application.
17833 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17834 ciphersuites has been selected (as required by the SSL 3/TLS 1
17835 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17836 is set, we interpret this as a request to violate the specification
17837 (the worst that can happen is a handshake failure, and 'correct'
17838 behaviour would result in a handshake failure anyway).
17842 * In SSL_CTX_add_session, take into account that there might be multiple
17843 SSL_SESSION structures with the same session ID (e.g. when two threads
17844 concurrently obtain them from an external cache).
17845 The internal cache can handle only one SSL_SESSION with a given ID,
17846 so if there's a conflict, we now throw out the old one to achieve
17851 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17852 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17853 some routines that use cipher OIDs: some ciphers do not have OIDs
17854 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17859 * Simplify the trust setting structure and code. Now we just have
17860 two sequences of OIDs for trusted and rejected settings. These will
17861 typically have values the same as the extended key usage extension
17862 and any application specific purposes.
17864 The trust checking code now has a default behaviour: it will just
17865 check for an object with the same NID as the passed id. Functions can
17866 be provided to override either the default behaviour or the behaviour
17867 for a given id. SSL client, server and email already have functions
17868 in place for compatibility: they check the NID and also return "trusted"
17869 if the certificate is self signed.
17873 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17874 traditional format into an EVP_PKEY structure.
17878 * Add a password callback function PEM_cb() which either prompts for
17879 a password if usr_data is NULL or otherwise assumes it is a null
17880 terminated password. Allow passwords to be passed on command line
17881 environment or config files in a few more utilities.
17885 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17886 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17887 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17888 Update documentation.
17892 * Support for ASN1 "NULL" type. This could be handled before by using
17893 ASN1_TYPE but there wasn't any function that would try to read a NULL
17894 and produce an error if it couldn't. For compatibility we also have
17895 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17896 don't allocate anything because they don't need to.
17900 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17903 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17905 * Rebuild of the memory allocation routines used by OpenSSL code and
17906 possibly others as well. The purpose is to make an interface that
17907 provide hooks so anyone can build a separate set of allocation and
17908 deallocation routines to be used by OpenSSL, for example memory
17909 pool implementations, or something else, which was previously hard
17910 since Malloc(), Realloc() and Free() were defined as macros having
17911 the values malloc, realloc and free, respectively (except for Win32
17912 compilations). The same is provided for memory debugging code.
17913 OpenSSL already comes with functionality to find memory leaks, but
17914 this gives people a chance to debug other memory problems.
17916 With these changes, a new set of functions and macros have appeared:
17918 CRYPTO_set_mem_debug_functions() [F]
17919 CRYPTO_get_mem_debug_functions() [F]
17920 CRYPTO_dbg_set_options() [F]
17921 CRYPTO_dbg_get_options() [F]
17922 CRYPTO_malloc_debug_init() [M]
17924 The memory debug functions are NULL by default, unless the library
17925 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17926 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17927 gives the standard debugging functions that come with OpenSSL) or
17928 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17929 provided by the library user) must be used. When the standard
17930 debugging functions are used, CRYPTO_dbg_set_options can be used to
17931 request additional information:
17932 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17933 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17935 Also, things like CRYPTO_set_mem_functions will always give the
17936 expected result (the new set of functions is used for allocation
17937 and deallocation) at all times, regardless of platform and compiler
17940 To finish it up, some functions that were never use in any other
17941 way than through macros have a new API and new semantic:
17943 CRYPTO_dbg_malloc()
17944 CRYPTO_dbg_realloc()
17947 All macros of value have retained their old syntax.
17949 *Richard Levitte and Bodo Moeller*
17951 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17952 ordering of SMIMECapabilities wasn't in "strength order" and there
17953 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17958 * Some ASN1 types with illegal zero length encoding (INTEGER,
17959 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17961 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17963 * Merge in my S/MIME library for OpenSSL. This provides a simple
17964 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17965 functionality to handle multipart/signed properly) and a utility
17966 called 'smime' to call all this stuff. This is based on code I
17967 originally wrote for Celo who have kindly allowed it to be
17968 included in OpenSSL.
17972 * Add variants des_set_key_checked and des_set_key_unchecked of
17973 des_set_key (aka des_key_sched). Global variable des_check_key
17974 decides which of these is called by des_set_key; this way
17975 des_check_key behaves as it always did, but applications and
17976 the library itself, which was buggy for des_check_key == 1,
17977 have a cleaner way to pick the version they need.
17981 * New function PKCS12_newpass() which changes the password of a
17986 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17987 dynamic mix. In both cases the ids can be used as an index into the
17988 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17989 functions so they accept a list of the field values and the
17990 application doesn't need to directly manipulate the X509_TRUST
17995 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
18000 * Modify the way the V3 extension code looks up extensions. This now
18001 works in a similar way to the object code: we have some "standard"
18002 extensions in a static table which is searched with OBJ_bsearch()
18003 and the application can add dynamic ones if needed. The file
18004 crypto/x509v3/ext_dat.h now has the info: this file needs to be
18005 updated whenever a new extension is added to the core code and kept
18006 in ext_nid order. There is a simple program 'tabtest.c' which checks
18007 this. New extensions are not added too often so this file can readily
18008 be maintained manually.
18010 There are two big advantages in doing things this way. The extensions
18011 can be looked up immediately and no longer need to be "added" using
18012 X509V3_add_standard_extensions(): this function now does nothing.
18013 Side note: I get *lots* of email saying the extension code doesn't
18014 work because people forget to call this function.
18015 Also no dynamic allocation is done unless new extensions are added:
18016 so if we don't add custom extensions there is no need to call
18017 X509V3_EXT_cleanup().
18021 * Modify enc utility's salting as follows: make salting the default. Add a
18022 magic header, so unsalted files fail gracefully instead of just decrypting
18023 to garbage. This is because not salting is a big security hole, so people
18024 should be discouraged from doing it.
18028 * Fixes and enhancements to the 'x509' utility. It allowed a message
18029 digest to be passed on the command line but it only used this
18030 parameter when signing a certificate. Modified so all relevant
18031 operations are affected by the digest parameter including the
18032 -fingerprint and -x509toreq options. Also -x509toreq choked if a
18033 DSA key was used because it didn't fix the digest.
18037 * Initial certificate chain verify code. Currently tests the untrusted
18038 certificates for consistency with the verify purpose (which is set
18039 when the X509_STORE_CTX structure is set up) and checks the pathlength.
18041 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
18042 this is because it will reject chains with invalid extensions whereas
18043 every previous version of OpenSSL and SSLeay made no checks at all.
18045 Trust code: checks the root CA for the relevant trust settings. Trust
18046 settings have an initial value consistent with the verify purpose: e.g.
18047 if the verify purpose is for SSL client use it expects the CA to be
18048 trusted for SSL client use. However the default value can be changed to
18049 permit custom trust settings: one example of this would be to only trust
18050 certificates from a specific "secure" set of CAs.
18052 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
18053 which should be used for version portability: especially since the
18054 verify structure is likely to change more often now.
18056 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
18057 to set them. If not set then assume SSL clients will verify SSL servers
18060 Two new options to the verify program: -untrusted allows a set of
18061 untrusted certificates to be passed in and -purpose which sets the
18062 intended purpose of the certificate. If a purpose is set then the
18063 new chain verify code is used to check extension consistency.
18067 * Support for the authority information access extension.
18071 * Modify RSA and DSA PEM read routines to transparently handle
18072 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18073 public keys in a format compatible with certificate
18074 SubjectPublicKeyInfo structures. Unfortunately there were already
18075 functions called *_PublicKey_* which used various odd formats so
18076 these are retained for compatibility: however the DSA variants were
18077 never in a public release so they have been deleted. Changed dsa/rsa
18078 utilities to handle the new format: note no releases ever handled public
18079 keys so we should be OK.
18081 The primary motivation for this change is to avoid the same fiasco
18082 that dogs private keys: there are several incompatible private key
18083 formats some of which are standard and some OpenSSL specific and
18084 require various evil hacks to allow partial transparent handling and
18085 even then it doesn't work with DER formats. Given the option anything
18086 other than PKCS#8 should be dumped: but the other formats have to
18087 stay in the name of compatibility.
18089 With public keys and the benefit of hindsight one standard format
18090 is used which works with EVP_PKEY, RSA or DSA structures: though
18091 it clearly returns an error if you try to read the wrong kind of key.
18093 Added a -pubkey option to the 'x509' utility to output the public key.
18094 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18095 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18096 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18097 that do the same as the `EVP_PKEY_assign_*()` except they up the
18098 reference count of the added key (they don't "swallow" the
18103 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18104 CRLs would fail if the file contained no certificates or no CRLs:
18105 added a new function to read in both types and return the number
18106 read: this means that if none are read it will be an error. The
18107 DER versions of the certificate and CRL reader would always fail
18108 because it isn't possible to mix certificates and CRLs in DER format
18109 without choking one or the other routine. Changed this to just read
18110 a certificate: this is the best we can do. Also modified the code
18111 in `apps/verify.c` to take notice of return codes: it was previously
18112 attempting to read in certificates from NULL pointers and ignoring
18113 any errors: this is one reason why the cert and CRL reader seemed
18114 to work. It doesn't check return codes from the default certificate
18115 routines: these may well fail if the certificates aren't installed.
18119 * Code to support otherName option in GeneralName.
18123 * First update to verify code. Change the verify utility
18124 so it warns if it is passed a self signed certificate:
18125 for consistency with the normal behaviour. X509_verify
18126 has been modified to it will now verify a self signed
18127 certificate if *exactly* the same certificate appears
18128 in the store: it was previously impossible to trust a
18129 single self signed certificate. This means that:
18130 openssl verify ss.pem
18131 now gives a warning about a self signed certificate but
18132 openssl verify -CAfile ss.pem ss.pem
18137 * For servers, store verify_result in SSL_SESSION data structure
18138 (and add it to external session representation).
18139 This is needed when client certificate verifications fails,
18140 but an application-provided verification callback (set by
18141 SSL_CTX_set_cert_verify_callback) allows accepting the session
18142 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18143 but returns 1): When the session is reused, we have to set
18144 ssl->verify_result to the appropriate error code to avoid
18147 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18149 * Fix a bug in the new PKCS#7 code: it didn't consider the
18150 case in PKCS7_dataInit() where the signed PKCS7 structure
18151 didn't contain any existing data because it was being created.
18153 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18155 * Add a salt to the key derivation routines in enc.c. This
18156 forms the first 8 bytes of the encrypted file. Also add a
18157 -S option to allow a salt to be input on the command line.
18161 * New function X509_cmp(). Oddly enough there wasn't a function
18162 to compare two certificates. We do this by working out the SHA1
18163 hash and comparing that. X509_cmp() will be needed by the trust
18168 * SSL_get1_session() is like SSL_get_session(), but increments
18169 the reference count in the SSL_SESSION returned.
18171 *Geoff Thorpe <geoff@eu.c2.net>*
18173 * Fix for 'req': it was adding a null to request attributes.
18174 Also change the X509_LOOKUP and X509_INFO code to handle
18175 certificate auxiliary information.
18179 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18184 * Add the possibility to add extra information to the memory leak
18185 detecting output, to form tracebacks, showing from where each
18186 allocation was originated: CRYPTO_push_info("constant string") adds
18187 the string plus current file name and line number to a per-thread
18188 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18189 is like calling CYRPTO_pop_info() until the stack is empty.
18190 Also updated memory leak detection code to be multi-thread-safe.
18194 * Add options -text and -noout to pkcs7 utility and delete the
18195 encryption options which never did anything. Update docs.
18199 * Add options to some of the utilities to allow the pass phrase
18200 to be included on either the command line (not recommended on
18201 OSes like Unix) or read from the environment. Update the
18202 manpages and fix a few bugs.
18206 * Add a few manpages for some of the openssl commands.
18210 * Fix the -revoke option in ca. It was freeing up memory twice,
18211 leaking and not finding already revoked certificates.
18215 * Extensive changes to support certificate auxiliary information.
18216 This involves the use of X509_CERT_AUX structure and X509_AUX
18217 functions. An X509_AUX function such as PEM_read_X509_AUX()
18218 can still read in a certificate file in the usual way but it
18219 will also read in any additional "auxiliary information". By
18220 doing things this way a fair degree of compatibility can be
18221 retained: existing certificates can have this information added
18222 using the new 'x509' options.
18224 Current auxiliary information includes an "alias" and some trust
18225 settings. The trust settings will ultimately be used in enhanced
18226 certificate chain verification routines: currently a certificate
18227 can only be trusted if it is self signed and then it is trusted
18232 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18233 The problem was that one of the replacement routines had not been working
18234 since SSLeay releases. For now the offending routine has been replaced
18235 with non-optimised assembler. Even so, this now gives around 95%
18236 performance improvement for 1024 bit RSA signs.
18240 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18241 handling. Most clients have the effective key size in bits equal to
18242 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18243 A few however don't do this and instead use the size of the decrypted key
18244 to determine the RC2 key length and the AlgorithmIdentifier to determine
18245 the effective key length. In this case the effective key length can still
18246 be 40 bits but the key length can be 168 bits for example. This is fixed
18247 by manually forcing an RC2 key into the EVP_PKEY structure because the
18248 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18249 the key length and effective key length are equal.
18253 * Add a bunch of functions that should simplify the creation of
18254 X509_NAME structures. Now you should be able to do:
18255 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18256 and have it automatically work out the correct field type and fill in
18257 the structures. The more adventurous can try:
18258 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18259 and it will (hopefully) work out the correct multibyte encoding.
18263 * Change the 'req' utility to use the new field handling and multibyte
18264 copy routines. Before the DN field creation was handled in an ad hoc
18265 way in req, ca, and x509 which was rather broken and didn't support
18266 BMPStrings or UTF8Strings. Since some software doesn't implement
18267 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18268 using the dirstring_type option. See the new comment in the default
18269 openssl.cnf for more info.
18273 * Make crypto/rand/md_rand.c more robust:
18274 - Assure unique random numbers after fork().
18275 - Make sure that concurrent threads access the global counter and
18276 md serializably so that we never lose entropy in them
18277 or use exactly the same state in multiple threads.
18278 Access to the large state is not always serializable because
18279 the additional locking could be a performance killer, and
18280 md should be large enough anyway.
18284 * New file `apps/app_rand.c` with commonly needed functionality
18285 for handling the random seed file.
18287 Use the random seed file in some applications that previously did not:
18289 dsaparam -genkey (which also ignored its '-rand' option),
18292 x509 (when signing).
18293 Except on systems with /dev/urandom, it is crucial to have a random
18294 seed file at least for key creation, DSA signing, and for DH exchanges;
18295 for RSA signatures we could do without one.
18297 gendh and gendsa (unlike genrsa) used to read only the first byte
18298 of each file listed in the '-rand' option. The function as previously
18299 found in genrsa is now in app_rand.c and is used by all programs
18300 that support '-rand'.
18304 * In RAND_write_file, use mode 0600 for creating files;
18305 don't just chmod when it may be too late.
18309 * Report an error from X509_STORE_load_locations
18310 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18314 * New function ASN1_mbstring_copy() this copies a string in either
18315 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18316 into an ASN1_STRING type. A mask of permissible types is passed
18317 and it chooses the "minimal" type to use or an error if not type
18322 * Add function equivalents to the various macros in asn1.h. The old
18323 macros are retained with an `M_` prefix. Code inside the library can
18324 use the `M_` macros. External code (including the openssl utility)
18325 should *NOT* in order to be "shared library friendly".
18329 * Add various functions that can check a certificate's extensions
18330 to see if it usable for various purposes such as SSL client,
18331 server or S/MIME and CAs of these types. This is currently
18332 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18333 verification. Also added a -purpose flag to x509 utility to
18334 print out all the purposes.
18338 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18343 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18344 for, obtain and decode and extension and obtain its critical flag.
18345 This allows all the necessary extension code to be handled in a
18346 single function call.
18350 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18351 platforms. See crypto/rc4/rc4_enc.c for further details.
18355 * New -noout option to asn1parse. This causes no output to be produced
18356 its main use is when combined with -strparse and -out to extract data
18357 from a file (which may not be in ASN.1 format).
18361 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18362 when producing the local key id.
18364 *Richard Levitte <levitte@stacken.kth.se>*
18366 * New option -dhparam in s_server. This allows a DH parameter file to be
18367 stated explicitly. If it is not stated then it tries the first server
18368 certificate file. The previous behaviour hard coded the filename
18373 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18374 a public key to be input or output. For example:
18375 openssl rsa -in key.pem -pubout -out pubkey.pem
18376 Also added necessary DSA public key functions to handle this.
18380 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18381 in the message. This was handled by allowing
18382 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18384 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18386 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18387 to the end of the strings whereas this didn't. This would cause problems
18388 if strings read with d2i_ASN1_bytes() were later modified.
18390 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18392 * Fix for base64 decode bug. When a base64 bio reads only one line of
18393 data and it contains EOF it will end up returning an error. This is
18394 caused by input 46 bytes long. The cause is due to the way base64
18395 BIOs find the start of base64 encoded data. They do this by trying a
18396 trial decode on each line until they find one that works. When they
18397 do a flag is set and it starts again knowing it can pass all the
18398 data directly through the decoder. Unfortunately it doesn't reset
18399 the context it uses. This means that if EOF is reached an attempt
18400 is made to pass two EOFs through the context and this causes the
18401 resulting error. This can also cause other problems as well. As is
18402 usual with these problems it takes *ages* to find and the fix is
18403 trivial: move one line.
18405 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18407 * Ugly workaround to get s_client and s_server working under Windows. The
18408 old code wouldn't work because it needed to select() on sockets and the
18409 tty (for keypresses and to see if data could be written). Win32 only
18410 supports select() on sockets so we select() with a 1s timeout on the
18411 sockets and then see if any characters are waiting to be read, if none
18412 are present then we retry, we also assume we can always write data to
18413 the tty. This isn't nice because the code then blocks until we've
18414 received a complete line of data and it is effectively polling the
18415 keyboard at 1s intervals: however it's quite a bit better than not
18416 working at all :-) A dedicated Windows application might handle this
18417 with an event loop for example.
18421 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18422 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18423 will be called when RSA_sign() and RSA_verify() are used. This is useful
18424 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18425 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18426 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18427 This necessitated the support of an extra signature type NID_md5_sha1
18428 for SSL signatures and modifications to the SSL library to use it instead
18429 of calling RSA_public_decrypt() and RSA_private_encrypt().
18433 * Add new -verify -CAfile and -CApath options to the crl program, these
18434 will lookup a CRL issuers certificate and verify the signature in a
18435 similar way to the verify program. Tidy up the crl program so it
18436 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18437 less strict. It will now permit CRL extensions even if it is not
18438 a V2 CRL: this will allow it to tolerate some broken CRLs.
18442 * Initialize all non-automatic variables each time one of the openssl
18443 sub-programs is started (this is necessary as they may be started
18444 multiple times from the "OpenSSL>" prompt).
18446 *Lennart Bang, Bodo Moeller*
18448 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18449 removing all other RSA functionality (this is what NO_RSA does). This
18450 is so (for example) those in the US can disable those operations covered
18451 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18456 * Non-copying interface to BIO pairs.
18457 (still largely untested)
18461 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18462 ASCII string. This was handled independently in various places before.
18466 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18467 UTF8 strings a character at a time.
18471 * Use client_version from client hello to select the protocol
18472 (s23_srvr.c) and for RSA client key exchange verification
18473 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18477 * Add various utility functions to handle SPKACs, these were previously
18478 handled by poking round in the structure internals. Added new function
18479 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18480 print, verify and generate SPKACs. Based on an original idea from
18481 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18485 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18489 * Allow the config file extension section to be overwritten on the
18490 command line. Based on an original idea from Massimiliano Pala
18491 <madwolf@comune.modena.it>. The new option is called -extensions
18492 and can be applied to ca, req and x509. Also -reqexts to override
18493 the request extensions in req and -crlexts to override the crl extensions
18498 * Add new feature to the SPKAC handling in ca. Now you can include
18499 the same field multiple times by preceding it by "XXXX." for example:
18502 this is the same syntax as used in the req config file.
18506 * Allow certificate extensions to be added to certificate requests. These
18507 are specified in a 'req_extensions' option of the req section of the
18508 config file. They can be printed out with the -text option to req but
18509 are otherwise ignored at present.
18513 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18514 data read consists of only the final block it would not decrypted because
18515 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18516 A misplaced 'break' also meant the decrypted final block might not be
18517 copied until the next read.
18521 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18522 a few extra parameters to the DH structure: these will be useful if
18523 for example we want the value of 'q' or implement X9.42 DH.
18527 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18528 provides hooks that allow the default DSA functions or functions on a
18529 "per key" basis to be replaced. This allows hardware acceleration and
18530 hardware key storage to be handled without major modification to the
18531 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18532 associated functions.
18536 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18537 as "read only": it can't be written to and the buffer it points to will
18538 not be freed. Reading from a read only BIO is much more efficient than
18539 a normal memory BIO. This was added because there are several times when
18540 an area of memory needs to be read from a BIO. The previous method was
18541 to create a memory BIO and write the data to it, this results in two
18542 copies of the data and an O(n^2) reading algorithm. There is a new
18543 function BIO_new_mem_buf() which creates a read only memory BIO from
18544 an area of memory. Also modified the PKCS#7 routines to use read only
18549 * Bugfix: ssl23_get_client_hello did not work properly when called in
18550 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18551 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18552 but a retry condition occurred while trying to read the rest.
18556 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18557 NID_pkcs7_encrypted by default: this was wrong since this should almost
18558 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18559 the encrypted data type: this is a more sensible place to put it and it
18560 allows the PKCS#12 code to be tidied up that duplicated this
18565 * Changed obj_dat.pl script so it takes its input and output files on
18566 the command line. This should avoid shell escape redirection problems
18571 * Initial support for certificate extension requests, these are included
18572 in things like Xenroll certificate requests. Included functions to allow
18573 extensions to be obtained and added.
18577 * -crlf option to s_client and s_server for sending newlines as
18578 CRLF (as required by many protocols).
18582 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18584 * Install libRSAglue.a when OpenSSL is built with RSAref.
18586 *Ralf S. Engelschall*
18588 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18590 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18592 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18597 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18598 DH parameters/keys (q is lost during that conversion, but the resulting
18599 DH parameters contain its length).
18601 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18602 much faster than DH_generate_parameters (which creates parameters
18603 where `p = 2*q + 1`), and also the smaller q makes DH computations
18604 much more efficient (160-bit exponentiation instead of 1024-bit
18605 exponentiation); so this provides a convenient way to support DHE
18606 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18607 utter importance to use
18608 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18610 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18611 when such DH parameters are used, because otherwise small subgroup
18612 attacks may become possible!
18616 * Avoid memory leak in i2d_DHparams.
18620 * Allow the -k option to be used more than once in the enc program:
18621 this allows the same encrypted message to be read by multiple recipients.
18625 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18626 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18627 it will always use the numerical form of the OID, even if it has a short
18632 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18633 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18634 otherwise bn_mod_exp was called. In the case of hardware keys for example
18635 no private key components need be present and it might store extra data
18636 in the RSA structure, which cannot be accessed from bn_mod_exp.
18637 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18638 private key operations.
18642 * Added support for SPARC Linux.
18646 * pem_password_cb function type incompatibly changed from
18647 typedef int pem_password_cb(char *buf, int size, int rwflag);
18649 ....(char *buf, int size, int rwflag, void *userdata);
18650 so that applications can pass data to their callbacks:
18651 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18652 additional void * argument, which is just handed through whenever
18653 the password callback is called.
18655 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18657 New function SSL_CTX_set_default_passwd_cb_userdata.
18659 Compatibility note: As many C implementations push function arguments
18660 onto the stack in reverse order, the new library version is likely to
18661 interoperate with programs that have been compiled with the old
18662 pem_password_cb definition (PEM_whatever takes some data that
18663 happens to be on the stack as its last argument, and the callback
18664 just ignores this garbage); but there is no guarantee whatsoever that
18667 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18668 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18669 problems not only on Windows, but also on some Unix platforms.
18670 To avoid problematic command lines, these definitions are now in an
18671 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18672 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18676 * MIPS III/IV assembler module is reimplemented.
18680 * More DES library cleanups: remove references to srand/rand and
18681 delete an unused file.
18685 * Add support for the free Netwide assembler (NASM) under Win32,
18686 since not many people have MASM (ml) and it can be hard to obtain.
18687 This is currently experimental but it seems to work OK and pass all
18688 the tests. Check out INSTALL.W32 for info.
18692 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18693 without temporary keys kept an extra copy of the server key,
18694 and connections with temporary keys did not free everything in case
18699 * New function RSA_check_key and new openssl rsa option -check
18700 for verifying the consistency of RSA keys.
18702 *Ulf Moeller, Bodo Moeller*
18704 * Various changes to make Win32 compile work:
18705 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18706 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18707 comparison" warnings.
18708 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18712 * Add a debugging option to PKCS#5 v2 key generation function: when
18713 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18714 derived keys are printed to stderr.
18718 * Copy the flags in ASN1_STRING_dup().
18720 *Roman E. Pavlov <pre@mo.msk.ru>*
18722 * The x509 application mishandled signing requests containing DSA
18723 keys when the signing key was also DSA and the parameters didn't match.
18725 It was supposed to omit the parameters when they matched the signing key:
18726 the verifying software was then supposed to automatically use the CA's
18727 parameters if they were absent from the end user certificate.
18729 Omitting parameters is no longer recommended. The test was also
18730 the wrong way round! This was probably due to unusual behaviour in
18731 EVP_cmp_parameters() which returns 1 if the parameters match.
18732 This meant that parameters were omitted when they *didn't* match and
18733 the certificate was useless. Certificates signed with 'ca' didn't have
18736 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18738 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18739 The interface is as follows:
18740 Applications can use
18741 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18742 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18743 "off" is now the default.
18744 The library internally uses
18745 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18746 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18747 to disable memory-checking temporarily.
18749 Some inconsistent states that previously were possible (and were
18750 even the default) are now avoided.
18752 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18753 with each memory chunk allocated; this is occasionally more helpful
18754 than just having a counter.
18756 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18758 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18763 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18764 which largely parallels "options", but is for changing API behaviour,
18765 whereas "options" are about protocol behaviour.
18766 Initial "mode" flags are:
18768 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18769 a single record has been written.
18770 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18771 retries use the same buffer location.
18772 (But all of the contents must be
18777 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18780 * Fix problems with no-hmac etc.
18782 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18784 * New functions RSA_get_default_method(), RSA_set_method() and
18785 RSA_get_method(). These allows replacement of RSA_METHODs without having
18786 to mess around with the internals of an RSA structure.
18790 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18791 Also really enable memory leak checks in openssl.c and in some
18794 *Chad C. Mulligan, Bodo Moeller*
18796 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18797 up the length of negative integers. This has now been simplified to just
18798 store the length when it is first determined and use it later, rather
18799 than trying to keep track of where data is copied and updating it to
18801 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18803 * Add a new function PKCS7_signatureVerify. This allows the verification
18804 of a PKCS#7 signature but with the signing certificate passed to the
18805 function itself. This contrasts with PKCS7_dataVerify which assumes the
18806 certificate is present in the PKCS#7 structure. This isn't always the
18807 case: certificates can be omitted from a PKCS#7 structure and be
18808 distributed by "out of band" means (such as a certificate database).
18812 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18813 function prototypes in pem.h, also change util/mkdef.pl to add the
18814 necessary function names.
18818 * mk1mf.pl (used by Windows builds) did not properly read the
18819 options set by Configure in the top level Makefile, and Configure
18820 was not even able to write more than one option correctly.
18821 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18825 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18826 file to be loaded from a BIO or FILE pointer. The BIO version will
18827 for example allow memory BIOs to contain config info.
18831 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18832 Whoever hopes to achieve shared-library compatibility across versions
18833 must use this, not the compile-time macro.
18834 (Exercise 0.9.4: Which is the minimum library version required by
18836 Note: All this applies only to multi-threaded programs, others don't
18841 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18842 through a BIO pair triggered the default case, i.e.
18843 SSLerr(...,SSL_R_UNKNOWN_STATE).
18847 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18848 can use the SSL library even if none of the specific BIOs is
18853 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18854 for the encoded length.
18856 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18858 * Add initial documentation of the X509V3 functions.
18862 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18863 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18864 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18865 secure PKCS#8 private key format with a high iteration count.
18869 * Fix determination of Perl interpreter: A perl or perl5
18870 *directory* in $PATH was also accepted as the interpreter.
18872 *Ralf S. Engelschall*
18874 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18875 wrong with it but it was very old and did things like calling
18876 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18877 unusual formatting.
18881 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18882 to use the new extension code.
18886 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18887 with macros. This should make it easier to change their form, add extra
18888 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18893 * Add to configuration table a new entry that can specify an alternative
18894 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18895 according to Mark Crispin <MRC@Panda.COM>.
18899 * DES CBC did not update the IV. Weird.
18903 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18904 Changing the behaviour of the former might break existing programs --
18905 where IV updating is needed, des_ncbc_encrypt can be used.
18908 * When bntest is run from "make test" it drives bc to check its
18909 calculations, as well as internally checking them. If an internal check
18910 fails, it needs to cause bc to give a non-zero result or make test carries
18911 on without noticing the failure. Fixed.
18915 * DES library cleanups.
18919 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18920 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18921 ciphers. NOTE: although the key derivation function has been verified
18922 against some published test vectors it has not been extensively tested
18923 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18928 * Instead of "mkdir -p", which is not fully portable, use new
18929 Perl script "util/mkdir-p.pl".
18933 * Rewrite the way password based encryption (PBE) is handled. It used to
18934 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18935 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18936 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18937 the 'parameter' field of the AlgorithmIdentifier is passed to the
18938 underlying key generation function so it must do its own ASN1 parsing.
18939 This has also changed the EVP_PBE_CipherInit() function which now has a
18940 'parameter' argument instead of literal salt and iteration count values
18941 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18945 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18946 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18947 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18948 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18949 value was just used as a "magic string" and not used directly its
18950 value doesn't matter.
18954 * Introduce some semblance of const correctness to BN. Shame C doesn't
18959 * "linux-sparc64" configuration (ultrapenguin).
18961 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18962 "linux-sparc" configuration.
18964 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18966 * config now generates no-xxx options for missing ciphers.
18970 * Support the EBCDIC character set (work in progress).
18971 File ebcdic.c not yet included because it has a different license.
18973 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18975 * Support BS2000/OSD-POSIX.
18977 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18979 * Make callbacks for key generation use `void *` instead of `char *`.
18983 * Make S/MIME samples compile (not yet tested).
18987 * Additional typesafe stacks.
18991 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18995 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18997 * New configuration variant "sco5-gcc".
18999 * Updated some demos.
19001 *Sean O Riordain, Wade Scholine*
19003 * Add missing BIO_free at exit of pkcs12 application.
19007 * Fix memory leak in conf.c.
19011 * Updates for Win32 to assembler version of MD5.
19015 * Set #! path to perl in `apps/der_chop` to where we found it
19016 instead of using a fixed path.
19020 * SHA library changes for irix64-mips4-cc.
19024 * Improvements for VMS support.
19028 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
19030 * Bignum library bug fix. IRIX 6 passes "make test" now!
19031 This also avoids the problems with SC4.2 and unpatched SC5.
19033 *Andy Polyakov <appro@fy.chalmers.se>*
19035 * New functions sk_num, sk_value and sk_set to replace the previous macros.
19036 These are required because of the typesafe stack would otherwise break
19037 existing code. If old code used a structure member which used to be STACK
19038 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
19039 sk_num or sk_value it would produce an error because the num, data members
19040 are not present in STACK_OF. Now it just produces a warning. sk_set
19041 replaces the old method of assigning a value to sk_value
19042 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
19043 that does this will no longer work (and should use sk_set instead) but
19044 this could be regarded as a "questionable" behaviour anyway.
19048 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
19049 correctly handle encrypted S/MIME data.
19053 * Change type of various DES function arguments from des_cblock
19054 (which means, in function argument declarations, pointer to char)
19055 to des_cblock * (meaning pointer to array with 8 char elements),
19056 which allows the compiler to do more typechecking; it was like
19057 that back in SSLeay, but with lots of ugly casts.
19059 Introduce new type const_des_cblock.
19063 * Reorganise the PKCS#7 library and get rid of some of the more obvious
19064 problems: find RecipientInfo structure that matches recipient certificate
19065 and initialise the ASN1 structures properly based on passed cipher.
19069 * Belatedly make the BN tests actually check the results.
19073 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19074 to and from BNs: it was completely broken. New compilation option
19075 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19076 key elements as negative integers.
19080 * Reorganize and speed up MD5.
19082 *Andy Polyakov <appro@fy.chalmers.se>*
19086 *Richard Levitte <richard@levitte.org>*
19088 * New option -out to asn1parse to allow the parsed structure to be
19089 output to a file. This is most useful when combined with the -strparse
19090 option to examine the output of things like OCTET STRINGS.
19094 * Make SSL library a little more fool-proof by not requiring any longer
19095 that `SSL_set_{accept,connect}_state` be called before
19096 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19097 in many applications because usually everything *appeared* to work as
19098 intended anyway -- now it really works as intended).
19102 * Move openssl.cnf out of lib/.
19106 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19107 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19108 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19110 *Ralf S. Engelschall*
19112 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19113 handle PKCS#7 enveloped data properly.
19115 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19117 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19118 copying pointers. The cert_st handling is changed by this in
19119 various ways (and thus what used to be known as ctx->default_cert
19120 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19121 any longer when s->cert does not give us what we need).
19122 ssl_cert_instantiate becomes obsolete by this change.
19123 As soon as we've got the new code right (possibly it already is?),
19124 we have solved a couple of bugs of the earlier code where s->cert
19125 was used as if it could not have been shared with other SSL structures.
19127 Note that using the SSL API in certain dirty ways now will result
19128 in different behaviour than observed with earlier library versions:
19129 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19130 does not influence s as it used to.
19132 In order to clean up things more thoroughly, inside SSL_SESSION
19133 we don't use CERT any longer, but a new structure SESS_CERT
19134 that holds per-session data (if available); currently, this is
19135 the peer's certificate chain and, for clients, the server's certificate
19136 and temporary key. CERT holds only those values that can have
19137 meaningful defaults in an SSL_CTX.
19141 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19142 from the internal representation. Various PKCS#7 fixes: remove some
19143 evil casts and set the enc_dig_alg field properly based on the signing
19148 * Allow PKCS#12 password to be set from the command line or the
19149 environment. Let 'ca' get its config file name from the environment
19150 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19155 * Allow certificate policies extension to use an IA5STRING for the
19156 organization field. This is contrary to the PKIX definition but
19157 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19162 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19163 without disallowing inline assembler and the like for non-pedantic builds.
19167 * Support Borland C++ builder.
19169 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19175 * SHA-1 cleanups and performance enhancements.
19177 *Andy Polyakov <appro@fy.chalmers.se>*
19179 * Sparc v8plus assembler for the bignum library.
19181 *Andy Polyakov <appro@fy.chalmers.se>*
19183 * Accept any -xxx and +xxx compiler options in Configure.
19187 * Update HPUX configuration.
19191 * Add missing `sk_<type>_unshift()` function to safestack.h
19193 *Ralf S. Engelschall*
19195 * New function SSL_CTX_use_certificate_chain_file that sets the
19196 "extra_cert"s in addition to the certificate. (This makes sense
19197 only for "PEM" format files, as chains as a whole are not
19202 * Support verify_depth from the SSL API.
19203 x509_vfy.c had what can be considered an off-by-one-error:
19204 Its depth (which was not part of the external interface)
19205 was actually counting the number of certificates in a chain;
19206 now it really counts the depth.
19210 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19211 instead of X509err, which often resulted in confusing error
19212 messages since the error codes are not globally unique
19213 (e.g. an alleged error in ssl3_accept when a certificate
19214 didn't match the private key).
19216 * New function SSL_CTX_set_session_id_context that allows to set a default
19217 value (so that you don't need SSL_set_session_id_context for each
19218 connection using the SSL_CTX).
19222 * OAEP decoding bug fix.
19226 * Support INSTALL_PREFIX for package builders, as proposed by
19231 * New Configure options "threads" and "no-threads". For systems
19232 where the proper compiler options are known (currently Solaris
19233 and Linux), "threads" is the default.
19237 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19241 * Install various scripts to $(OPENSSLDIR)/misc, not to
19242 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19243 such as /usr/local/bin.
19247 * "make linux-shared" to build shared libraries.
19249 *Niels Poppe <niels@netbox.org>*
19251 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19255 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19256 extension adding in x509 utility.
19260 * Remove NOPROTO sections and error code comments.
19264 * Partial rewrite of the DEF file generator to now parse the ANSI
19269 * New Configure options --prefix=DIR and --openssldir=DIR.
19273 * Complete rewrite of the error code script(s). It is all now handled
19274 by one script at the top level which handles error code gathering,
19275 header rewriting and C source file generation. It should be much better
19276 than the old method: it now uses a modified version of Ulf's parser to
19277 read the ANSI prototypes in all header files (thus the old K&R definitions
19278 aren't needed for error creation any more) and do a better job of
19279 translating function codes into names. The old 'ASN1 error code embedded
19280 in a comment' is no longer necessary and it doesn't use .err files which
19281 have now been deleted. Also the error code call doesn't have to appear all
19282 on one line (which resulted in some large lines...).
19286 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19290 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19291 0 (which usually indicates a closed connection), but continue reading.
19295 * Fix some race conditions.
19299 * Add support for CRL distribution points extension. Add Certificate
19300 Policies and CRL distribution points documentation.
19304 * Move the autogenerated header file parts to crypto/opensslconf.h.
19308 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19309 8 of keying material. Merlin has also confirmed interop with this fix
19310 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19312 *Merlin Hughes <merlin@baltimore.ie>*
19314 * Fix lots of warnings.
19316 *Richard Levitte <levitte@stacken.kth.se>*
19318 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19319 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19321 *Richard Levitte <levitte@stacken.kth.se>*
19323 * Fix problems with sizeof(long) == 8.
19325 *Andy Polyakov <appro@fy.chalmers.se>*
19327 * Change functions to ANSI C.
19331 * Fix typos in error codes.
19333 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19335 * Remove defunct assembler files from Configure.
19339 * SPARC v8 assembler BIGNUM implementation.
19341 *Andy Polyakov <appro@fy.chalmers.se>*
19343 * Support for Certificate Policies extension: both print and set.
19344 Various additions to support the r2i method this uses.
19348 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19349 return a const string when you are expecting an allocated buffer.
19353 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19354 types DirectoryString and DisplayText.
19358 * Add code to allow r2i extensions to access the configuration database,
19359 add an LHASH database driver and add several ctx helper functions.
19363 * Fix an evil bug in bn_expand2() which caused various BN functions to
19364 fail when they extended the size of a BIGNUM.
19368 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19369 support typesafe stack.
19373 * Fix typo in SSL_[gs]et_options().
19375 *Nils Frostberg <nils@medcom.se>*
19377 * Delete various functions and files that belonged to the (now obsolete)
19378 old X509V3 handling code.
19382 * New Configure option "rsaref".
19386 * Don't auto-generate pem.h.
19390 * Introduce type-safe ASN.1 SETs.
19394 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19396 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19398 * Introduce type-safe STACKs. This will almost certainly break lots of code
19399 that links with OpenSSL (well at least cause lots of warnings), but fear
19400 not: the conversion is trivial, and it eliminates loads of evil casts. A
19401 few STACKed things have been converted already. Feel free to convert more.
19402 In the fullness of time, I'll do away with the STACK type altogether.
19406 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19407 specified in `<certfile>` by updating the entry in the index.txt file.
19408 This way one no longer has to edit the index.txt file manually for
19409 revoking a certificate. The -revoke option does the gory details now.
19411 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19413 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19414 `-text` option at all and this way the `-noout -text` combination was
19415 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19417 *Ralf S. Engelschall*
19419 * Make sure a corresponding plain text error message exists for the
19420 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19421 verify callback function determined that a certificate was revoked.
19423 *Ralf S. Engelschall*
19425 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19426 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19427 all available ciphers including rc5, which was forgotten until now.
19428 In order to let the testing shell script know which algorithms
19429 are available, a new (up to now undocumented) command
19430 `openssl list-cipher-commands` is used.
19434 * Bugfix: s_client occasionally would sleep in select() when
19435 it should have checked SSL_pending() first.
19439 * New functions DSA_do_sign and DSA_do_verify to provide access to
19440 the raw DSA values prior to ASN.1 encoding.
19444 * Tweaks to Configure
19446 *Niels Poppe <niels@netbox.org>*
19448 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19453 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19457 * New config option to avoid instructions that are illegal on the 80386.
19458 The default code is faster, but requires at least a 486.
19462 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19463 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19464 same as SSL2_VERSION anyway.
19468 * New "-showcerts" option for s_client.
19472 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19473 application. Various cleanups and fixes.
19477 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19478 modify error routines to work internally. Add error codes and PBE init
19479 to library startup routines.
19483 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19484 packing functions to asn1 and evp. Changed function names and error
19485 codes along the way.
19489 * PKCS12 integration: and so it begins... First of several patches to
19490 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19491 objects to objects.h
19495 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19496 and display support for Thawte strong extranet extension.
19500 * Add LinuxPPC support.
19502 *Jeff Dubrule <igor@pobox.org>*
19504 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19505 bn_div_words in alpha.s.
19507 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19509 * Make sure the RSA OAEP test is skipped under -DRSAref because
19510 OAEP isn't supported when OpenSSL is built with RSAref.
19512 *Ulf Moeller <ulf@fitug.de>*
19514 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19515 so they no longer are missing under -DNOPROTO.
19517 *Soren S. Jorvang <soren@t.dk>*
19519 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19521 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19522 doesn't work when the session is reused. Coming soon!
19526 * Fix a security hole, that allows sessions to be reused in the wrong
19527 context thus bypassing client cert protection! All software that uses
19528 client certs and session caches in multiple contexts NEEDS PATCHING to
19529 allow session reuse! A fuller solution is in the works.
19531 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19533 * Some more source tree cleanups (removed obsolete files
19534 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19535 permission on "config" script to be executable) and a fix for the INSTALL
19538 *Ulf Moeller <ulf@fitug.de>*
19540 * Remove some legacy and erroneous uses of malloc, free instead of
19543 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19545 * Make rsa_oaep_test return non-zero on error.
19547 *Ulf Moeller <ulf@fitug.de>*
19549 * Add support for native Solaris shared libraries. Configure
19550 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19551 if someone would make that last step automatic.
19553 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19555 * ctx_size was not built with the right compiler during "make links". Fixed.
19559 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19560 except NULL ciphers". This means the default cipher list will no longer
19561 enable NULL ciphers. They need to be specifically enabled e.g. with
19562 the string "DEFAULT:eNULL".
19566 * Fix to RSA private encryption routines: if p < q then it would
19567 occasionally produce an invalid result. This will only happen with
19568 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19572 * Be less restrictive and allow also `perl util/perlpath.pl
19573 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19574 because this way one can also use an interpreter named `perl5` (which is
19575 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19576 installed as `perl`).
19578 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19580 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19582 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19584 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19585 advapi32.lib to Win32 build and change the pem test comparison
19586 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19587 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19588 and crypto/des/ede_cbcm_enc.c.
19592 * DES quad checksum was broken on big-endian architectures. Fixed.
19596 * Comment out two functions in bio.h that aren't implemented. Fix up the
19597 Win32 test batch file so it (might) work again. The Win32 test batch file
19598 is horrible: I feel ill....
19602 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19603 in e_os.h. Audit of header files to check ANSI and non ANSI
19604 sections: 10 functions were absent from non ANSI section and not exported
19605 from Windows DLLs. Fixed up libeay.num for new functions.
19609 * Make `openssl version` output lines consistent.
19611 *Ralf S. Engelschall*
19613 * Fix Win32 symbol export lists for BIO functions: Added
19614 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19615 to ms/libeay{16,32}.def.
19617 *Ralf S. Engelschall*
19619 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19620 fine under Unix and passes some trivial tests I've now added. But the
19621 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19622 added to make sure no one expects that this stuff really works in the
19623 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19624 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19627 *Ralf S. Engelschall*
19629 * Fix the generation of two part addresses in perl.
19631 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19633 * Add config entry for Linux on MIPS.
19635 *John Tobey <jtobey@channel1.com>*
19637 * Make links whenever Configure is run, unless we are on Windoze.
19641 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19642 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19647 * Add a useful kludge to allow package maintainers to specify compiler and
19648 other platforms details on the command line without having to patch the
19649 Configure script every time: One now can use
19650 `perl Configure <id>:<details>`,
19651 i.e. platform ids are allowed to have details appended
19652 to them (separated by colons). This is treated as there would be a static
19653 pre-configured entry in Configure's %table under key `<id>` with value
19654 `<details>` and `perl Configure <id>` is called. So, when you want to
19655 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19656 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19657 now, which overrides the FreeBSD-elf entry on-the-fly.
19659 *Ralf S. Engelschall*
19661 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19665 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19666 on the `perl Configure ...` command line. This way one can compile
19667 OpenSSL libraries with Position Independent Code (PIC) which is needed
19668 for linking it into DSOs.
19670 *Ralf S. Engelschall*
19672 * Remarkably, export ciphers were totally broken and no-one had noticed!
19677 * Cleaned up the LICENSE document: The official contact for any license
19678 questions now is the OpenSSL core team under openssl-core@openssl.org.
19679 And add a paragraph about the dual-license situation to make sure people
19680 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19681 to the OpenSSL toolkit.
19683 *Ralf S. Engelschall*
19685 * General source tree makefile cleanups: Made `making xxx in yyy...`
19686 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19687 Additionally cleaned up the `make links` target: Remove unnecessary
19688 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19689 to speed processing and no longer clutter the display with confusing
19690 stuff. Instead only the actually done links are displayed.
19692 *Ralf S. Engelschall*
19694 * Permit null encryption ciphersuites, used for authentication only. It used
19695 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19696 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19701 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19702 signed attributes when verifying signatures (this would break them),
19703 the detached data encoding was wrong and public keys obtained using
19704 X509_get_pubkey() weren't freed.
19708 * Add text documentation for the BUFFER functions. Also added a work around
19709 to a Win95 console bug. This was triggered by the password read stuff: the
19710 last character typed gets carried over to the next fread(). If you were
19711 generating a new cert request using 'req' for example then the last
19712 character of the passphrase would be CR which would then enter the first
19717 * Added the new 'Includes OpenSSL Cryptography Software' button as
19718 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19719 button and can be used by applications based on OpenSSL to show the
19720 relationship to the OpenSSL project.
19722 *Ralf S. Engelschall*
19724 * Remove confusing variables in function signatures in files
19725 ssl/ssl_lib.c and ssl/ssl.h.
19727 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19729 * Don't install bss_file.c under PREFIX/include/
19731 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19733 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19734 functions that return function pointers and has support for NT specific
19735 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19736 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19737 unsigned to signed types: this was killing the Win32 compile.
19741 * Add new certificate file to stack functions,
19742 SSL_add_dir_cert_subjects_to_stack() and
19743 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19744 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19745 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19746 This means that Apache-SSL and similar packages don't have to mess around
19747 to add as many CAs as they want to the preferred list.
19751 * Experiment with doxygen documentation. Currently only partially applied to
19753 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19754 openssl.doxy as the configuration file.
19758 * Get rid of remaining C++-style comments which strict C compilers hate.
19760 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19762 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19763 compiled in by default: it has problems with large keys.
19767 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19768 DH private keys and/or callback functions which directly correspond to
19769 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19770 is needed for applications which have to configure certificates on a
19771 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19773 For the RSA certificate situation is makes no difference, but
19774 for the DSA certificate situation this fixes the "no shared cipher"
19775 problem where the OpenSSL cipher selection procedure failed because the
19776 temporary keys were not overtaken from the context and the API provided
19777 no way to reconfigure them.
19778 The new functions now let applications reconfigure the stuff and they
19779 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19780 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19781 non-public-API function ssl_cert_instantiate() is used as a helper
19782 function and also to reduce code redundancy inside ssl_rsa.c.
19784 *Ralf S. Engelschall*
19786 * Move s_server -dcert and -dkey options out of the undocumented feature
19787 area because they are useful for the DSA situation and should be
19788 recognized by the users.
19790 *Ralf S. Engelschall*
19792 * Fix the cipher decision scheme for export ciphers: the export bits are
19793 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19794 SSL_EXP_MASK. So, the original variable has to be used instead of the
19795 already masked variable.
19797 *Richard Levitte <levitte@stacken.kth.se>*
19799 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19801 *Richard Levitte <levitte@stacken.kth.se>*
19803 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19804 from `int` to `unsigned int` because it is a length and initialized by
19805 EVP_DigestFinal() which expects an `unsigned int *`.
19807 *Richard Levitte <levitte@stacken.kth.se>*
19809 * Don't hard-code path to Perl interpreter on shebang line of Configure
19810 script. Instead use the usual Shell->Perl transition trick.
19812 *Ralf S. Engelschall*
19814 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19815 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19816 -noout -modulus` as it's already the case for `openssl rsa -noout
19817 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19818 currently the public key is printed (a decision which was already done by
19819 `openssl dsa -modulus` in the past) which serves a similar purpose.
19820 Additionally the NO_RSA no longer completely removes the whole -modulus
19821 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19824 *Ralf S. Engelschall*
19826 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19827 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19829 *Arne Ansper <arne@ats.cyber.ee>*
19831 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19832 to be added. Now both 'req' and 'ca' can use new objects defined in the
19837 * Add cool BIO that does syslog (or event log on NT).
19839 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19841 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19842 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19843 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19844 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19848 * Add preliminary config info for new extension code.
19852 * Make RSA_NO_PADDING really use no padding.
19854 *Ulf Moeller <ulf@fitug.de>*
19856 * Generate errors when private/public key check is done.
19860 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19861 for some CRL extensions and new objects added.
19865 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19866 key usage extension and fuller support for authority key id.
19870 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19871 padding method for RSA, which is recommended for new applications in PKCS
19872 #1 v2.0 (RFC 2437, October 1998).
19873 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19874 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19875 against Bleichbacher's attack on RSA.
19876 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19879 * Updates to the new SSL compression code
19881 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19883 * Fix so that the version number in the master secret, when passed
19884 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19885 (because the server will not accept higher), that the version number
19886 is 0x03,0x01, not 0x03,0x00
19888 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19890 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19891 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19892 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19896 * Support for RAW extensions where an arbitrary extension can be
19897 created by including its DER encoding. See `apps/openssl.cnf` for
19902 * Make sure latest Perl versions don't interpret some generated C array
19903 code as Perl array code in the crypto/err/err_genc.pl script.
19905 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19907 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19908 not many people have the assembler. Various Win32 compilation fixes and
19909 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19910 build instructions.
19914 * Modify configure script 'Configure' to automatically create crypto/date.h
19915 file under Win32 and also build pem.h from pem.org. New script
19916 util/mkfiles.pl to create the MINFO file on environments that can't do a
19917 'make files': perl util/mkfiles.pl >MINFO should work.
19921 * Major rework of DES function declarations, in the pursuit of correctness
19922 and purity. As a result, many evil casts evaporated, and some weirdness,
19923 too. You may find this causes warnings in your code. Zapping your evil
19924 casts will probably fix them. Mostly.
19928 * Fix for a typo in asn1.h. Bug fix to object creation script
19929 obj_dat.pl. It considered a zero in an object definition to mean
19930 "end of object": none of the objects in objects.h have any zeros
19931 so it wasn't spotted.
19933 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19935 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19936 Masking (CBCM). In the absence of test vectors, the best I have been able
19937 to do is check that the decrypt undoes the encrypt, so far. Send me test
19938 vectors if you have them.
19942 * Correct calculation of key length for export ciphers (too much space was
19943 allocated for null ciphers). This has not been tested!
19947 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19948 message is now correct (it understands "crypto" and "ssl" on its
19949 command line). There is also now an "update" option. This will update
19950 the util/ssleay.num and util/libeay.num files with any new functions.
19952 perl util/mkdef.pl crypto ssl update
19953 it will update them.
19957 * Overhauled the Perl interface:
19958 - ported BN stuff to OpenSSL's different BN library
19959 - made the perl/ source tree CVS-aware
19960 - renamed the package from SSLeay to OpenSSL (the files still contain
19961 their history because I've copied them in the repository)
19962 - removed obsolete files (the test scripts will be replaced
19963 by better Test::Harness variants in the future)
19965 *Ralf S. Engelschall*
19967 * First cut for a very conservative source tree cleanup:
19968 1. merge various obsolete readme texts into doc/ssleay.txt
19969 where we collect the old documents and readme texts.
19970 2. remove the first part of files where I'm already sure that we no
19971 longer need them because of three reasons: either they are just temporary
19972 files which were left by Eric or they are preserved original files where
19973 I've verified that the diff is also available in the CVS via "cvs diff
19974 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19975 the crypto/md/ stuff).
19977 *Ralf S. Engelschall*
19979 * More extension code. Incomplete support for subject and issuer alt
19980 name, issuer and authority key id. Change the i2v function parameters
19981 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19982 what that's for :-) Fix to ASN1 macro which messed up
19983 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19987 * Preliminary support for ENUMERATED type. This is largely copied from the
19992 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19994 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19996 * Make sure `make rehash` target really finds the `openssl` program.
19998 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
20000 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
20001 like to hear about it if this slows down other processors.
20005 * Add CygWin32 platform information to Configure script.
20007 *Alan Batie <batie@aahz.jf.intel.com>*
20009 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
20011 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
20013 * New program nseq to manipulate netscape certificate sequences
20017 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
20022 * Fixes to BN code. Previously the default was to define BN_RECURSION
20023 but the BN code had some problems that would cause failures when
20024 doing certificate verification and some other functions.
20026 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20028 * Add ASN1 and PEM code to support netscape certificate sequences.
20032 * Add ASN1 and PEM code to support netscape certificate sequences.
20036 * Add several PKIX and private extended key usage OIDs.
20040 * Modify the 'ca' program to handle the new extension code. Modify
20041 openssl.cnf for new extension format, add comments.
20045 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
20046 and add a sample to openssl.cnf so req -x509 now adds appropriate
20051 * Continued X509 V3 changes. Add to other makefiles, integrate with the
20052 error code, add initial support to X509_print() and x509 application.
20056 * Takes a deep breath and start adding X509 V3 extension support code. Add
20057 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
20058 stuff is currently isolated and isn't even compiled yet.
20062 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
20063 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
20064 Removed the versions check from X509 routines when loading extensions:
20065 this allows certain broken certificates that don't set the version
20066 properly to be processed.
20070 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20071 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20072 can still be regenerated with "make depend".
20076 * Spelling mistake in C version of CAST-128.
20078 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20080 * Changes to the error generation code. The perl script err-code.pl
20081 now reads in the old error codes and retains the old numbers, only
20082 adding new ones if necessary. It also only changes the .err files if new
20083 codes are added. The makefiles have been modified to only insert errors
20084 when needed (to avoid needlessly modifying header files). This is done
20085 by only inserting errors if the .err file is newer than the auto generated
20086 C file. To rebuild all the error codes from scratch (the old behaviour)
20087 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20088 or delete all the .err files.
20092 * CAST-128 was incorrectly implemented for short keys. The C version has
20093 been fixed, but is untested. The assembler versions are also fixed, but
20094 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20095 to regenerate it if needed.
20096 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20097 Hagino <itojun@kame.net>*
20099 * File was opened incorrectly in randfile.c.
20101 *Ulf Möller <ulf@fitug.de>*
20103 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20104 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20105 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20106 al: it's just almost always a UTCTime. Note this patch adds new error
20107 codes so do a "make errors" if there are problems.
20111 * Correct Linux 1 recognition in config.
20113 *Ulf Möller <ulf@fitug.de>*
20115 * Remove pointless MD5 hash when using DSA keys in ca.
20117 *Anonymous <nobody@replay.com>*
20119 * Generate an error if given an empty string as a cert directory. Also
20120 generate an error if handed NULL (previously returned 0 to indicate an
20121 error, but didn't set one).
20123 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20125 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20129 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20130 parameters. This was causing a warning which killed off the Win32 compile.
20134 * Remove C++ style comments from crypto/bn/bn_local.h.
20136 *Neil Costigan <neil.costigan@celocom.com>*
20138 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20139 based on a text string, looking up short and long names and finally
20140 "dot" format. The "dot" format stuff didn't work. Added new function
20141 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20142 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20143 OID is not part of the table.
20147 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20148 X509_LOOKUP_by_alias().
20152 * Sort openssl functions by name.
20156 * Get the `gendsa` command working and add it to the `list` command. Remove
20157 encryption from sample DSA keys (in case anyone is interested the password
20162 * Make *all* `*_free` functions accept a NULL pointer.
20164 *Frans Heymans <fheymans@isaserver.be>*
20166 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20169 *Anonymous <nobody@replay.com>*
20171 * s_server should send the CAfile as acceptable CAs, not its own cert.
20173 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20175 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20177 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20179 * Temp key "for export" tests were wrong in s3_srvr.c.
20181 *Anonymous <nobody@replay.com>*
20183 * Add prototype for temp key callback functions
20184 SSL_CTX_set_tmp_{rsa,dh}_callback().
20188 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20189 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20193 * X509_name_add_entry() freed the wrong thing after an error.
20195 *Arne Ansper <arne@ats.cyber.ee>*
20197 * rsa_eay.c would attempt to free a NULL context.
20199 *Arne Ansper <arne@ats.cyber.ee>*
20201 * BIO_s_socket() had a broken should_retry() on Windoze.
20203 *Arne Ansper <arne@ats.cyber.ee>*
20205 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20207 *Arne Ansper <arne@ats.cyber.ee>*
20209 * Make sure the already existing X509_STORE->depth variable is initialized
20210 in X509_STORE_new(), but document the fact that this variable is still
20211 unused in the certificate verification process.
20213 *Ralf S. Engelschall*
20215 * Fix the various library and `apps/` files to free up pkeys obtained from
20216 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20220 * Fix reference counting in X509_PUBKEY_get(). This makes
20221 demos/maurice/example2.c work, amongst others, probably.
20223 *Steve Henson and Ben Laurie*
20225 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20226 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20227 are no longer created. This way we have a single and consistent command
20228 line interface `openssl <command>`, similar to `cvs <command>`.
20230 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20232 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20233 BIT STRING wrapper always have zero unused bits.
20237 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20241 * Make the top-level INSTALL documentation easier to understand.
20245 * Makefiles updated to exit if an error occurs in a sub-directory
20246 make (including if user presses ^C) [Paul Sutton]
20248 * Make Montgomery context stuff explicit in RSA data structure.
20252 * Fix build order of pem and err to allow for generated pem.h.
20256 * Fix renumbering bug in X509_NAME_delete_entry().
20260 * Enhanced the err-ins.pl script so it makes the error library number
20261 global and can add a library name. This is needed for external ASN1 and
20262 other error libraries.
20266 * Fixed sk_insert which never worked properly.
20270 * Fix ASN1 macros so they can handle indefinite length constructed
20271 EXPLICIT tags. Some non standard certificates use these: they can now
20276 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20277 into a single doc/ssleay.txt bundle. This way the information is still
20278 preserved but no longer messes up this directory. Now it's new room for
20279 the new set of documentation files.
20281 *Ralf S. Engelschall*
20283 * SETs were incorrectly DER encoded. This was a major pain, because they
20284 shared code with SEQUENCEs, which aren't coded the same. This means that
20285 almost everything to do with SETs or SEQUENCEs has either changed name or
20286 number of arguments.
20288 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20290 * Fix test data to work with the above.
20294 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20295 was already fixed by Eric for 0.9.1 it seems.
20297 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20299 * Autodetect FreeBSD3.
20303 * Fix various bugs in Configure. This affects the following platforms:
20307 unixware-2.0-pentium
20312 * Eliminate generated files from CVS. Reorder tests to regenerate files
20313 before they are needed.
20317 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20321 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20323 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20324 changed SSLeay to OpenSSL in version strings.
20326 *Ralf S. Engelschall*
20328 * Some fixups to the top-level documents.
20332 * Fixed the nasty bug where rsaref.h was not found under compile-time
20333 because the symlink to include/ was missing.
20335 *Ralf S. Engelschall*
20337 * Incorporated the popular no-RSA/DSA-only patches
20338 which allow to compile an RSA-free SSLeay.
20340 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20342 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20343 when "ssleay" is still not found.
20345 *Ralf S. Engelschall*
20347 * Added more platforms to Configure: Cray T3E, HPUX 11,
20349 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20351 * Updated the README file.
20353 *Ralf S. Engelschall*
20355 * Added various .cvsignore files in the CVS repository subdirs
20356 to make a "cvs update" really silent.
20358 *Ralf S. Engelschall*
20360 * Recompiled the error-definition header files and added
20361 missing symbols to the Win32 linker tables.
20363 *Ralf S. Engelschall*
20365 * Cleaned up the top-level documents;
20366 o new files: CHANGES and LICENSE
20367 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20368 o merged COPYRIGHT into LICENSE
20369 o removed obsolete TODO file
20370 o renamed MICROSOFT to INSTALL.W32
20372 *Ralf S. Engelschall*
20374 * Removed dummy files from the 0.9.1b source tree:
20375 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20376 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20377 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20378 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20379 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20381 *Ralf S. Engelschall*
20383 * Added various platform portability fixes.
20387 * The Genesis of the OpenSSL rpject:
20388 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20389 Young and Tim J. Hudson created while they were working for C2Net until
20392 *The OpenSSL Project*
20394 ### Changes between 0.9.0b and 0.9.1b [not released]
20396 * Updated a few CA certificates under certs/
20400 * Changed some BIGNUM api stuff.
20404 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20405 DGUX x86, Linux Alpha, etc.
20409 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20410 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20415 * Add -strparse option to asn1pars program which parses nested
20418 *Dr Stephen Henson <shenson@bigfoot.com>*
20420 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20424 * DSA fix for "ca" program.
20428 * Added "-genkey" option to "dsaparam" program.
20432 * Added RIPE MD160 (rmd160) message digest.
20436 * Added -a (all) option to "ssleay version" command.
20440 * Added PLATFORM define which is the id given to Configure.
20444 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20448 * Extended the ASN.1 parser routines.
20452 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20456 * Added a BN_CTX to the BN library.
20460 * Fixed the weak key values in DES library
20464 * Changed API in EVP library for cipher aliases.
20468 * Added support for RC2/64bit cipher.
20472 * Converted the lhash library to the crypto/mem.c functions.
20476 * Added more recognized ASN.1 object ids.
20480 * Added more RSA padding checks for SSL/TLS.
20484 * Added BIO proxy/filter functionality.
20488 * Added extra_certs to SSL_CTX which can be used
20489 send extra CA certificates to the client in the CA cert chain sending
20490 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20494 * Now Fortezza is denied in the authentication phase because
20495 this is key exchange mechanism is not supported by SSLeay at all.
20499 * Additional PKCS1 checks.
20503 * Support the string "TLSv1" for all TLS v1 ciphers.
20507 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20508 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20512 * Fixed a few memory leaks.
20516 * Fixed various code and comment typos.
20520 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20521 bytes sent in the client random.
20523 *Edward Bishop <ebishop@spyglass.com>*
20527 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20528 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20529 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20530 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20531 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20532 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20533 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20534 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20535 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20536 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20537 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20538 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20539 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20540 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20541 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20542 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20543 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20544 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20545 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20546 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20547 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20548 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20549 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20550 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20551 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20552 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20553 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20554 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20555 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20556 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20557 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20558 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20559 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20560 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20561 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20562 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20563 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20564 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20565 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20566 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20567 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20568 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20569 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20570 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20571 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20572 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20573 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20574 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20575 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20576 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20577 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20578 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20579 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20580 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20581 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20582 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20583 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20584 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20585 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20586 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20587 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20588 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20589 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20590 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20591 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20592 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20593 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20594 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20595 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20596 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20597 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20598 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20599 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20600 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20601 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20602 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20603 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20604 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20605 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20606 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20607 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20608 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20609 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20610 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20611 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20612 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20613 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20614 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20615 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20616 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20617 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20618 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20619 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20620 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20621 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20622 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20623 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20624 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20625 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20626 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20627 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20628 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20629 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20630 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20631 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20632 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20633 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20634 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20635 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20636 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20637 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20638 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20639 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20640 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20641 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20642 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20643 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20644 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20645 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20646 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20647 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20648 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20649 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20650 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20651 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20652 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20653 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20654 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20655 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20656 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20657 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20658 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20659 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20660 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20661 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20662 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20663 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20664 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20665 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20666 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20667 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20668 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20669 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20670 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20671 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20672 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20673 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20674 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20675 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20676 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20677 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20678 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20679 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20680 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20681 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20682 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20683 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20684 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20685 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20686 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20687 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20688 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20689 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20690 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20691 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20692 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20693 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20694 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20695 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20696 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20697 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20698 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20699 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20700 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20701 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20702 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20703 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20704 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20705 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20706 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20707 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20708 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20709 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20710 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20711 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20712 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20713 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20714 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655