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 * The `-verify` option to the `openssl crl` and `openssl req` will make
32 the program exit with 1 on failure.
36 * The BIO_get_new_index() function can only be called 127 times before it
37 reaches its upper bound of BIO_TYPE_MASK. It will now correctly return an
38 error of -1 once it is exhausted. Users may need to reserve using this
39 function for cases where BIO_find_type() is required. Either BIO_TYPE_NONE
40 or BIO_get_new_index() can be used to supply a type to BIO_meth_new().
44 * Added API functions SSL_SESSION_get_time_ex(), SSL_SESSION_set_time_ex()
45 using time_t which is Y2038 safe on 32 bit systems when 64 bit time
46 is enabled (e.g via setting glibc macro _TIME_BITS=64).
50 * The d2i_ASN1_GENERALIZEDTIME(), d2i_ASN1_UTCTIME(), ASN1_TIME_check(), and
51 related functions have been augmented to check for a minimum length of
52 the input string, in accordance with ITU-T X.690 section 11.7 and 11.8.
56 * Unknown entries in TLS SignatureAlgorithms, ClientSignatureAlgorithms
57 config options and the respective calls to SSL[_CTX]_set1_sigalgs() and
58 SSL[_CTX]_set1_client_sigalgs() that start with `?` character are
59 ignored and the configuration will still be used.
61 Similarly unknown entries that start with `?` character in a TLS
62 Groups config option or set with SSL[_CTX]_set1_groups_list() are ignored
63 and the configuration will still be used.
65 In both cases if the resulting list is empty, an error is returned.
69 * The EVP_PKEY_fromdata function has been augmented to allow for the derivation
70 of CRT (Chinese Remainder Theorem) parameters when requested. See the
71 OSSL_PKEY_PARAM_RSA_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
75 * The activate and soft_load configuration settings for providers in
76 openssl.cnf have been updated to require a value of [1|yes|true|on]
77 (in lower or UPPER case) to enable the setting. Conversely a value
78 of [0|no|false|off] will disable the setting. All other values, or the
79 omission of a value for these settings will result in an error.
83 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
84 override the Issuer and Subject when creating a certificate. The `-subj`
85 option now is an alias for `-set_subject`.
87 *Job Snijders, George Michaelson*
89 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
90 if called with a NULL stack argument.
94 * In `openssl speed`, changed the default hash function used with `hmac` from
99 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
100 - `certProfile` request message header and respective `-profile` CLI option
101 - support for delayed delivery of all types of response messages
105 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
106 be less hard coded in the build file templates, and to allow easier
107 addition of more exporters. With that, an exporter for CMake is also
112 * The BLAKE2s hash algorithm matches BLAKE2b's support
113 for configurable output length.
115 *Ahelenia Ziemiańska*
117 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
118 server to prefer session resumption using PSK-only key exchange over PSK
119 with DHE, if both are available.
121 *Markus Minichmayr, Tapkey GmbH*
123 * New API `SSL_write_ex2`, which can be used to send an end-of-stream (FIN)
124 condition in an optimised way when using QUIC.
128 * New atexit configuration switch, which controls whether the OPENSSL_cleanup
129 is registered when libcrypto is unloaded. This is turned off on NonStop
130 configurations because of loader differences on that platform compared to
135 * Support for qlog for tracing QUIC connections has been added.
137 The qlog output from OpenSSL currently uses a pre-standard draft version of
138 qlog. The output from OpenSSL will change in incompatible ways in future
139 releases, and is not subject to any format stability or compatibility
140 guarantees at this time. This functionality can be
141 disabled with the build-time option `no-unstable-qlog`. See the
142 openssl-qlog(7) manpage for details.
146 * Added APIs to allow configuring the negotiated idle timeout for QUIC
147 connections, and to allow determining the number of additional streams
148 that can currently be created for a QUIC connection.
152 * Added APIs to allow disabling implicit QUIC event processing for
153 QUIC SSL objects, allowing applications to control when event handling
154 occurs. Refer to the SSL_get_value_uint(3) manpage for details.
158 * Added APIs to allow querying the size and utilisation of a QUIC stream's
159 write buffer. Refer to the SSL_get_value_uint(3) manpage for details.
163 * New limit on HTTP response headers is introduced to HTTP client. The
164 default limit is set to 256 header lines. If limit is exceeded the
165 response processing stops with error HTTP_R_RESPONSE_TOO_MANY_HDRLINES.
166 Application may call OSSL_HTTP_REQ_CTX_set_max_response_hdr_lines(3)
167 to change the default. Setting the value to 0 disables the limit.
169 *Alexandr Nedvedicky*
174 ### Changes between 3.2.1 and 3.2.2 [xx XXX xxxx]
176 * Fixed bug where SSL_export_keying_material() could not be used with QUIC
177 connections. (#23560)
181 ### Changes between 3.2.0 and 3.2.1 [30 Jan 2024]
183 * A file in PKCS12 format can contain certificates and keys and may come from
184 an untrusted source. The PKCS12 specification allows certain fields to be
185 NULL, but OpenSSL did not correctly check for this case. A fix has been
186 applied to prevent a NULL pointer dereference that results in OpenSSL
187 crashing. If an application processes PKCS12 files from an untrusted source
188 using the OpenSSL APIs then that application will be vulnerable to this
189 issue prior to this fix.
191 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
192 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
193 and PKCS12_newpass().
195 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
196 function is related to writing data we do not consider it security
203 * When function EVP_PKEY_public_check() is called on RSA public keys,
204 a computation is done to confirm that the RSA modulus, n, is composite.
205 For valid RSA keys, n is a product of two or more large primes and this
206 computation completes quickly. However, if n is an overly large prime,
207 then this computation would take a long time.
209 An application that calls EVP_PKEY_public_check() and supplies an RSA key
210 obtained from an untrusted source could be vulnerable to a Denial of Service
213 The function EVP_PKEY_public_check() is not called from other OpenSSL
214 functions however it is called from the OpenSSL pkey command line
215 application. For that reason that application is also vulnerable if used
216 with the "-pubin" and "-check" options on untrusted data.
218 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
219 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
225 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
226 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
231 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
232 for PowerPC CPUs saves the contents of vector registers in different
233 order than they are restored. Thus the contents of some of these vector
234 registers is corrupted when returning to the caller. The vulnerable code is
235 used only on newer PowerPC processors supporting the PowerISA 2.07
238 The consequences of this kind of internal application state corruption can
239 be various - from no consequences, if the calling application does not
240 depend on the contents of non-volatile XMM registers at all, to the worst
241 consequences, where the attacker could get complete control of the
242 application process. However unless the compiler uses the vector registers
243 for storing pointers, the most likely consequence, if any, would be an
244 incorrect result of some application dependent calculations or a crash
245 leading to a denial of service.
251 * Disable building QUIC server utility when OpenSSL is configured with
256 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
258 * Fix excessive time spent in DH check / generation with large Q parameter
261 Applications that use the functions DH_generate_key() to generate an
262 X9.42 DH key may experience long delays. Likewise, applications that use
263 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
264 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
265 Where the key or parameters that are being checked have been obtained from
266 an untrusted source this may lead to a Denial of Service.
272 * The BLAKE2b hash algorithm supports a configurable output length
273 by setting the "size" parameter.
275 *Čestmír Kalina and Tomáš Mráz*
277 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
281 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
282 and the corresponding provider-storemgmt API function
283 OSSL_FUNC_store_delete().
287 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
288 a passphrase callback when opening a store.
292 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
293 from 8 bytes to 16 bytes.
294 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
295 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
296 requires a salt length of 128 bits. This affects OpenSSL command line
297 applications such as "genrsa" and "pkcs8" and API's such as
298 PEM_write_bio_PrivateKey() that are reliant on the default value.
299 The additional commandline option 'saltlen' has been added to the
300 OpenSSL command line applications for "pkcs8" and "enc" to allow the
301 salt length to be set to a non default value.
305 * Changed the default value of the `ess_cert_id_alg` configuration
306 option which is used to calculate the TSA's public key certificate
307 identifier. The default algorithm is updated to be sha256 instead
310 *Małgorzata Olszówka*
312 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
313 table for point multiplication of the base point, which increases the size of
314 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
315 been added to disable the precomputed table.
319 * Added client side support for QUIC
321 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
323 * Added multiple tutorials on the OpenSSL library and in particular
324 on writing various clients (using TLS and QUIC protocols) with libssl.
328 * Added secp384r1 implementation using Solinas' reduction to improve
329 speed of the NIST P-384 elliptic curve. To enable the implementation
330 the build option `enable-ec_nistp_64_gcc_128` must be used.
334 * Improved RFC7468 compliance of the asn1parse command.
336 *Matthias St. Pierre*
338 * Added SHA256/192 algorithm support.
342 * Added support for securely getting root CA certificate update in
347 * Improved contention on global write locks by using more read locks where
352 * Improved performance of OSSL_PARAM lookups in performance critical
357 * Added the SSL_get0_group_name() function to provide access to the
358 name of the group used for the TLS key exchange.
362 * Provide a new configure option `no-http` that can be used to disable the
363 HTTP support. Provide new configure options `no-apps` and `no-docs` to
364 disable building the openssl command line application and the documentation.
368 * Provide a new configure option `no-ecx` that can be used to disable the
369 X25519, X448, and EdDSA support.
373 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
374 the EVP_KDF_CTX_set_params() function they are now concatenated not just
375 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
379 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
380 the provider context as a parameter.
384 * TLS round-trip time calculation was added by a Brigham Young University
385 Capstone team partnering with Sandia National Laboratories. A new function
386 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
391 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
392 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
393 option. Use of the "advanced" s_client command command via the "-adv" option
398 * Added an "advanced" command mode to s_client. Use this with the "-adv"
399 option. The old "basic" command mode recognises certain letters that must
400 always appear at the start of a line and cannot be escaped. The advanced
401 command mode enables commands to be entered anywhere and there is an
402 escaping mechanism. After starting s_client with "-adv" type "{help}"
403 to show a list of available commands.
407 * Add Raw Public Key (RFC7250) support. Authentication is supported
408 by matching keys against either local policy (TLSA records synthesised
409 from the expected keys) or DANE (TLSA records obtained by the
410 application from DNS). TLSA records will also match the same key in
411 the server certificate, should RPK use not happen to be negotiated.
415 * Added support for modular exponentiation and CRT offloading for the
420 * Added further assembler code for the RISC-V architecture.
424 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
425 from a given EC_GROUP.
429 * Improved support for non-default library contexts and property queries
430 when parsing PKCS#12 files.
434 * Implemented support for all five instances of EdDSA from RFC8032:
435 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
436 The streaming is not yet supported for the HashEdDSA variants
437 (Ed25519ph and Ed448ph).
441 * Added SM4 optimization for ARM processors using ASIMD and AES HW
446 * Implemented SM4-XTS support.
450 * Added platform-agnostic OSSL_sleep() function.
454 * Implemented deterministic ECDSA signatures (RFC6979) support.
458 * Implemented AES-GCM-SIV (RFC8452) support.
462 * Added support for pluggable (provider-based) TLS signature algorithms.
463 This enables TLS 1.3 authentication operations with algorithms embedded
464 in providers not included by default in OpenSSL. In combination with
465 the already available pluggable KEM and X.509 support, this enables
466 for example suitable providers to deliver post-quantum or quantum-safe
467 cryptography to OpenSSL users.
471 * Added support for pluggable (provider-based) CMS signature algorithms.
472 This enables CMS sign and verify operations with algorithms embedded
473 in providers not included by default in OpenSSL.
477 * Added support for Hybrid Public Key Encryption (HPKE) as defined
478 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
479 Message Layer Security (MLS) and other IETF specifications.
480 HPKE can also be used by other applications that require
481 encrypting "to" an ECDH public key. External APIs are defined in
482 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
486 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
491 * Add support for certificate compression (RFC8879), including
492 library support for Brotli and Zstandard compression.
496 * Add the ability to add custom attributes to PKCS12 files. Add a new API
497 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
498 for a user specified callback and optional argument.
499 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
500 added to the existing STACK_OF attrs.
504 * Major refactor of the libssl record layer.
508 * Add a mac salt length option for the pkcs12 command.
512 * Add more SRTP protection profiles from RFC8723 and RFC8269.
516 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
518 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
520 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
521 supported and enabled.
525 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
526 to the list of ciphersuites providing Perfect Forward Secrecy as
527 required by SECLEVEL >= 3.
529 *Dmitry Belyavskiy, Nicola Tuveri*
531 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
532 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
533 SSL_get0_iana_groups() function-like macro, retrieves the list of
534 supported groups sent by the peer.
535 The function SSL_client_hello_get_extension_order() populates
536 a caller-supplied array with the list of extension types present in the
537 ClientHello, in order of appearance.
541 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
542 to make it possible to use empty passphrase strings.
546 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
550 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
551 to change functions used for allocating the memory of asynchronous call stack.
555 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
559 * A failure exit code is returned when using the openssl x509 command to check
560 certificate attributes and the checks fail.
564 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
565 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
566 of 160 bits and above and less than 224 bits were previously accepted by
567 default but are now no longer allowed. By default TLS compression was
568 already disabled in previous OpenSSL versions. At security level 2 it cannot
573 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
578 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
579 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
580 will need to load the legacy crypto provider.
584 * CCM8 cipher suites in TLS have been downgraded to security level zero
585 because they use a short authentication tag which lowers their strength.
589 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
590 by default. Also spaces surrounding `=` in DN output are removed.
594 * Add X.509 certificate codeSigning purpose and related checks on key usage and
595 extended key usage of the leaf certificate according to the CA/Browser Forum.
599 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
600 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
601 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
602 X.509 version 3 if the certificate information includes X.509 extensions.
606 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
607 such as adding a trace facility for debugging certificate chain building.
611 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
612 in particular supporting requests for central key generation, generalized
613 polling, and various types of genm/genp exchanges defined in CMP Updates.
617 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
618 like correcting the TLS and proxy support and adding tracing for debugging.
622 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
626 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
627 a certificate to be added is already present. `CMS_sign_ex()` and
628 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
629 and no longer throw an error for them.
633 * Fixed and extended `util/check-format.pl` for checking adherence to the
634 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
635 The checks are meanwhile more complete and yield fewer false positives.
639 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
640 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
641 calls. They can be used as the transport BIOs for QUIC.
643 *Hugo Landau, Matt Caswell and Tomáš Mráz*
645 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
646 sending and receiving multiple messages in a single call. An implementation
647 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
651 * Support for loading root certificates from the Windows certificate store
652 has been added. The support is in the form of a store which recognises the
653 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
654 arguments. This store is built by default and can be disabled using the new
655 compile-time option `no-winstore`. This store is not currently used by
656 default and must be loaded explicitly using the above store URI. It is
657 expected to be loaded by default in the future.
661 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
662 kernel versions that support KTLS have a known bug in CCM processing. That
663 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
664 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
669 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
674 * Zerocopy KTLS sendfile() support on Linux.
678 * The OBJ_ calls are now thread safe using a global lock.
682 * New parameter `-digest` for openssl cms command allowing signing
683 pre-computed digests and new CMS API functions supporting that
688 * OPENSSL_malloc() and other allocation functions now raise errors on
689 allocation failures. The callers do not need to explicitly raise errors
690 unless they want to for tracing purposes.
694 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
695 decryption as a protection against Bleichenbacher-like attacks.
696 The RSA decryption API will now return a randomly generated deterministic
697 message instead of an error in case it detects an error when checking
698 padding during PKCS#1 v1.5 decryption. This is a general protection against
699 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
701 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
702 on the RSA decryption context.
706 * Added support for Brainpool curves in TLS-1.3.
708 *Bernd Edlinger and Matt Caswell*
710 * Added OpenBSD specific build targets.
714 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
715 a basic thread pool implementation for select platforms.
722 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
724 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
725 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
726 that alter the key or IV length ([CVE-2023-5363]).
730 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
732 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
734 The POLY1305 MAC (message authentication code) implementation in OpenSSL
735 does not save the contents of non-volatile XMM registers on Windows 64
736 platform when calculating the MAC of data larger than 64 bytes. Before
737 returning to the caller all the XMM registers are set to zero rather than
738 restoring their previous content. The vulnerable code is used only on newer
739 x86_64 processors supporting the AVX512-IFMA instructions.
741 The consequences of this kind of internal application state corruption can
742 be various - from no consequences, if the calling application does not
743 depend on the contents of non-volatile XMM registers at all, to the worst
744 consequences, where the attacker could get complete control of the
745 application process. However given the contents of the registers are just
746 zeroized so the attacker cannot put arbitrary values inside, the most likely
747 consequence, if any, would be an incorrect result of some application
748 dependent calculations or a crash leading to a denial of service.
754 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
756 * Fix excessive time spent checking DH q parameter value.
758 The function DH_check() performs various checks on DH parameters. After
759 fixing CVE-2023-3446 it was discovered that a large q parameter value can
760 also trigger an overly long computation during some of these checks.
761 A correct q value, if present, cannot be larger than the modulus p
762 parameter, thus it is unnecessary to perform these checks if q is larger
765 If DH_check() is called with such q parameter value,
766 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
767 intensive checks are skipped.
773 * Fix DH_check() excessive time with over sized modulus.
775 The function DH_check() performs various checks on DH parameters. One of
776 those checks confirms that the modulus ("p" parameter) is not too large.
777 Trying to use a very large modulus is slow and OpenSSL will not normally use
778 a modulus which is over 10,000 bits in length.
780 However the DH_check() function checks numerous aspects of the key or
781 parameters that have been supplied. Some of those checks use the supplied
782 modulus value even if it has already been found to be too large.
784 A new limit has been added to DH_check of 32,768 bits. Supplying a
785 key/parameters with a modulus over this size will simply cause DH_check() to
792 * Do not ignore empty associated data entries with AES-SIV.
794 The AES-SIV algorithm allows for authentication of multiple associated
795 data entries along with the encryption. To authenticate empty data the
796 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
797 with NULL pointer as the output buffer and 0 as the input buffer length.
798 The AES-SIV implementation in OpenSSL just returns success for such call
799 instead of performing the associated data authentication operation.
800 The empty data thus will not be authenticated. ([CVE-2023-2975])
802 Thanks to Juerg Wullschleger (Google) for discovering the issue.
804 The fix changes the authentication tag value and the ciphertext for
805 applications that use empty associated data entries with AES-SIV.
806 To decrypt data encrypted with previous versions of OpenSSL the application
807 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
812 * When building with the `enable-fips` option and using the resulting
813 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
814 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
815 not operate with truncated digests (FIPS 140-3 IG G.R).
819 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
821 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
822 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
824 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
825 numeric text form. For gigantic sub-identifiers, this would take a very
826 long time, the time complexity being O(n^2) where n is the size of that
827 sub-identifier. ([CVE-2023-2650])
829 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
830 IDENTIFIER to canonical numeric text form if the size of that OBJECT
831 IDENTIFIER is 586 bytes or less, and fail otherwise.
833 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
834 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
835 most 128 sub-identifiers, and that the maximum value that each sub-
836 identifier may have is 2^32-1 (4294967295 decimal).
838 For each byte of every sub-identifier, only the 7 lower bits are part of
839 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
840 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
845 * Multiple algorithm implementation fixes for ARM BE platforms.
849 * Added a -pedantic option to fipsinstall that adjusts the various
850 settings to ensure strict FIPS compliance rather than backwards
855 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
856 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
857 trigger a crash of an application using AES-XTS decryption if the memory
858 just after the buffer being decrypted is not mapped.
859 Thanks to Anton Romanov (Amazon) for discovering the issue.
864 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
865 The previous fix for this timing side channel turned out to cause
866 a severe 2-3x performance regression in the typical use case
867 compared to 3.0.7. The new fix uses existing constant time
868 code paths, and restores the previous performance level while
869 fully eliminating all existing timing side channels.
870 The fix was developed by Bernd Edlinger with testing support
875 * Add FIPS provider configuration option to disallow the use of
876 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
877 The option '-no_drbg_truncated_digests' can optionally be
878 supplied to 'openssl fipsinstall'.
882 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
883 that it does not enable policy checking. Thanks to David Benjamin for
884 discovering this issue.
889 * Fixed an issue where invalid certificate policies in leaf certificates are
890 silently ignored by OpenSSL and other certificate policy checks are skipped
891 for that certificate. A malicious CA could use this to deliberately assert
892 invalid certificate policies in order to circumvent policy checking on the
893 certificate altogether.
898 * Limited the number of nodes created in a policy tree to mitigate
899 against CVE-2023-0464. The default limit is set to 1000 nodes, which
900 should be sufficient for most installations. If required, the limit
901 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
902 time define to a desired maximum number of nodes or zero to allow
908 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
910 * Add FIPS provider configuration option to enforce the
911 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
912 The option '-ems-check' can optionally be supplied to
913 'openssl fipsinstall'.
917 * The FIPS provider includes a few non-approved algorithms for
918 backward compatibility purposes and the "fips=yes" property query
919 must be used for all algorithm fetches to ensure FIPS compliance.
921 The algorithms that are included but not approved are Triple DES ECB,
922 Triple DES CBC and EdDSA.
926 * Added support for KMAC in KBKDF.
930 * RNDR and RNDRRS support in provider functions to provide
931 random number generation for Arm CPUs (aarch64).
935 * s_client and s_server apps now explicitly say when the TLS version
936 does not include the renegotiation mechanism. This avoids confusion
937 between that scenario versus when the TLS version includes secure
938 renegotiation but the peer lacks support for it.
942 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
944 *Tomasz Kantecki, Andrey Matyukov*
946 * The various OBJ_* functions have been made thread safe.
950 * Parallel dual-prime 1536/2048-bit modular exponentiation for
951 AVX512_IFMA capable processors.
953 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
955 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
956 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
957 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
958 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
959 `OPENSSL_NO_DEPRECATED_3_1`.
961 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
962 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
963 definitions for these functions regardless of whether
964 `OPENSSL_NO_DEPRECATED_3_1` is defined.
966 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
967 functions regardless of whether they are using them. It is recommended that
968 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
972 * When generating safe-prime DH parameters set the recommended private key
973 length equivalent to minimum key lengths as in RFC 7919.
977 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
978 maximum size that is smaller or equal to the digest length to comply with
979 FIPS 186-4 section 5. This is implemented by a new option
980 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
981 `rsa_pss_saltlen` parameter, which is now the default. Signature
982 verification is not affected by this change and continues to work as before.
989 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
990 listed here are only a brief description.
991 The migration guide contains more detailed information related to new features,
992 breaking changes, and mappings for the large list of deprecated functions.
994 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
996 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
998 * Fixed NULL dereference during PKCS7 data verification.
1000 A NULL pointer can be dereferenced when signatures are being
1001 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
1002 algorithm used for the signature is known to the OpenSSL library but
1003 the implementation of the hash algorithm is not available the digest
1004 initialization will fail. There is a missing check for the return
1005 value from the initialization function which later leads to invalid
1006 usage of the digest API most likely leading to a crash.
1009 PKCS7 data is processed by the SMIME library calls and also by the
1010 time stamp (TS) library calls. The TLS implementation in OpenSSL does
1011 not call these functions however third party applications would be
1012 affected if they call these functions to verify signatures on untrusted
1017 * Fixed X.400 address type confusion in X.509 GeneralName.
1019 There is a type confusion vulnerability relating to X.400 address processing
1020 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
1021 but the public structure definition for GENERAL_NAME incorrectly specified
1022 the type of the x400Address field as ASN1_TYPE. This field is subsequently
1023 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
1024 than an ASN1_STRING.
1026 When CRL checking is enabled (i.e. the application sets the
1027 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
1028 pass arbitrary pointers to a memcmp call, enabling them to read memory
1029 contents or enact a denial of service.
1034 * Fixed NULL dereference validating DSA public key.
1036 An invalid pointer dereference on read can be triggered when an
1037 application tries to check a malformed DSA public key by the
1038 EVP_PKEY_public_check() function. This will most likely lead
1039 to an application crash. This function can be called on public
1040 keys supplied from untrusted sources which could allow an attacker
1041 to cause a denial of service attack.
1043 The TLS implementation in OpenSSL does not call this function
1044 but applications might call the function if there are additional
1045 security requirements imposed by standards such as FIPS 140-3.
1048 *Shane Lontis, Tomáš Mráz*
1050 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
1052 An invalid pointer dereference on read can be triggered when an
1053 application tries to load malformed PKCS7 data with the
1054 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
1056 The result of the dereference is an application crash which could
1057 lead to a denial of service attack. The TLS implementation in OpenSSL
1058 does not call this function however third party applications might
1059 call these functions on untrusted data.
1064 * Fixed Use-after-free following BIO_new_NDEF.
1066 The public API function BIO_new_NDEF is a helper function used for
1067 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
1068 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
1069 be called directly by end user applications.
1071 The function receives a BIO from the caller, prepends a new BIO_f_asn1
1072 filter BIO onto the front of it to form a BIO chain, and then returns
1073 the new head of the BIO chain to the caller. Under certain conditions,
1074 for example if a CMS recipient public key is invalid, the new filter BIO
1075 is freed and the function returns a NULL result indicating a failure.
1076 However, in this case, the BIO chain is not properly cleaned up and the
1077 BIO passed by the caller still retains internal pointers to the previously
1078 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
1079 then a use-after-free will occur. This will most likely result in a crash.
1082 *Viktor Dukhovni, Matt Caswell*
1084 * Fixed Double free after calling PEM_read_bio_ex.
1086 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
1087 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1088 data. If the function succeeds then the "name_out", "header" and "data"
1089 arguments are populated with pointers to buffers containing the relevant
1090 decoded data. The caller is responsible for freeing those buffers. It is
1091 possible to construct a PEM file that results in 0 bytes of payload data.
1092 In this case PEM_read_bio_ex() will return a failure code but will populate
1093 the header argument with a pointer to a buffer that has already been freed.
1094 If the caller also frees this buffer then a double free will occur. This
1095 will most likely lead to a crash.
1097 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1098 PEM_read_bio_ex() and therefore these functions are also directly affected.
1100 These functions are also called indirectly by a number of other OpenSSL
1101 functions including PEM_X509_INFO_read_bio_ex() and
1102 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1103 internal uses of these functions are not vulnerable because the caller does
1104 not free the header argument if PEM_read_bio_ex() returns a failure code.
1107 *Kurt Roeckx, Matt Caswell*
1109 * Fixed Timing Oracle in RSA Decryption.
1111 A timing based side channel exists in the OpenSSL RSA Decryption
1112 implementation which could be sufficient to recover a plaintext across
1113 a network in a Bleichenbacher style attack. To achieve a successful
1114 decryption an attacker would have to be able to send a very large number
1115 of trial messages for decryption. The vulnerability affects all RSA padding
1116 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1119 *Dmitry Belyavsky, Hubert Kario*
1121 * Fixed X.509 Name Constraints Read Buffer Overflow.
1123 A read buffer overrun can be triggered in X.509 certificate verification,
1124 specifically in name constraint checking. The read buffer overrun might
1125 result in a crash which could lead to a denial of service attack.
1126 In a TLS client, this can be triggered by connecting to a malicious
1127 server. In a TLS server, this can be triggered if the server requests
1128 client authentication and a malicious client connects.
1133 * Fixed X.509 Policy Constraints Double Locking security issue.
1135 If an X.509 certificate contains a malformed policy constraint and
1136 policy processing is enabled, then a write lock will be taken twice
1137 recursively. On some operating systems (most widely: Windows) this
1138 results in a denial of service when the affected process hangs. Policy
1139 processing being enabled on a publicly facing server is not considered
1140 to be a common setup.
1145 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1146 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1147 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1148 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1149 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1150 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1151 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1152 for legacy EC and SM2 keys is also changed similarly to honor the
1153 equivalent conversion format flag as specified in the underlying
1154 `EC_KEY` object being exported to a provider, when this function is
1155 called through `EVP_PKEY_export()`.
1159 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1161 * Fixed two buffer overflows in punycode decoding functions.
1163 A buffer overrun can be triggered in X.509 certificate verification,
1164 specifically in name constraint checking. Note that this occurs after
1165 certificate chain signature verification and requires either a CA to
1166 have signed the malicious certificate or for the application to continue
1167 certificate verification despite failure to construct a path to a trusted
1170 In a TLS client, this can be triggered by connecting to a malicious
1171 server. In a TLS server, this can be triggered if the server requests
1172 client authentication and a malicious client connects.
1174 An attacker can craft a malicious email address to overflow
1175 an arbitrary number of bytes containing the `.` character (decimal 46)
1176 on the stack. This buffer overflow could result in a crash (causing a
1180 An attacker can craft a malicious email address to overflow four
1181 attacker-controlled bytes on the stack. This buffer overflow could
1182 result in a crash (causing a denial of service) or potentially remote code
1183 execution depending on stack layout for any given platform/compiler.
1188 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1189 parameters in OpenSSL code.
1190 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1191 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1192 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1193 Using these invalid names may cause algorithms to use slower methods
1194 that ignore the CRT parameters.
1198 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1203 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1204 data to be signed before signing the certificate.
1208 * Added RIPEMD160 to the default provider.
1212 * Ensured that the key share group sent or accepted for the key exchange
1213 is allowed for the protocol version.
1217 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1219 * OpenSSL supports creating a custom cipher via the legacy
1220 EVP_CIPHER_meth_new() function and associated function calls. This function
1221 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1222 to use the new provider mechanism in order to implement custom ciphers.
1224 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1225 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1226 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1227 and decryption initialisation functions). Instead of using the custom cipher
1228 directly it incorrectly tries to fetch an equivalent cipher from the
1229 available providers. An equivalent cipher is found based on the NID passed to
1230 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1231 given cipher. However it is possible for an application to incorrectly pass
1232 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1233 is used in this way the OpenSSL encryption/decryption initialisation function
1234 will match the NULL cipher as being equivalent and will fetch this from the
1235 available providers. This will succeed if the default provider has been
1236 loaded (or if a third party provider has been loaded that offers this
1237 cipher). Using the NULL cipher means that the plaintext is emitted as the
1240 Applications are only affected by this issue if they call
1241 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1242 encryption/decryption initialisation function. Applications that only use
1243 SSL/TLS are not impacted by this issue.
1248 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1253 * Fixed the linux-mips64 Configure target which was missing the
1254 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1259 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1264 * Correctly handle a retransmitted ClientHello in DTLS
1268 * Fixed detection of ktls support in cross-compile environment on Linux
1272 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1277 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1278 report correct results in some cases
1282 * Fix UWP builds by defining VirtualLock
1286 * For known safe primes use the minimum key length according to RFC 7919.
1287 Longer private key sizes unnecessarily raise the cycles needed to compute the
1288 shared secret without any increase of the real security. This fixes a
1289 regression from 1.1.1 where these shorter keys were generated for the known
1294 * Added the loongarch64 target
1298 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1299 only passed to the FIPS provider and not to the default or legacy provider.
1303 * Fixed reported performance degradation on aarch64. Restored the
1304 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1305 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1306 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1307 The new algorithm is still used for 32 bit targets.
1311 * Added a missing header for memcmp that caused compilation failure on some
1316 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1318 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1319 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1320 This issue makes the RSA implementation with 2048 bit private keys
1321 incorrect on such machines and memory corruption will happen during
1322 the computation. As a consequence of the memory corruption an attacker
1323 may be able to trigger a remote code execution on the machine performing
1326 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1327 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1328 are affected by this issue.
1333 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1334 implementation would not encrypt the entirety of the data under some
1335 circumstances. This could reveal sixteen bytes of data that was
1336 preexisting in the memory that wasn't written. In the special case of
1337 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1339 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1340 they are both unaffected.
1343 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1345 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1347 * In addition to the c_rehash shell command injection identified in
1348 CVE-2022-1292, further bugs where the c_rehash script does not
1349 properly sanitise shell metacharacters to prevent command injection have been
1352 When the CVE-2022-1292 was fixed it was not discovered that there
1353 are other places in the script where the file names of certificates
1354 being hashed were possibly passed to a command executed through the shell.
1356 This script is distributed by some operating systems in a manner where
1357 it is automatically executed. On such operating systems, an attacker
1358 could execute arbitrary commands with the privileges of the script.
1360 Use of the c_rehash script is considered obsolete and should be replaced
1361 by the OpenSSL rehash command line tool.
1364 *Daniel Fiala, Tomáš Mráz*
1366 * Case insensitive string comparison no longer uses locales. It has instead
1367 been directly implemented.
1371 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1373 * Case insensitive string comparison is reimplemented via new locale-agnostic
1374 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1375 comparison. The previous implementation had problems when the Turkish locale
1380 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1381 metacharacters to prevent command injection. This script is distributed by
1382 some operating systems in a manner where it is automatically executed. On
1383 such operating systems, an attacker could execute arbitrary commands with the
1384 privileges of the script.
1386 Use of the c_rehash script is considered obsolete and should be replaced
1387 by the OpenSSL rehash command line tool.
1392 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1393 certificate on an OCSP response. The bug caused the function in the case
1394 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1395 response (meaning a successful verification) even in the case where the
1396 response signing certificate fails to verify.
1398 It is anticipated that most users of `OCSP_basic_verify` will not use the
1399 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1400 a negative value (indicating a fatal error) in the case of a certificate
1401 verification failure. The normal expected return value in this case would be
1404 This issue also impacts the command line OpenSSL "ocsp" application. When
1405 verifying an ocsp response with the "-no_cert_checks" option the command line
1406 application will report that the verification is successful even though it
1407 has in fact failed. In this case the incorrect successful response will also
1408 be accompanied by error messages showing the failure and contradicting the
1409 apparently successful result.
1414 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1415 AAD data as the MAC key. This made the MAC key trivially predictable.
1417 An attacker could exploit this issue by performing a man-in-the-middle attack
1418 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1419 that the modified data would still pass the MAC integrity check.
1421 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1422 endpoint will always be rejected by the recipient and the connection will
1423 fail at that point. Many application protocols require data to be sent from
1424 the client to the server first. Therefore, in such a case, only an OpenSSL
1425 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1427 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1428 sent in both directions. In this case both clients and servers could be
1429 affected, regardless of the application protocol.
1431 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1432 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1433 the handshake when using this ciphersuite.
1435 The confidentiality of data is not impacted by this issue, i.e. an attacker
1436 cannot decrypt data that has been encrypted using this ciphersuite - they can
1439 In order for this attack to work both endpoints must legitimately negotiate
1440 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1441 OpenSSL 3.0, and is not available within the default provider or the default
1442 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1443 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1444 following must have occurred:
1446 1) OpenSSL must have been compiled with the (non-default) compile time option
1447 enable-weak-ssl-ciphers
1449 2) OpenSSL must have had the legacy provider explicitly loaded (either
1450 through application code or via configuration)
1452 3) The ciphersuite must have been explicitly added to the ciphersuite list
1454 4) The libssl security level must have been set to 0 (default is 1)
1456 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1458 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1459 others that both endpoints have in common
1464 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1465 occupied by the removed hash table entries.
1467 This function is used when decoding certificates or keys. If a long lived
1468 process periodically decodes certificates or keys its memory usage will
1469 expand without bounds and the process might be terminated by the operating
1470 system causing a denial of service. Also traversing the empty hash table
1471 entries will take increasingly more time.
1473 Typically such long lived processes might be TLS clients or TLS servers
1474 configured to accept client certificate authentication.
1477 *Hugo Landau, Aliaksei Levin*
1479 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1480 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1481 statistics are no longer supported. For compatibility, these statistics are
1482 still listed in the output but are now always reported as zero.
1486 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1488 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1489 for non-prime moduli.
1491 Internally this function is used when parsing certificates that contain
1492 elliptic curve public keys in compressed form or explicit elliptic curve
1493 parameters with a base point encoded in compressed form.
1495 It is possible to trigger the infinite loop by crafting a certificate that
1496 has invalid explicit curve parameters.
1498 Since certificate parsing happens prior to verification of the certificate
1499 signature, any process that parses an externally supplied certificate may thus
1500 be subject to a denial of service attack. The infinite loop can also be
1501 reached when parsing crafted private keys as they can contain explicit
1502 elliptic curve parameters.
1504 Thus vulnerable situations include:
1506 - TLS clients consuming server certificates
1507 - TLS servers consuming client certificates
1508 - Hosting providers taking certificates or private keys from customers
1509 - Certificate authorities parsing certification requests from subscribers
1510 - Anything else which parses ASN.1 elliptic curve parameters
1512 Also any other applications that use the BN_mod_sqrt() where the attacker
1513 can control the parameter values are vulnerable to this DoS issue.
1518 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1519 to the list of ciphersuites providing Perfect Forward Secrecy as
1520 required by SECLEVEL >= 3.
1522 *Dmitry Belyavskiy, Nicola Tuveri*
1524 * Made the AES constant time code for no-asm configurations
1525 optional due to the resulting 95% performance degradation.
1526 The AES constant time code can be enabled, for no assembly
1527 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1531 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1536 * The negative return value handling of the certificate verification callback
1537 was reverted. The replacement is to set the verification retry state with
1538 the SSL_set_retry_verify() function.
1542 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1544 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1545 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1546 verify a certificate supplied by a server. That function may return a
1547 negative return value to indicate an internal error (for example out of
1548 memory). Such a negative return value is mishandled by OpenSSL and will cause
1549 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1550 success and a subsequent call to SSL_get_error() to return the value
1551 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1552 returned by OpenSSL if the application has previously called
1553 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1554 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1555 totally unexpected and applications may not behave correctly as a result. The
1556 exact behaviour will depend on the application but it could result in
1557 crashes, infinite loops or other similar incorrect responses.
1559 This issue is made more serious in combination with a separate bug in OpenSSL
1560 3.0 that will cause X509_verify_cert() to indicate an internal error when
1561 processing a certificate chain. This will occur where a certificate does not
1562 include the Subject Alternative Name extension but where a Certificate
1563 Authority has enforced name constraints. This issue can occur even with valid
1569 * Corrected a few file name and file reference bugs in the build,
1570 installation and setup scripts, which lead to installation verification
1571 failures. Slightly enhanced the installation verification script.
1575 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1580 * Fixed PVK encoder to properly query for the passphrase.
1584 * Multiple fixes in the OSSL_HTTP API functions.
1588 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1589 OSSL_PARAM_INTEGER data type and return error on negative numbers
1590 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1591 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1595 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1599 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1603 * Multiple threading fixes.
1607 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1611 * Allow fetching an operation from the provider that owns an unexportable key
1612 as a fallback if that is still allowed by the property query.
1616 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1618 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1623 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1624 S390X capability vector to zero. This simplifies testing of different code
1625 paths on S390X architecture.
1629 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1630 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1631 SP 800-38D". The communication will fail at this point.
1635 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1636 confidential in EC_GROUP data.
1640 * The byte order mark (BOM) character is ignored if encountered at the
1641 beginning of a PEM-formatted file.
1645 * Added CMS support for the Russian GOST algorithms.
1649 * Due to move of the implementation of cryptographic operations
1650 to the providers, validation of various operation parameters can
1651 be postponed until the actual operation is executed where previously
1652 it happened immediately when an operation parameter was set.
1654 For example when setting an unsupported curve with
1655 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1656 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1658 *OpenSSL team members and many third party contributors*
1660 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1661 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1662 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1663 instead to retrieve these algorithms from a provider.
1667 * On build targets where the multilib postfix is set in the build
1668 configuration the libdir directory was changing based on whether
1669 the lib directory with the multilib postfix exists on the system
1670 or not. This unpredictable behavior was removed and eventual
1671 multilib postfix is now always added to the default libdir. Use
1672 `--libdir=lib` to override the libdir if adding the postfix is
1677 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1678 no longer interoperable with OpenSSL 1.1.1.
1682 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1683 function codes, this function can only cause problems for calling
1688 * Add a configurable flag to output date formats as ISO 8601. Does not
1689 change the default date format.
1693 * Version of MSVC earlier than 1300 could get link warnings, which could
1694 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1695 Support for this flag has been removed.
1699 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1700 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1701 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1702 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1703 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1707 * The signatures of the functions to get and set options on SSL and
1708 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1709 Some source code changes may be required.
1713 * The public definitions of conf_method_st and conf_st have been
1714 deprecated. They will be made opaque in a future release.
1716 *Rich Salz and Tomáš Mráz*
1718 * Client-initiated renegotiation is disabled by default. To allow it, use
1719 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1720 flag, or the "ClientRenegotiation" config parameter as appropriate.
1724 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1725 or modify relative pathname inclusion.
1729 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1730 validated. Please consult the README-FIPS and
1731 README-PROVIDERS files, as well as the migration guide.
1733 *OpenSSL team members and many third party contributors*
1735 * For the key types DH and DHX the allowed settable parameters are now different.
1739 * The openssl commands that read keys, certificates, and CRLs now
1740 automatically detect the PEM or DER format of the input files.
1742 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1744 * Added enhanced PKCS#12 APIs which accept a library context.
1748 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1752 * Added support for Kernel TLS (KTLS).
1754 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1756 * Support for RFC 5746 secure renegotiation is now required by default for
1757 SSL or TLS connections to succeed.
1761 * The signature of the `copy` functional parameter of the
1762 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1763 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1764 the signature of the `pub_decode` functional parameter of the
1765 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1766 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1770 * The error return values from some control calls (ctrl) have changed.
1774 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1778 * Many functions in the EVP_ namespace that are getters of values from
1779 implementations or contexts were renamed to include get or get0 in their
1780 names. Old names are provided as macro aliases for compatibility and
1785 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1786 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1787 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1792 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1795 * The output from the command line applications may have minor
1800 * The output from numerous "printing" may have minor changes.
1804 * Windows thread synchronization uses read/write primitives (SRWLock) when
1805 supported by the OS, otherwise CriticalSection continues to be used.
1809 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1810 work on read only BIO source/sinks that do not support these functions.
1811 This allows piping or redirection of a file BIO using stdin to be buffered
1812 into memory. This is used internally in OSSL_DECODER_from_bio().
1816 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1817 this function would return one of the values OSSL_STORE_INFO_NAME,
1818 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1819 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1820 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1821 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1822 using this function should be amended to handle the changed return value.
1826 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1827 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1828 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1829 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1830 contains more than one certificate identifier: This means that all
1831 certificates referenced there MUST be part of the validation chain.
1835 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1836 RC5, DESX and DES have been moved to the legacy provider.
1840 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1841 RIPEMD-160 have been moved to the legacy provider.
1845 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1850 * The deprecated functions EVP_PKEY_get0_RSA(),
1851 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1852 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1853 well as the similarly named "get1" functions behave differently in
1858 * A number of functions handling low-level keys or engines were deprecated
1859 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1860 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1861 EVP_PKEY_get0_siphash().
1865 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1866 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1867 will need to load the legacy crypto provider. This includes these PBE
1868 algorithms which use this KDF:
1869 - NID_pbeWithMD2AndDES_CBC
1870 - NID_pbeWithMD5AndDES_CBC
1871 - NID_pbeWithSHA1AndRC2_CBC
1872 - NID_pbeWithMD2AndRC2_CBC
1873 - NID_pbeWithMD5AndRC2_CBC
1874 - NID_pbeWithSHA1AndDES_CBC
1878 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1879 BIO_debug_callback() functions.
1883 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1884 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1888 * The RAND_METHOD APIs have been deprecated.
1892 * The SRP APIs have been deprecated.
1896 * Add a compile time option to prevent the caching of provider fetched
1897 algorithms. This is enabled by including the no-cached-fetch option
1898 at configuration time.
1902 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1903 count of PKCS12_DEFAULT_ITER.
1905 *Tomáš Mráz and Sahana Prasad*
1907 * The openssl speed command does not use low-level API calls anymore.
1911 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1914 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1916 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1920 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1921 providers may supply their own group implementations (using either the "key
1922 exchange" or the "key encapsulation" methods) which will automatically be
1923 detected and used by libssl.
1925 *Matt Caswell, Nicola Tuveri*
1927 * The undocumented function X509_certificate_type() has been deprecated;
1931 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1935 * Removed RSA padding mode for SSLv23 (which was only used for
1936 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1937 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1942 * Deprecated the obsolete X9.31 RSA key generation related functions.
1944 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1945 is not allowed to return a value > 1, this is no more taken as failure.
1947 *Viktor Dukhovni and David von Oheimb*
1949 * Deprecated the obsolete X9.31 RSA key generation related functions
1950 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1951 BN_X931_generate_prime_ex().
1955 * The default key generation method for the regular 2-prime RSA keys was
1956 changed to the FIPS 186-4 B.3.6 method.
1960 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1964 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1968 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1969 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1971 *Rich Salz, Richard Levitte, and David von Oheimb*
1973 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1977 * Deprecated `OCSP_parse_url()`.
1981 * Validation of SM2 keys has been separated from the validation of regular EC
1986 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1987 switches: a validation failure triggers an early exit, returning a failure
1988 exit status to the parent process.
1992 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1993 to ignore unknown ciphers.
1997 * The `-cipher-commands` and `-digest-commands` options
1998 of the command line utility `list` have been deprecated.
1999 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
2003 * Added convenience functions for generating asymmetric key pairs:
2004 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
2005 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
2009 * All of the low-level EC_KEY functions have been deprecated.
2011 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
2013 * Deprecated all the libcrypto and libssl error string loading
2018 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
2019 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
2024 * The `-crypt` option to the `passwd` command line tool has been removed.
2028 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
2033 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
2037 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
2038 EVP_PKEY_get1_tls_encodedpoint().
2042 * The security callback, which can be customised by application code, supports
2043 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
2044 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
2048 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
2049 interface. Their functionality remains unchanged.
2053 * Added new option for 'openssl list', '-providers', which will display the
2054 list of loaded providers, their names, version and status. It optionally
2055 displays their gettable parameters.
2059 * Removed EVP_PKEY_set_alias_type().
2063 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
2064 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
2068 * Changed all "STACK" functions to be macros instead of inline functions. Macro
2069 parameters are still checked for type safety at compile time via helper
2074 * Remove the RAND_DRBG API
2076 *Paul Dale and Matthias St. Pierre*
2078 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
2079 as well as actual hostnames.
2083 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2084 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2085 conversely, silently ignore DTLS protocol version bounds when configuring
2086 TLS-based contexts. The commands can be repeated to set bounds of both
2087 types. The same applies with the corresponding "min_protocol" and
2088 "max_protocol" command-line switches, in case some application uses both TLS
2091 SSL_CTX instances that are created for a fixed protocol version (e.g.
2092 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2093 attempts to apply bounds to these protocol versions would result in an
2094 error. Now only the "version-flexible" SSL_CTX instances are subject to
2095 limits in configuration files in command-line options.
2099 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2104 * Reworked the recorded ERR codes to make better space for system errors.
2105 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2106 given code is a system error (true) or an OpenSSL error (false).
2110 * Reworked the test perl framework to better allow parallel testing.
2112 *Nicola Tuveri and David von Oheimb*
2114 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2115 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2119 * 'Configure' has been changed to figure out the configuration target if
2120 none is given on the command line. Consequently, the 'config' script is
2121 now only a mere wrapper. All documentation is changed to only mention
2124 *Rich Salz and Richard Levitte*
2126 * Added a library context `OSSL_LIB_CTX` that applications as well as
2127 other libraries can use to form a separate context within which
2128 libcrypto operations are performed.
2132 * Added various `_ex` functions to the OpenSSL API that support using
2133 a non-default `OSSL_LIB_CTX`.
2137 * Handshake now fails if Extended Master Secret extension is dropped
2142 * Dropped interactive mode from the `openssl` program.
2146 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2148 *David von Oheimb and Shane Lontis*
2150 * Deprecated `EC_METHOD_get_field_type()`.
2154 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2155 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2156 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2160 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2164 * Add CAdES-BES signature verification support, mostly derived
2165 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2167 *Filipe Raimundo da Silva*
2169 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2173 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2174 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2178 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2182 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2183 EC_KEY_precompute_mult().
2187 * Deprecated EC_POINTs_mul().
2191 * Removed FIPS_mode() and FIPS_mode_set().
2195 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2199 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2200 EC_POINT_get_Jprojective_coordinates_GFp().
2204 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2205 arrays to be more easily constructed via a series of utility functions.
2206 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2207 the various push functions and finally convert to a passable OSSL_PARAM
2208 array using OSSL_PARAM_BLD_to_param().
2212 * The security strength of SHA1 and MD5 based signatures in TLS has been
2217 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2218 contain a provider side internal key.
2222 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2226 * Project text documents not yet having a proper file name extension
2227 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2228 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2232 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2233 have been converted to Markdown with the goal to produce documents
2234 which not only look pretty when viewed online in the browser, but
2235 remain well readable inside a plain text editor.
2237 To achieve this goal, a 'minimalistic' Markdown style has been applied
2238 which avoids formatting elements that interfere too much with the
2239 reading flow in the text file. For example, it
2241 * avoids [ATX headings][] and uses [setext headings][] instead
2242 (which works for `<h1>` and `<h2>` headings only).
2243 * avoids [inline links][] and uses [reference links][] instead.
2244 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2246 [ATX headings]: https://github.github.com/gfm/#atx-headings
2247 [setext headings]: https://github.github.com/gfm/#setext-headings
2248 [inline links]: https://github.github.com/gfm/#inline-link
2249 [reference links]: https://github.github.com/gfm/#reference-link
2250 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2251 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2253 *Matthias St. Pierre*
2255 * The test suite is changed to preserve results of each test recipe.
2256 A new directory test-runs/ with subdirectories named like the
2257 test recipes are created in the build tree for this purpose.
2261 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2262 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2263 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2265 *David von Oheimb, Martin Peylo*
2267 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2268 It supports arbitrary request and response content types, GET redirection,
2269 TLS, connections via HTTP(S) proxies, connections and exchange via
2270 user-defined BIOs (allowing implicit connections), persistent connections,
2271 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2272 The legacy OCSP-focused (and only partly documented) API
2273 is retained for backward compatibility, while most of it is deprecated.
2277 * Added `util/check-format.pl`, a tool for checking adherence to the
2278 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2279 The checks performed are incomplete and yield some false positives.
2280 Still the tool should be useful for detecting most typical glitches.
2284 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2285 If domain name resolution yields multiple IP addresses all of them are tried
2286 after `connect()` failures.
2290 * All of the low-level RSA functions have been deprecated.
2294 * X509 certificates signed using SHA1 are no longer allowed at security
2299 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2300 modified to use PKEY APIs. These commands are now in maintenance mode
2301 and no new features will be added to them.
2305 * The command line utility rsautl has been deprecated.
2309 * The command line utilities genrsa and rsa have been modified to use PKEY
2310 APIs. They now write PKCS#8 keys by default. These commands are now in
2311 maintenance mode and no new features will be added to them.
2315 * All of the low-level DH functions have been deprecated.
2317 *Paul Dale and Matt Caswell*
2319 * All of the low-level DSA functions have been deprecated.
2323 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2324 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2328 * Deprecated low-level ECDH and ECDSA functions.
2332 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2336 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2337 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2338 a new formulation to include all the things it can be used for,
2339 as well as words of caution.
2343 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2347 * All of the low-level HMAC functions have been deprecated.
2349 *Paul Dale and David von Oheimb*
2351 * Over two thousand fixes were made to the documentation, including:
2352 - Common options (such as -rand/-writerand, TLS version control, etc)
2353 were refactored and point to newly-enhanced descriptions in openssl.pod.
2354 - Added style conformance for all options (with help from Richard Levitte),
2355 documented all reported missing options, added a CI build to check
2356 that all options are documented and that no unimplemented options
2358 - Documented some internals, such as all use of environment variables.
2359 - Addressed all internal broken L<> references.
2363 * All of the low-level CMAC functions have been deprecated.
2367 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2368 functions have been deprecated.
2370 *Paul Dale and David von Oheimb*
2372 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2373 set of functions. The documentation mentioned negative values for some
2374 errors, but this was never the case, so the mention of negative values
2377 Code that followed the documentation and thereby check with something
2378 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2382 * All of the low-level cipher functions have been deprecated.
2384 *Matt Caswell and Paul Dale*
2386 * Removed include/openssl/opensslconf.h.in and replaced it with
2387 include/openssl/configuration.h.in, which differs in not including
2388 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2389 was added to include both.
2391 This allows internal hacks where one might need to modify the set
2392 of configured macros, for example this if deprecated symbols are
2393 still supposed to be available internally:
2395 #include <openssl/configuration.h>
2397 #undef OPENSSL_NO_DEPRECATED
2398 #define OPENSSL_SUPPRESS_DEPRECATED
2400 #include <openssl/macros.h>
2402 This should not be used by applications that use the exported
2403 symbols, as that will lead to linking errors.
2407 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2408 used in exponentiation with 512-bit moduli. No EC algorithms are
2409 affected. Analysis suggests that attacks against 2-prime RSA1024,
2410 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2411 difficult to perform and are not believed likely. Attacks against DH512
2412 are considered just feasible. However, for an attack the target would
2413 have to reuse the DH512 private key, which is not recommended anyway.
2414 Also applications directly using the low-level API BN_mod_exp may be
2415 affected if they use BN_FLG_CONSTTIME.
2420 * Most memory-debug features have been deprecated, and the functionality
2421 replaced with no-ops.
2425 * Added documentation for the STACK API.
2429 * Introduced a new method type and API, OSSL_ENCODER, to represent
2430 generic encoders. These do the same sort of job that PEM writers
2431 and d2i functions do, but with support for methods supplied by
2432 providers, and the possibility for providers to support other
2437 * Introduced a new method type and API, OSSL_DECODER, to represent
2438 generic decoders. These do the same sort of job that PEM readers
2439 and i2d functions do, but with support for methods supplied by
2440 providers, and the possibility for providers to support other
2445 * Added a .pragma directive to the syntax of configuration files, to
2446 allow varying behavior in a supported and predictable manner.
2447 Currently added pragma:
2451 This allows dollar signs to be a keyword character unless it's
2452 followed by a opening brace or parenthesis. This is useful for
2453 platforms where dollar signs are commonly used in names, such as
2454 volume names and system directory names on VMS.
2458 * Added functionality to create an EVP_PKEY from user data.
2462 * Change the interpretation of the '--api' configuration option to
2463 mean that this is a desired API compatibility level with no
2464 further meaning. The previous interpretation, that this would
2465 also mean to remove all deprecated symbols up to and including
2466 the given version, no requires that 'no-deprecated' is also used
2467 in the configuration.
2469 When building applications, the desired API compatibility level
2470 can be set with the OPENSSL_API_COMPAT macro like before. For
2471 API compatibility version below 3.0, the old style numerical
2472 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2473 For version 3.0 and on, the value is expected to be the decimal
2474 value calculated from the major and minor version like this:
2476 MAJOR * 10000 + MINOR * 100
2480 -DOPENSSL_API_COMPAT=30000 For 3.0
2481 -DOPENSSL_API_COMPAT=30200 For 3.2
2483 To hide declarations that are deprecated up to and including the
2484 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2485 given when building the application as well.
2489 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2490 access to certificate and CRL stores via URIs and OSSL_STORE
2493 This adds the following functions:
2495 - X509_LOOKUP_store()
2496 - X509_STORE_load_file()
2497 - X509_STORE_load_path()
2498 - X509_STORE_load_store()
2499 - SSL_add_store_cert_subjects_to_stack()
2500 - SSL_CTX_set_default_verify_store()
2501 - SSL_CTX_load_verify_file()
2502 - SSL_CTX_load_verify_dir()
2503 - SSL_CTX_load_verify_store()
2507 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2508 The presence of this system service is determined at run-time.
2512 * Added functionality to create an EVP_PKEY context based on data
2513 for methods from providers. This takes an algorithm name and a
2514 property query string and simply stores them, with the intent
2515 that any operation that uses this context will use those strings
2516 to fetch the needed methods implicitly, thereby making the port
2517 of application written for pre-3.0 OpenSSL easier.
2521 * The undocumented function NCONF_WIN32() has been deprecated; for
2522 conversion details see the HISTORY section of doc/man5/config.pod
2526 * Introduced the new functions EVP_DigestSignInit_ex() and
2527 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2528 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2529 pages for further details.
2533 * Over two thousand fixes were made to the documentation, including:
2534 adding missing command flags, better style conformance, documentation
2537 *Rich Salz, Richard Levitte*
2539 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2540 X25519, X448, Ed25519 and Ed448.
2544 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2549 * Deprecated the public definition of `ERR_STATE` as well as the function
2550 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2555 * Added ERR functionality to give callers access to the stored function
2556 names that have replaced the older function code based functions.
2558 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2559 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2560 ERR_peek_error_all() and ERR_peek_last_error_all().
2562 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2563 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2564 ERR_func_error_string().
2568 * Extended testing to be verbose for failing tests only. The make variables
2569 VERBOSE_FAILURE or VF can be used to enable this:
2571 $ make VF=1 test # Unix
2572 $ mms /macro=(VF=1) test ! OpenVMS
2573 $ nmake VF=1 test # Windows
2577 * Added the `-copy_extensions` option to the `x509` command for use with
2578 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2579 all extensions in the request are copied to the certificate or vice versa.
2581 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2583 * Added the `-copy_extensions` option to the `req` command for use with
2584 `-x509`. When given with the `copy` or `copyall` argument,
2585 all extensions in the certification request are copied to the certificate.
2589 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2590 they generate are by default RFC 5280 compliant in the following sense:
2591 There is a subjectKeyIdentifier extension with a hash value of the public key
2592 and for not self-signed certs there is an authorityKeyIdentifier extension
2593 with a keyIdentifier field or issuer information identifying the signing key.
2594 This is done unless some configuration overrides the new default behavior,
2595 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2599 * Added several checks to `X509_verify_cert()` according to requirements in
2600 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2601 (which may be done by using the CLI option `-x509_strict`):
2602 * The basicConstraints of CA certificates must be marked critical.
2603 * CA certificates must explicitly include the keyUsage extension.
2604 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2605 * The issuer name of any certificate must not be empty.
2606 * The subject name of CA certs, certs with keyUsage crlSign,
2607 and certs without subjectAlternativeName must not be empty.
2608 * If a subjectAlternativeName extension is given it must not be empty.
2609 * The signatureAlgorithm field and the cert signature must be consistent.
2610 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2611 must not be marked critical.
2612 * The authorityKeyIdentifier must be given for X.509v3 certs
2613 unless they are self-signed.
2614 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2618 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2619 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2623 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2624 used even when parsing explicit parameters, when loading a encoded key
2625 or calling `EC_GROUP_new_from_ecpkparameters()`/
2626 `EC_GROUP_new_from_ecparameters()`.
2627 This prevents bypass of security hardening and performance gains,
2628 especially for curves with specialized EC_METHODs.
2629 By default, if a key encoded with explicit parameters is loaded and later
2630 encoded, the output is still encoded with explicit parameters, even if
2631 internally a "named" EC_GROUP is used for computation.
2635 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2636 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2637 NULL. After this change, only the cofactor parameter can be NULL. It also
2638 does some minimal sanity checks on the passed order.
2643 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2644 An attack is simple, if the first CMS_recipientInfo is valid but the
2645 second CMS_recipientInfo is chosen ciphertext. If the second
2646 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2647 encryption key will be replaced by garbage, and the message cannot be
2648 decoded, but if the RSA decryption fails, the correct encryption key is
2649 used and the recipient will not notice the attack.
2650 As a work around for this potential attack the length of the decrypted
2651 key must be equal to the cipher default key length, in case the
2652 certificate is not given and all recipientInfo are tried out.
2653 The old behaviour can be re-enabled in the CMS code by setting the
2654 CMS_DEBUG_DECRYPT flag.
2658 * Early start up entropy quality from the DEVRANDOM seed source has been
2659 improved for older Linux systems. The RAND subsystem will wait for
2660 /dev/random to be producing output before seeding from /dev/urandom.
2661 The seeded state is stored for future library initialisations using
2662 a system global shared memory segment. The shared memory identifier
2663 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2664 the desired value. The default identifier is 114.
2668 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2669 when primes for RSA keys are computed.
2670 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2671 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2672 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2673 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2674 This avoids possible fingerprinting of newly generated RSA modules.
2678 * Correct the extended master secret constant on EBCDIC systems. Without this
2679 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2680 negotiate EMS will fail. Unfortunately this also means that TLS connections
2681 between EBCDIC systems with this fix, and EBCDIC systems without this
2682 fix will fail if they negotiate EMS.
2686 * Changed the library initialisation so that the config file is now loaded
2687 by default. This was already the case for libssl. It now occurs for both
2688 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2689 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2693 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2694 where the former acts as a replacement for `ERR_put_error()`, and the
2695 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2696 `ERR_raise_data()` adds more flexibility by taking a format string and
2697 an arbitrary number of arguments following it, to be processed with
2702 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2703 to check if a named provider is loaded and available. When called, it
2704 will also activate all fallback providers if such are still present.
2708 * Enforce a minimum DH modulus size of 512 bits.
2712 * Changed DH parameters to generate the order q subgroup instead of 2q.
2713 Previously generated DH parameters are still accepted by DH_check
2714 but DH_generate_key works around that by clearing bit 0 of the
2715 private key for those. This avoids leaking bit 0 of the private key.
2719 * Significantly reduce secure memory usage by the randomness pools.
2723 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2728 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2729 algorithms. An implementation of a key exchange algorithm can be obtained
2730 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2731 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2732 the older EVP_PKEY_derive_init() function. See the man pages for the new
2733 functions for further details.
2737 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2741 * Removed the function names from error messages and deprecated the
2746 * Removed NextStep support and the macro OPENSSL_UNISTD
2750 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2751 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2752 Also removed "export var as function" capability; we do not export
2753 variables, only functions.
2757 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2758 an error and 1 indicating success. In previous versions of OpenSSL this
2759 was a void type. If a key was set longer than the maximum possible this
2764 * Support SM2 signing and verification schemes with X509 certificate.
2768 * Use SHA256 as the default digest for TS query in the `ts` app.
2772 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2776 * Default cipher lists/suites are now available via a function, the
2777 #defines are deprecated.
2781 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2782 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2783 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2787 * Join the directories crypto/x509 and crypto/x509v3
2791 * Added command 'openssl kdf' that uses the EVP_KDF API.
2795 * Added command 'openssl mac' that uses the EVP_MAC API.
2799 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2800 as default directories. Also added the command 'openssl info'
2801 for scripting purposes.
2805 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2810 * Add prediction resistance to the DRBG reseeding process.
2814 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2815 mandated by IEEE Std 1619-2018.
2819 * Added newline escaping functionality to a filename when using openssl dgst.
2820 This output format is to replicate the output format found in the `*sum`
2821 checksum programs. This aims to preserve backward compatibility.
2823 *Matt Eaton, Richard Levitte, and Paul Dale*
2825 * Removed the heartbeat message in DTLS feature, as it has very
2826 little usage and doesn't seem to fulfill a valuable purpose.
2827 The configuration option is now deprecated.
2831 * Changed the output of 'openssl {digestname} < file' to display the
2832 digest name in its output.
2836 * Added a new generic trace API which provides support for enabling
2837 instrumentation through trace output.
2839 *Richard Levitte & Matthias St. Pierre*
2841 * Added build tests for C++. These are generated files that only do one
2842 thing, to include one public OpenSSL head file each. This tests that
2843 the public header files can be usefully included in a C++ application.
2845 This test isn't enabled by default. It can be enabled with the option
2846 'enable-buildtest-c++'.
2850 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2854 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2858 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2862 * Added KMAC to EVP_MAC.
2866 * Added property based algorithm implementation selection framework to
2871 * Added SCA hardening for modular field inversion in EC_GROUP through
2872 a new dedicated field_inv() pointer in EC_METHOD.
2873 This also addresses a leakage affecting conversions from projective
2874 to affine coordinates.
2876 *Billy Bob Brumley, Nicola Tuveri*
2878 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2879 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2880 those algorithms that were already supported through the EVP_PKEY API
2881 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2882 and scrypt are now wrappers that call EVP_KDF.
2886 * Build devcrypto engine as a dynamic engine.
2888 *Eneas U de Queiroz*
2890 * Add keyed BLAKE2 to EVP_MAC.
2894 * Fix a bug in the computation of the endpoint-pair shared secret used
2895 by DTLS over SCTP. This breaks interoperability with older versions
2896 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2897 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2898 interoperability with such broken implementations. However, enabling
2899 this switch breaks interoperability with correct implementations.
2901 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2902 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2906 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2910 * Changed the license to the Apache License v2.0.
2914 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2916 - Major releases (indicated by incrementing the MAJOR release number)
2917 may introduce incompatible API/ABI changes.
2918 - Minor releases (indicated by incrementing the MINOR release number)
2919 may introduce new features but retain API/ABI compatibility.
2920 - Patch releases (indicated by incrementing the PATCH number)
2921 are intended for bug fixes and other improvements of existing
2922 features only (like improving performance or adding documentation)
2923 and retain API/ABI compatibility.
2927 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2931 * Remove the 'dist' target and add a tarball building script. The
2932 'dist' target has fallen out of use, and it shouldn't be
2933 necessary to configure just to create a source distribution.
2937 * Recreate the OS390-Unix config target. It no longer relies on a
2938 special script like it did for OpenSSL pre-1.1.0.
2942 * Instead of having the source directories listed in Configure, add
2943 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2948 * Add GMAC to EVP_MAC.
2952 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2956 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2957 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2958 to facilitate the continued use of MACs through raw private keys in
2959 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2963 * Deprecate ECDH_KDF_X9_62().
2967 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2968 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2969 are retained for backwards compatibility.
2973 * AES-XTS mode now enforces that its two keys are different to mitigate
2974 the attacked described in "Efficient Instantiations of Tweakable
2975 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2976 Details of this attack can be obtained from:
2977 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2981 * Rename the object files, i.e. give them other names than in previous
2982 versions. Their names now include the name of the final product, as
2983 well as its type mnemonic (bin, lib, shlib).
2987 * Added new option for 'openssl list', '-objects', which will display the
2988 list of built in objects, i.e. OIDs with names.
2992 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2993 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2998 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2999 improves application performance by removing data copies and providing
3000 applications with zero-copy system calls such as sendfile and splice.
3004 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
3008 * `PKCS12_parse` now maintains the order of the parsed certificates
3009 when outputting them via `*ca` (rather than reversing it).
3013 * Deprecated pthread fork support methods.
3017 * Added support for FFDHE key exchange in TLS 1.3.
3021 * Added a new concept for OpenSSL plugability: providers. This
3022 functionality is designed to replace the ENGINE API and ENGINE
3023 implementations, and to be much more dynamic, allowing provider
3024 authors to introduce new algorithms among other things, as long as
3025 there's an API that supports the algorithm type.
3027 With this concept comes a new core API for interaction between
3028 libcrypto and provider implementations. Public libcrypto functions
3029 that want to use providers do so through this core API.
3031 The main documentation for this core API is found in
3032 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
3033 refer to other manuals describing the API specific for supported
3034 algorithm types (also called operations).
3041 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
3043 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
3045 * Avoid loading of a dynamic engine twice.
3049 * Prioritise DANE TLSA issuer certs over peer certs
3053 * Fixed random API for MacOS prior to 10.12
3055 These MacOS versions don't support the CommonCrypto APIs
3059 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
3061 * Fixed an SM2 Decryption Buffer Overflow.
3063 In order to decrypt SM2 encrypted data an application is expected to
3064 call the API function EVP_PKEY_decrypt(). Typically an application will
3065 call this function twice. The first time, on entry, the "out" parameter
3066 can be NULL and, on exit, the "outlen" parameter is populated with the
3067 buffer size required to hold the decrypted plaintext. The application
3068 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
3069 again, but this time passing a non-NULL value for the "out" parameter.
3071 A bug in the implementation of the SM2 decryption code means that the
3072 calculation of the buffer size required to hold the plaintext returned
3073 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
3074 size required by the second call. This can lead to a buffer overflow
3075 when EVP_PKEY_decrypt() is called by the application a second time with
3076 a buffer that is too small.
3078 A malicious attacker who is able present SM2 content for decryption to
3079 an application could cause attacker chosen data to overflow the buffer
3080 by up to a maximum of 62 bytes altering the contents of other data held
3081 after the buffer, possibly changing application behaviour or causing
3082 the application to crash. The location of the buffer is application
3083 dependent but is typically heap allocated.
3088 * Fixed various read buffer overruns processing ASN.1 strings
3090 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3091 structure which contains a buffer holding the string data and a field
3092 holding the buffer length. This contrasts with normal C strings which
3093 are represented as a buffer for the string data which is terminated
3094 with a NUL (0) byte.
3096 Although not a strict requirement, ASN.1 strings that are parsed using
3097 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3098 well as any string whose value has been set with the ASN1_STRING_set()
3099 function will additionally NUL terminate the byte array in the
3100 ASN1_STRING structure.
3102 However, it is possible for applications to directly construct valid
3103 ASN1_STRING structures which do not NUL terminate the byte array by
3104 directly setting the "data" and "length" fields in the ASN1_STRING
3105 array. This can also happen by using the ASN1_STRING_set0() function.
3107 Numerous OpenSSL functions that print ASN.1 data have been found to
3108 assume that the ASN1_STRING byte array will be NUL terminated, even
3109 though this is not guaranteed for strings that have been directly
3110 constructed. Where an application requests an ASN.1 structure to be
3111 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3112 been directly constructed by the application without NUL terminating
3113 the "data" field, then a read buffer overrun can occur.
3115 The same thing can also occur during name constraints processing
3116 of certificates (for example if a certificate has been directly
3117 constructed by the application instead of loading it via the OpenSSL
3118 parsing functions, and the certificate contains non NUL terminated
3119 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3120 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3122 If a malicious actor can cause an application to directly construct an
3123 ASN1_STRING and then process it through one of the affected OpenSSL
3124 functions then this issue could be hit. This might result in a crash
3125 (causing a Denial of Service attack). It could also result in the
3126 disclosure of private memory contents (such as private keys, or
3127 sensitive plaintext).
3132 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3134 * Fixed a problem with verifying a certificate chain when using the
3135 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3136 the certificates present in a certificate chain. It is not set by default.
3138 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3139 the chain that have explicitly encoded elliptic curve parameters was added
3140 as an additional strict check.
3142 An error in the implementation of this check meant that the result of a
3143 previous check to confirm that certificates in the chain are valid CA
3144 certificates was overwritten. This effectively bypasses the check
3145 that non-CA certificates must not be able to issue other certificates.
3147 If a "purpose" has been configured then there is a subsequent opportunity
3148 for checks that the certificate is a valid CA. All of the named "purpose"
3149 values implemented in libcrypto perform this check. Therefore, where
3150 a purpose is set the certificate chain will still be rejected even when the
3151 strict flag has been used. A purpose is set by default in libssl client and
3152 server certificate verification routines, but it can be overridden or
3153 removed by an application.
3155 In order to be affected, an application must explicitly set the
3156 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3157 for the certificate verification or, in the case of TLS client or server
3158 applications, override the default purpose.
3163 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3164 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3165 renegotiation ClientHello omits the signature_algorithms extension (where it
3166 was present in the initial ClientHello), but includes a
3167 signature_algorithms_cert extension then a NULL pointer dereference will
3168 result, leading to a crash and a denial of service attack.
3170 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3171 (which is the default configuration). OpenSSL TLS clients are not impacted by
3175 *Peter Kästle and Samuel Sapalski*
3177 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3179 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3180 create a unique hash value based on the issuer and serial number data
3181 contained within an X509 certificate. However, it was failing to correctly
3182 handle any errors that may occur while parsing the issuer field (which might
3183 occur if the issuer field is maliciously constructed). This may subsequently
3184 result in a NULL pointer deref and a crash leading to a potential denial of
3190 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3191 padding mode to correctly check for rollback attacks. This is considered a
3192 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3197 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3198 functions. Previously they could overflow the output length argument in some
3199 cases where the input length is close to the maximum permissible length for
3200 an integer on the platform. In such cases the return value from the function
3201 call would be 1 (indicating success), but the output length value would be
3202 negative. This could cause applications to behave incorrectly or crash.
3207 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3208 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3209 could be exploited in a side channel attack to recover the password. Since
3210 the attack is local host only this is outside of the current OpenSSL
3211 threat model and therefore no CVE is assigned.
3213 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3218 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3220 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3221 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3222 If an attacker can control both items being compared then this could lead
3223 to a possible denial of service attack. OpenSSL itself uses the
3224 GENERAL_NAME_cmp function for two purposes:
3225 1) Comparing CRL distribution point names between an available CRL and a
3226 CRL distribution point embedded in an X509 certificate
3227 2) When verifying that a timestamp response token signer matches the
3228 timestamp authority name (exposed via the API functions
3229 TS_RESP_verify_response and TS_RESP_verify_token)
3234 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3236 * Certificates with explicit curve parameters are now disallowed in
3237 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3241 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3242 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3243 conversely, silently ignore DTLS protocol version bounds when configuring
3244 TLS-based contexts. The commands can be repeated to set bounds of both
3245 types. The same applies with the corresponding "min_protocol" and
3246 "max_protocol" command-line switches, in case some application uses both TLS
3249 SSL_CTX instances that are created for a fixed protocol version (e.g.
3250 TLSv1_server_method()) also silently ignore version bounds. Previously
3251 attempts to apply bounds to these protocol versions would result in an
3252 error. Now only the "version-flexible" SSL_CTX instances are subject to
3253 limits in configuration files in command-line options.
3257 * Handshake now fails if Extended Master Secret extension is dropped
3262 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3264 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3266 * Fixed segmentation fault in SSL_check_chain()
3267 Server or client applications that call the SSL_check_chain() function
3268 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3269 dereference as a result of incorrect handling of the
3270 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3271 or unrecognised signature algorithm is received from the peer. This could
3272 be exploited by a malicious peer in a Denial of Service attack.
3277 * Added AES consttime code for no-asm configurations
3278 an optional constant time support for AES was added
3279 when building openssl for no-asm.
3280 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3281 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3282 At this time this feature is by default disabled.
3283 It will be enabled by default in 3.0.
3287 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3289 * Revert the change of EOF detection while reading in libssl to avoid
3290 regressions in applications depending on the current way of reporting
3291 the EOF. As the existing method is not fully accurate the change to
3292 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3293 branch and will be present in the 3.0 release.
3297 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3298 when primes for RSA keys are computed.
3299 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3300 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3301 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3302 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3303 This avoids possible fingerprinting of newly generated RSA modules.
3307 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3309 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3310 while reading in libssl then we would report an error back to the
3311 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3312 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3313 therefore give a hint as to what went wrong.
3317 * Check that ed25519 and ed448 are allowed by the security level. Previously
3318 signature algorithms not using an MD were not being checked that they were
3319 allowed by the security level.
3323 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3324 was not quite right. The behaviour was not consistent between resumption
3325 and normal handshakes, and also not quite consistent with historical
3326 behaviour. The behaviour in various scenarios has been clarified and
3327 it has been updated to make it match historical behaviour as closely as
3332 * *[VMS only]* The header files that the VMS compilers include automatically,
3333 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3334 that the C++ compiler doesn't understand. This is a shortcoming in the
3335 compiler, but can be worked around with `__cplusplus` guards.
3337 C++ applications that use OpenSSL libraries must be compiled using the
3338 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3339 functions. Otherwise, only functions with symbols of less than 31
3340 characters can be used, as the linker will not be able to successfully
3341 resolve symbols with longer names.
3345 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3346 The presence of this system service is determined at run-time.
3350 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3355 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3357 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3358 number generator (RNG). This was intended to include protection in the
3359 event of a fork() system call in order to ensure that the parent and child
3360 processes did not share the same RNG state. However, this protection was not
3361 being used in the default case.
3363 A partial mitigation for this issue is that the output from a high
3364 precision timer is mixed into the RNG state so the likelihood of a parent
3365 and child process sharing state is significantly reduced.
3367 If an application already calls OPENSSL_init_crypto() explicitly using
3368 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3371 *Matthias St. Pierre*
3373 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3374 used even when parsing explicit parameters, when loading a encoded key
3375 or calling `EC_GROUP_new_from_ecpkparameters()`/
3376 `EC_GROUP_new_from_ecparameters()`.
3377 This prevents bypass of security hardening and performance gains,
3378 especially for curves with specialized EC_METHODs.
3379 By default, if a key encoded with explicit parameters is loaded and later
3380 encoded, the output is still encoded with explicit parameters, even if
3381 internally a "named" EC_GROUP is used for computation.
3385 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3386 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3387 NULL. After this change, only the cofactor parameter can be NULL. It also
3388 does some minimal sanity checks on the passed order.
3393 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3394 An attack is simple, if the first CMS_recipientInfo is valid but the
3395 second CMS_recipientInfo is chosen ciphertext. If the second
3396 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3397 encryption key will be replaced by garbage, and the message cannot be
3398 decoded, but if the RSA decryption fails, the correct encryption key is
3399 used and the recipient will not notice the attack.
3400 As a work around for this potential attack the length of the decrypted
3401 key must be equal to the cipher default key length, in case the
3402 certificate is not given and all recipientInfo are tried out.
3403 The old behaviour can be re-enabled in the CMS code by setting the
3404 CMS_DEBUG_DECRYPT flag.
3409 * Early start up entropy quality from the DEVRANDOM seed source has been
3410 improved for older Linux systems. The RAND subsystem will wait for
3411 /dev/random to be producing output before seeding from /dev/urandom.
3412 The seeded state is stored for future library initialisations using
3413 a system global shared memory segment. The shared memory identifier
3414 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3415 the desired value. The default identifier is 114.
3419 * Correct the extended master secret constant on EBCDIC systems. Without this
3420 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3421 negotiate EMS will fail. Unfortunately this also means that TLS connections
3422 between EBCDIC systems with this fix, and EBCDIC systems without this
3423 fix will fail if they negotiate EMS.
3427 * Use Windows installation paths in the mingw builds
3429 Mingw isn't a POSIX environment per se, which means that Windows
3430 paths should be used for installation.
3435 * Changed DH_check to accept parameters with order q and 2q subgroups.
3436 With order 2q subgroups the bit 0 of the private key is not secret
3437 but DH_generate_key works around that by clearing bit 0 of the
3438 private key for those. This avoids leaking bit 0 of the private key.
3442 * Significantly reduce secure memory usage by the randomness pools.
3446 * Revert the DEVRANDOM_WAIT feature for Linux systems
3448 The DEVRANDOM_WAIT feature added a select() call to wait for the
3449 /dev/random device to become readable before reading from the
3450 /dev/urandom device.
3452 It turned out that this change had negative side effects on
3453 performance which were not acceptable. After some discussion it
3454 was decided to revert this feature and leave it up to the OS
3455 resp. the platform maintainer to ensure a proper initialization
3456 during early boot time.
3458 *Matthias St. Pierre*
3460 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3462 * Add build tests for C++. These are generated files that only do one
3463 thing, to include one public OpenSSL head file each. This tests that
3464 the public header files can be usefully included in a C++ application.
3466 This test isn't enabled by default. It can be enabled with the option
3467 'enable-buildtest-c++'.
3471 * Enable SHA3 pre-hashing for ECDSA and DSA.
3475 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3476 This changes the size when using the `genpkey` command when no size is given.
3477 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3478 generation commands to use 2048 bits by default.
3482 * Reorganize the manual pages to consistently have RETURN VALUES,
3483 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3484 util/fix-doc-nits accordingly.
3486 *Paul Yang, Joshua Lock*
3488 * Add the missing accessor EVP_PKEY_get0_engine()
3492 * Have commands like `s_client` and `s_server` output the signature scheme
3493 along with other cipher suite parameters when debugging.
3497 * Make OPENSSL_config() error agnostic again.
3501 * Do the error handling in RSA decryption constant time.
3505 * Prevent over long nonces in ChaCha20-Poly1305.
3507 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3508 for every encryption operation. RFC 7539 specifies that the nonce value
3509 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3510 and front pads the nonce with 0 bytes if it is less than 12
3511 bytes. However it also incorrectly allows a nonce to be set of up to 16
3512 bytes. In this case only the last 12 bytes are significant and any
3513 additional leading bytes are ignored.
3515 It is a requirement of using this cipher that nonce values are
3516 unique. Messages encrypted using a reused nonce value are susceptible to
3517 serious confidentiality and integrity attacks. If an application changes
3518 the default nonce length to be longer than 12 bytes and then makes a
3519 change to the leading bytes of the nonce expecting the new value to be a
3520 new unique nonce then such an application could inadvertently encrypt
3521 messages with a reused nonce.
3523 Additionally the ignored bytes in a long nonce are not covered by the
3524 integrity guarantee of this cipher. Any application that relies on the
3525 integrity of these ignored leading bytes of a long nonce may be further
3526 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3527 is safe because no such use sets such a long nonce value. However user
3528 applications that use this cipher directly and set a non-default nonce
3529 length to be longer than 12 bytes may be vulnerable.
3531 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3537 * Add DEVRANDOM_WAIT feature for Linux systems
3539 On older Linux systems where the getrandom() system call is not available,
3540 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3541 Contrary to getrandom(), the /dev/urandom device will not block during
3542 early boot when the kernel CSPRNG has not been seeded yet.
3544 To mitigate this known weakness, use select() to wait for /dev/random to
3545 become readable before reading from /dev/urandom.
3547 * Ensure that SM2 only uses SM3 as digest algorithm
3551 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3553 * Change the info callback signals for the start and end of a post-handshake
3554 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3555 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3556 confused by this and assume that a TLSv1.2 renegotiation has started. This
3557 can break KeyUpdate handling. Instead we no longer signal the start and end
3558 of a post handshake message exchange (although the messages themselves are
3559 still signalled). This could break some applications that were expecting
3560 the old signals. However without this KeyUpdate is not usable for many
3565 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3567 * Timing vulnerability in DSA signature generation
3569 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3570 timing side channel attack. An attacker could use variations in the signing
3571 algorithm to recover the private key.
3573 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3578 * Timing vulnerability in ECDSA signature generation
3580 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3581 timing side channel attack. An attacker could use variations in the signing
3582 algorithm to recover the private key.
3584 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3589 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3590 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3591 of two gigabytes and the error handling improved.
3593 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3594 categorized as a normal bug, not a security issue, because the DRBG reseeds
3595 automatically and is fully functional even without additional randomness
3596 provided by the application.
3598 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3600 * Add a new ClientHello callback. Provides a callback interface that gives
3601 the application the ability to adjust the nascent SSL object at the
3602 earliest stage of ClientHello processing, immediately after extensions have
3603 been collected but before they have been processed. In particular, this
3604 callback can adjust the supported TLS versions in response to the contents
3609 * Add SM2 base algorithm support.
3613 * s390x assembly pack: add (improved) hardware-support for the following
3614 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3615 aes-cfb/cfb8, aes-ecb.
3619 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3620 parameter is no longer accepted, as it leads to a corrupt table. NULL
3621 pem_str is reserved for alias entries only.
3625 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3626 step for prime curves. The new implementation is based on formulae from
3627 differential addition-and-doubling in homogeneous projective coordinates
3628 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3629 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3630 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3631 to work in projective coordinates.
3633 *Billy Bob Brumley, Nicola Tuveri*
3635 * Change generating and checking of primes so that the error rate of not
3636 being prime depends on the intended use based on the size of the input.
3637 For larger primes this will result in more rounds of Miller-Rabin.
3638 The maximal error rate for primes with more than 1080 bits is lowered
3641 *Kurt Roeckx, Annie Yousar*
3643 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3647 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3648 moving between systems, and to avoid confusion when a Windows build is
3649 done with mingw vs with MSVC. For POSIX installs, there's still a
3650 symlink or copy named 'tsget' to avoid that confusion as well.
3654 * Revert blinding in ECDSA sign and instead make problematic addition
3655 length-invariant. Switch even to fixed-length Montgomery multiplication.
3659 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3660 step for binary curves. The new implementation is based on formulae from
3661 differential addition-and-doubling in mixed Lopez-Dahab projective
3662 coordinates, modified to independently blind the operands.
3664 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3666 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3667 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3668 EC_METHODs to implement their own specialized "ladder step", to take
3669 advantage of more favorable coordinate systems or more efficient
3670 differential addition-and-doubling algorithms.
3672 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3674 * Modified the random device based seed sources to keep the relevant
3675 file descriptors open rather than reopening them on each access.
3676 This allows such sources to operate in a chroot() jail without
3677 the associated device nodes being available. This behaviour can be
3678 controlled using RAND_keep_random_devices_open().
3682 * Numerous side-channel attack mitigations have been applied. This may have
3683 performance impacts for some algorithms for the benefit of improved
3684 security. Specific changes are noted in this change log by their respective
3689 * AIX shared library support overhaul. Switch to AIX "natural" way of
3690 handling shared libraries, which means collecting shared objects of
3691 different versions and bitnesses in one common archive. This allows to
3692 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3693 doesn't affect the way 3rd party applications are linked, only how
3694 multi-version installation is managed.
3698 * Make ec_group_do_inverse_ord() more robust and available to other
3699 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3700 mitigations are applied to the fallback BN_mod_inverse().
3701 When using this function rather than BN_mod_inverse() directly, new
3702 EC cryptosystem implementations are then safer-by-default.
3706 * Add coordinate blinding for EC_POINT and implement projective
3707 coordinate blinding for generic prime curves as a countermeasure to
3708 chosen point SCA attacks.
3710 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3712 * Add blinding to ECDSA and DSA signatures to protect against side channel
3713 attacks discovered by Keegan Ryan (NCC Group).
3717 * Enforce checking in the `pkeyutl` command to ensure that the input
3718 length does not exceed the maximum supported digest length when performing
3719 a sign, verify or verifyrecover operation.
3723 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3724 I/O in combination with something like select() or poll() will hang. This
3725 can be turned off again using SSL_CTX_clear_mode().
3726 Many applications do not properly handle non-application data records, and
3727 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3728 around the problems in those applications, but can also break some.
3729 It's recommended to read the manpages about SSL_read(), SSL_write(),
3730 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3731 SSL_CTX_set_read_ahead() again.
3735 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3736 now allow empty (zero character) pass phrases.
3740 * Apply blinding to binary field modular inversion and remove patent
3741 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3745 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3746 binary and prime elliptic curves.
3750 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3751 constant time fixed point multiplication.
3755 * Revise elliptic curve scalar multiplication with timing attack
3756 defenses: ec_wNAF_mul redirects to a constant time implementation
3757 when computing fixed point and variable point multiplication (which
3758 in OpenSSL are mostly used with secret scalars in keygen, sign,
3759 ECDH derive operations).
3760 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3763 * Updated CONTRIBUTING
3767 * Updated DRBG / RAND to request nonce and additional low entropy
3768 randomness from the system.
3770 *Matthias St. Pierre*
3772 * Updated 'openssl rehash' to use OpenSSL consistent default.
3776 * Moved the load of the ssl_conf module to libcrypto, which helps
3777 loading engines that libssl uses before libssl is initialised.
3781 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3785 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3787 *Ingo Schwarze, Rich Salz*
3789 * Added output of accepting IP address and port for 'openssl s_server'
3793 * Added a new API for TLSv1.3 ciphersuites:
3794 SSL_CTX_set_ciphersuites()
3795 SSL_set_ciphersuites()
3799 * Memory allocation failures consistently add an error to the error
3804 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3805 in libcrypto when run as setuid/setgid.
3809 * Load any config file by default when libssl is used.
3813 * Added new public header file <openssl/rand_drbg.h> and documentation
3814 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3816 *Matthias St. Pierre*
3818 * QNX support removed (cannot find contributors to get their approval
3819 for the license change).
3823 * TLSv1.3 replay protection for early data has been implemented. See the
3824 SSL_read_early_data() man page for further details.
3828 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3829 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3830 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3831 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3832 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3833 configuration has been separated out. See the ciphers man page or the
3834 SSL_CTX_set_ciphersuites() man page for more information.
3838 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3839 in responder mode now supports the new "-multi" option, which
3840 spawns the specified number of child processes to handle OCSP
3841 requests. The "-timeout" option now also limits the OCSP
3842 responder's patience to wait to receive the full client request
3843 on a newly accepted connection. Child processes are respawned
3844 as needed, and the CA index file is automatically reloaded
3845 when changed. This makes it possible to run the "ocsp" responder
3846 as a long-running service, making the OpenSSL CA somewhat more
3847 feature-complete. In this mode, most diagnostic messages logged
3848 after entering the event loop are logged via syslog(3) rather than
3853 * Added support for X448 and Ed448. Heavily based on original work by
3858 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3859 objects loaded. This adds the functions OSSL_STORE_expect() and
3860 OSSL_STORE_find() as well as needed tools to construct searches and
3861 get the search data out of them.
3865 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3866 version of OpenSSL should review their configuration settings to ensure
3867 that they are still appropriate for TLSv1.3. For further information see:
3868 <https://wiki.openssl.org/index.php/TLS1.3>
3872 * Grand redesign of the OpenSSL random generator
3874 The default RAND method now utilizes an AES-CTR DRBG according to
3875 NIST standard SP 800-90Ar1. The new random generator is essentially
3876 a port of the default random generator from the OpenSSL FIPS 2.0
3877 object module. It is a hybrid deterministic random bit generator
3878 using an AES-CTR bit stream and which seeds and reseeds itself
3879 automatically using trusted system entropy sources.
3881 Some of its new features are:
3882 - Support for multiple DRBG instances with seed chaining.
3883 - The default RAND method makes use of a DRBG.
3884 - There is a public and private DRBG instance.
3885 - The DRBG instances are fork-safe.
3886 - Keep all global DRBG instances on the secure heap if it is enabled.
3887 - The public and private DRBG instance are per thread for lock free
3890 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3892 * Changed Configure so it only says what it does and doesn't dump
3893 so much data. Instead, ./configdata.pm should be used as a script
3894 to display all sorts of configuration data.
3898 * Added processing of "make variables" to Configure.
3902 * Added SHA512/224 and SHA512/256 algorithm support.
3906 * The last traces of Netware support, first removed in 1.1.0, have
3911 * Get rid of Makefile.shared, and in the process, make the processing
3912 of certain files (rc.obj, or the .def/.map/.opt files produced from
3913 the ordinal files) more visible and hopefully easier to trace and
3914 debug (or make silent).
3918 * Make it possible to have environment variable assignments as
3919 arguments to config / Configure.
3923 * Add multi-prime RSA (RFC 8017) support.
3927 * Add SM3 implemented according to GB/T 32905-2016
3928 *Jack Lloyd <jack.lloyd@ribose.com>,*
3929 *Ronald Tse <ronald.tse@ribose.com>,*
3930 *Erick Borsboom <erick.borsboom@ribose.com>*
3932 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3933 as documented in RFC6066.
3934 Based on a patch from Tomasz Moń
3936 *Filipe Raimundo da Silva*
3938 * Add SM4 implemented according to GB/T 32907-2016.
3939 *Jack Lloyd <jack.lloyd@ribose.com>,*
3940 *Ronald Tse <ronald.tse@ribose.com>,*
3941 *Erick Borsboom <erick.borsboom@ribose.com>*
3943 * Reimplement -newreq-nodes and ERR_error_string_n; the
3944 original author does not agree with the license change.
3948 * Add ARIA AEAD TLS support.
3952 * Some macro definitions to support VS6 have been removed. Visual
3953 Studio 6 has not worked since 1.1.0
3957 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3958 without clearing the errors.
3962 * Add "atfork" functions. If building on a system that without
3963 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3964 requirements. The RAND facility now uses/requires this.
3972 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3973 not possible to disable entirely. However, it's still possible to
3974 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3977 To disable, configure with 'no-ui-console'. 'no-ui' is still
3978 possible to use as an alias. Check at compile time with the
3979 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3980 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3984 * Add a STORE module, which implements a uniform and URI based reader of
3985 stores that can contain keys, certificates, CRLs and numerous other
3986 objects. The main API is loosely based on a few stdio functions,
3987 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3988 OSSL_STORE_error and OSSL_STORE_close.
3989 The implementation uses backends called "loaders" to implement arbitrary
3990 URI schemes. There is one built in "loader" for the 'file' scheme.
3994 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3995 then adjusted to work on FreeBSD 8.4 as well.
3996 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3997 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
4001 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
4002 util/mkerr.pl, which is adapted to allow those prefixes, leading to
4003 error code calls like this:
4005 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
4007 With this change, we claim the namespaces OSSL and OPENSSL in a manner
4008 that can be encoded in C. For the foreseeable future, this will only
4011 *Richard Levitte and Tim Hudson*
4013 * Removed BSD cryptodev engine.
4017 * Add a build target 'build_all_generated', to build all generated files
4018 and only that. This can be used to prepare everything that requires
4019 things like perl for a system that lacks perl and then move everything
4020 to that system and do the rest of the build there.
4024 * In the UI interface, make it possible to duplicate the user data. This
4025 can be used by engines that need to retain the data for a longer time
4026 than just the call where this user data is passed.
4030 * Ignore the '-named_curve auto' value for compatibility of applications
4033 *Tomáš Mráz <tmraz@fedoraproject.org>*
4035 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
4036 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
4037 alerts across multiple records (some of which could be empty). In practice
4038 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
4039 prohibits this altogether and other libraries (BoringSSL, NSS) do not
4040 support this at all. Supporting it adds significant complexity to the
4041 record layer, and its removal is unlikely to cause interoperability
4046 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
4047 with Z. These are meant to replace LONG and ZLONG and to be size safe.
4048 The use of LONG and ZLONG is discouraged and scheduled for deprecation
4053 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
4054 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
4056 *Richard Levitte, Andy Polyakov*
4058 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
4063 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4064 platform rather than 'mingw'.
4068 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
4069 success if they are asked to add an object which already exists
4070 in the store. This change cascades to other functions which load
4071 certificates and CRLs.
4075 * x86_64 assembly pack: annotate code with DWARF CFI directives to
4076 facilitate stack unwinding even from assembly subroutines.
4080 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
4081 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
4085 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4086 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4087 which is the minimum version we support.
4091 * Certificate time validation (X509_cmp_time) enforces stricter
4092 compliance with RFC 5280. Fractional seconds and timezone offsets
4093 are no longer allowed.
4097 * Add support for ARIA
4101 * s_client will now send the Server Name Indication (SNI) extension by
4102 default unless the new "-noservername" option is used. The server name is
4103 based on the host provided to the "-connect" option unless overridden by
4104 using "-servername".
4108 * Add support for SipHash
4112 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4113 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4114 prevent issues where no progress is being made and the peer continually
4115 sends unrecognised record types, using up resources processing them.
4119 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4120 using the algorithm defined in
4121 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4125 * Heartbeat support has been removed; the ABI is changed for now.
4127 *Richard Levitte, Rich Salz*
4129 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4133 * The RSA "null" method, which was partially supported to avoid patent
4134 issues, has been replaced to always returns NULL.
4141 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4143 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4144 used even when parsing explicit parameters, when loading a encoded key
4145 or calling `EC_GROUP_new_from_ecpkparameters()`/
4146 `EC_GROUP_new_from_ecparameters()`.
4147 This prevents bypass of security hardening and performance gains,
4148 especially for curves with specialized EC_METHODs.
4149 By default, if a key encoded with explicit parameters is loaded and later
4150 encoded, the output is still encoded with explicit parameters, even if
4151 internally a "named" EC_GROUP is used for computation.
4155 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4156 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4157 NULL. After this change, only the cofactor parameter can be NULL. It also
4158 does some minimal sanity checks on the passed order.
4163 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4164 An attack is simple, if the first CMS_recipientInfo is valid but the
4165 second CMS_recipientInfo is chosen ciphertext. If the second
4166 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4167 encryption key will be replaced by garbage, and the message cannot be
4168 decoded, but if the RSA decryption fails, the correct encryption key is
4169 used and the recipient will not notice the attack.
4170 As a work around for this potential attack the length of the decrypted
4171 key must be equal to the cipher default key length, in case the
4172 certificate is not given and all recipientInfo are tried out.
4173 The old behaviour can be re-enabled in the CMS code by setting the
4174 CMS_DEBUG_DECRYPT flag.
4179 * Use Windows installation paths in the mingw builds
4181 Mingw isn't a POSIX environment per se, which means that Windows
4182 paths should be used for installation.
4187 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4189 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4190 This changes the size when using the `genpkey` command when no size is given.
4191 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4192 generation commands to use 2048 bits by default.
4196 * Prevent over long nonces in ChaCha20-Poly1305.
4198 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4199 for every encryption operation. RFC 7539 specifies that the nonce value
4200 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4201 and front pads the nonce with 0 bytes if it is less than 12
4202 bytes. However it also incorrectly allows a nonce to be set of up to 16
4203 bytes. In this case only the last 12 bytes are significant and any
4204 additional leading bytes are ignored.
4206 It is a requirement of using this cipher that nonce values are
4207 unique. Messages encrypted using a reused nonce value are susceptible to
4208 serious confidentiality and integrity attacks. If an application changes
4209 the default nonce length to be longer than 12 bytes and then makes a
4210 change to the leading bytes of the nonce expecting the new value to be a
4211 new unique nonce then such an application could inadvertently encrypt
4212 messages with a reused nonce.
4214 Additionally the ignored bytes in a long nonce are not covered by the
4215 integrity guarantee of this cipher. Any application that relies on the
4216 integrity of these ignored leading bytes of a long nonce may be further
4217 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4218 is safe because no such use sets such a long nonce value. However user
4219 applications that use this cipher directly and set a non-default nonce
4220 length to be longer than 12 bytes may be vulnerable.
4222 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4228 * Added SCA hardening for modular field inversion in EC_GROUP through
4229 a new dedicated field_inv() pointer in EC_METHOD.
4230 This also addresses a leakage affecting conversions from projective
4231 to affine coordinates.
4233 *Billy Bob Brumley, Nicola Tuveri*
4235 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4236 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4240 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4244 * Remove the 'dist' target and add a tarball building script. The
4245 'dist' target has fallen out of use, and it shouldn't be
4246 necessary to configure just to create a source distribution.
4250 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4252 * Timing vulnerability in DSA signature generation
4254 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4255 timing side channel attack. An attacker could use variations in the signing
4256 algorithm to recover the private key.
4258 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4263 * Timing vulnerability in ECDSA signature generation
4265 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4266 timing side channel attack. An attacker could use variations in the signing
4267 algorithm to recover the private key.
4269 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4274 * Add coordinate blinding for EC_POINT and implement projective
4275 coordinate blinding for generic prime curves as a countermeasure to
4276 chosen point SCA attacks.
4278 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4280 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4282 * Client DoS due to large DH parameter
4284 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4285 malicious server can send a very large prime value to the client. This will
4286 cause the client to spend an unreasonably long period of time generating a
4287 key for this prime resulting in a hang until the client has finished. This
4288 could be exploited in a Denial Of Service attack.
4290 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4295 * Cache timing vulnerability in RSA Key Generation
4297 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4298 a cache timing side channel attack. An attacker with sufficient access to
4299 mount cache timing attacks during the RSA key generation process could
4300 recover the private key.
4302 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4303 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4308 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4309 parameter is no longer accepted, as it leads to a corrupt table. NULL
4310 pem_str is reserved for alias entries only.
4314 * Revert blinding in ECDSA sign and instead make problematic addition
4315 length-invariant. Switch even to fixed-length Montgomery multiplication.
4319 * Change generating and checking of primes so that the error rate of not
4320 being prime depends on the intended use based on the size of the input.
4321 For larger primes this will result in more rounds of Miller-Rabin.
4322 The maximal error rate for primes with more than 1080 bits is lowered
4325 *Kurt Roeckx, Annie Yousar*
4327 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4331 * Add blinding to ECDSA and DSA signatures to protect against side channel
4332 attacks discovered by Keegan Ryan (NCC Group).
4336 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4337 now allow empty (zero character) pass phrases.
4341 * Certificate time validation (X509_cmp_time) enforces stricter
4342 compliance with RFC 5280. Fractional seconds and timezone offsets
4343 are no longer allowed.
4347 * Fixed a text canonicalisation bug in CMS
4349 Where a CMS detached signature is used with text content the text goes
4350 through a canonicalisation process first prior to signing or verifying a
4351 signature. This process strips trailing space at the end of lines, converts
4352 line terminators to CRLF and removes additional trailing line terminators
4353 at the end of a file. A bug in the canonicalisation process meant that
4354 some characters, such as form-feed, were incorrectly treated as whitespace
4355 and removed. This is contrary to the specification (RFC5485). This fix
4356 could mean that detached text data signed with an earlier version of
4357 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4358 signed with a fixed OpenSSL may fail to verify with an earlier version of
4359 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4360 and use the "-binary" flag (for the "cms" command line application) or set
4361 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4365 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4367 * Constructed ASN.1 types with a recursive definition could exceed the stack
4369 Constructed ASN.1 types with a recursive definition (such as can be found
4370 in PKCS7) could eventually exceed the stack given malicious input with
4371 excessive recursion. This could result in a Denial Of Service attack. There
4372 are no such structures used within SSL/TLS that come from untrusted sources
4373 so this is considered safe.
4375 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4381 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4383 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4384 effectively reduced to only comparing the least significant bit of each
4385 byte. This allows an attacker to forge messages that would be considered as
4386 authenticated in an amount of tries lower than that guaranteed by the
4387 security claims of the scheme. The module can only be compiled by the
4388 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4390 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4396 * Add a build target 'build_all_generated', to build all generated files
4397 and only that. This can be used to prepare everything that requires
4398 things like perl for a system that lacks perl and then move everything
4399 to that system and do the rest of the build there.
4403 * Backport SSL_OP_NO_RENGOTIATION
4405 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4406 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4407 changes this is no longer possible in 1.1.0. Therefore, the new
4408 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4409 1.1.0 to provide equivalent functionality.
4411 Note that if an application built against 1.1.0h headers (or above) is run
4412 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4413 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4417 * Removed the OS390-Unix config target. It relied on a script that doesn't
4422 * rsaz_1024_mul_avx2 overflow bug on x86_64
4424 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4425 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4426 Analysis suggests that attacks against RSA and DSA as a result of this
4427 defect would be very difficult to perform and are not believed likely.
4428 Attacks against DH1024 are considered just feasible, because most of the
4429 work necessary to deduce information about a private key may be performed
4430 offline. The amount of resources required for such an attack would be
4431 significant. However, for an attack on TLS to be meaningful, the server
4432 would have to share the DH1024 private key among multiple clients, which is
4433 no longer an option since CVE-2016-0701.
4435 This only affects processors that support the AVX2 but not ADX extensions
4436 like Intel Haswell (4th generation).
4438 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4439 was originally found via the OSS-Fuzz project.
4444 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4446 * bn_sqrx8x_internal carry bug on x86_64
4448 There is a carry propagating bug in the x86_64 Montgomery squaring
4449 procedure. No EC algorithms are affected. Analysis suggests that attacks
4450 against RSA and DSA as a result of this defect would be very difficult to
4451 perform and are not believed likely. Attacks against DH are considered just
4452 feasible (although very difficult) because most of the work necessary to
4453 deduce information about a private key may be performed offline. The amount
4454 of resources required for such an attack would be very significant and
4455 likely only accessible to a limited number of attackers. An attacker would
4456 additionally need online access to an unpatched system using the target
4457 private key in a scenario with persistent DH parameters and a private
4458 key that is shared between multiple clients.
4460 This only affects processors that support the BMI1, BMI2 and ADX extensions
4461 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4463 This issue was reported to OpenSSL by the OSS-Fuzz project.
4468 * Malformed X.509 IPAddressFamily could cause OOB read
4470 If an X.509 certificate has a malformed IPAddressFamily extension,
4471 OpenSSL could do a one-byte buffer overread. The most likely result
4472 would be an erroneous display of the certificate in text format.
4474 This issue was reported to OpenSSL by the OSS-Fuzz project.
4479 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4481 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4482 platform rather than 'mingw'.
4486 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4487 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4488 which is the minimum version we support.
4492 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4494 * Encrypt-Then-Mac renegotiation crash
4496 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4497 negotiated where it was not in the original handshake (or vice-versa) then
4498 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4499 and servers are affected.
4501 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4506 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4508 * Truncated packet could crash via OOB read
4510 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4511 cipher is being used, then a truncated packet can cause that host to
4512 perform an out-of-bounds read, usually resulting in a crash.
4514 This issue was reported to OpenSSL by Robert Święcki of Google.
4519 * Bad (EC)DHE parameters cause a client crash
4521 If a malicious server supplies bad parameters for a DHE or ECDHE key
4522 exchange then this can result in the client attempting to dereference a
4523 NULL pointer leading to a client crash. This could be exploited in a Denial
4526 This issue was reported to OpenSSL by Guido Vranken.
4531 * BN_mod_exp may produce incorrect results on x86_64
4533 There is a carry propagating bug in the x86_64 Montgomery squaring
4534 procedure. No EC algorithms are affected. Analysis suggests that attacks
4535 against RSA and DSA as a result of this defect would be very difficult to
4536 perform and are not believed likely. Attacks against DH are considered just
4537 feasible (although very difficult) because most of the work necessary to
4538 deduce information about a private key may be performed offline. The amount
4539 of resources required for such an attack would be very significant and
4540 likely only accessible to a limited number of attackers. An attacker would
4541 additionally need online access to an unpatched system using the target
4542 private key in a scenario with persistent DH parameters and a private
4543 key that is shared between multiple clients. For example this can occur by
4544 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4545 similar to CVE-2015-3193 but must be treated as a separate problem.
4547 This issue was reported to OpenSSL by the OSS-Fuzz project.
4552 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4554 * ChaCha20/Poly1305 heap-buffer-overflow
4556 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4557 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4558 crash. This issue is not considered to be exploitable beyond a DoS.
4560 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4565 * CMS Null dereference
4567 Applications parsing invalid CMS structures can crash with a NULL pointer
4568 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4569 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4570 structure callback if an attempt is made to free certain invalid encodings.
4571 Only CHOICE structures using a callback which do not handle NULL value are
4574 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4579 * Montgomery multiplication may produce incorrect results
4581 There is a carry propagating bug in the Broadwell-specific Montgomery
4582 multiplication procedure that handles input lengths divisible by, but
4583 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4584 and DH private keys are impossible. This is because the subroutine in
4585 question is not used in operations with the private key itself and an input
4586 of the attacker's direct choice. Otherwise the bug can manifest itself as
4587 transient authentication and key negotiation failures or reproducible
4588 erroneous outcome of public-key operations with specially crafted input.
4589 Among EC algorithms only Brainpool P-512 curves are affected and one
4590 presumably can attack ECDH key negotiation. Impact was not analyzed in
4591 detail, because pre-requisites for attack are considered unlikely. Namely
4592 multiple clients have to choose the curve in question and the server has to
4593 share the private key among them, neither of which is default behaviour.
4594 Even then only clients that chose the curve will be affected.
4596 This issue was publicly reported as transient failures and was not
4597 initially recognized as a security issue. Thanks to Richard Morgan for
4598 providing reproducible case.
4603 * Removed automatic addition of RPATH in shared libraries and executables,
4604 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4608 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4610 * Fix Use After Free for large message sizes
4612 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4613 message larger than approx 16k is received then the underlying buffer to
4614 store the incoming message is reallocated and moved. Unfortunately a
4615 dangling pointer to the old location is left which results in an attempt to
4616 write to the previously freed location. This is likely to result in a
4617 crash, however it could potentially lead to execution of arbitrary code.
4619 This issue only affects OpenSSL 1.1.0a.
4621 This issue was reported to OpenSSL by Robert Święcki.
4626 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4628 * OCSP Status Request extension unbounded memory growth
4630 A malicious client can send an excessively large OCSP Status Request
4631 extension. If that client continually requests renegotiation, sending a
4632 large OCSP Status Request extension each time, then there will be unbounded
4633 memory growth on the server. This will eventually lead to a Denial Of
4634 Service attack through memory exhaustion. Servers with a default
4635 configuration are vulnerable even if they do not support OCSP. Builds using
4636 the "no-ocsp" build time option are not affected.
4638 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4643 * SSL_peek() hang on empty record
4645 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4646 sends an empty record. This could be exploited by a malicious peer in a
4647 Denial Of Service attack.
4649 This issue was reported to OpenSSL by Alex Gaynor.
4654 * Excessive allocation of memory in tls_get_message_header() and
4655 dtls1_preprocess_fragment()
4657 A (D)TLS message includes 3 bytes for its length in the header for the
4658 message. This would allow for messages up to 16Mb in length. Messages of
4659 this length are excessive and OpenSSL includes a check to ensure that a
4660 peer is sending reasonably sized messages in order to avoid too much memory
4661 being consumed to service a connection. A flaw in the logic of version
4662 1.1.0 means that memory for the message is allocated too early, prior to
4663 the excessive message length check. Due to way memory is allocated in
4664 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4665 to service a connection. This could lead to a Denial of Service through
4666 memory exhaustion. However, the excessive message length check still takes
4667 place, and this would cause the connection to immediately fail. Assuming
4668 that the application calls SSL_free() on the failed connection in a timely
4669 manner then the 21Mb of allocated memory will then be immediately freed
4670 again. Therefore, the excessive memory allocation will be transitory in
4671 nature. This then means that there is only a security impact if:
4673 1) The application does not call SSL_free() in a timely manner in the event
4674 that the connection fails
4676 2) The application is working in a constrained environment where there is
4677 very little free memory
4679 3) The attacker initiates multiple connection attempts such that there are
4680 multiple connections in a state where memory has been allocated for the
4681 connection; SSL_free() has not yet been called; and there is insufficient
4682 memory to service the multiple requests.
4684 Except in the instance of (1) above any Denial Of Service is likely to be
4685 transitory because as soon as the connection fails the memory is
4686 subsequently freed again in the SSL_free() call. However there is an
4687 increased risk during this period of application crashes due to the lack of
4688 memory - which would then mean a more serious Denial of Service.
4690 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4691 (CVE-2016-6307 and CVE-2016-6308)
4695 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4696 had to be removed. Primary reason is that vendor assembler can't
4697 assemble our modules with -KPIC flag. As result it, assembly
4698 support, was not even available as option. But its lack means
4699 lack of side-channel resistant code, which is incompatible with
4700 security by todays standards. Fortunately gcc is readily available
4701 prepackaged option, which we firmly point at...
4705 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4707 * Windows command-line tool supports UTF-8 opt-in option for arguments
4708 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4709 (to any value) allows Windows user to access PKCS#12 file generated
4710 with Windows CryptoAPI and protected with non-ASCII password, as well
4711 as files generated under UTF-8 locale on Linux also protected with
4716 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4717 have been disabled by default and removed from DEFAULT, just like RC4.
4718 See the RC4 item below to re-enable both.
4722 * The method for finding the storage location for the Windows RAND seed file
4723 has changed. First we check %RANDFILE%. If that is not set then we check
4724 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4725 all else fails we fall back to C:\.
4729 * The EVP_EncryptUpdate() function has had its return type changed from void
4730 to int. A return of 0 indicates and error while a return of 1 indicates
4735 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4736 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4737 off the constant time implementation for RSA, DSA and DH have been made
4738 no-ops and deprecated.
4742 * Windows RAND implementation was simplified to only get entropy by
4743 calling CryptGenRandom(). Various other RAND-related tickets
4746 *Joseph Wylie Yandle, Rich Salz*
4748 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4749 and `OPENSSL_LH_`, respectively. The old names are available
4750 with API compatibility. They new names are now completely documented.
4754 * Unify TYPE_up_ref(obj) methods signature.
4755 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4756 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4757 int (instead of void) like all others TYPE_up_ref() methods.
4758 So now these methods also check the return value of CRYPTO_atomic_add(),
4759 and the validity of object reference counter.
4761 *fdasilvayy@gmail.com*
4763 * With Windows Visual Studio builds, the .pdb files are installed
4764 alongside the installed libraries and executables. For a static
4765 library installation, ossl_static.pdb is the associate compiler
4766 generated .pdb file to be used when linking programs.
4770 * Remove openssl.spec. Packaging files belong with the packagers.
4774 * Automatic Darwin/OSX configuration has had a refresh, it will now
4775 recognise x86_64 architectures automatically. You can still decide
4776 to build for a different bitness with the environment variable
4777 KERNEL_BITS (can be 32 or 64), for example:
4779 KERNEL_BITS=32 ./config
4783 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4784 256 bit AES and HMAC with SHA256.
4788 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4792 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4796 * To enable users to have their own config files and build file templates,
4797 Configure looks in the directory indicated by the environment variable
4798 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4799 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4800 name and is used as is.
4804 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4805 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4806 X509_CERT_FILE_CTX was removed.
4810 * "shared" builds are now the default. To create only static libraries use
4811 the "no-shared" Configure option.
4815 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4816 All of these option have not worked for some while and are fundamental
4821 * Make various cleanup routines no-ops and mark them as deprecated. Most
4822 global cleanup functions are no longer required because they are handled
4823 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4824 Explicitly de-initing can cause problems (e.g. where a library that uses
4825 OpenSSL de-inits, but an application is still using it). The affected
4826 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4827 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4828 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4829 COMP_zlib_cleanup().
4833 * --strict-warnings no longer enables runtime debugging options
4834 such as REF_DEBUG. Instead, debug options are automatically
4835 enabled with '--debug' builds.
4837 *Andy Polyakov, Emilia Käsper*
4839 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4840 have been moved out of the public header files. New functions for managing
4841 these have been added.
4845 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4846 objects have been moved out of the public header files. New
4847 functions for managing these have been added.
4851 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4852 have been moved out of the public header files. New functions for managing
4853 these have been added.
4857 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4858 moved out of the public header files. New functions for managing these
4863 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4867 * Removed the mk1mf build scripts.
4871 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4872 it is always safe to #include a header now.
4876 * Removed the aged BC-32 config and all its supporting scripts
4880 * Removed support for Ultrix, Netware, and OS/2.
4884 * Add support for HKDF.
4886 *Alessandro Ghedini*
4888 * Add support for blake2b and blake2s
4892 * Added support for "pipelining". Ciphers that have the
4893 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4894 encryptions/decryptions simultaneously. There are currently no built-in
4895 ciphers with this property but the expectation is that engines will be able
4896 to offer it to significantly improve throughput. Support has been extended
4897 into libssl so that multiple records for a single connection can be
4898 processed in one go (for >=TLS 1.1).
4902 * Added the AFALG engine. This is an async capable engine which is able to
4903 offload work to the Linux kernel. In this initial version it only supports
4904 AES128-CBC. The kernel must be version 4.1.0 or greater.
4908 * OpenSSL now uses a new threading API. It is no longer necessary to
4909 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4910 are two supported threading models: pthreads and windows threads. It is
4911 also possible to configure OpenSSL at compile time for "no-threads". The
4912 old threading API should no longer be used. The functions have been
4913 replaced with "no-op" compatibility macros.
4915 *Alessandro Ghedini, Matt Caswell*
4917 * Modify behavior of ALPN to invoke callback after SNI/servername
4918 callback, such that updates to the SSL_CTX affect ALPN.
4922 * Add SSL_CIPHER queries for authentication and key-exchange.
4926 * Changes to the DEFAULT cipherlist:
4927 - Prefer (EC)DHE handshakes over plain RSA.
4928 - Prefer AEAD ciphers over legacy ciphers.
4929 - Prefer ECDSA over RSA when both certificates are available.
4930 - Prefer TLSv1.2 ciphers/PRF.
4931 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4936 * Change the ECC default curve list to be this, in order: x25519,
4937 secp256r1, secp521r1, secp384r1.
4941 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4942 disabled by default. They can be re-enabled using the
4943 enable-weak-ssl-ciphers option to Configure.
4947 * If the server has ALPN configured, but supports no protocols that the
4948 client advertises, send a fatal "no_application_protocol" alert.
4949 This behaviour is SHALL in RFC 7301, though it isn't universally
4950 implemented by other servers.
4954 * Add X25519 support.
4955 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4956 for public and private key encoding using the format documented in
4957 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4958 key generation and key derivation.
4960 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4965 * Deprecate SRP_VBASE_get_by_user.
4966 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4967 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4968 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4969 seed, even if the seed is configured.
4971 Users should use SRP_VBASE_get1_by_user instead. Note that in
4972 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4973 also that even though configuring the SRP seed attempts to hide
4974 invalid usernames by continuing the handshake with fake
4975 credentials, this behaviour is not constant time and no strong
4976 guarantees are made that the handshake is indistinguishable from
4977 that of a valid user.
4981 * Configuration change; it's now possible to build dynamic engines
4982 without having to build shared libraries and vice versa. This
4983 only applies to the engines in `engines/`, those in `crypto/engine/`
4984 will always be built into libcrypto (i.e. "static").
4986 Building dynamic engines is enabled by default; to disable, use
4987 the configuration option "disable-dynamic-engine".
4989 The only requirements for building dynamic engines are the
4990 presence of the DSO module and building with position independent
4991 code, so they will also automatically be disabled if configuring
4992 with "disable-dso" or "disable-pic".
4994 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4995 are also taken away from openssl/opensslconf.h, as they are
5000 * Configuration change; if there is a known flag to compile
5001 position independent code, it will always be applied on the
5002 libcrypto and libssl object files, and never on the application
5003 object files. This means other libraries that use routines from
5004 libcrypto / libssl can be made into shared libraries regardless
5005 of how OpenSSL was configured.
5007 If this isn't desirable, the configuration options "disable-pic"
5008 or "no-pic" can be used to disable the use of PIC. This will
5009 also disable building shared libraries and dynamic engines.
5013 * Removed JPAKE code. It was experimental and has no wide use.
5017 * The INSTALL_PREFIX Makefile variable has been renamed to
5018 DESTDIR. That makes for less confusion on what this variable
5019 is for. Also, the configuration option --install_prefix is
5024 * Heartbeat for TLS has been removed and is disabled by default
5025 for DTLS; configure with enable-heartbeats. Code that uses the
5026 old #define's might need to be updated.
5028 *Emilia Käsper, Rich Salz*
5030 * Rename REF_CHECK to REF_DEBUG.
5034 * New "unified" build system
5036 The "unified" build system is aimed to be a common system for all
5037 platforms we support. With it comes new support for VMS.
5039 This system builds supports building in a different directory tree
5040 than the source tree. It produces one Makefile (for unix family
5041 or lookalikes), or one descrip.mms (for VMS).
5043 The source of information to make the Makefile / descrip.mms is
5044 small files called 'build.info', holding the necessary
5045 information for each directory with source to compile, and a
5046 template in Configurations, like unix-Makefile.tmpl or
5049 With this change, the library names were also renamed on Windows
5050 and on VMS. They now have names that are closer to the standard
5051 on Unix, and include the major version number, and in certain
5052 cases, the architecture they are built for. See "Notes on shared
5053 libraries" in INSTALL.
5055 We rely heavily on the perl module Text::Template.
5059 * Added support for auto-initialisation and de-initialisation of the library.
5060 OpenSSL no longer requires explicit init or deinit routines to be called,
5061 except in certain circumstances. See the OPENSSL_init_crypto() and
5062 OPENSSL_init_ssl() man pages for further information.
5066 * The arguments to the DTLSv1_listen function have changed. Specifically the
5067 "peer" argument is now expected to be a BIO_ADDR object.
5069 * Rewrite of BIO networking library. The BIO library lacked consistent
5070 support of IPv6, and adding it required some more extensive
5071 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
5072 which hold all types of addresses and chains of address information.
5073 It also introduces a new API, with functions like BIO_socket,
5074 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
5075 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
5076 have been adapted accordingly.
5080 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
5085 * CRIME protection: disable compression by default, even if OpenSSL is
5086 compiled with zlib enabled. Applications can still enable compression
5087 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5088 using the SSL_CONF library to configure compression.
5092 * The signature of the session callback configured with
5093 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5094 was explicitly marked as `const unsigned char*` instead of
5099 * Always DPURIFY. Remove the use of uninitialized memory in the
5100 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5104 * Removed many obsolete configuration items, including
5105 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5106 MD2_CHAR, MD2_INT, MD2_LONG
5108 IDEA_SHORT, IDEA_LONG
5109 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5111 *Rich Salz, with advice from Andy Polyakov*
5113 * Many BN internals have been moved to an internal header file.
5115 *Rich Salz with help from Andy Polyakov*
5117 * Configuration and writing out the results from it has changed.
5118 Files such as Makefile include/openssl/opensslconf.h and are now
5119 produced through general templates, such as Makefile.in and
5120 crypto/opensslconf.h.in and some help from the perl module
5123 Also, the center of configuration information is no longer
5124 Makefile. Instead, Configure produces a perl module in
5125 configdata.pm which holds most of the config data (in the hash
5126 table %config), the target data that comes from the target
5127 configuration in one of the `Configurations/*.conf` files (in
5132 * To clarify their intended purposes, the Configure options
5133 --prefix and --openssldir change their semantics, and become more
5134 straightforward and less interdependent.
5136 --prefix shall be used exclusively to give the location INSTALLTOP
5137 where programs, scripts, libraries, include files and manuals are
5138 going to be installed. The default is now /usr/local.
5140 --openssldir shall be used exclusively to give the default
5141 location OPENSSLDIR where certificates, private keys, CRLs are
5142 managed. This is also where the default openssl.cnf gets
5144 If the directory given with this option is a relative path, the
5145 values of both the --prefix value and the --openssldir value will
5146 be combined to become OPENSSLDIR.
5147 The default for --openssldir is INSTALLTOP/ssl.
5149 Anyone who uses --openssldir to specify where OpenSSL is to be
5150 installed MUST change to use --prefix instead.
5154 * The GOST engine was out of date and therefore it has been removed. An up
5155 to date GOST engine is now being maintained in an external repository.
5156 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5157 support for GOST ciphersuites (these are only activated if a GOST engine
5162 * EGD is no longer supported by default; use enable-egd when
5165 *Ben Kaduk and Rich Salz*
5167 * The distribution now has Makefile.in files, which are used to
5168 create Makefile's when Configure is run. *Configure must be run
5169 before trying to build now.*
5173 * The return value for SSL_CIPHER_description() for error conditions
5178 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5180 Obtaining and performing DNSSEC validation of TLSA records is
5181 the application's responsibility. The application provides
5182 the TLSA records of its choice to OpenSSL, and these are then
5183 used to authenticate the peer.
5185 The TLSA records need not even come from DNS. They can, for
5186 example, be used to implement local end-entity certificate or
5187 trust-anchor "pinning", where the "pin" data takes the form
5188 of TLSA records, which can augment or replace verification
5189 based on the usual WebPKI public certification authorities.
5193 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5194 continues to support deprecated interfaces in default builds.
5195 However, applications are strongly advised to compile their
5196 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5197 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5198 or the 1.1.0 releases.
5200 In environments in which all applications have been ported to
5201 not use any deprecated interfaces OpenSSL's Configure script
5202 should be used with the --api=1.1.0 option to entirely remove
5203 support for the deprecated features from the library and
5204 unconditionally disable them in the installed headers.
5205 Essentially the same effect can be achieved with the "no-deprecated"
5206 argument to Configure, except that this will always restrict
5207 the build to just the latest API, rather than a fixed API
5210 As applications are ported to future revisions of the API,
5211 they should update their compile-time OPENSSL_API_COMPAT define
5212 accordingly, but in most cases should be able to continue to
5213 compile with later releases.
5215 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5216 0x10000000L and 0x00908000L, respectively. However those
5217 versions did not support the OPENSSL_API_COMPAT feature, and
5218 so applications are not typically tested for explicit support
5219 of just the undeprecated features of either release.
5223 * Add support for setting the minimum and maximum supported protocol.
5224 It can bet set via the SSL_set_min_proto_version() and
5225 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5226 MaxProtocol. It's recommended to use the new APIs to disable
5227 protocols instead of disabling individual protocols using
5228 SSL_set_options() or SSL_CONF's Protocol. This change also
5229 removes support for disabling TLS 1.2 in the OpenSSL TLS
5230 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5234 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5238 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5239 and integrates ECDSA and ECDH functionality into EC. Implementations can
5240 now redirect key generation and no longer need to convert to or from
5243 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5244 include the ec.h header file instead.
5248 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5249 ciphers who are no longer supported and drops support the ephemeral RSA key
5250 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5254 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5255 opaque. For HMAC_CTX, the following constructors and destructors
5258 HMAC_CTX *HMAC_CTX_new(void);
5259 void HMAC_CTX_free(HMAC_CTX *ctx);
5261 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5262 destroy such methods has been added. See EVP_MD_meth_new(3) and
5263 EVP_CIPHER_meth_new(3) for documentation.
5266 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5267 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5268 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5269 an already created structure.
5270 2) For consistency with the majority of our object creators and
5271 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5272 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5273 for deprecated builds.
5277 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5278 cryptographic operations to be performed asynchronously as long as an
5279 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5280 further details. Libssl has also had this capability integrated with the
5281 introduction of the new mode SSL_MODE_ASYNC and associated error
5282 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5283 pages. This work was developed in partnership with Intel Corp.
5287 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5288 always enabled now. If you want to disable the support you should
5289 exclude it using the list of supported ciphers. This also means that the
5290 "-no_ecdhe" option has been removed from s_server.
5294 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5295 SSL_{CTX_}set1_curves() which can set a list.
5299 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5300 curve you want to support using SSL_{CTX_}set1_curves().
5304 * State machine rewrite. The state machine code has been significantly
5305 refactored in order to remove much duplication of code and solve issues
5306 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5307 further details). This change does have some associated API changes.
5308 Notably the SSL_state() function has been removed and replaced by
5309 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5310 SSL_set_state() has been removed altogether. The previous handshake states
5311 defined in ssl.h and ssl3.h have also been removed.
5315 * All instances of the string "ssleay" in the public API were replaced
5316 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5317 Some error codes related to internal RSA_eay API's were renamed.
5321 * The demo files in crypto/threads were moved to demo/threads.
5325 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5328 *Matt Caswell, Rich Salz*
5330 * New ASN.1 embed macro.
5332 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5333 structure is not allocated: it is part of the parent. That is instead of
5341 This reduces memory fragmentation and make it impossible to accidentally
5342 set a mandatory field to NULL.
5344 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5345 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5346 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5351 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5355 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5356 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5357 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5358 DES and RC4 ciphersuites.
5362 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5363 This changes the decoding behaviour for some invalid messages,
5364 though the change is mostly in the more lenient direction, and
5365 legacy behaviour is preserved as much as possible.
5369 * Fix no-stdio build.
5370 *David Woodhouse <David.Woodhouse@intel.com> and also*
5371 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5373 * New testing framework
5374 The testing framework has been largely rewritten and is now using
5375 perl and the perl modules Test::Harness and an extended variant of
5376 Test::More called OpenSSL::Test to do its work. All test scripts in
5377 test/ have been rewritten into test recipes, and all direct calls to
5378 executables in test/Makefile have become individual recipes using the
5379 simplified testing OpenSSL::Test::Simple.
5381 For documentation on our testing modules, do:
5383 perldoc test/testlib/OpenSSL/Test/Simple.pm
5384 perldoc test/testlib/OpenSSL/Test.pm
5388 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5389 are used; the latter aborts on memory leaks (usually checked on exit).
5390 Some undocumented "set malloc, etc., hooks" functions were removed
5391 and others were changed. All are now documented.
5395 * In DSA_generate_parameters_ex, if the provided seed is too short,
5398 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5400 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5401 from RFC4279, RFC4785, RFC5487, RFC5489.
5403 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5404 original RSA_PSK patch.
5408 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5409 era flag was never set throughout the codebase (only read). Also removed
5410 SSL3_FLAGS_POP_BUFFER which was only used if
5411 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5415 * Changed the default name options in the "ca", "crl", "req" and "x509"
5416 to be "oneline" instead of "compat".
5420 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5421 not aware of clients that still exhibit this bug, and the workaround
5422 hasn't been working properly for a while.
5426 * The return type of BIO_number_read() and BIO_number_written() as well as
5427 the corresponding num_read and num_write members in the BIO structure has
5428 changed from unsigned long to uint64_t. On platforms where an unsigned
5429 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5434 * Given the pervasive nature of TLS extensions it is inadvisable to run
5435 OpenSSL without support for them. It also means that maintaining
5436 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5437 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5441 * Removed support for the two export grade static DH ciphersuites
5442 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5443 were newly added (along with a number of other static DH ciphersuites) to
5444 1.0.2. However the two export ones have *never* worked since they were
5445 introduced. It seems strange in any case to be adding new export
5446 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5450 * Version negotiation has been rewritten. In particular SSLv23_method(),
5451 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5452 and turned into macros which simply call the new preferred function names
5453 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5454 should use the new names instead. Also as part of this change the ssl23.h
5455 header file has been removed.
5459 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5460 code and the associated standard is no longer considered fit-for-purpose.
5464 * RT2547 was closed. When generating a private key, try to make the
5465 output file readable only by the owner. This behavior change might
5466 be noticeable when interacting with other software.
5468 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5473 * Added HTTP GET support to the ocsp command.
5477 * Changed default digest for the dgst and enc commands from MD5 to
5482 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5486 * Added support for TLS extended master secret from
5487 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5488 initial patch which was a great help during development.
5492 * All libssl internal structures have been removed from the public header
5493 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5494 now redundant). Users should not attempt to access internal structures
5495 directly. Instead they should use the provided API functions.
5499 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5500 Access to deprecated functions can be re-enabled by running config with
5501 "enable-deprecated". In addition applications wishing to use deprecated
5502 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5503 will, by default, disable some transitive includes that previously existed
5504 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5508 * Added support for OCB mode. OpenSSL has been granted a patent license
5509 compatible with the OpenSSL license for use of OCB. Details are available
5510 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5511 for OCB can be removed by calling config with no-ocb.
5515 * SSLv2 support has been removed. It still supports receiving an SSLv2
5516 compatible client hello.
5520 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5521 done while fixing the error code for the key-too-small case.
5523 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5525 * CA.sh has been removed; use CA.pl instead.
5529 * Removed old DES API.
5533 * Remove various unsupported platforms:
5539 Sinix/ReliantUNIX RM400
5544 16-bit platforms such as WIN16
5548 * Clean up OPENSSL_NO_xxx #define's
5549 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5550 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5551 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5552 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5553 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5554 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5555 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5556 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5557 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5558 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5562 * Cleaned up dead code
5563 Remove all but one '#ifdef undef' which is to be looked at.
5567 * Clean up calling of xxx_free routines.
5568 Just like free(), fix most of the xxx_free routines to accept
5569 NULL. Remove the non-null checks from callers. Save much code.
5573 * Add secure heap for storage of private keys (when possible).
5574 Add BIO_s_secmem(), CBIGNUM, etc.
5575 Contributed by Akamai Technologies under our Corporate CLA.
5579 * Experimental support for a new, fast, unbiased prime candidate generator,
5580 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5582 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5584 * New output format NSS in the sess_id command line tool. This allows
5585 exporting the session id and the master key in NSS keylog format.
5587 *Martin Kaiser <martin@kaiser.cx>*
5589 * Harmonize version and its documentation. -f flag is used to display
5592 *mancha <mancha1@zoho.com>*
5594 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5595 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5597 *mancha <mancha1@zoho.com>*
5599 * Fix some double frees. These are not thought to be exploitable.
5601 *mancha <mancha1@zoho.com>*
5603 * A missing bounds check in the handling of the TLS heartbeat extension
5604 can be used to reveal up to 64k of memory to a connected client or
5607 Thanks for Neel Mehta of Google Security for discovering this bug and to
5608 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5609 preparing the fix ([CVE-2014-0160])
5611 *Adam Langley, Bodo Moeller*
5613 * Fix for the attack described in the paper "Recovering OpenSSL
5614 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5615 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5616 <http://eprint.iacr.org/2014/140>
5618 Thanks to Yuval Yarom and Naomi Benger for discovering this
5619 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5621 *Yuval Yarom and Naomi Benger*
5623 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5624 this fixes a limitation in previous versions of OpenSSL.
5628 * Experimental encrypt-then-mac support.
5630 Experimental support for encrypt then mac from
5631 draft-gutmann-tls-encrypt-then-mac-02.txt
5633 To enable it set the appropriate extension number (0x42 for the test
5634 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5636 For non-compliant peers (i.e. just about everything) this should have no
5639 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5643 * Add EVP support for key wrapping algorithms, to avoid problems with
5644 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5645 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5646 algorithms and include tests cases.
5650 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5655 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5656 MGF1 digest and OAEP label.
5660 * Make openssl verify return errors.
5662 *Chris Palmer <palmer@google.com> and Ben Laurie*
5664 * New function ASN1_TIME_diff to calculate the difference between two
5665 ASN1_TIME structures or one structure and the current time.
5669 * Update fips_test_suite to support multiple command line options. New
5670 test to induce all self test errors in sequence and check expected
5675 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5676 sign or verify all in one operation.
5680 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5681 test programs and fips_test_suite. Includes functionality to parse
5682 the minimal script output of fipsalgest.pl directly.
5686 * Add authorisation parameter to FIPS_module_mode_set().
5690 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5694 * Use separate DRBG fields for internal and external flags. New function
5695 FIPS_drbg_health_check() to perform on demand health checking. Add
5696 generation tests to fips_test_suite with reduced health check interval to
5697 demonstrate periodic health checking. Add "nodh" option to
5698 fips_test_suite to skip very slow DH test.
5702 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5707 * More extensive health check for DRBG checking many more failure modes.
5708 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5709 combination: call this in fips_test_suite.
5713 * Add support for canonical generation of DSA parameter 'g'. See
5716 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5717 POST to handle HMAC cases.
5721 * Add functions FIPS_module_version() and FIPS_module_version_text()
5722 to return numerical and string versions of the FIPS module number.
5726 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5727 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5728 outside the validated module in the FIPS capable OpenSSL.
5732 * Minor change to DRBG entropy callback semantics. In some cases
5733 there is no multiple of the block length between min_len and
5734 max_len. Allow the callback to return more than max_len bytes
5735 of entropy but discard any extra: it is the callback's responsibility
5736 to ensure that the extra data discarded does not impact the
5737 requested amount of entropy.
5741 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5742 information in FIPS186-3, SP800-57 and SP800-131A.
5746 * CCM support via EVP. Interface is very similar to GCM case except we
5747 must supply all data in one chunk (i.e. no update, final) and the
5748 message length must be supplied if AAD is used. Add algorithm test
5753 * Initial version of POST overhaul. Add POST callback to allow the status
5754 of POST to be monitored and/or failures induced. Modify fips_test_suite
5755 to use callback. Always run all selftests even if one fails.
5759 * XTS support including algorithm test driver in the fips_gcmtest program.
5760 Note: this does increase the maximum key length from 32 to 64 bytes but
5761 there should be no binary compatibility issues as existing applications
5762 will never use XTS mode.
5766 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5767 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5768 performs algorithm blocking for unapproved PRNG types. Also do not
5769 set PRNG type in FIPS_mode_set(): leave this to the application.
5770 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5771 the standard OpenSSL PRNG: set additional data to a date time vector.
5775 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5776 This shouldn't present any incompatibility problems because applications
5777 shouldn't be using these directly and any that are will need to rethink
5778 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5782 * Extensive self tests and health checking required by SP800-90 DRBG.
5783 Remove strength parameter from FIPS_drbg_instantiate and always
5784 instantiate at maximum supported strength.
5788 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5792 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5796 * New function DH_compute_key_padded() to compute a DH key and pad with
5797 leading zeroes if needed: this complies with SP800-56A et al.
5801 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5802 anything, incomplete, subject to change and largely untested at present.
5806 * Modify fipscanisteronly build option to only build the necessary object
5807 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5811 * Add experimental option FIPSSYMS to give all symbols in
5812 fipscanister.o and FIPS or fips prefix. This will avoid
5813 conflicts with future versions of OpenSSL. Add perl script
5814 util/fipsas.pl to preprocess assembly language source files
5815 and rename any affected symbols.
5819 * Add selftest checks and algorithm block of non-fips algorithms in
5820 FIPS mode. Remove DES2 from selftests.
5824 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5825 return internal method without any ENGINE dependencies. Add new
5826 tiny fips sign and verify functions.
5830 * New build option no-ec2m to disable characteristic 2 code.
5834 * New build option "fipscanisteronly". This only builds fipscanister.o
5835 and (currently) associated fips utilities. Uses the file Makefile.fips
5836 instead of Makefile.org as the prototype.
5840 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5841 Update fips_gcmtest to use IV generator.
5845 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5846 setting output buffer to NULL. The `*Final` function must be
5847 called although it will not retrieve any additional data. The tag
5848 can be set or retrieved with a ctrl. The IV length is by default 12
5849 bytes (96 bits) but can be set to an alternative value. If the IV
5850 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5855 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5856 underlying do_cipher function handles all cipher semantics itself
5857 including padding and finalisation. This is useful if (for example)
5858 an ENGINE cipher handles block padding itself. The behaviour of
5859 do_cipher is subtly changed if this flag is set: the return value
5860 is the number of characters written to the output buffer (zero is
5861 no longer an error code) or a negative error code. Also if the
5862 input buffer is NULL and length 0 finalisation should be performed.
5866 * If a candidate issuer certificate is already part of the constructed
5867 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5871 * Improve forward-security support: add functions
5873 void SSL_CTX_set_not_resumable_session_callback(
5874 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5875 void SSL_set_not_resumable_session_callback(
5876 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5878 for use by SSL/TLS servers; the callback function will be called whenever a
5879 new session is created, and gets to decide whether the session may be
5880 cached to make it resumable (return 0) or not (return 1). (As by the
5881 SSL/TLS protocol specifications, the session_id sent by the server will be
5882 empty to indicate that the session is not resumable; also, the server will
5883 not generate RFC 4507 (RFC 5077) session tickets.)
5885 A simple reasonable callback implementation is to return is_forward_secure.
5886 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5887 by the SSL/TLS server library, indicating whether it can provide forward
5890 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5892 * New -verify_name option in command line utilities to set verification
5897 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5898 Add CMAC pkey methods.
5902 * Experimental renegotiation in s_server -www mode. If the client
5903 browses /reneg connection is renegotiated. If /renegcert it is
5904 renegotiated requesting a certificate.
5908 * Add an "external" session cache for debugging purposes to s_server. This
5909 should help trace issues which normally are only apparent in deployed
5910 multi-process servers.
5914 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5915 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5916 BIO_set_cipher() and some obscure PEM functions were changed so they
5917 can now return an error. The RAND changes required a change to the
5918 RAND_METHOD structure.
5922 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5923 a gcc attribute to warn if the result of a function is ignored. This
5924 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5925 whose return value is often ignored.
5929 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5930 These allow SCTs (signed certificate timestamps) to be requested and
5931 validated when establishing a connection.
5933 *Rob Percival <robpercival@google.com>*
5938 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5940 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5941 used even when parsing explicit parameters, when loading a encoded key
5942 or calling `EC_GROUP_new_from_ecpkparameters()`/
5943 `EC_GROUP_new_from_ecparameters()`.
5944 This prevents bypass of security hardening and performance gains,
5945 especially for curves with specialized EC_METHODs.
5946 By default, if a key encoded with explicit parameters is loaded and later
5947 encoded, the output is still encoded with explicit parameters, even if
5948 internally a "named" EC_GROUP is used for computation.
5952 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5953 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5954 NULL. After this change, only the cofactor parameter can be NULL. It also
5955 does some minimal sanity checks on the passed order.
5960 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5961 An attack is simple, if the first CMS_recipientInfo is valid but the
5962 second CMS_recipientInfo is chosen ciphertext. If the second
5963 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5964 encryption key will be replaced by garbage, and the message cannot be
5965 decoded, but if the RSA decryption fails, the correct encryption key is
5966 used and the recipient will not notice the attack.
5967 As a work around for this potential attack the length of the decrypted
5968 key must be equal to the cipher default key length, in case the
5969 certificate is not given and all recipientInfo are tried out.
5970 The old behaviour can be re-enabled in the CMS code by setting the
5971 CMS_DEBUG_DECRYPT flag.
5976 * Document issue with installation paths in diverse Windows builds
5978 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5979 binaries and run-time config file.
5984 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5986 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5987 This changes the size when using the `genpkey` command when no size is given.
5988 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5989 generation commands to use 2048 bits by default.
5993 * Add FIPS support for Android Arm 64-bit
5995 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5996 Module in Version 2.0.10. For some reason, the corresponding target
5997 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5998 built with FIPS support on Android Arm 64-bit. This omission has been
6001 *Matthias St. Pierre*
6003 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
6005 * 0-byte record padding oracle
6007 If an application encounters a fatal protocol error and then calls
6008 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
6009 then OpenSSL can respond differently to the calling application if a 0 byte
6010 record is received with invalid padding compared to if a 0 byte record is
6011 received with an invalid MAC. If the application then behaves differently
6012 based on that in a way that is detectable to the remote peer, then this
6013 amounts to a padding oracle that could be used to decrypt data.
6015 In order for this to be exploitable "non-stitched" ciphersuites must be in
6016 use. Stitched ciphersuites are optimised implementations of certain
6017 commonly used ciphersuites. Also the application must call SSL_shutdown()
6018 twice even if a protocol error has occurred (applications should not do
6019 this but some do anyway).
6021 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
6022 Aviram, with additional investigation by Steven Collison and Andrew
6023 Hourselt. It was reported to OpenSSL on 10th December 2018.
6028 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
6032 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
6034 * Microarchitecture timing vulnerability in ECC scalar multiplication
6036 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
6037 shown to be vulnerable to a microarchitecture timing side channel attack.
6038 An attacker with sufficient access to mount local timing attacks during
6039 ECDSA signature generation could recover the private key.
6041 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
6042 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
6048 * Timing vulnerability in DSA signature generation
6050 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
6051 timing side channel attack. An attacker could use variations in the signing
6052 algorithm to recover the private key.
6054 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
6059 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
6060 Module, accidentally introduced while backporting security fixes from the
6061 development branch and hindering the use of ECC in FIPS mode.
6065 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
6067 * Client DoS due to large DH parameter
6069 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
6070 malicious server can send a very large prime value to the client. This will
6071 cause the client to spend an unreasonably long period of time generating a
6072 key for this prime resulting in a hang until the client has finished. This
6073 could be exploited in a Denial Of Service attack.
6075 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
6080 * Cache timing vulnerability in RSA Key Generation
6082 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
6083 a cache timing side channel attack. An attacker with sufficient access to
6084 mount cache timing attacks during the RSA key generation process could
6085 recover the private key.
6087 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6088 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6093 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6094 parameter is no longer accepted, as it leads to a corrupt table. NULL
6095 pem_str is reserved for alias entries only.
6099 * Revert blinding in ECDSA sign and instead make problematic addition
6100 length-invariant. Switch even to fixed-length Montgomery multiplication.
6104 * Change generating and checking of primes so that the error rate of not
6105 being prime depends on the intended use based on the size of the input.
6106 For larger primes this will result in more rounds of Miller-Rabin.
6107 The maximal error rate for primes with more than 1080 bits is lowered
6110 *Kurt Roeckx, Annie Yousar*
6112 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6116 * Add blinding to ECDSA and DSA signatures to protect against side channel
6117 attacks discovered by Keegan Ryan (NCC Group).
6121 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6122 now allow empty (zero character) pass phrases.
6126 * Certificate time validation (X509_cmp_time) enforces stricter
6127 compliance with RFC 5280. Fractional seconds and timezone offsets
6128 are no longer allowed.
6132 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6134 * Constructed ASN.1 types with a recursive definition could exceed the stack
6136 Constructed ASN.1 types with a recursive definition (such as can be found
6137 in PKCS7) could eventually exceed the stack given malicious input with
6138 excessive recursion. This could result in a Denial Of Service attack. There
6139 are no such structures used within SSL/TLS that come from untrusted sources
6140 so this is considered safe.
6142 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6148 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6150 * Read/write after SSL object in error state
6152 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6153 mechanism. The intent was that if a fatal error occurred during a handshake
6154 then OpenSSL would move into the error state and would immediately fail if
6155 you attempted to continue the handshake. This works as designed for the
6156 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6157 SSL_connect()), however due to a bug it does not work correctly if
6158 SSL_read() or SSL_write() is called directly. In that scenario, if the
6159 handshake fails then a fatal error will be returned in the initial function
6160 call. If SSL_read()/SSL_write() is subsequently called by the application
6161 for the same SSL object then it will succeed and the data is passed without
6162 being decrypted/encrypted directly from the SSL/TLS record layer.
6164 In order to exploit this issue an application bug would have to be present
6165 that resulted in a call to SSL_read()/SSL_write() being issued after having
6166 already received a fatal error.
6168 This issue was reported to OpenSSL by David Benjamin (Google).
6173 * rsaz_1024_mul_avx2 overflow bug on x86_64
6175 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6176 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6177 Analysis suggests that attacks against RSA and DSA as a result of this
6178 defect would be very difficult to perform and are not believed likely.
6179 Attacks against DH1024 are considered just feasible, because most of the
6180 work necessary to deduce information about a private key may be performed
6181 offline. The amount of resources required for such an attack would be
6182 significant. However, for an attack on TLS to be meaningful, the server
6183 would have to share the DH1024 private key among multiple clients, which is
6184 no longer an option since CVE-2016-0701.
6186 This only affects processors that support the AVX2 but not ADX extensions
6187 like Intel Haswell (4th generation).
6189 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6190 was originally found via the OSS-Fuzz project.
6195 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6197 * bn_sqrx8x_internal carry bug on x86_64
6199 There is a carry propagating bug in the x86_64 Montgomery squaring
6200 procedure. No EC algorithms are affected. Analysis suggests that attacks
6201 against RSA and DSA as a result of this defect would be very difficult to
6202 perform and are not believed likely. Attacks against DH are considered just
6203 feasible (although very difficult) because most of the work necessary to
6204 deduce information about a private key may be performed offline. The amount
6205 of resources required for such an attack would be very significant and
6206 likely only accessible to a limited number of attackers. An attacker would
6207 additionally need online access to an unpatched system using the target
6208 private key in a scenario with persistent DH parameters and a private
6209 key that is shared between multiple clients.
6211 This only affects processors that support the BMI1, BMI2 and ADX extensions
6212 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6214 This issue was reported to OpenSSL by the OSS-Fuzz project.
6219 * Malformed X.509 IPAddressFamily could cause OOB read
6221 If an X.509 certificate has a malformed IPAddressFamily extension,
6222 OpenSSL could do a one-byte buffer overread. The most likely result
6223 would be an erroneous display of the certificate in text format.
6225 This issue was reported to OpenSSL by the OSS-Fuzz project.
6229 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6231 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6232 platform rather than 'mingw'.
6236 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6238 * Truncated packet could crash via OOB read
6240 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6241 cipher is being used, then a truncated packet can cause that host to
6242 perform an out-of-bounds read, usually resulting in a crash.
6244 This issue was reported to OpenSSL by Robert Święcki of Google.
6249 * BN_mod_exp may produce incorrect results on x86_64
6251 There is a carry propagating bug in the x86_64 Montgomery squaring
6252 procedure. No EC algorithms are affected. Analysis suggests that attacks
6253 against RSA and DSA as a result of this defect would be very difficult to
6254 perform and are not believed likely. Attacks against DH are considered just
6255 feasible (although very difficult) because most of the work necessary to
6256 deduce information about a private key may be performed offline. The amount
6257 of resources required for such an attack would be very significant and
6258 likely only accessible to a limited number of attackers. An attacker would
6259 additionally need online access to an unpatched system using the target
6260 private key in a scenario with persistent DH parameters and a private
6261 key that is shared between multiple clients. For example this can occur by
6262 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6263 similar to CVE-2015-3193 but must be treated as a separate problem.
6265 This issue was reported to OpenSSL by the OSS-Fuzz project.
6270 * Montgomery multiplication may produce incorrect results
6272 There is a carry propagating bug in the Broadwell-specific Montgomery
6273 multiplication procedure that handles input lengths divisible by, but
6274 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6275 and DH private keys are impossible. This is because the subroutine in
6276 question is not used in operations with the private key itself and an input
6277 of the attacker's direct choice. Otherwise the bug can manifest itself as
6278 transient authentication and key negotiation failures or reproducible
6279 erroneous outcome of public-key operations with specially crafted input.
6280 Among EC algorithms only Brainpool P-512 curves are affected and one
6281 presumably can attack ECDH key negotiation. Impact was not analyzed in
6282 detail, because pre-requisites for attack are considered unlikely. Namely
6283 multiple clients have to choose the curve in question and the server has to
6284 share the private key among them, neither of which is default behaviour.
6285 Even then only clients that chose the curve will be affected.
6287 This issue was publicly reported as transient failures and was not
6288 initially recognized as a security issue. Thanks to Richard Morgan for
6289 providing reproducible case.
6294 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6295 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6296 prevent issues where no progress is being made and the peer continually
6297 sends unrecognised record types, using up resources processing them.
6301 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6303 * Missing CRL sanity check
6305 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6306 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6307 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6309 This issue only affects the OpenSSL 1.0.2i
6314 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6316 * OCSP Status Request extension unbounded memory growth
6318 A malicious client can send an excessively large OCSP Status Request
6319 extension. If that client continually requests renegotiation, sending a
6320 large OCSP Status Request extension each time, then there will be unbounded
6321 memory growth on the server. This will eventually lead to a Denial Of
6322 Service attack through memory exhaustion. Servers with a default
6323 configuration are vulnerable even if they do not support OCSP. Builds using
6324 the "no-ocsp" build time option are not affected.
6326 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6331 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6334 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6340 * OOB write in MDC2_Update()
6342 An overflow can occur in MDC2_Update() either if called directly or
6343 through the EVP_DigestUpdate() function using MDC2. If an attacker
6344 is able to supply very large amounts of input data after a previous
6345 call to EVP_EncryptUpdate() with a partial block then a length check
6346 can overflow resulting in a heap corruption.
6348 The amount of data needed is comparable to SIZE_MAX which is impractical
6351 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6356 * Malformed SHA512 ticket DoS
6358 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6359 DoS attack where a malformed ticket will result in an OOB read which will
6362 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6363 a custom server callback and ticket lookup mechanism.
6365 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6370 * OOB write in BN_bn2dec()
6372 The function BN_bn2dec() does not check the return value of BN_div_word().
6373 This can cause an OOB write if an application uses this function with an
6374 overly large BIGNUM. This could be a problem if an overly large certificate
6375 or CRL is printed out from an untrusted source. TLS is not affected because
6376 record limits will reject an oversized certificate before it is parsed.
6378 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6383 * OOB read in TS_OBJ_print_bio()
6385 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6386 the total length the OID text representation would use and not the amount
6387 of data written. This will result in OOB reads when large OIDs are
6390 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6395 * Pointer arithmetic undefined behaviour
6397 Avoid some undefined pointer arithmetic
6399 A common idiom in the codebase is to check limits in the following manner:
6402 Where "p" points to some malloc'd data of SIZE bytes and
6405 "len" here could be from some externally supplied data (e.g. from a TLS
6408 The rules of C pointer arithmetic are such that "p + len" is only well
6409 defined where len <= SIZE. Therefore the above idiom is actually
6410 undefined behaviour.
6412 For example this could cause problems if some malloc implementation
6413 provides an address for "p" such that "p + len" actually overflows for
6414 values of len that are too big and therefore p + len < limit.
6416 This issue was reported to OpenSSL by Guido Vranken
6421 * Constant time flag not preserved in DSA signing
6423 Operations in the DSA signing algorithm should run in constant time in
6424 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6425 implementation means that a non-constant time codepath is followed for
6426 certain operations. This has been demonstrated through a cache-timing
6427 attack to be sufficient for an attacker to recover the private DSA key.
6429 This issue was reported by César Pereida (Aalto University), Billy Brumley
6430 (Tampere University of Technology), and Yuval Yarom (The University of
6431 Adelaide and NICTA).
6436 * DTLS buffered message DoS
6438 In a DTLS connection where handshake messages are delivered out-of-order
6439 those messages that OpenSSL is not yet ready to process will be buffered
6440 for later use. Under certain circumstances, a flaw in the logic means that
6441 those messages do not get removed from the buffer even though the handshake
6442 has been completed. An attacker could force up to approx. 15 messages to
6443 remain in the buffer when they are no longer required. These messages will
6444 be cleared when the DTLS connection is closed. The default maximum size for
6445 a message is 100k. Therefore, the attacker could force an additional 1500k
6446 to be consumed per connection. By opening many simultaneous connections an
6447 attacker could cause a DoS attack through memory exhaustion.
6449 This issue was reported to OpenSSL by Quan Luo.
6454 * DTLS replay protection DoS
6456 A flaw in the DTLS replay attack protection mechanism means that records
6457 that arrive for future epochs update the replay protection "window" before
6458 the MAC for the record has been validated. This could be exploited by an
6459 attacker by sending a record for the next epoch (which does not have to
6460 decrypt or have a valid MAC), with a very large sequence number. This means
6461 that all subsequent legitimate packets are dropped causing a denial of
6462 service for a specific DTLS connection.
6464 This issue was reported to OpenSSL by the OCAP audit team.
6469 * Certificate message OOB reads
6471 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6472 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6473 theoretical DoS risk but this has not been observed in practice on common
6476 The messages affected are client certificate, client certificate request
6477 and server certificate. As a result the attack can only be performed
6478 against a client or a server which enables client authentication.
6480 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6485 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6487 * Prevent padding oracle in AES-NI CBC MAC check
6489 A MITM attacker can use a padding oracle attack to decrypt traffic
6490 when the connection uses an AES CBC cipher and the server support
6493 This issue was introduced as part of the fix for Lucky 13 padding
6494 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6495 constant time by making sure that always the same bytes are read and
6496 compared against either the MAC or padding bytes. But it no longer
6497 checked that there was enough data to have both the MAC and padding
6500 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6504 * Fix EVP_EncodeUpdate overflow
6506 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6507 Base64 encoding of binary data. If an attacker is able to supply very large
6508 amounts of input data then a length check can overflow resulting in a heap
6511 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6512 the `PEM_write_bio*` family of functions. These are mainly used within the
6513 OpenSSL command line applications, so any application which processes data
6514 from an untrusted source and outputs it as a PEM file should be considered
6515 vulnerable to this issue. User applications that call these APIs directly
6516 with large amounts of untrusted data may also be vulnerable.
6518 This issue was reported by Guido Vranken.
6523 * Fix EVP_EncryptUpdate overflow
6525 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6526 is able to supply very large amounts of input data after a previous call to
6527 EVP_EncryptUpdate() with a partial block then a length check can overflow
6528 resulting in a heap corruption. Following an analysis of all OpenSSL
6529 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6530 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6531 the first called function after an EVP_EncryptInit(), and therefore that
6532 specific call must be safe. The second form is where the length passed to
6533 EVP_EncryptUpdate() can be seen from the code to be some small value and
6534 therefore there is no possibility of an overflow. Since all instances are
6535 one of these two forms, it is believed that there can be no overflows in
6536 internal code due to this problem. It should be noted that
6537 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6538 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6539 of these calls have also been analysed too and it is believed there are no
6540 instances in internal usage where an overflow could occur.
6542 This issue was reported by Guido Vranken.
6547 * Prevent ASN.1 BIO excessive memory allocation
6549 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6550 a short invalid encoding can cause allocation of large amounts of memory
6551 potentially consuming excessive resources or exhausting memory.
6553 Any application parsing untrusted data through d2i BIO functions is
6554 affected. The memory based functions such as d2i_X509() are *not* affected.
6555 Since the memory based functions are used by the TLS library, TLS
6556 applications are not affected.
6558 This issue was reported by Brian Carpenter.
6565 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6566 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6567 in arbitrary stack data being returned in the buffer.
6569 This issue was reported by Guido Vranken.
6574 * Modify behavior of ALPN to invoke callback after SNI/servername
6575 callback, such that updates to the SSL_CTX affect ALPN.
6579 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6584 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6585 methods are enabled and ssl2 is disabled the methods return NULL.
6589 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6591 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6592 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6593 provide any "EXPORT" or "LOW" strength ciphers.
6597 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6598 is by default disabled at build-time. Builds that are not configured with
6599 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6600 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6601 will need to explicitly call either of:
6603 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6605 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6607 as appropriate. Even if either of those is used, or the application
6608 explicitly uses the version-specific SSLv2_method() or its client and
6609 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6610 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6611 ciphers, and SSLv2 56-bit DES are no longer available.
6616 * Fix a double-free in DSA code
6618 A double free bug was discovered when OpenSSL parses malformed DSA private
6619 keys and could lead to a DoS attack or memory corruption for applications
6620 that receive DSA private keys from untrusted sources. This scenario is
6623 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6629 * Disable SRP fake user seed to address a server memory leak.
6631 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6633 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6634 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6635 was changed to ignore the "fake user" SRP seed, even if the seed
6638 Users should use SRP_VBASE_get1_by_user instead. Note that in
6639 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6640 also that even though configuring the SRP seed attempts to hide
6641 invalid usernames by continuing the handshake with fake
6642 credentials, this behaviour is not constant time and no strong
6643 guarantees are made that the handshake is indistinguishable from
6644 that of a valid user.
6649 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6651 In the BN_hex2bn function the number of hex digits is calculated using an
6652 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6653 large values of `i` this can result in `bn_expand` not allocating any
6654 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6655 field as NULL leading to a subsequent NULL ptr deref. For very large values
6656 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6657 In this case memory is allocated to the internal BIGNUM data field, but it
6658 is insufficiently sized leading to heap corruption. A similar issue exists
6659 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6660 is ever called by user applications with very large untrusted hex/dec data.
6661 This is anticipated to be a rare occurrence.
6663 All OpenSSL internal usage of these functions use data that is not expected
6664 to be untrusted, e.g. config file data or application command line
6665 arguments. If user developed applications generate config file data based
6666 on untrusted data then it is possible that this could also lead to security
6667 consequences. This is also anticipated to be rare.
6669 This issue was reported to OpenSSL by Guido Vranken.
6674 * Fix memory issues in `BIO_*printf` functions
6676 The internal `fmtstr` function used in processing a "%s" format string in
6677 the `BIO_*printf` functions could overflow while calculating the length of a
6678 string and cause an OOB read when printing very long strings.
6680 Additionally the internal `doapr_outch` function can attempt to write to an
6681 OOB memory location (at an offset from the NULL pointer) in the event of a
6682 memory allocation failure. In 1.0.2 and below this could be caused where
6683 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6684 could be in processing a very long "%s" format string. Memory leaks can
6687 The first issue may mask the second issue dependent on compiler behaviour.
6688 These problems could enable attacks where large amounts of untrusted data
6689 is passed to the `BIO_*printf` functions. If applications use these functions
6690 in this way then they could be vulnerable. OpenSSL itself uses these
6691 functions when printing out human-readable dumps of ASN.1 data. Therefore
6692 applications that print this data could be vulnerable if the data is from
6693 untrusted sources. OpenSSL command line applications could also be
6694 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6695 as command line arguments.
6697 Libssl is not considered directly vulnerable. Additionally certificates etc
6698 received via remote connections via libssl are also unlikely to be able to
6699 trigger these issues because of message size limits enforced within libssl.
6701 This issue was reported to OpenSSL Guido Vranken.
6706 * Side channel attack on modular exponentiation
6708 A side-channel attack was found which makes use of cache-bank conflicts on
6709 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6710 of RSA keys. The ability to exploit this issue is limited as it relies on
6711 an attacker who has control of code in a thread running on the same
6712 hyper-threaded core as the victim thread which is performing decryptions.
6714 This issue was reported to OpenSSL by Yuval Yarom, The University of
6715 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6716 Nadia Heninger, University of Pennsylvania with more information at
6717 <http://cachebleed.info>.
6722 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6723 if no keysize is specified with default_bits. This fixes an
6724 omission in an earlier change that changed all RSA/DSA key generation
6725 commands to use 2048 bits by default.
6729 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6731 * DH small subgroups
6733 Historically OpenSSL only ever generated DH parameters based on "safe"
6734 primes. More recently (in version 1.0.2) support was provided for
6735 generating X9.42 style parameter files such as those required for RFC 5114
6736 support. The primes used in such files may not be "safe". Where an
6737 application is using DH configured with parameters based on primes that are
6738 not "safe" then an attacker could use this fact to find a peer's private
6739 DH exponent. This attack requires that the attacker complete multiple
6740 handshakes in which the peer uses the same private DH exponent. For example
6741 this could be used to discover a TLS server's private DH exponent if it's
6742 reusing the private DH exponent or it's using a static DH ciphersuite.
6744 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6745 TLS. It is not on by default. If the option is not set then the server
6746 reuses the same private DH exponent for the life of the server process and
6747 would be vulnerable to this attack. It is believed that many popular
6748 applications do set this option and would therefore not be at risk.
6750 The fix for this issue adds an additional check where a "q" parameter is
6751 available (as is the case in X9.42 based parameters). This detects the
6752 only known attack, and is the only possible defense for static DH
6753 ciphersuites. This could have some performance impact.
6755 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6756 default and cannot be disabled. This could have some performance impact.
6758 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6763 * SSLv2 doesn't block disabled ciphers
6765 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6766 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6767 been disabled, provided that the SSLv2 protocol was not also disabled via
6770 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6771 and Sebastian Schinzel.
6776 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6778 * BN_mod_exp may produce incorrect results on x86_64
6780 There is a carry propagating bug in the x86_64 Montgomery squaring
6781 procedure. No EC algorithms are affected. Analysis suggests that attacks
6782 against RSA and DSA as a result of this defect would be very difficult to
6783 perform and are not believed likely. Attacks against DH are considered just
6784 feasible (although very difficult) because most of the work necessary to
6785 deduce information about a private key may be performed offline. The amount
6786 of resources required for such an attack would be very significant and
6787 likely only accessible to a limited number of attackers. An attacker would
6788 additionally need online access to an unpatched system using the target
6789 private key in a scenario with persistent DH parameters and a private
6790 key that is shared between multiple clients. For example this can occur by
6791 default in OpenSSL DHE based SSL/TLS ciphersuites.
6793 This issue was reported to OpenSSL by Hanno Böck.
6798 * Certificate verify crash with missing PSS parameter
6800 The signature verification routines will crash with a NULL pointer
6801 dereference if presented with an ASN.1 signature using the RSA PSS
6802 algorithm and absent mask generation function parameter. Since these
6803 routines are used to verify certificate signature algorithms this can be
6804 used to crash any certificate verification operation and exploited in a
6805 DoS attack. Any application which performs certificate verification is
6806 vulnerable including OpenSSL clients and servers which enable client
6809 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6814 * X509_ATTRIBUTE memory leak
6816 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6817 memory. This structure is used by the PKCS#7 and CMS routines so any
6818 application which reads PKCS#7 or CMS data from untrusted sources is
6819 affected. SSL/TLS is not affected.
6821 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6827 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6828 This changes the decoding behaviour for some invalid messages,
6829 though the change is mostly in the more lenient direction, and
6830 legacy behaviour is preserved as much as possible.
6834 * In DSA_generate_parameters_ex, if the provided seed is too short,
6837 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6839 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6841 * Alternate chains certificate forgery
6843 During certificate verification, OpenSSL will attempt to find an
6844 alternative certificate chain if the first attempt to build such a chain
6845 fails. An error in the implementation of this logic can mean that an
6846 attacker could cause certain checks on untrusted certificates to be
6847 bypassed, such as the CA flag, enabling them to use a valid leaf
6848 certificate to act as a CA and "issue" an invalid certificate.
6850 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6855 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6857 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6858 incompatibility in the handling of HMAC. The previous ABI has now been
6863 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6865 * Malformed ECParameters causes infinite loop
6867 When processing an ECParameters structure OpenSSL enters an infinite loop
6868 if the curve specified is over a specially malformed binary polynomial
6871 This can be used to perform denial of service against any
6872 system which processes public keys, certificate requests or
6873 certificates. This includes TLS clients and TLS servers with
6874 client authentication enabled.
6876 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6881 * Exploitable out-of-bounds read in X509_cmp_time
6883 X509_cmp_time does not properly check the length of the ASN1_TIME
6884 string and can read a few bytes out of bounds. In addition,
6885 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6888 An attacker can use this to craft malformed certificates and CRLs of
6889 various sizes and potentially cause a segmentation fault, resulting in
6890 a DoS on applications that verify certificates or CRLs. TLS clients
6891 that verify CRLs are affected. TLS clients and servers with client
6892 authentication enabled may be affected if they use custom verification
6895 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6896 independently by Hanno Böck.
6901 * PKCS7 crash with missing EnvelopedContent
6903 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6904 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6905 with missing content and trigger a NULL pointer dereference on parsing.
6907 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6908 structures from untrusted sources are affected. OpenSSL clients and
6909 servers are not affected.
6911 This issue was reported to OpenSSL by Michal Zalewski (Google).
6916 * CMS verify infinite loop with unknown hash function
6918 When verifying a signedData message the CMS code can enter an infinite loop
6919 if presented with an unknown hash function OID. This can be used to perform
6920 denial of service against any system which verifies signedData messages using
6922 This issue was reported to OpenSSL by Johannes Bauer.
6927 * Race condition handling NewSessionTicket
6929 If a NewSessionTicket is received by a multi-threaded client when attempting to
6930 reuse a previous ticket then a race condition can occur potentially leading to
6931 a double free of the ticket data.
6936 * Only support 256-bit or stronger elliptic curves with the
6937 'ecdh_auto' setting (server) or by default (client). Of supported
6938 curves, prefer P-256 (both).
6942 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6944 * ClientHello sigalgs DoS fix
6946 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6947 invalid signature algorithms extension a NULL pointer dereference will
6948 occur. This can be exploited in a DoS attack against the server.
6950 This issue was was reported to OpenSSL by David Ramos of Stanford
6954 *Stephen Henson and Matt Caswell*
6956 * Multiblock corrupted pointer fix
6958 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6959 feature only applies on 64 bit x86 architecture platforms that support AES
6960 NI instructions. A defect in the implementation of "multiblock" can cause
6961 OpenSSL's internal write buffer to become incorrectly set to NULL when
6962 using non-blocking IO. Typically, when the user application is using a
6963 socket BIO for writing, this will only result in a failed connection.
6964 However if some other BIO is used then it is likely that a segmentation
6965 fault will be triggered, thus enabling a potential DoS attack.
6967 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6972 * Segmentation fault in DTLSv1_listen fix
6974 The DTLSv1_listen function is intended to be stateless and processes the
6975 initial ClientHello from many peers. It is common for user code to loop
6976 over the call to DTLSv1_listen until a valid ClientHello is received with
6977 an associated cookie. A defect in the implementation of DTLSv1_listen means
6978 that state is preserved in the SSL object from one invocation to the next
6979 that can lead to a segmentation fault. Errors processing the initial
6980 ClientHello can trigger this scenario. An example of such an error could be
6981 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6984 This issue was reported to OpenSSL by Per Allansson.
6989 * Segmentation fault in ASN1_TYPE_cmp fix
6991 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6992 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6993 certificate signature algorithm consistency this can be used to crash any
6994 certificate verification operation and exploited in a DoS attack. Any
6995 application which performs certificate verification is vulnerable including
6996 OpenSSL clients and servers which enable client authentication.
7001 * Segmentation fault for invalid PSS parameters fix
7003 The signature verification routines will crash with a NULL pointer
7004 dereference if presented with an ASN.1 signature using the RSA PSS
7005 algorithm and invalid parameters. Since these routines are used to verify
7006 certificate signature algorithms this can be used to crash any
7007 certificate verification operation and exploited in a DoS attack. Any
7008 application which performs certificate verification is vulnerable including
7009 OpenSSL clients and servers which enable client authentication.
7011 This issue was was reported to OpenSSL by Brian Carpenter.
7016 * ASN.1 structure reuse memory corruption fix
7018 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7019 memory corruption via an invalid write. Such reuse is and has been
7020 strongly discouraged and is believed to be rare.
7022 Applications that parse structures containing CHOICE or ANY DEFINED BY
7023 components may be affected. Certificate parsing (d2i_X509 and related
7024 functions) are however not affected. OpenSSL clients and servers are
7030 * PKCS7 NULL pointer dereferences fix
7032 The PKCS#7 parsing code does not handle missing outer ContentInfo
7033 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7034 missing content and trigger a NULL pointer dereference on parsing.
7036 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7037 otherwise parse PKCS#7 structures from untrusted sources are
7038 affected. OpenSSL clients and servers are not affected.
7040 This issue was reported to OpenSSL by Michal Zalewski (Google).
7045 * DoS via reachable assert in SSLv2 servers fix
7047 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7048 servers that both support SSLv2 and enable export cipher suites by sending
7049 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7051 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7052 (OpenSSL development team).
7057 * Empty CKE with client auth and DHE fix
7059 If client auth is used then a server can seg fault in the event of a DHE
7060 ciphersuite being selected and a zero length ClientKeyExchange message
7061 being sent by the client. This could be exploited in a DoS attack.
7066 * Handshake with unseeded PRNG fix
7068 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
7069 with an unseeded PRNG. The conditions are:
7070 - The client is on a platform where the PRNG has not been seeded
7071 automatically, and the user has not seeded manually
7072 - A protocol specific client method version has been used (i.e. not
7073 SSL_client_methodv23)
7074 - A ciphersuite is used that does not require additional random data from
7075 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
7077 If the handshake succeeds then the client random that has been used will
7078 have been generated from a PRNG with insufficient entropy and therefore the
7079 output may be predictable.
7081 For example using the following command with an unseeded openssl will
7082 succeed on an unpatched platform:
7084 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7089 * Use After Free following d2i_ECPrivatekey error fix
7091 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7092 could cause a use after free condition. This, in turn, could cause a double
7093 free in several private key parsing functions (such as d2i_PrivateKey
7094 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7095 for applications that receive EC private keys from untrusted
7096 sources. This scenario is considered rare.
7098 This issue was discovered by the BoringSSL project and fixed in their
7104 * X509_to_X509_REQ NULL pointer deref fix
7106 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7107 the certificate key is invalid. This function is rarely used in practice.
7109 This issue was discovered by Brian Carpenter.
7114 * Removed the export ciphers from the DEFAULT ciphers
7118 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7120 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7121 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7122 So far those who have to target multiple platforms would compromise
7123 and argue that binary targeting say ARMv5 would still execute on
7124 ARMv8. "Universal" build resolves this compromise by providing
7125 near-optimal performance even on newer platforms.
7129 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7130 (other platforms pending).
7132 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7134 * Add support for the SignedCertificateTimestampList certificate and
7135 OCSP response extensions from RFC6962.
7139 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7140 for corner cases. (Certain input points at infinity could lead to
7141 bogus results, with non-infinity inputs mapped to infinity too.)
7145 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7146 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7147 common cases are optimized and there still is room for further
7148 improvements. Vector Permutation AES for Altivec is also added.
7152 * Add support for little-endian ppc64 Linux target.
7154 *Marcelo Cerri (IBM)*
7156 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7157 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7158 are optimized and there still is room for further improvements.
7159 Both 32- and 64-bit modes are supported.
7161 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7163 * Improved ARMv7 NEON support.
7167 * Support for SPARC Architecture 2011 crypto extensions, first
7168 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7169 SHA256/512, MD5, GHASH and modular exponentiation.
7171 *Andy Polyakov, David Miller*
7173 * Accelerated modular exponentiation for Intel processors, a.k.a.
7176 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7178 * Support for new and upcoming Intel processors, including AVX2,
7179 BMI and SHA ISA extensions. This includes additional "stitched"
7180 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7183 This work was sponsored by Intel Corp.
7187 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7188 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7189 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7193 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7194 this fixes a limitation in previous versions of OpenSSL.
7198 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7199 MGF1 digest and OAEP label.
7203 * Add EVP support for key wrapping algorithms, to avoid problems with
7204 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7205 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7206 algorithms and include tests cases.
7210 * Add functions to allocate and set the fields of an ECDSA_METHOD
7213 *Douglas E. Engert, Steve Henson*
7215 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7216 difference in days and seconds between two tm or ASN1_TIME structures.
7220 * Add -rev test option to s_server to just reverse order of characters
7221 received by client and send back to server. Also prints an abbreviated
7222 summary of the connection parameters.
7226 * New option -brief for s_client and s_server to print out a brief summary
7227 of connection parameters.
7231 * Add callbacks for arbitrary TLS extensions.
7233 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7235 * New option -crl_download in several openssl utilities to download CRLs
7236 from CRLDP extension in certificates.
7240 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7244 * New function X509_CRL_diff to generate a delta CRL from the difference
7245 of two full CRLs. Add support to "crl" utility.
7249 * New functions to set lookup_crls function and to retrieve
7250 X509_STORE from X509_STORE_CTX.
7254 * Print out deprecated issuer and subject unique ID fields in
7259 * Extend OCSP I/O functions so they can be used for simple general purpose
7260 HTTP as well as OCSP. New wrapper function which can be used to download
7261 CRLs using the OCSP API.
7265 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7269 * `SSL_CONF*` functions. These provide a common framework for application
7270 configuration using configuration files or command lines.
7274 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7275 message callback and prints the results. Needs compile time option
7276 "enable-ssl-trace". New options to s_client and s_server to enable
7281 * New ctrl and macro to retrieve supported points extensions.
7282 Print out extension in s_server and s_client.
7286 * New functions to retrieve certificate signature and signature
7291 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7296 * New Suite B modes for TLS code. These use and enforce the requirements
7297 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7298 only use Suite B curves. The Suite B modes can be set by using the
7299 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7303 * New chain verification flags for Suite B levels of security. Check
7304 algorithms are acceptable when flags are set in X509_verify_cert.
7308 * Make tls1_check_chain return a set of flags indicating checks passed
7309 by a certificate chain. Add additional tests to handle client
7310 certificates: checks for matching certificate type and issuer name
7315 * If an attempt is made to use a signature algorithm not in the peer
7316 preference list abort the handshake. If client has no suitable
7317 signature algorithms in response to a certificate request do not
7318 use the certificate.
7322 * If server EC tmp key is not in client preference list abort handshake.
7326 * Add support for certificate stores in CERT structure. This makes it
7327 possible to have different stores per SSL structure or one store in
7328 the parent SSL_CTX. Include distinct stores for certificate chain
7329 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7330 to build and store a certificate chain in CERT structure: returning
7331 an error if the chain cannot be built: this will allow applications
7332 to test if a chain is correctly configured.
7334 Note: if the CERT based stores are not set then the parent SSL_CTX
7335 store is used to retain compatibility with existing behaviour.
7339 * New function ssl_set_client_disabled to set a ciphersuite disabled
7340 mask based on the current session, check mask when sending client
7341 hello and checking the requested ciphersuite.
7345 * New ctrls to retrieve and set certificate types in a certificate
7346 request message. Print out received values in s_client. If certificate
7347 types is not set with custom values set sensible values based on
7348 supported signature algorithms.
7352 * Support for distinct client and server supported signature algorithms.
7356 * Add certificate callback. If set this is called whenever a certificate
7357 is required by client or server. An application can decide which
7358 certificate chain to present based on arbitrary criteria: for example
7359 supported signature algorithms. Add very simple example to s_server.
7360 This fixes many of the problems and restrictions of the existing client
7361 certificate callback: for example you can now clear an existing
7362 certificate and specify the whole chain.
7366 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7367 the certificate can be used for (if anything). Set valid_flags field
7368 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7369 to have similar checks in it.
7371 Add new "cert_flags" field to CERT structure and include a "strict mode".
7372 This enforces some TLS certificate requirements (such as only permitting
7373 certificate signature algorithms contained in the supported algorithms
7374 extension) which some implementations ignore: this option should be used
7375 with caution as it could cause interoperability issues.
7379 * Update and tidy signature algorithm extension processing. Work out
7380 shared signature algorithms based on preferences and peer algorithms
7381 and print them out in s_client and s_server. Abort handshake if no
7382 shared signature algorithms.
7386 * Add new functions to allow customised supported signature algorithms
7387 for SSL and SSL_CTX structures. Add options to s_client and s_server
7392 * New function SSL_certs_clear() to delete all references to certificates
7393 from an SSL structure. Before this once a certificate had been added
7394 it couldn't be removed.
7398 * Integrate hostname, email address and IP address checking with certificate
7399 verification. New verify options supporting checking in openssl utility.
7403 * Fixes and wildcard matching support to hostname and email checking
7404 functions. Add manual page.
7406 *Florian Weimer (Red Hat Product Security Team)*
7408 * New functions to check a hostname email or IP address against a
7409 certificate. Add options x509 utility to print results of checks against
7414 * Fix OCSP checking.
7416 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7418 * Initial experimental support for explicitly trusted non-root CAs.
7419 OpenSSL still tries to build a complete chain to a root but if an
7420 intermediate CA has a trust setting included that is used. The first
7421 setting is used: whether to trust (e.g., -addtrust option to the x509
7426 * Add -trusted_first option which attempts to find certificates in the
7427 trusted store even if an untrusted chain is also supplied.
7431 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7432 platform support for Linux and Android.
7436 * Support for linux-x32, ILP32 environment in x86_64 framework.
7440 * Experimental multi-implementation support for FIPS capable OpenSSL.
7441 When in FIPS mode the approved implementations are used as normal,
7442 when not in FIPS mode the internal unapproved versions are used instead.
7443 This means that the FIPS capable OpenSSL isn't forced to use the
7444 (often lower performance) FIPS implementations outside FIPS mode.
7448 * Transparently support X9.42 DH parameters when calling
7449 PEM_read_bio_DHparameters. This means existing applications can handle
7450 the new parameter format automatically.
7454 * Initial experimental support for X9.42 DH parameter format: mainly
7455 to support use of 'q' parameter for RFC5114 parameters.
7459 * Add DH parameters from RFC5114 including test data to dhtest.
7463 * Support for automatic EC temporary key parameter selection. If enabled
7464 the most preferred EC parameters are automatically used instead of
7465 hardcoded fixed parameters. Now a server just has to call:
7466 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7467 support ECDH and use the most appropriate parameters.
7471 * Enhance and tidy EC curve and point format TLS extension code. Use
7472 static structures instead of allocation if default values are used.
7473 New ctrls to set curves we wish to support and to retrieve shared curves.
7474 Print out shared curves in s_server. New options to s_server and s_client
7475 to set list of supported curves.
7479 * New ctrls to retrieve supported signature algorithms and
7480 supported curve values as an array of NIDs. Extend openssl utility
7481 to print out received values.
7485 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7486 between NIDs and the more common NIST names such as "P-256". Enhance
7487 ecparam utility and ECC method to recognise the NIST names for curves.
7491 * Enhance SSL/TLS certificate chain handling to support different
7492 chains for each certificate instead of one chain in the parent SSL_CTX.
7496 * Support for fixed DH ciphersuite client authentication: where both
7497 server and client use DH certificates with common parameters.
7501 * Support for fixed DH ciphersuites: those requiring DH server
7506 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7508 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7509 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7510 X509_CINF_get_signature were reverted post internal team review.
7515 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7517 * OCSP Status Request extension unbounded memory growth
7519 A malicious client can send an excessively large OCSP Status Request
7520 extension. If that client continually requests renegotiation, sending a
7521 large OCSP Status Request extension each time, then there will be unbounded
7522 memory growth on the server. This will eventually lead to a Denial Of
7523 Service attack through memory exhaustion. Servers with a default
7524 configuration are vulnerable even if they do not support OCSP. Builds using
7525 the "no-ocsp" build time option are not affected.
7527 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7532 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7535 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7541 * OOB write in MDC2_Update()
7543 An overflow can occur in MDC2_Update() either if called directly or
7544 through the EVP_DigestUpdate() function using MDC2. If an attacker
7545 is able to supply very large amounts of input data after a previous
7546 call to EVP_EncryptUpdate() with a partial block then a length check
7547 can overflow resulting in a heap corruption.
7549 The amount of data needed is comparable to SIZE_MAX which is impractical
7552 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7557 * Malformed SHA512 ticket DoS
7559 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7560 DoS attack where a malformed ticket will result in an OOB read which will
7563 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7564 a custom server callback and ticket lookup mechanism.
7566 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7571 * OOB write in BN_bn2dec()
7573 The function BN_bn2dec() does not check the return value of BN_div_word().
7574 This can cause an OOB write if an application uses this function with an
7575 overly large BIGNUM. This could be a problem if an overly large certificate
7576 or CRL is printed out from an untrusted source. TLS is not affected because
7577 record limits will reject an oversized certificate before it is parsed.
7579 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7584 * OOB read in TS_OBJ_print_bio()
7586 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7587 the total length the OID text representation would use and not the amount
7588 of data written. This will result in OOB reads when large OIDs are
7591 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7596 * Pointer arithmetic undefined behaviour
7598 Avoid some undefined pointer arithmetic
7600 A common idiom in the codebase is to check limits in the following manner:
7603 Where "p" points to some malloc'd data of SIZE bytes and
7606 "len" here could be from some externally supplied data (e.g. from a TLS
7609 The rules of C pointer arithmetic are such that "p + len" is only well
7610 defined where len <= SIZE. Therefore, the above idiom is actually
7611 undefined behaviour.
7613 For example this could cause problems if some malloc implementation
7614 provides an address for "p" such that "p + len" actually overflows for
7615 values of len that are too big and therefore p + len < limit.
7617 This issue was reported to OpenSSL by Guido Vranken
7622 * Constant time flag not preserved in DSA signing
7624 Operations in the DSA signing algorithm should run in constant time in
7625 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7626 implementation means that a non-constant time codepath is followed for
7627 certain operations. This has been demonstrated through a cache-timing
7628 attack to be sufficient for an attacker to recover the private DSA key.
7630 This issue was reported by César Pereida (Aalto University), Billy Brumley
7631 (Tampere University of Technology), and Yuval Yarom (The University of
7632 Adelaide and NICTA).
7637 * DTLS buffered message DoS
7639 In a DTLS connection where handshake messages are delivered out-of-order
7640 those messages that OpenSSL is not yet ready to process will be buffered
7641 for later use. Under certain circumstances, a flaw in the logic means that
7642 those messages do not get removed from the buffer even though the handshake
7643 has been completed. An attacker could force up to approx. 15 messages to
7644 remain in the buffer when they are no longer required. These messages will
7645 be cleared when the DTLS connection is closed. The default maximum size for
7646 a message is 100k. Therefore, the attacker could force an additional 1500k
7647 to be consumed per connection. By opening many simultaneous connections an
7648 attacker could cause a DoS attack through memory exhaustion.
7650 This issue was reported to OpenSSL by Quan Luo.
7655 * DTLS replay protection DoS
7657 A flaw in the DTLS replay attack protection mechanism means that records
7658 that arrive for future epochs update the replay protection "window" before
7659 the MAC for the record has been validated. This could be exploited by an
7660 attacker by sending a record for the next epoch (which does not have to
7661 decrypt or have a valid MAC), with a very large sequence number. This means
7662 that all subsequent legitimate packets are dropped causing a denial of
7663 service for a specific DTLS connection.
7665 This issue was reported to OpenSSL by the OCAP audit team.
7670 * Certificate message OOB reads
7672 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7673 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7674 theoretical DoS risk but this has not been observed in practice on common
7677 The messages affected are client certificate, client certificate request
7678 and server certificate. As a result the attack can only be performed
7679 against a client or a server which enables client authentication.
7681 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7686 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7688 * Prevent padding oracle in AES-NI CBC MAC check
7690 A MITM attacker can use a padding oracle attack to decrypt traffic
7691 when the connection uses an AES CBC cipher and the server support
7694 This issue was introduced as part of the fix for Lucky 13 padding
7695 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7696 constant time by making sure that always the same bytes are read and
7697 compared against either the MAC or padding bytes. But it no longer
7698 checked that there was enough data to have both the MAC and padding
7701 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7706 * Fix EVP_EncodeUpdate overflow
7708 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7709 Base64 encoding of binary data. If an attacker is able to supply very large
7710 amounts of input data then a length check can overflow resulting in a heap
7713 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7714 the `PEM_write_bio*` family of functions. These are mainly used within the
7715 OpenSSL command line applications, so any application which processes data
7716 from an untrusted source and outputs it as a PEM file should be considered
7717 vulnerable to this issue. User applications that call these APIs directly
7718 with large amounts of untrusted data may also be vulnerable.
7720 This issue was reported by Guido Vranken.
7725 * Fix EVP_EncryptUpdate overflow
7727 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7728 is able to supply very large amounts of input data after a previous call to
7729 EVP_EncryptUpdate() with a partial block then a length check can overflow
7730 resulting in a heap corruption. Following an analysis of all OpenSSL
7731 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7732 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7733 the first called function after an EVP_EncryptInit(), and therefore that
7734 specific call must be safe. The second form is where the length passed to
7735 EVP_EncryptUpdate() can be seen from the code to be some small value and
7736 therefore there is no possibility of an overflow. Since all instances are
7737 one of these two forms, it is believed that there can be no overflows in
7738 internal code due to this problem. It should be noted that
7739 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7740 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7741 of these calls have also been analysed too and it is believed there are no
7742 instances in internal usage where an overflow could occur.
7744 This issue was reported by Guido Vranken.
7749 * Prevent ASN.1 BIO excessive memory allocation
7751 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7752 a short invalid encoding can casuse allocation of large amounts of memory
7753 potentially consuming excessive resources or exhausting memory.
7755 Any application parsing untrusted data through d2i BIO functions is
7756 affected. The memory based functions such as d2i_X509() are *not* affected.
7757 Since the memory based functions are used by the TLS library, TLS
7758 applications are not affected.
7760 This issue was reported by Brian Carpenter.
7767 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7768 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7769 in arbitrary stack data being returned in the buffer.
7771 This issue was reported by Guido Vranken.
7776 * Modify behavior of ALPN to invoke callback after SNI/servername
7777 callback, such that updates to the SSL_CTX affect ALPN.
7781 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7786 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7787 methods are enabled and ssl2 is disabled the methods return NULL.
7791 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7793 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7794 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7795 provide any "EXPORT" or "LOW" strength ciphers.
7799 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7800 is by default disabled at build-time. Builds that are not configured with
7801 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7802 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7803 will need to explicitly call either of:
7805 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7807 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7809 as appropriate. Even if either of those is used, or the application
7810 explicitly uses the version-specific SSLv2_method() or its client and
7811 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7812 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7813 ciphers, and SSLv2 56-bit DES are no longer available.
7818 * Fix a double-free in DSA code
7820 A double free bug was discovered when OpenSSL parses malformed DSA private
7821 keys and could lead to a DoS attack or memory corruption for applications
7822 that receive DSA private keys from untrusted sources. This scenario is
7825 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7831 * Disable SRP fake user seed to address a server memory leak.
7833 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7835 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7836 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7837 was changed to ignore the "fake user" SRP seed, even if the seed
7840 Users should use SRP_VBASE_get1_by_user instead. Note that in
7841 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7842 also that even though configuring the SRP seed attempts to hide
7843 invalid usernames by continuing the handshake with fake
7844 credentials, this behaviour is not constant time and no strong
7845 guarantees are made that the handshake is indistinguishable from
7846 that of a valid user.
7851 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7853 In the BN_hex2bn function the number of hex digits is calculated using an
7854 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7855 large values of `i` this can result in `bn_expand` not allocating any
7856 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7857 field as NULL leading to a subsequent NULL ptr deref. For very large values
7858 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7859 In this case memory is allocated to the internal BIGNUM data field, but it
7860 is insufficiently sized leading to heap corruption. A similar issue exists
7861 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7862 is ever called by user applications with very large untrusted hex/dec data.
7863 This is anticipated to be a rare occurrence.
7865 All OpenSSL internal usage of these functions use data that is not expected
7866 to be untrusted, e.g. config file data or application command line
7867 arguments. If user developed applications generate config file data based
7868 on untrusted data then it is possible that this could also lead to security
7869 consequences. This is also anticipated to be rare.
7871 This issue was reported to OpenSSL by Guido Vranken.
7876 * Fix memory issues in `BIO_*printf` functions
7878 The internal `fmtstr` function used in processing a "%s" format string in
7879 the `BIO_*printf` functions could overflow while calculating the length of a
7880 string and cause an OOB read when printing very long strings.
7882 Additionally the internal `doapr_outch` function can attempt to write to an
7883 OOB memory location (at an offset from the NULL pointer) in the event of a
7884 memory allocation failure. In 1.0.2 and below this could be caused where
7885 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7886 could be in processing a very long "%s" format string. Memory leaks can
7889 The first issue may mask the second issue dependent on compiler behaviour.
7890 These problems could enable attacks where large amounts of untrusted data
7891 is passed to the `BIO_*printf` functions. If applications use these functions
7892 in this way then they could be vulnerable. OpenSSL itself uses these
7893 functions when printing out human-readable dumps of ASN.1 data. Therefore
7894 applications that print this data could be vulnerable if the data is from
7895 untrusted sources. OpenSSL command line applications could also be
7896 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7897 as command line arguments.
7899 Libssl is not considered directly vulnerable. Additionally certificates etc
7900 received via remote connections via libssl are also unlikely to be able to
7901 trigger these issues because of message size limits enforced within libssl.
7903 This issue was reported to OpenSSL Guido Vranken.
7908 * Side channel attack on modular exponentiation
7910 A side-channel attack was found which makes use of cache-bank conflicts on
7911 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7912 of RSA keys. The ability to exploit this issue is limited as it relies on
7913 an attacker who has control of code in a thread running on the same
7914 hyper-threaded core as the victim thread which is performing decryptions.
7916 This issue was reported to OpenSSL by Yuval Yarom, The University of
7917 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7918 Nadia Heninger, University of Pennsylvania with more information at
7919 <http://cachebleed.info>.
7924 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7925 if no keysize is specified with default_bits. This fixes an
7926 omission in an earlier change that changed all RSA/DSA key generation
7927 commands to use 2048 bits by default.
7931 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7933 * Protection for DH small subgroup attacks
7935 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7936 switched on by default and cannot be disabled. This could have some
7941 * SSLv2 doesn't block disabled ciphers
7943 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7944 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7945 been disabled, provided that the SSLv2 protocol was not also disabled via
7948 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7949 and Sebastian Schinzel.
7954 * Reject DH handshakes with parameters shorter than 1024 bits.
7958 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7960 * Certificate verify crash with missing PSS parameter
7962 The signature verification routines will crash with a NULL pointer
7963 dereference if presented with an ASN.1 signature using the RSA PSS
7964 algorithm and absent mask generation function parameter. Since these
7965 routines are used to verify certificate signature algorithms this can be
7966 used to crash any certificate verification operation and exploited in a
7967 DoS attack. Any application which performs certificate verification is
7968 vulnerable including OpenSSL clients and servers which enable client
7971 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7976 * X509_ATTRIBUTE memory leak
7978 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7979 memory. This structure is used by the PKCS#7 and CMS routines so any
7980 application which reads PKCS#7 or CMS data from untrusted sources is
7981 affected. SSL/TLS is not affected.
7983 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7989 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7990 This changes the decoding behaviour for some invalid messages,
7991 though the change is mostly in the more lenient direction, and
7992 legacy behaviour is preserved as much as possible.
7996 * In DSA_generate_parameters_ex, if the provided seed is too short,
7997 use a random seed, as already documented.
7999 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
8001 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
8003 * Alternate chains certificate forgery
8005 During certificate verification, OpenSSL will attempt to find an
8006 alternative certificate chain if the first attempt to build such a chain
8007 fails. An error in the implementation of this logic can mean that an
8008 attacker could cause certain checks on untrusted certificates to be
8009 bypassed, such as the CA flag, enabling them to use a valid leaf
8010 certificate to act as a CA and "issue" an invalid certificate.
8012 This issue was reported to OpenSSL by Adam Langley/David Benjamin
8018 * Race condition handling PSK identify hint
8020 If PSK identity hints are received by a multi-threaded client then
8021 the values are wrongly updated in the parent SSL_CTX structure. This can
8022 result in a race condition potentially leading to a double free of the
8028 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
8030 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
8031 incompatibility in the handling of HMAC. The previous ABI has now been
8034 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
8036 * Malformed ECParameters causes infinite loop
8038 When processing an ECParameters structure OpenSSL enters an infinite loop
8039 if the curve specified is over a specially malformed binary polynomial
8042 This can be used to perform denial of service against any
8043 system which processes public keys, certificate requests or
8044 certificates. This includes TLS clients and TLS servers with
8045 client authentication enabled.
8047 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8052 * Exploitable out-of-bounds read in X509_cmp_time
8054 X509_cmp_time does not properly check the length of the ASN1_TIME
8055 string and can read a few bytes out of bounds. In addition,
8056 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8059 An attacker can use this to craft malformed certificates and CRLs of
8060 various sizes and potentially cause a segmentation fault, resulting in
8061 a DoS on applications that verify certificates or CRLs. TLS clients
8062 that verify CRLs are affected. TLS clients and servers with client
8063 authentication enabled may be affected if they use custom verification
8066 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8067 independently by Hanno Böck.
8072 * PKCS7 crash with missing EnvelopedContent
8074 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8075 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8076 with missing content and trigger a NULL pointer dereference on parsing.
8078 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8079 structures from untrusted sources are affected. OpenSSL clients and
8080 servers are not affected.
8082 This issue was reported to OpenSSL by Michal Zalewski (Google).
8087 * CMS verify infinite loop with unknown hash function
8089 When verifying a signedData message the CMS code can enter an infinite loop
8090 if presented with an unknown hash function OID. This can be used to perform
8091 denial of service against any system which verifies signedData messages using
8093 This issue was reported to OpenSSL by Johannes Bauer.
8098 * Race condition handling NewSessionTicket
8100 If a NewSessionTicket is received by a multi-threaded client when attempting to
8101 reuse a previous ticket then a race condition can occur potentially leading to
8102 a double free of the ticket data.
8107 * Reject DH handshakes with parameters shorter than 768 bits.
8109 *Kurt Roeckx and Emilia Kasper*
8111 * dhparam: generate 2048-bit parameters by default.
8113 *Kurt Roeckx and Emilia Kasper*
8115 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8117 * Segmentation fault in ASN1_TYPE_cmp fix
8119 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8120 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8121 certificate signature algorithm consistency this can be used to crash any
8122 certificate verification operation and exploited in a DoS attack. Any
8123 application which performs certificate verification is vulnerable including
8124 OpenSSL clients and servers which enable client authentication.
8129 * ASN.1 structure reuse memory corruption fix
8131 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8132 memory corruption via an invalid write. Such reuse is and has been
8133 strongly discouraged and is believed to be rare.
8135 Applications that parse structures containing CHOICE or ANY DEFINED BY
8136 components may be affected. Certificate parsing (d2i_X509 and related
8137 functions) are however not affected. OpenSSL clients and servers are
8143 * PKCS7 NULL pointer dereferences fix
8145 The PKCS#7 parsing code does not handle missing outer ContentInfo
8146 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8147 missing content and trigger a NULL pointer dereference on parsing.
8149 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8150 otherwise parse PKCS#7 structures from untrusted sources are
8151 affected. OpenSSL clients and servers are not affected.
8153 This issue was reported to OpenSSL by Michal Zalewski (Google).
8158 * DoS via reachable assert in SSLv2 servers fix
8160 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8161 servers that both support SSLv2 and enable export cipher suites by sending
8162 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8164 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8165 (OpenSSL development team).
8170 * Use After Free following d2i_ECPrivatekey error fix
8172 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8173 could cause a use after free condition. This, in turn, could cause a double
8174 free in several private key parsing functions (such as d2i_PrivateKey
8175 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8176 for applications that receive EC private keys from untrusted
8177 sources. This scenario is considered rare.
8179 This issue was discovered by the BoringSSL project and fixed in their
8185 * X509_to_X509_REQ NULL pointer deref fix
8187 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8188 the certificate key is invalid. This function is rarely used in practice.
8190 This issue was discovered by Brian Carpenter.
8195 * Removed the export ciphers from the DEFAULT ciphers
8199 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8201 * Build fixes for the Windows and OpenVMS platforms
8203 *Matt Caswell and Richard Levitte*
8205 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8207 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8208 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8209 dereference. This could lead to a Denial Of Service attack. Thanks to
8210 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8215 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8216 dtls1_buffer_record function under certain conditions. In particular this
8217 could occur if an attacker sent repeated DTLS records with the same
8218 sequence number but for the next epoch. The memory leak could be exploited
8219 by an attacker in a Denial of Service attack through memory exhaustion.
8220 Thanks to Chris Mueller for reporting this issue.
8225 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8226 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8227 method would be set to NULL which could later result in a NULL pointer
8228 dereference. Thanks to Frank Schmirler for reporting this issue.
8233 * Abort handshake if server key exchange message is omitted for ephemeral
8236 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8237 reporting this issue.
8242 * Remove non-export ephemeral RSA code on client and server. This code
8243 violated the TLS standard by allowing the use of temporary RSA keys in
8244 non-export ciphersuites and could be used by a server to effectively
8245 downgrade the RSA key length used to a value smaller than the server
8246 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8247 INRIA or reporting this issue.
8252 * Fixed issue where DH client certificates are accepted without verification.
8253 An OpenSSL server will accept a DH certificate for client authentication
8254 without the certificate verify message. This effectively allows a client to
8255 authenticate without the use of a private key. This only affects servers
8256 which trust a client certificate authority which issues certificates
8257 containing DH keys: these are extremely rare and hardly ever encountered.
8258 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8264 * Ensure that the session ID context of an SSL is updated when its
8265 SSL_CTX is updated via SSL_set_SSL_CTX.
8267 The session ID context is typically set from the parent SSL_CTX,
8268 and can vary with the CTX.
8272 * Fix various certificate fingerprint issues.
8274 By using non-DER or invalid encodings outside the signed portion of a
8275 certificate the fingerprint can be changed without breaking the signature.
8276 Although no details of the signed portion of the certificate can be changed
8277 this can cause problems with some applications: e.g. those using the
8278 certificate fingerprint for blacklists.
8280 1. Reject signatures with non zero unused bits.
8282 If the BIT STRING containing the signature has non zero unused bits reject
8283 the signature. All current signature algorithms require zero unused bits.
8285 2. Check certificate algorithm consistency.
8287 Check the AlgorithmIdentifier inside TBS matches the one in the
8288 certificate signature. NB: this will result in signature failure
8289 errors for some broken certificates.
8291 Thanks to Konrad Kraszewski from Google for reporting this issue.
8293 3. Check DSA/ECDSA signatures use DER.
8295 Re-encode DSA/ECDSA signatures and compare with the original received
8296 signature. Return an error if there is a mismatch.
8298 This will reject various cases including garbage after signature
8299 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8300 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8301 (negative or with leading zeroes).
8303 Further analysis was conducted and fixes were developed by Stephen Henson
8304 of the OpenSSL core team.
8310 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8311 results on some platforms, including x86_64. This bug occurs at random
8312 with a very low probability, and is not known to be exploitable in any
8313 way, though its exact impact is difficult to determine. Thanks to Pieter
8314 Wuille (Blockstream) who reported this issue and also suggested an initial
8315 fix. Further analysis was conducted by the OpenSSL development team and
8316 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8317 the OpenSSL core team.
8322 * Do not resume sessions on the server if the negotiated protocol
8323 version does not match the session's version. Resuming with a different
8324 version, while not strictly forbidden by the RFC, is of questionable
8325 sanity and breaks all known clients.
8327 *David Benjamin, Emilia Käsper*
8329 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8330 early CCS messages during renegotiation. (Note that because
8331 renegotiation is encrypted, this early CCS was not exploitable.)
8335 * Tighten client-side session ticket handling during renegotiation:
8336 ensure that the client only accepts a session ticket if the server sends
8337 the extension anew in the ServerHello. Previously, a TLS client would
8338 reuse the old extension state and thus accept a session ticket if one was
8339 announced in the initial ServerHello.
8341 Similarly, ensure that the client requires a session ticket if one
8342 was advertised in the ServerHello. Previously, a TLS client would
8343 ignore a missing NewSessionTicket message.
8347 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8351 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8352 sends a carefully crafted handshake message, to cause OpenSSL to fail
8353 to free up to 64k of memory causing a memory leak. This could be
8354 exploited in a Denial Of Service attack. This issue affects OpenSSL
8355 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8356 whether SRTP is used or configured. Implementations of OpenSSL that
8357 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8359 The fix was developed by the OpenSSL team.
8364 * Session Ticket Memory Leak.
8366 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8367 integrity of that ticket is first verified. In the event of a session
8368 ticket integrity check failing, OpenSSL will fail to free memory
8369 causing a memory leak. By sending a large number of invalid session
8370 tickets an attacker could exploit this issue in a Denial Of Service
8376 * Build option no-ssl3 is incomplete.
8378 When OpenSSL is configured with "no-ssl3" as a build option, servers
8379 could accept and complete an SSL 3.0 handshake, and clients could be
8380 configured to send them.
8383 *Akamai and the OpenSSL team*
8385 * Add support for TLS_FALLBACK_SCSV.
8386 Client applications doing fallback retries should call
8387 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8390 *Adam Langley, Bodo Moeller*
8392 * Add additional DigestInfo checks.
8394 Re-encode DigestInto in DER and check against the original when
8395 verifying RSA signature: this will reject any improperly encoded
8396 DigestInfo structures.
8398 Note: this is a precautionary measure and no attacks are currently known.
8402 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8404 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8405 SRP code can be overrun an internal buffer. Add sanity check that
8406 g, A, B < N to SRP code.
8408 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8409 Group for discovering this issue.
8414 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8415 TLS 1.0 instead of higher protocol versions when the ClientHello message
8416 is badly fragmented. This allows a man-in-the-middle attacker to force a
8417 downgrade to TLS 1.0 even if both the server and the client support a
8418 higher protocol version, by modifying the client's TLS records.
8420 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8421 researching this issue.
8426 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8427 to a denial of service attack. A malicious server can crash the client
8428 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8429 ciphersuite and sending carefully crafted handshake messages.
8431 Thanks to Felix Gröbert (Google) for discovering and researching this
8437 * By sending carefully crafted DTLS packets an attacker could cause openssl
8438 to leak memory. This can be exploited through a Denial of Service attack.
8439 Thanks to Adam Langley for discovering and researching this issue.
8444 * An attacker can force openssl to consume large amounts of memory whilst
8445 processing DTLS handshake messages. This can be exploited through a
8446 Denial of Service attack.
8447 Thanks to Adam Langley for discovering and researching this issue.
8452 * An attacker can force an error condition which causes openssl to crash
8453 whilst processing DTLS packets due to memory being freed twice. This
8454 can be exploited through a Denial of Service attack.
8455 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8461 * If a multithreaded client connects to a malicious server using a resumed
8462 session and the server sends an ec point format extension it could write
8463 up to 255 bytes to freed memory.
8465 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8471 * A malicious server can crash an OpenSSL client with a null pointer
8472 dereference (read) by specifying an SRP ciphersuite even though it was not
8473 properly negotiated with the client. This can be exploited through a
8474 Denial of Service attack.
8476 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8477 discovering and researching this issue.
8482 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8483 X509_name_oneline, X509_name_print_ex et al. to leak some information
8484 from the stack. Applications may be affected if they echo pretty printing
8485 output to the attacker.
8487 Thanks to Ivan Fratric (Google) for discovering this issue.
8490 *Emilia Käsper, and Steve Henson*
8492 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8493 for corner cases. (Certain input points at infinity could lead to
8494 bogus results, with non-infinity inputs mapped to infinity too.)
8498 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8500 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8501 handshake can force the use of weak keying material in OpenSSL
8502 SSL/TLS clients and servers.
8504 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8505 researching this issue. ([CVE-2014-0224])
8507 *KIKUCHI Masashi, Steve Henson*
8509 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8510 OpenSSL DTLS client the code can be made to recurse eventually crashing
8513 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8516 *Imre Rad, Steve Henson*
8518 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8519 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8520 client or server. This is potentially exploitable to run arbitrary
8521 code on a vulnerable client or server.
8523 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8525 *Jüri Aedla, Steve Henson*
8527 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8528 are subject to a denial of service attack.
8530 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8531 this issue. ([CVE-2014-3470])
8533 *Felix Gröbert, Ivan Fratric, Steve Henson*
8535 * Harmonize version and its documentation. -f flag is used to display
8538 *mancha <mancha1@zoho.com>*
8540 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8541 in i2d_ECPrivateKey.
8543 *mancha <mancha1@zoho.com>*
8545 * Fix some double frees. These are not thought to be exploitable.
8547 *mancha <mancha1@zoho.com>*
8549 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8551 * A missing bounds check in the handling of the TLS heartbeat extension
8552 can be used to reveal up to 64k of memory to a connected client or
8555 Thanks for Neel Mehta of Google Security for discovering this bug and to
8556 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8557 preparing the fix ([CVE-2014-0160])
8559 *Adam Langley, Bodo Moeller*
8561 * Fix for the attack described in the paper "Recovering OpenSSL
8562 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8563 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8564 <http://eprint.iacr.org/2014/140>
8566 Thanks to Yuval Yarom and Naomi Benger for discovering this
8567 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8569 *Yuval Yarom and Naomi Benger*
8571 * TLS pad extension: draft-agl-tls-padding-03
8573 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8574 TLS client Hello record length value would otherwise be > 255 and
8575 less that 512 pad with a dummy extension containing zeroes so it
8576 is at least 512 bytes long.
8578 *Adam Langley, Steve Henson*
8580 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8582 * Fix for TLS record tampering bug. A carefully crafted invalid
8583 handshake could crash OpenSSL with a NULL pointer exception.
8584 Thanks to Anton Johansson for reporting this issues.
8587 * Keep original DTLS digest and encryption contexts in retransmission
8588 structures so we can use the previous session parameters if they need
8589 to be resent. ([CVE-2013-6450])
8593 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8594 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8595 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8596 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8597 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8598 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8600 *Rob Stradling, Adam Langley*
8602 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8604 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8605 supporting platforms or when small records were transferred.
8607 *Andy Polyakov, Steve Henson*
8609 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8611 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8613 This addresses the flaw in CBC record processing discovered by
8614 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8615 at: <http://www.isg.rhul.ac.uk/tls/>
8617 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8618 Security Group at Royal Holloway, University of London
8619 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8620 Emilia Käsper for the initial patch.
8623 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8625 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8626 ciphersuites which can be exploited in a denial of service attack.
8627 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8628 and detecting this bug and to Wolfgang Ettlinger
8629 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8634 * Return an error when checking OCSP signatures when key is NULL.
8635 This fixes a DoS attack. ([CVE-2013-0166])
8639 * Make openssl verify return errors.
8641 *Chris Palmer <palmer@google.com> and Ben Laurie*
8643 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8644 the right response is stapled. Also change SSL_get_certificate()
8645 so it returns the certificate actually sent.
8646 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8648 *Rob Stradling <rob.stradling@comodo.com>*
8650 * Fix possible deadlock when decoding public keys.
8654 * Don't use TLS 1.0 record version number in initial client hello
8659 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8661 * Sanity check record length before skipping explicit IV in TLS
8662 1.2, 1.1 and DTLS to fix DoS attack.
8664 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8665 fuzzing as a service testing platform.
8670 * Initialise tkeylen properly when encrypting CMS messages.
8671 Thanks to Solar Designer of Openwall for reporting this issue.
8675 * In FIPS mode don't try to use composite ciphers as they are not
8680 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8682 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8683 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8684 mean any application compiled against OpenSSL 1.0.0 headers setting
8685 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8686 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8687 0x10000000L Any application which was previously compiled against
8688 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8689 will need to be recompiled as a result. Letting be results in
8690 inability to disable specifically TLS 1.1 and in client context,
8691 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8695 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8696 disable just protocol X, but all protocols above X *if* there are
8697 protocols *below* X still enabled. In more practical terms it means
8698 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8699 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8700 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8705 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8707 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8708 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8709 in CRYPTO_realloc_clean.
8711 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8712 issue and to Adam Langley <agl@chromium.org> for fixing it.
8715 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8717 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8721 * Workarounds for some broken servers that "hang" if a client hello
8722 record length exceeds 255 bytes.
8724 1. Do not use record version number > TLS 1.0 in initial client
8725 hello: some (but not all) hanging servers will now work.
8726 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8727 the number of ciphers sent in the client hello. This should be
8728 set to an even number, such as 50, for example by passing:
8729 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8730 Most broken servers should now work.
8731 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8732 TLS 1.2 client support entirely.
8736 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8740 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8742 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8743 STRING form instead of a DigestInfo.
8747 * The format used for MDC2 RSA signatures is inconsistent between EVP
8748 and the RSA_sign/RSA_verify functions. This was made more apparent when
8749 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8750 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8751 the correct format in RSA_verify so both forms transparently work.
8755 * Some servers which support TLS 1.0 can choke if we initially indicate
8756 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8757 encrypted premaster secret. As a workaround use the maximum permitted
8758 client version in client hello, this should keep such servers happy
8759 and still work with previous versions of OpenSSL.
8763 * Add support for TLS/DTLS heartbeats.
8765 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8767 * Add support for SCTP.
8769 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8771 * Improved PRNG seeding for VOS.
8773 *Paul Green <Paul.Green@stratus.com>*
8775 * Extensive assembler packs updates, most notably:
8777 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8778 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8779 - x86_64: bit-sliced AES implementation;
8780 - ARM: NEON support, contemporary platforms optimizations;
8781 - s390x: z196 support;
8782 - `*`: GHASH and GF(2^m) multiplication implementations;
8786 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8787 (removal of unnecessary code)
8789 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8791 * Add TLS key material exporter from RFC 5705.
8795 * Add DTLS-SRTP negotiation from RFC 5764.
8799 * Add Next Protocol Negotiation,
8800 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8801 disabled with a no-npn flag to config or Configure. Code donated
8804 *Adam Langley <agl@google.com> and Ben Laurie*
8806 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8807 NIST-P256, NIST-P521, with constant-time single point multiplication on
8808 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8809 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8810 Code made available under Apache License version 2.0.
8812 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8813 line to include this in your build of OpenSSL, and run "make depend" (or
8814 "make update"). This enables the following EC_METHODs:
8816 EC_GFp_nistp224_method()
8817 EC_GFp_nistp256_method()
8818 EC_GFp_nistp521_method()
8820 EC_GROUP_new_by_curve_name() will automatically use these (while
8821 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8824 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8826 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8827 all platforms. Move ssize_t definition from e_os.h to the public
8828 header file e_os2.h as it now appears in public header file cms.h
8832 * New -sigopt option to the ca, req and x509 utilities. Additional
8833 signature parameters can be passed using this option and in
8838 * Add RSA PSS signing function. This will generate and set the
8839 appropriate AlgorithmIdentifiers for PSS based on those in the
8840 corresponding EVP_MD_CTX structure. No application support yet.
8844 * Support for companion algorithm specific ASN1 signing routines.
8845 New function ASN1_item_sign_ctx() signs a pre-initialised
8846 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8847 the appropriate parameters.
8851 * Add new algorithm specific ASN1 verification initialisation function
8852 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8853 handling will be the same no matter what EVP_PKEY_METHOD is used.
8854 Add a PSS handler to support verification of PSS signatures: checked
8855 against a number of sample certificates.
8859 * Add signature printing for PSS. Add PSS OIDs.
8861 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8863 * Add algorithm specific signature printing. An individual ASN1 method
8864 can now print out signatures instead of the standard hex dump.
8866 More complex signatures (e.g. PSS) can print out more meaningful
8867 information. Include DSA version that prints out the signature
8872 * Password based recipient info support for CMS library: implementing
8877 * Split password based encryption into PBES2 and PBKDF2 functions. This
8878 neatly separates the code into cipher and PBE sections and is required
8879 for some algorithms that split PBES2 into separate pieces (such as
8880 password based CMS).
8884 * Session-handling fixes:
8885 - Fix handling of connections that are resuming with a session ID,
8886 but also support Session Tickets.
8887 - Fix a bug that suppressed issuing of a new ticket if the client
8888 presented a ticket with an expired session.
8889 - Try to set the ticket lifetime hint to something reasonable.
8890 - Make tickets shorter by excluding irrelevant information.
8891 - On the client side, don't ignore renewed tickets.
8893 *Adam Langley, Bodo Moeller (Google)*
8895 * Fix PSK session representation.
8899 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8901 This work was sponsored by Intel.
8905 * Add GCM support to TLS library. Some custom code is needed to split
8906 the IV between the fixed (from PRF) and explicit (from TLS record)
8907 portions. This adds all GCM ciphersuites supported by RFC5288 and
8908 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8909 add a special AESGCM string for GCM only.
8913 * Expand range of ctrls for AES GCM. Permit setting invocation
8914 field on decrypt and retrieval of invocation field only on encrypt.
8918 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8919 As required by RFC5289 these ciphersuites cannot be used if for
8920 versions of TLS earlier than 1.2.
8924 * For FIPS capable OpenSSL interpret a NULL default public key method
8925 as unset and return the appropriate default but do *not* set the default.
8926 This means we can return the appropriate method in applications that
8927 switch between FIPS and non-FIPS modes.
8931 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8932 ENGINE is used then we cannot handle that in the FIPS module so we
8933 keep original code iff non-FIPS operations are allowed.
8937 * Add -attime option to openssl utilities.
8939 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8941 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8945 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8946 FIPS EC methods unconditionally for now.
8950 * New build option no-ec2m to disable characteristic 2 code.
8954 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8955 all cases can be covered as some introduce binary incompatibilities.
8959 * Redirect RSA operations to FIPS module including keygen,
8960 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8964 * Add similar low-level API blocking to ciphers.
8968 * low-level digest APIs are not approved in FIPS mode: any attempt
8969 to use these will cause a fatal error. Applications that *really* want
8970 to use them can use the `private_*` version instead.
8974 * Redirect cipher operations to FIPS module for FIPS builds.
8978 * Redirect digest operations to FIPS module for FIPS builds.
8982 * Update build system to add "fips" flag which will link in fipscanister.o
8983 for static and shared library builds embedding a signature if needed.
8987 * Output TLS supported curves in preference order instead of numerical
8988 order. This is currently hardcoded for the highest order curves first.
8989 This should be configurable so applications can judge speed vs strength.
8993 * Add TLS v1.2 server support for client authentication.
8997 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
9002 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
9003 FIPS modules versions.
9007 * Add TLS v1.2 client side support for client authentication. Keep cache
9008 of handshake records longer as we don't know the hash algorithm to use
9009 until after the certificate request message is received.
9013 * Initial TLS v1.2 client support. Add a default signature algorithms
9014 extension including all the algorithms we support. Parse new signature
9015 format in client key exchange. Relax some ECC signing restrictions for
9016 TLS v1.2 as indicated in RFC5246.
9020 * Add server support for TLS v1.2 signature algorithms extension. Switch
9021 to new signature format when needed using client digest preference.
9022 All server ciphersuites should now work correctly in TLS v1.2. No client
9023 support yet and no support for client certificates.
9027 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
9028 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
9029 ciphersuites. At present only RSA key exchange ciphersuites work with
9030 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
9031 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
9032 and version checking.
9036 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
9037 with this defined it will not be affected by any changes to ssl internal
9038 structures. Add several utility functions to allow openssl application
9039 to work with OPENSSL_NO_SSL_INTERN defined.
9043 * A long standing patch to add support for SRP from EdelWeb (Peter
9044 Sylvester and Christophe Renou) was integrated.
9045 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
9046 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
9049 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
9053 * Permit abbreviated handshakes when renegotiating using the function
9054 SSL_renegotiate_abbreviated().
9056 *Robin Seggelmann <seggelmann@fh-muenster.de>*
9058 * Add call to ENGINE_register_all_complete() to
9059 ENGINE_load_builtin_engines(), so some implementations get used
9060 automatically instead of needing explicit application support.
9064 * Add support for TLS key exporter as described in RFC5705.
9066 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
9068 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
9069 a few changes are required:
9071 Add SSL_OP_NO_TLSv1_1 flag.
9072 Add TLSv1_1 methods.
9073 Update version checking logic to handle version 1.1.
9074 Add explicit IV handling (ported from DTLS code).
9075 Add command line options to s_client/s_server.
9082 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
9084 * X509_ATTRIBUTE memory leak
9086 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
9087 memory. This structure is used by the PKCS#7 and CMS routines so any
9088 application which reads PKCS#7 or CMS data from untrusted sources is
9089 affected. SSL/TLS is not affected.
9091 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9097 * Race condition handling PSK identify hint
9099 If PSK identity hints are received by a multi-threaded client then
9100 the values are wrongly updated in the parent SSL_CTX structure. This can
9101 result in a race condition potentially leading to a double free of the
9107 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9109 * Malformed ECParameters causes infinite loop
9111 When processing an ECParameters structure OpenSSL enters an infinite loop
9112 if the curve specified is over a specially malformed binary polynomial
9115 This can be used to perform denial of service against any
9116 system which processes public keys, certificate requests or
9117 certificates. This includes TLS clients and TLS servers with
9118 client authentication enabled.
9120 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9125 * Exploitable out-of-bounds read in X509_cmp_time
9127 X509_cmp_time does not properly check the length of the ASN1_TIME
9128 string and can read a few bytes out of bounds. In addition,
9129 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9132 An attacker can use this to craft malformed certificates and CRLs of
9133 various sizes and potentially cause a segmentation fault, resulting in
9134 a DoS on applications that verify certificates or CRLs. TLS clients
9135 that verify CRLs are affected. TLS clients and servers with client
9136 authentication enabled may be affected if they use custom verification
9139 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9140 independently by Hanno Böck.
9145 * PKCS7 crash with missing EnvelopedContent
9147 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9148 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9149 with missing content and trigger a NULL pointer dereference on parsing.
9151 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9152 structures from untrusted sources are affected. OpenSSL clients and
9153 servers are not affected.
9155 This issue was reported to OpenSSL by Michal Zalewski (Google).
9160 * CMS verify infinite loop with unknown hash function
9162 When verifying a signedData message the CMS code can enter an infinite loop
9163 if presented with an unknown hash function OID. This can be used to perform
9164 denial of service against any system which verifies signedData messages using
9166 This issue was reported to OpenSSL by Johannes Bauer.
9171 * Race condition handling NewSessionTicket
9173 If a NewSessionTicket is received by a multi-threaded client when attempting to
9174 reuse a previous ticket then a race condition can occur potentially leading to
9175 a double free of the ticket data.
9180 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9182 * Segmentation fault in ASN1_TYPE_cmp fix
9184 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9185 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9186 certificate signature algorithm consistency this can be used to crash any
9187 certificate verification operation and exploited in a DoS attack. Any
9188 application which performs certificate verification is vulnerable including
9189 OpenSSL clients and servers which enable client authentication.
9194 * ASN.1 structure reuse memory corruption fix
9196 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9197 memory corruption via an invalid write. Such reuse is and has been
9198 strongly discouraged and is believed to be rare.
9200 Applications that parse structures containing CHOICE or ANY DEFINED BY
9201 components may be affected. Certificate parsing (d2i_X509 and related
9202 functions) are however not affected. OpenSSL clients and servers are
9208 * PKCS7 NULL pointer dereferences fix
9210 The PKCS#7 parsing code does not handle missing outer ContentInfo
9211 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9212 missing content and trigger a NULL pointer dereference on parsing.
9214 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9215 otherwise parse PKCS#7 structures from untrusted sources are
9216 affected. OpenSSL clients and servers are not affected.
9218 This issue was reported to OpenSSL by Michal Zalewski (Google).
9223 * DoS via reachable assert in SSLv2 servers fix
9225 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9226 servers that both support SSLv2 and enable export cipher suites by sending
9227 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9229 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9230 (OpenSSL development team).
9235 * Use After Free following d2i_ECPrivatekey error fix
9237 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9238 could cause a use after free condition. This, in turn, could cause a double
9239 free in several private key parsing functions (such as d2i_PrivateKey
9240 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9241 for applications that receive EC private keys from untrusted
9242 sources. This scenario is considered rare.
9244 This issue was discovered by the BoringSSL project and fixed in their
9250 * X509_to_X509_REQ NULL pointer deref fix
9252 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9253 the certificate key is invalid. This function is rarely used in practice.
9255 This issue was discovered by Brian Carpenter.
9260 * Removed the export ciphers from the DEFAULT ciphers
9264 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9266 * Build fixes for the Windows and OpenVMS platforms
9268 *Matt Caswell and Richard Levitte*
9270 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9272 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9273 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9274 dereference. This could lead to a Denial Of Service attack. Thanks to
9275 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9280 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9281 dtls1_buffer_record function under certain conditions. In particular this
9282 could occur if an attacker sent repeated DTLS records with the same
9283 sequence number but for the next epoch. The memory leak could be exploited
9284 by an attacker in a Denial of Service attack through memory exhaustion.
9285 Thanks to Chris Mueller for reporting this issue.
9290 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9291 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9292 method would be set to NULL which could later result in a NULL pointer
9293 dereference. Thanks to Frank Schmirler for reporting this issue.
9298 * Abort handshake if server key exchange message is omitted for ephemeral
9301 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9302 reporting this issue.
9307 * Remove non-export ephemeral RSA code on client and server. This code
9308 violated the TLS standard by allowing the use of temporary RSA keys in
9309 non-export ciphersuites and could be used by a server to effectively
9310 downgrade the RSA key length used to a value smaller than the server
9311 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9312 INRIA or reporting this issue.
9317 * Fixed issue where DH client certificates are accepted without verification.
9318 An OpenSSL server will accept a DH certificate for client authentication
9319 without the certificate verify message. This effectively allows a client to
9320 authenticate without the use of a private key. This only affects servers
9321 which trust a client certificate authority which issues certificates
9322 containing DH keys: these are extremely rare and hardly ever encountered.
9323 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9329 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9330 results on some platforms, including x86_64. This bug occurs at random
9331 with a very low probability, and is not known to be exploitable in any
9332 way, though its exact impact is difficult to determine. Thanks to Pieter
9333 Wuille (Blockstream) who reported this issue and also suggested an initial
9334 fix. Further analysis was conducted by the OpenSSL development team and
9335 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9336 the OpenSSL core team.
9341 * Fix various certificate fingerprint issues.
9343 By using non-DER or invalid encodings outside the signed portion of a
9344 certificate the fingerprint can be changed without breaking the signature.
9345 Although no details of the signed portion of the certificate can be changed
9346 this can cause problems with some applications: e.g. those using the
9347 certificate fingerprint for blacklists.
9349 1. Reject signatures with non zero unused bits.
9351 If the BIT STRING containing the signature has non zero unused bits reject
9352 the signature. All current signature algorithms require zero unused bits.
9354 2. Check certificate algorithm consistency.
9356 Check the AlgorithmIdentifier inside TBS matches the one in the
9357 certificate signature. NB: this will result in signature failure
9358 errors for some broken certificates.
9360 Thanks to Konrad Kraszewski from Google for reporting this issue.
9362 3. Check DSA/ECDSA signatures use DER.
9364 Re-encode DSA/ECDSA signatures and compare with the original received
9365 signature. Return an error if there is a mismatch.
9367 This will reject various cases including garbage after signature
9368 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9369 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9370 (negative or with leading zeroes).
9372 Further analysis was conducted and fixes were developed by Stephen Henson
9373 of the OpenSSL core team.
9379 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9381 * Session Ticket Memory Leak.
9383 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9384 integrity of that ticket is first verified. In the event of a session
9385 ticket integrity check failing, OpenSSL will fail to free memory
9386 causing a memory leak. By sending a large number of invalid session
9387 tickets an attacker could exploit this issue in a Denial Of Service
9393 * Build option no-ssl3 is incomplete.
9395 When OpenSSL is configured with "no-ssl3" as a build option, servers
9396 could accept and complete an SSL 3.0 handshake, and clients could be
9397 configured to send them.
9400 *Akamai and the OpenSSL team*
9402 * Add support for TLS_FALLBACK_SCSV.
9403 Client applications doing fallback retries should call
9404 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9407 *Adam Langley, Bodo Moeller*
9409 * Add additional DigestInfo checks.
9411 Re-encode DigestInto in DER and check against the original when
9412 verifying RSA signature: this will reject any improperly encoded
9413 DigestInfo structures.
9415 Note: this is a precautionary measure and no attacks are currently known.
9419 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9421 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9422 to a denial of service attack. A malicious server can crash the client
9423 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9424 ciphersuite and sending carefully crafted handshake messages.
9426 Thanks to Felix Gröbert (Google) for discovering and researching this
9432 * By sending carefully crafted DTLS packets an attacker could cause openssl
9433 to leak memory. This can be exploited through a Denial of Service attack.
9434 Thanks to Adam Langley for discovering and researching this issue.
9439 * An attacker can force openssl to consume large amounts of memory whilst
9440 processing DTLS handshake messages. This can be exploited through a
9441 Denial of Service attack.
9442 Thanks to Adam Langley for discovering and researching this issue.
9447 * An attacker can force an error condition which causes openssl to crash
9448 whilst processing DTLS packets due to memory being freed twice. This
9449 can be exploited through a Denial of Service attack.
9450 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9456 * If a multithreaded client connects to a malicious server using a resumed
9457 session and the server sends an ec point format extension it could write
9458 up to 255 bytes to freed memory.
9460 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9466 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9467 X509_name_oneline, X509_name_print_ex et al. to leak some information
9468 from the stack. Applications may be affected if they echo pretty printing
9469 output to the attacker.
9471 Thanks to Ivan Fratric (Google) for discovering this issue.
9474 *Emilia Käsper, and Steve Henson*
9476 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9477 for corner cases. (Certain input points at infinity could lead to
9478 bogus results, with non-infinity inputs mapped to infinity too.)
9482 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9484 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9485 handshake can force the use of weak keying material in OpenSSL
9486 SSL/TLS clients and servers.
9488 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9489 researching this issue. ([CVE-2014-0224])
9491 *KIKUCHI Masashi, Steve Henson*
9493 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9494 OpenSSL DTLS client the code can be made to recurse eventually crashing
9497 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9500 *Imre Rad, Steve Henson*
9502 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9503 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9504 client or server. This is potentially exploitable to run arbitrary
9505 code on a vulnerable client or server.
9507 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9509 *Jüri Aedla, Steve Henson*
9511 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9512 are subject to a denial of service attack.
9514 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9515 this issue. ([CVE-2014-3470])
9517 *Felix Gröbert, Ivan Fratric, Steve Henson*
9519 * Harmonize version and its documentation. -f flag is used to display
9522 *mancha <mancha1@zoho.com>*
9524 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9525 in i2d_ECPrivateKey.
9527 *mancha <mancha1@zoho.com>*
9529 * Fix some double frees. These are not thought to be exploitable.
9531 *mancha <mancha1@zoho.com>*
9533 * Fix for the attack described in the paper "Recovering OpenSSL
9534 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9535 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9536 <http://eprint.iacr.org/2014/140>
9538 Thanks to Yuval Yarom and Naomi Benger for discovering this
9539 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9541 *Yuval Yarom and Naomi Benger*
9543 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9545 * Keep original DTLS digest and encryption contexts in retransmission
9546 structures so we can use the previous session parameters if they need
9547 to be resent. ([CVE-2013-6450])
9551 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9552 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9553 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9554 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9555 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9556 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9558 *Rob Stradling, Adam Langley*
9560 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9562 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9564 This addresses the flaw in CBC record processing discovered by
9565 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9566 at: <http://www.isg.rhul.ac.uk/tls/>
9568 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9569 Security Group at Royal Holloway, University of London
9570 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9571 Emilia Käsper for the initial patch.
9574 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9576 * Return an error when checking OCSP signatures when key is NULL.
9577 This fixes a DoS attack. ([CVE-2013-0166])
9581 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9582 the right response is stapled. Also change SSL_get_certificate()
9583 so it returns the certificate actually sent.
9584 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9585 (This is a backport)
9587 *Rob Stradling <rob.stradling@comodo.com>*
9589 * Fix possible deadlock when decoding public keys.
9593 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9595 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9598 * Sanity check record length before skipping explicit IV in DTLS
9601 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9602 fuzzing as a service testing platform.
9607 * Initialise tkeylen properly when encrypting CMS messages.
9608 Thanks to Solar Designer of Openwall for reporting this issue.
9612 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9614 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9615 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9616 in CRYPTO_realloc_clean.
9618 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9619 issue and to Adam Langley <agl@chromium.org> for fixing it.
9622 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9624 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9626 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9627 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9628 content decryption and always return the same error. Note: this attack
9629 needs on average 2^20 messages so it only affects automated senders. The
9630 old behaviour can be re-enabled in the CMS code by setting the
9631 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9632 an MMA defence is not necessary.
9633 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9634 this issue. ([CVE-2012-0884])
9638 * Fix CVE-2011-4619: make sure we really are receiving a
9639 client hello before rejecting multiple SGC restarts. Thanks to
9640 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9644 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9646 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9647 Thanks to Antonio Martin, Enterprise Secure Access Research and
9648 Development, Cisco Systems, Inc. for discovering this bug and
9649 preparing a fix. ([CVE-2012-0050])
9653 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9655 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9656 of the Vaudenay padding oracle attack on CBC mode encryption
9657 which enables an efficient plaintext recovery attack against
9658 the OpenSSL implementation of DTLS. Their attack exploits timing
9659 differences arising during decryption processing. A research
9660 paper describing this attack can be found at:
9661 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9662 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9663 Security Group at Royal Holloway, University of London
9664 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9665 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9666 for preparing the fix. ([CVE-2011-4108])
9668 *Robin Seggelmann, Michael Tuexen*
9670 * Clear bytes used for block padding of SSL 3.0 records.
9673 *Adam Langley (Google)*
9675 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9676 Kadianakis <desnacked@gmail.com> for discovering this issue and
9677 Adam Langley for preparing the fix. ([CVE-2011-4619])
9679 *Adam Langley (Google)*
9681 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9683 *Andrey Kulikov <amdeich@gmail.com>*
9685 * Prevent malformed RFC3779 data triggering an assertion failure.
9686 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9687 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9689 *Rob Austein <sra@hactrn.net>*
9691 * Improved PRNG seeding for VOS.
9693 *Paul Green <Paul.Green@stratus.com>*
9695 * Fix ssl_ciph.c set-up race.
9697 *Adam Langley (Google)*
9699 * Fix spurious failures in ecdsatest.c.
9701 *Emilia Käsper (Google)*
9703 * Fix the BIO_f_buffer() implementation (which was mixing different
9704 interpretations of the `..._len` fields).
9706 *Adam Langley (Google)*
9708 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9709 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9710 threads won't reuse the same blinding coefficients.
9712 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9713 lock to call BN_BLINDING_invert_ex, and avoids one use of
9714 BN_BLINDING_update for each BN_BLINDING structure (previously,
9715 the last update always remained unused).
9717 *Emilia Käsper (Google)*
9719 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9721 *Bob Buckholz (Google)*
9723 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9725 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9726 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9728 *Kaspar Brand <ossl@velox.ch>*
9730 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9731 for multi-threaded use of ECDH. ([CVE-2011-3210])
9733 *Adam Langley (Google)*
9735 * Fix x509_name_ex_d2i memory leak on bad inputs.
9739 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9740 signature public key algorithm by using OID xref utilities instead.
9741 Before this you could only use some ECC ciphersuites with SHA1 only.
9745 * Add protection against ECDSA timing attacks as mentioned in the paper
9746 by Billy Bob Brumley and Nicola Tuveri, see:
9747 <http://eprint.iacr.org/2011/232.pdf>
9749 *Billy Bob Brumley and Nicola Tuveri*
9751 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9753 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9755 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9757 * Fix bug in string printing code: if *any* escaping is enabled we must
9758 escape the escape character (backslash) or the resulting string is
9763 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9765 * Disable code workaround for ancient and obsolete Netscape browsers
9766 and servers: an attacker can use it in a ciphersuite downgrade attack.
9767 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9771 * Fixed J-PAKE implementation error, originally discovered by
9772 Sebastien Martini, further info and confirmation from Stefan
9773 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9777 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9779 * Fix extension code to avoid race conditions which can result in a buffer
9780 overrun vulnerability: resumed sessions must not be modified as they can
9781 be shared by multiple threads. CVE-2010-3864
9785 * Fix WIN32 build system to correctly link an ENGINE directory into
9790 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9792 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9795 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9797 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9799 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9800 context. The operation can be customised via the ctrl mechanism in
9801 case ENGINEs want to include additional functionality.
9805 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9809 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9810 output hashes compatible with older versions of OpenSSL.
9812 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9814 * Fix compression algorithm handling: if resuming a session use the
9815 compression algorithm of the resumed session instead of determining
9816 it from client hello again. Don't allow server to change algorithm.
9820 * Add load_crls() function to commands tidying load_certs() too. Add option
9821 to verify utility to allow additional CRLs to be included.
9825 * Update OCSP request code to permit adding custom headers to the request:
9826 some responders need this.
9830 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9833 *Julia Lawall <julia@diku.dk>*
9835 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9836 needlessly dereferenced structures, used obsolete functions and
9837 didn't handle all updated verify codes correctly.
9841 * Disable MD2 in the default configuration.
9845 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9846 indicate the initial BIO being pushed or popped. This makes it possible
9847 to determine whether the BIO is the one explicitly called or as a result
9848 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9849 it handles reference counts correctly and doesn't zero out the I/O bio
9850 when it is not being explicitly popped. WARNING: applications which
9851 included workarounds for the old buggy behaviour will need to be modified
9852 or they could free up already freed BIOs.
9856 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9857 renaming to all platforms (within the 0.9.8 branch, this was
9858 done conditionally on Netware platforms to avoid a name clash).
9860 *Guenter <lists@gknw.net>*
9862 * Add ECDHE and PSK support to DTLS.
9864 *Michael Tuexen <tuexen@fh-muenster.de>*
9866 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9871 * Add "missing" function EVP_MD_flags() (without this the only way to
9872 retrieve a digest flags is by accessing the structure directly. Update
9873 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9874 or cipher is registered as in the "from" argument. Print out all
9875 registered digests in the dgst usage message instead of manually
9876 attempting to work them out.
9880 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9881 this allows the use of compression and extensions. Change default cipher
9882 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9883 by default unless an application cipher string requests it.
9887 * Alter match criteria in PKCS12_parse(). It used to try to use local
9888 key ids to find matching certificates and keys but some PKCS#12 files
9889 don't follow the (somewhat unwritten) rules and this strategy fails.
9890 Now just gather all certificates together and the first private key
9891 then look for the first certificate that matches the key.
9895 * Support use of registered digest and cipher names for dgst and cipher
9896 commands instead of having to add each one as a special case. So now
9903 openssl dgst -sha256 foo
9905 and this works for ENGINE based algorithms too.
9909 * Update Gost ENGINE to support parameter files.
9911 *Victor B. Wagner <vitus@cryptocom.ru>*
9913 * Support GeneralizedTime in ca utility.
9915 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9917 * Enhance the hash format used for certificate directory links. The new
9918 form uses the canonical encoding (meaning equivalent names will work
9919 even if they aren't identical) and uses SHA1 instead of MD5. This form
9920 is incompatible with the older format and as a result c_rehash should
9921 be used to rebuild symbolic links.
9925 * Make PKCS#8 the default write format for private keys, replacing the
9926 traditional format. This form is standardised, more secure and doesn't
9927 include an implicit MD5 dependency.
9931 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9932 committed to OpenSSL should pass this lot as a minimum.
9936 * Add session ticket override functionality for use by EAP-FAST.
9938 *Jouni Malinen <j@w1.fi>*
9940 * Modify HMAC functions to return a value. Since these can be implemented
9941 in an ENGINE errors can occur.
9945 * Type-checked OBJ_bsearch_ex.
9949 * Type-checked OBJ_bsearch. Also some constification necessitated
9950 by type-checking. Still to come: TXT_DB, bsearch(?),
9951 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9956 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9957 seconds to a tm structure directly, instead of going through OS
9958 specific date routines. This avoids any issues with OS routines such
9959 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9960 and X509_time_adj_ex() to cover the extended range. The existing
9961 X509_time_adj() is still usable and will no longer have any date issues.
9965 * Delta CRL support. New use deltas option which will attempt to locate
9966 and search any appropriate delta CRLs available.
9968 This work was sponsored by Google.
9972 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9973 code and add additional score elements. Validate alternate CRL paths
9974 as part of the CRL checking and indicate a new error "CRL path validation
9975 error" in this case. Applications wanting additional details can use
9976 the verify callback and check the new "parent" field. If this is not
9977 NULL CRL path validation is taking place. Existing applications won't
9978 see this because it requires extended CRL support which is off by
9981 This work was sponsored by Google.
9985 * Support for freshest CRL extension.
9987 This work was sponsored by Google.
9991 * Initial indirect CRL support. Currently only supported in the CRLs
9992 passed directly and not via lookup. Process certificate issuer
9993 CRL entry extension and lookup CRL entries by bother issuer name
9994 and serial number. Check and process CRL issuer entry in IDP extension.
9996 This work was sponsored by Google.
10000 * Add support for distinct certificate and CRL paths. The CRL issuer
10001 certificate is validated separately in this case. Only enabled if
10002 an extended CRL support flag is set: this flag will enable additional
10003 CRL functionality in future.
10005 This work was sponsored by Google.
10009 * Add support for policy mappings extension.
10011 This work was sponsored by Google.
10015 * Fixes to pathlength constraint, self issued certificate handling,
10016 policy processing to align with RFC3280 and PKITS tests.
10018 This work was sponsored by Google.
10022 * Support for name constraints certificate extension. DN, email, DNS
10023 and URI types are currently supported.
10025 This work was sponsored by Google.
10029 * To cater for systems that provide a pointer-based thread ID rather
10030 than numeric, deprecate the current numeric thread ID mechanism and
10031 replace it with a structure and associated callback type. This
10032 mechanism allows a numeric "hash" to be extracted from a thread ID in
10033 either case, and on platforms where pointers are larger than 'long',
10034 mixing is done to help ensure the numeric 'hash' is usable even if it
10035 can't be guaranteed unique. The default mechanism is to use "&errno"
10036 as a pointer-based thread ID to distinguish between threads.
10038 Applications that want to provide their own thread IDs should now use
10039 CRYPTO_THREADID_set_callback() to register a callback that will call
10040 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
10042 Note that ERR_remove_state() is now deprecated, because it is tied
10043 to the assumption that thread IDs are numeric. ERR_remove_state(0)
10044 to free the current thread's error state should be replaced by
10045 ERR_remove_thread_state(NULL).
10047 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
10048 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
10049 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
10050 application was previously providing a numeric thread callback that
10051 was inappropriate for distinguishing threads, then uniqueness might
10052 have been obtained with &errno that happened immediately in the
10053 intermediate development versions of OpenSSL; this is no longer the
10054 case, the numeric thread callback will now override the automatic use
10057 *Geoff Thorpe, with help from Bodo Moeller*
10059 * Initial support for different CRL issuing certificates. This covers a
10060 simple case where the self issued certificates in the chain exist and
10061 the real CRL issuer is higher in the existing chain.
10063 This work was sponsored by Google.
10067 * Removed effectively defunct crypto/store from the build.
10071 * Revamp of STACK to provide stronger type-checking. Still to come:
10072 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
10073 ASN1_STRING, CONF_VALUE.
10077 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
10078 RAM on SSL connections. This option can save about 34k per idle SSL.
10082 * Revamp of LHASH to provide stronger type-checking. Still to come:
10083 STACK, TXT_DB, bsearch, qsort.
10087 * Initial support for Cryptographic Message Syntax (aka CMS) based
10088 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10089 support for data, signedData, compressedData, digestedData and
10090 encryptedData, envelopedData types included. Scripts to check against
10091 RFC4134 examples draft and interop and consistency checks of many
10092 content types and variants.
10096 * Add options to enc utility to support use of zlib compression BIO.
10100 * Extend mk1mf to support importing of options and assembly language
10101 files from Configure script, currently only included in VC-WIN32.
10102 The assembly language rules can now optionally generate the source
10103 files from the associated perl scripts.
10107 * Implement remaining functionality needed to support GOST ciphersuites.
10108 Interop testing has been performed using CryptoPro implementations.
10110 *Victor B. Wagner <vitus@cryptocom.ru>*
10112 * s390x assembler pack.
10116 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10121 * Implement Opaque PRF Input TLS extension as specified in
10122 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10123 official specification yet and no extension type assignment by
10124 IANA exists, this extension (for now) will have to be explicitly
10125 enabled when building OpenSSL by providing the extension number
10126 to use. For example, specify an option
10128 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10130 to the "config" or "Configure" script to enable the extension,
10131 assuming extension number 0x9527 (which is a completely arbitrary
10132 and unofficial assignment based on the MD5 hash of the Internet
10133 Draft). Note that by doing so, you potentially lose
10134 interoperability with other TLS implementations since these might
10135 be using the same extension number for other purposes.
10137 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10138 opaque PRF input value to use in the handshake. This will create
10139 an internal copy of the length-'len' string at 'src', and will
10140 return non-zero for success.
10142 To get more control and flexibility, provide a callback function
10145 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10146 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10150 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10153 Callback function 'cb' will be called in handshakes, and is
10154 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10155 Argument 'arg' is for application purposes (the value as given to
10156 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10157 be provided to the callback function). The callback function
10158 has to return non-zero to report success: usually 1 to use opaque
10159 PRF input just if possible, or 2 to enforce use of the opaque PRF
10160 input. In the latter case, the library will abort the handshake
10161 if opaque PRF input is not successfully negotiated.
10163 Arguments 'peerinput' and 'len' given to the callback function
10164 will always be NULL and 0 in the case of a client. A server will
10165 see the client's opaque PRF input through these variables if
10166 available (NULL and 0 otherwise). Note that if the server
10167 provides an opaque PRF input, the length must be the same as the
10168 length of the client's opaque PRF input.
10170 Note that the callback function will only be called when creating
10171 a new session (session resumption can resume whatever was
10172 previously negotiated), and will not be called in SSL 2.0
10173 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10174 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10175 for applications that need to enforce opaque PRF input.
10179 * Update ssl code to support digests other than SHA1+MD5 for handshake
10182 *Victor B. Wagner <vitus@cryptocom.ru>*
10184 * Add RFC4507 support to OpenSSL. This includes the corrections in
10185 RFC4507bis. The encrypted ticket format is an encrypted encoded
10186 SSL_SESSION structure, that way new session features are automatically
10189 If a client application caches session in an SSL_SESSION structure
10190 support is transparent because tickets are now stored in the encoded
10193 The SSL_CTX structure automatically generates keys for ticket
10194 protection in servers so again support should be possible
10195 with no application modification.
10197 If a client or server wishes to disable RFC4507 support then the option
10198 SSL_OP_NO_TICKET can be set.
10200 Add a TLS extension debugging callback to allow the contents of any client
10201 or server extensions to be examined.
10203 This work was sponsored by Google.
10207 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10208 OpenSSL should now compile cleanly on gcc 4.2
10210 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10212 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10213 support including streaming MAC support: this is required for GOST
10214 ciphersuite support.
10216 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10218 * Add option -stream to use PKCS#7 streaming in smime utility. New
10219 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10220 to output in BER and PEM format.
10224 * Experimental support for use of HMAC via EVP_PKEY interface. This
10225 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10226 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10227 ENGINE support for HMAC keys which are unextractable. New -mac and
10228 -macopt options to dgst utility.
10232 * New option -sigopt to dgst utility. Update dgst to use
10233 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10234 alternative signing parameters such as X9.31 or PSS in the dgst
10239 * Change ssl_cipher_apply_rule(), the internal function that does
10240 the work each time a ciphersuite string requests enabling
10241 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10242 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10243 the order of disabled ciphersuites such that those ciphersuites
10244 that most recently went from enabled to disabled not only stay
10245 in order with respect to each other, but also have higher priority
10246 than other disabled ciphersuites the next time ciphersuites are
10249 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10250 the same ciphersuites as with "HIGH" alone, but in a specific
10251 order where the PSK ciphersuites come first (since they are the
10252 most recently disabled ciphersuites when "HIGH" is parsed).
10254 Also, change ssl_create_cipher_list() (using this new
10255 functionality) such that between otherwise identical
10256 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10261 * Change ssl_create_cipher_list() so that it automatically
10262 arranges the ciphersuites in reasonable order before starting
10263 to process the rule string. Thus, the definition for "DEFAULT"
10264 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10265 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10266 This makes it much easier to arrive at a reasonable default order
10267 in applications for which anonymous ciphers are OK (meaning
10268 that you can't actually use DEFAULT).
10270 *Bodo Moeller; suggested by Victor Duchovni*
10272 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10273 processing) into multiple integers instead of setting
10274 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10275 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10276 (These masks as well as the individual bit definitions are hidden
10277 away into the non-exported interface ssl/ssl_locl.h, so this
10278 change to the definition of the SSL_CIPHER structure shouldn't
10279 affect applications.) This give us more bits for each of these
10280 categories, so there is no longer a need to coagulate AES128 and
10281 AES256 into a single algorithm bit, and to coagulate Camellia128
10282 and Camellia256 into a single algorithm bit, which has led to all
10285 Thus, among other things, the kludge introduced in 0.9.7m and
10286 0.9.8e for masking out AES256 independently of AES128 or masking
10287 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10289 With the change, we also introduce new ciphersuite aliases that
10290 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10295 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10296 Use the leftmost N bytes of the signature input if the input is
10297 larger than the prime q (with N being the size in bytes of q).
10301 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10302 it yet and it is largely untested.
10306 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10310 * Initial incomplete changes to avoid need for function casts in OpenSSL
10311 some compilers (gcc 4.2 and later) reject their use. Safestack is
10312 reimplemented. Update ASN1 to avoid use of legacy functions.
10316 * Win32/64 targets are linked with Winsock2.
10320 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10321 to external functions. This can be used to increase CRL handling
10322 efficiency especially when CRLs are very large by (for example) storing
10323 the CRL revoked certificates in a database.
10327 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10328 new CRLs added to a directory can be used. New command line option
10329 -verify_return_error to s_client and s_server. This causes real errors
10330 to be returned by the verify callback instead of carrying on no matter
10331 what. This reflects the way a "real world" verify callback would behave.
10335 * GOST engine, supporting several GOST algorithms and public key formats.
10336 Kindly donated by Cryptocom.
10340 * Partial support for Issuing Distribution Point CRL extension. CRLs
10341 partitioned by DP are handled but no indirect CRL or reason partitioning
10342 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10343 selected via a scoring technique which handles IDP and AKID in CRLs.
10347 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10348 will ultimately be used for all verify operations: this will remove the
10349 X509_STORE dependency on certificate verification and allow alternative
10350 lookup methods. X509_STORE based implementations of these two callbacks.
10354 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10355 Modify get_crl() to find a valid (unexpired) CRL if possible.
10359 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10360 this would be called X509_CRL_cmp() but that name is already used by
10361 a function that just compares CRL issuer names. Cache several CRL
10362 extensions in X509_CRL structure and cache CRLDP in X509.
10366 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10367 this maps equivalent X509_NAME structures into a consistent structure.
10368 Name comparison can then be performed rapidly using memcmp().
10372 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10377 * Allow digests to supply their own micalg string for S/MIME type using
10378 the ctrl EVP_MD_CTRL_MICALG.
10382 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10383 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10384 ctrl. It can then customise the structure before and/or after signing
10389 * New function OBJ_add_sigid() to allow application defined signature OIDs
10390 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10391 to free up any added signature OIDs.
10395 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10396 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10397 digest and cipher tables. New options added to openssl utility:
10398 list-message-digest-algorithms and list-cipher-algorithms.
10402 * Change the array representation of binary polynomials: the list
10403 of degrees of non-zero coefficients is now terminated with -1.
10404 Previously it was terminated with 0, which was also part of the
10405 value; thus, the array representation was not applicable to
10406 polynomials where t^0 has coefficient zero. This change makes
10407 the array representation useful in a more general context.
10411 * Various modifications and fixes to SSL/TLS cipher string
10412 handling. For ECC, the code now distinguishes between fixed ECDH
10413 with RSA certificates on the one hand and with ECDSA certificates
10414 on the other hand, since these are separate ciphersuites. The
10415 unused code for Fortezza ciphersuites has been removed.
10417 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10418 (not "ECDHE"). For consistency with the code for DH
10419 certificates, use of ECDH certificates is now considered ECDH
10420 authentication, not RSA or ECDSA authentication (the latter is
10421 merely the CA's signing algorithm and not actively used in the
10424 The temporary ciphersuite alias "ECCdraft" is no longer
10425 available, and ECC ciphersuites are no longer excluded from "ALL"
10426 and "DEFAULT". The following aliases now exist for RFC 4492
10427 ciphersuites, most of these by analogy with the DH case:
10429 kECDHr - ECDH cert, signed with RSA
10430 kECDHe - ECDH cert, signed with ECDSA
10431 kECDH - ECDH cert (signed with either RSA or ECDSA)
10432 kEECDH - ephemeral ECDH
10433 ECDH - ECDH cert or ephemeral ECDH
10436 aECDSA - ECDSA cert
10439 AECDH - anonymous ECDH
10440 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10444 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10445 Use correct micalg parameters depending on digest(s) in signed message.
10449 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10450 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10454 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10455 an engine to register a method. Add ENGINE lookups for methods and
10456 functional reference processing.
10460 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10461 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10466 * New -resign option to smime utility. This adds one or more signers
10467 to an existing PKCS#7 signedData structure. Also -md option to use an
10468 alternative message digest algorithm for signing.
10472 * Tidy up PKCS#7 routines and add new functions to make it easier to
10473 create PKCS7 structures containing multiple signers. Update smime
10474 application to support multiple signers.
10478 * New -macalg option to pkcs12 utility to allow setting of an alternative
10483 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10484 Reorganize PBE internals to lookup from a static table using NIDs,
10485 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10486 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10487 PRF which will be automatically used with PBES2.
10491 * Replace the algorithm specific calls to generate keys in "req" with the
10496 * Update PKCS#7 enveloped data routines to use new API. This is now
10497 supported by any public key method supporting the encrypt operation. A
10498 ctrl is added to allow the public key algorithm to examine or modify
10499 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10504 * Add a ctrl to asn1 method to allow a public key algorithm to express
10505 a default digest type to use. In most cases this will be SHA1 but some
10506 algorithms (such as GOST) need to specify an alternative digest. The
10507 return value indicates how strong the preference is 1 means optional and
10508 2 is mandatory (that is it is the only supported type). Modify
10509 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10510 use the default md. Update openssl utilities to use the default digest
10511 type for signing if it is not explicitly indicated.
10515 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10516 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10517 signing method from the key type. This effectively removes the link
10518 between digests and public key types.
10522 * Add an OID cross reference table and utility functions. Its purpose is to
10523 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10524 rsaEncryption. This will allow some of the algorithm specific hackery
10525 needed to use the correct OID to be removed.
10529 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10530 structures for PKCS7_sign(). They are now set up by the relevant public
10535 * Add provisional EC pkey method with support for ECDSA and ECDH.
10539 * Add support for key derivation (agreement) in the API, DH method and
10544 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10545 public and private key formats. As a side effect these add additional
10546 command line functionality not previously available: DSA signatures can be
10547 generated and verified using pkeyutl and DH key support and generation in
10554 *Oliver Tappe <zooey@hirschkaefer.de>*
10556 * New make target "install_html_docs" installs HTML renditions of the
10559 *Oliver Tappe <zooey@hirschkaefer.de>*
10561 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10562 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10563 support key and parameter generation and add initial key generation
10564 functionality for RSA.
10568 * Add functions for main EVP_PKEY_method operations. The undocumented
10569 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10570 `EVP_PKEY_{encrypt,decrypt}_old`.
10574 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10575 key API, doesn't do much yet.
10579 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10580 public key algorithms. New option to openssl utility:
10581 "list-public-key-algorithms" to print out info.
10585 * Implement the Supported Elliptic Curves Extension for
10586 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10590 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10591 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10595 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10596 utilities such as rsa, dsa, dsaparam etc except they process any key
10601 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10602 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10603 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10608 * Initial support for pluggable public key ASN1.
10609 De-spaghettify the public key ASN1 handling. Move public and private
10610 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10611 algorithm specific handling to a single module within the relevant
10612 algorithm directory. Add functions to allow (near) opaque processing
10613 of public and private key structures.
10617 * Implement the Supported Point Formats Extension for
10618 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10622 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10623 for the psk identity [hint] and the psk callback functions to the
10624 SSL_SESSION, SSL and SSL_CTX structure.
10627 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10631 SSL_CTX_use_psk_identity_hint
10632 SSL_get_psk_identity_hint
10633 SSL_get_psk_identity
10634 SSL_use_psk_identity_hint
10636 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10638 * Add RFC 3161 compliant time stamp request creation, response generation
10639 and response verification functionality.
10641 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10643 * Add initial support for TLS extensions, specifically for the server_name
10644 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10645 have new members for a hostname. The SSL data structure has an
10646 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10647 stored in that context to allow for session resumption, even after the
10648 SSL has been switched to a new SSL_CTX in reaction to a client's
10649 server_name extension.
10651 New functions (subject to change):
10653 SSL_get_servername()
10654 SSL_get_servername_type()
10657 New CTRL codes and macros (subject to change):
10659 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10660 - SSL_CTX_set_tlsext_servername_callback()
10661 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10662 - SSL_CTX_set_tlsext_servername_arg()
10663 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10665 openssl s_client has a new '-servername ...' option.
10667 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10668 '-key2 ...', '-servername_fatal' (subject to change). This allows
10669 testing the HostName extension for a specific single hostname ('-cert'
10670 and '-key' remain fallbacks for handshakes without HostName
10671 negotiation). If the unrecognized_name alert has to be sent, this by
10672 default is a warning; it becomes fatal with the '-servername_fatal'
10675 *Peter Sylvester, Remy Allais, Christophe Renou*
10677 * Whirlpool hash implementation is added.
10681 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10682 bn(64,32). Because of instruction set limitations it doesn't have
10683 any negative impact on performance. This was done mostly in order
10684 to make it possible to share assembler modules, such as bn_mul_mont
10685 implementations, between 32- and 64-bit builds without hassle.
10689 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10690 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10695 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10696 dedicated Montgomery multiplication procedure, is introduced.
10697 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10698 "64-bit" performance on certain 32-bit targets.
10702 * New option SSL_OP_NO_COMP to disable use of compression selectively
10703 in SSL structures. New SSL ctrl to set maximum send fragment size.
10704 Save memory by setting the I/O buffer sizes dynamically instead of
10705 using the maximum available value.
10709 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10710 in addition to the text details.
10714 * Very, very preliminary EXPERIMENTAL support for printing of general
10715 ASN1 structures. This currently produces rather ugly output and doesn't
10716 handle several customised structures at all.
10720 * Integrated support for PVK file format and some related formats such
10721 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10722 these in the 'rsa' and 'dsa' utilities.
10726 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10730 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10731 place for the (very old) "NETSCAPE" format certificates which are now
10732 handled using new ASN1 code equivalents.
10736 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10737 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10738 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10742 * Modify CRL distribution points extension code to print out previously
10743 unsupported fields. Enhance extension setting code to allow setting of
10748 * Add print and set support for Issuing Distribution Point CRL extension.
10752 * Change 'Configure' script to enable Camellia by default.
10759 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10761 * When rejecting SSL/TLS records due to an incorrect version number, never
10762 update s->server with a new major version number. As of
10763 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10764 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10765 the previous behavior could result in a read attempt at NULL when
10766 receiving specific incorrect SSL/TLS records once record payload
10767 protection is active. ([CVE-2010-0740])
10769 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10771 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10772 could be crashed if the relevant tables were not present (e.g. chrooted).
10774 *Tomas Hoger <thoger@redhat.com>*
10776 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10778 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10780 *Martin Olsson, Neel Mehta*
10782 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10783 accommodate for stack sorting, always a write lock!).
10787 * On some versions of WIN32 Heap32Next is very slow. This can cause
10788 excessive delays in the RAND_poll(): over a minute. As a workaround
10789 include a time check in the inner Heap32Next loop too.
10793 * The code that handled flushing of data in SSL/TLS originally used the
10794 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10795 the problem outlined in PR#1949. The fix suggested there however can
10796 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10797 of Apache). So instead simplify the code to flush unconditionally.
10798 This should be fine since flushing with no data to flush is a no op.
10802 * Handle TLS versions 2.0 and later properly and correctly use the
10803 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10804 off ancient servers have a habit of sticking around for a while...
10808 * Modify compression code so it frees up structures without using the
10809 ex_data callbacks. This works around a problem where some applications
10810 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10811 restarting) then use compression (e.g. SSL with compression) later.
10812 This results in significant per-connection memory leaks and
10813 has caused some security issues including CVE-2008-1678 and
10818 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10819 change when encrypting or decrypting.
10823 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10824 connect and renegotiate with servers which do not support RI.
10825 Until RI is more widely deployed this option is enabled by default.
10829 * Add "missing" ssl ctrls to clear options and mode.
10833 * If client attempts to renegotiate and doesn't support RI respond with
10834 a no_renegotiation alert as required by RFC5746. Some renegotiating
10835 TLS clients will continue a connection gracefully when they receive
10836 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10837 waiting for a server hello which it will never receive. Now we treat a
10838 received no_renegotiation alert as a fatal error. This is because
10839 applications requesting a renegotiation might well expect it to succeed
10840 and would have no code in place to handle the server denying it so the
10841 only safe thing to do is to terminate the connection.
10845 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10846 peer supports secure renegotiation and 0 otherwise. Print out peer
10847 renegotiation support in s_client/s_server.
10851 * Replace the highly broken and deprecated SPKAC certification method with
10852 the updated NID creation version. This should correctly handle UTF8.
10856 * Implement RFC5746. Re-enable renegotiation but require the extension
10857 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10858 turns out to be a bad idea. It has been replaced by
10859 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10860 SSL_CTX_set_options(). This is really not recommended unless you
10861 know what you are doing.
10863 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10865 * Fixes to stateless session resumption handling. Use initial_ctx when
10866 issuing and attempting to decrypt tickets in case it has changed during
10867 servername handling. Use a non-zero length session ID when attempting
10868 stateless session resumption: this makes it possible to determine if
10869 a resumption has occurred immediately after receiving server hello
10870 (several places in OpenSSL subtly assume this) instead of later in
10875 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10876 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10877 fixes for a few places where the return code is not checked
10880 *Julia Lawall <julia@diku.dk>*
10882 * Add --strict-warnings option to Configure script to include devteam
10883 warnings in other configurations.
10887 * Add support for --libdir option and LIBDIR variable in makefiles. This
10888 makes it possible to install openssl libraries in locations which
10889 have names other than "lib", for example "/usr/lib64" which some
10892 *Steve Henson, based on patch from Jeremy Utley*
10894 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10895 X690 8.9.12 and can produce some misleading textual output of OIDs.
10897 *Steve Henson, reported by Dan Kaminsky*
10899 * Delete MD2 from algorithm tables. This follows the recommendation in
10900 several standards that it is not used in new applications due to
10901 several cryptographic weaknesses. For binary compatibility reasons
10902 the MD2 API is still compiled in by default.
10906 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10911 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10912 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10915 *Guenter <lists@gknw.net>*
10917 * Fix the server certificate chain building code to use X509_verify_cert(),
10918 it used to have an ad-hoc builder which was unable to cope with anything
10919 other than a simple chain.
10921 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10923 * Don't check self signed certificate signatures in X509_verify_cert()
10924 by default (a flag can override this): it just wastes time without
10925 adding any security. As a useful side effect self signed root CAs
10926 with non-FIPS digests are now usable in FIPS mode.
10930 * In dtls1_process_out_of_seq_message() the check if the current message
10931 is already buffered was missing. For every new message was memory
10932 allocated, allowing an attacker to perform an denial of service attack
10933 with sending out of seq handshake messages until there is no memory
10934 left. Additionally every future message was buffered, even if the
10935 sequence number made no sense and would be part of another handshake.
10936 So only messages with sequence numbers less than 10 in advance will be
10937 buffered. ([CVE-2009-1378])
10939 *Robin Seggelmann, discovered by Daniel Mentz*
10941 * Records are buffered if they arrive with a future epoch to be
10942 processed after finishing the corresponding handshake. There is
10943 currently no limitation to this buffer allowing an attacker to perform
10944 a DOS attack with sending records with future epochs until there is no
10945 memory left. This patch adds the pqueue_size() function to determine
10946 the size of a buffer and limits the record buffer to 100 entries.
10949 *Robin Seggelmann, discovered by Daniel Mentz*
10951 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10952 parent structure is freed. ([CVE-2009-1379])
10956 * Handle non-blocking I/O properly in SSL_shutdown() call.
10958 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10960 * Add `2.5.4.*` OIDs
10962 *Ilya O. <vrghost@gmail.com>*
10964 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10966 * Disable renegotiation completely - this fixes a severe security
10967 problem ([CVE-2009-3555]) at the cost of breaking all
10968 renegotiation. Renegotiation can be re-enabled by setting
10969 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10970 run-time. This is really not recommended unless you know what
10975 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10977 * Don't set val to NULL when freeing up structures, it is freed up by
10978 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10979 zeroing past the valid field. ([CVE-2009-0789])
10981 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10983 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10984 checked correctly. This would allow some invalid signed attributes to
10985 appear to verify correctly. ([CVE-2009-0591])
10987 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10989 * Reject UniversalString and BMPString types with invalid lengths. This
10990 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10991 a legal length. ([CVE-2009-0590])
10995 * Set S/MIME signing as the default purpose rather than setting it
10996 unconditionally. This allows applications to override it at the store
11001 * Permit restricted recursion of ASN1 strings. This is needed in practice
11002 to handle some structures.
11006 * Improve efficiency of mem_gets: don't search whole buffer each time
11009 *Jeremy Shapiro <jnshapir@us.ibm.com>*
11011 * New -hex option for openssl rand.
11015 * Print out UTF8String and NumericString when parsing ASN1.
11019 * Support NumericString type for name components.
11023 * Allow CC in the environment to override the automatically chosen
11024 compiler. Note that nothing is done to ensure flags work with the
11029 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
11031 * Properly check EVP_VerifyFinal() and similar return values
11034 *Ben Laurie, Bodo Moeller, Google Security Team*
11036 * Enable TLS extensions by default.
11040 * Allow the CHIL engine to be loaded, whether the application is
11041 multithreaded or not. (This does not release the developer from the
11042 obligation to set up the dynamic locking callbacks.)
11044 *Sander Temme <sander@temme.net>*
11046 * Use correct exit code if there is an error in dgst command.
11048 *Steve Henson; problem pointed out by Roland Dirlewanger*
11050 * Tweak Configure so that you need to say "experimental-jpake" to enable
11051 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
11055 * Add experimental JPAKE support, including demo authentication in
11056 s_client and s_server.
11060 * Set the comparison function in v3_addr_canonize().
11062 *Rob Austein <sra@hactrn.net>*
11064 * Add support for XMPP STARTTLS in s_client.
11066 *Philip Paeps <philip@freebsd.org>*
11068 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
11069 to ensure that even with this option, only ciphersuites in the
11070 server's preference list will be accepted. (Note that the option
11071 applies only when resuming a session, so the earlier behavior was
11072 just about the algorithm choice for symmetric cryptography.)
11076 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
11078 * Fix NULL pointer dereference if a DTLS server received
11079 ChangeCipherSpec as first record ([CVE-2009-1386]).
11083 * Fix a state transition in s3_srvr.c and d1_srvr.c
11084 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
11086 *Nagendra Modadugu*
11088 * The fix in 0.9.8c that supposedly got rid of unsafe
11089 double-checked locking was incomplete for RSA blinding,
11090 addressing just one layer of what turns out to have been
11091 doubly unsafe triple-checked locking.
11093 So now fix this for real by retiring the MONT_HELPER macro
11094 in crypto/rsa/rsa_eay.c.
11096 *Bodo Moeller; problem pointed out by Marius Schilder*
11098 * Various precautionary measures:
11100 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11102 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11103 (NB: This would require knowledge of the secret session ticket key
11104 to exploit, in which case you'd be SOL either way.)
11106 - Change bn_nist.c so that it will properly handle input BIGNUMs
11107 outside the expected range.
11109 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11112 *Neel Mehta, Bodo Moeller*
11114 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11115 the load fails. Useful for distros.
11117 *Ben Laurie and the FreeBSD team*
11119 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11123 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11127 * Expand ENGINE to support engine supplied SSL client certificate functions.
11129 This work was sponsored by Logica.
11133 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11134 keystores. Support for SSL/TLS client authentication too.
11135 Not compiled unless enable-capieng specified to Configure.
11137 This work was sponsored by Logica.
11141 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11142 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11143 attribute creation routines such as certificate requests and PKCS#12
11148 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11150 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11151 handshake which could lead to a client crash as found using the
11152 Codenomicon TLS test suite ([CVE-2008-1672])
11154 *Steve Henson, Mark Cox*
11156 * Fix double free in TLS server name extensions which could lead to
11157 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11161 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11163 Clear the error queue to ensure that error entries left from
11164 older function calls do not interfere with the correct operation.
11166 *Lutz Jaenicke, Erik de Castro Lopo*
11168 * Remove root CA certificates of commercial CAs:
11170 The OpenSSL project does not recommend any specific CA and does not
11171 have any policy with respect to including or excluding any CA.
11172 Therefore, it does not make any sense to ship an arbitrary selection
11173 of root CA certificates with the OpenSSL software.
11177 * RSA OAEP patches to fix two separate invalid memory reads.
11178 The first one involves inputs when 'lzero' is greater than
11179 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11180 before the beginning of from). The second one involves inputs where
11181 the 'db' section contains nothing but zeroes (there is a one-byte
11182 invalid read after the end of 'db').
11184 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11186 * Partial backport from 0.9.9-dev:
11188 Introduce bn_mul_mont (dedicated Montgomery multiplication
11189 procedure) as a candidate for BIGNUM assembler implementation.
11190 While 0.9.9-dev uses assembler for various architectures, only
11191 x86_64 is available by default here in the 0.9.8 branch, and
11192 32-bit x86 is available through a compile-time setting.
11194 To try the 32-bit x86 assembler implementation, use Configure
11195 option "enable-montasm" (which exists only for this backport).
11197 As "enable-montasm" for 32-bit x86 disclaims code stability
11198 anyway, in this constellation we activate additional code
11199 backported from 0.9.9-dev for further performance improvements,
11200 namely BN_from_montgomery_word. (To enable this otherwise,
11201 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11203 *Andy Polyakov (backport partially by Bodo Moeller)*
11205 * Add TLS session ticket callback. This allows an application to set
11206 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11207 values. This is useful for key rollover for example where several key
11208 sets may exist with different names.
11212 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11213 This was broken until now in 0.9.8 releases, such that the only way
11214 a registered ENGINE could be used (assuming it initialises
11215 successfully on the host) was to explicitly set it as the default
11216 for the relevant algorithms. This is in contradiction with 0.9.7
11217 behaviour and the documentation. With this fix, when an ENGINE is
11218 registered into a given algorithm's table of implementations, the
11219 'uptodate' flag is reset so that auto-discovery will be used next
11220 time a new context for that algorithm attempts to select an
11223 *Ian Lister (tweaked by Geoff Thorpe)*
11225 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11226 implementation in the following ways:
11228 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11231 Lack of BER streaming support means one pass streaming processing is
11232 only supported if data is detached: setting the streaming flag is
11233 ignored for embedded content.
11235 CMS support is disabled by default and must be explicitly enabled
11236 with the enable-cms configuration option.
11240 * Update the GMP engine glue to do direct copies between BIGNUM and
11241 mpz_t when openssl and GMP use the same limb size. Otherwise the
11242 existing "conversion via a text string export" trick is still used.
11244 *Paul Sheer <paulsheer@gmail.com>*
11246 * Zlib compression BIO. This is a filter BIO which compressed and
11247 uncompresses any data passed through it.
11251 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11252 RFC3394 compatible AES key wrapping.
11256 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11257 sets string data without copying. X509_ALGOR_set0() and
11258 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11259 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11260 from an X509_ATTRIBUTE structure optionally checking it occurs only
11261 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11266 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11267 to get the expected BN_FLG_CONSTTIME behavior.
11269 *Bodo Moeller (Google)*
11273 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11274 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11275 - added some more tests to do_tests.pl
11276 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11277 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11278 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11279 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11280 - various changes to netware.pl to enable gcc-cross builds on Win32
11282 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11283 - various changes to fix missing prototype warnings
11284 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11285 - added AES, WHIRLPOOL and CPUID assembler code to build files
11286 - added missing AES assembler make rules to mk1mf.pl
11287 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11289 *Guenter Knauf <eflash@gmx.net>*
11291 * Implement certificate status request TLS extension defined in RFC3546.
11292 A client can set the appropriate parameters and receive the encoded
11293 OCSP response via a callback. A server can query the supplied parameters
11294 and set the encoded OCSP response in the callback. Add simplified examples
11295 to s_client and s_server.
11299 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11301 * Fix various bugs:
11302 + Binary incompatibility of ssl_ctx_st structure
11303 + DTLS interoperation with non-compliant servers
11304 + Don't call get_session_cb() without proposed session
11305 + Fix ia64 assembler code
11307 *Andy Polyakov, Steve Henson*
11309 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11311 * DTLS Handshake overhaul. There were longstanding issues with
11312 OpenSSL DTLS implementation, which were making it impossible for
11313 RFC 4347 compliant client to communicate with OpenSSL server.
11314 Unfortunately just fixing these incompatibilities would "cut off"
11315 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11316 server keeps tolerating non RFC compliant syntax. The opposite is
11317 not true, 0.9.8f client can not communicate with earlier server.
11318 This update even addresses CVE-2007-4995.
11322 * Changes to avoid need for function casts in OpenSSL: some compilers
11323 (gcc 4.2 and later) reject their use.
11324 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11327 * Add RFC4507 support to OpenSSL. This includes the corrections in
11328 RFC4507bis. The encrypted ticket format is an encrypted encoded
11329 SSL_SESSION structure, that way new session features are automatically
11332 If a client application caches session in an SSL_SESSION structure
11333 support is transparent because tickets are now stored in the encoded
11336 The SSL_CTX structure automatically generates keys for ticket
11337 protection in servers so again support should be possible
11338 with no application modification.
11340 If a client or server wishes to disable RFC4507 support then the option
11341 SSL_OP_NO_TICKET can be set.
11343 Add a TLS extension debugging callback to allow the contents of any client
11344 or server extensions to be examined.
11346 This work was sponsored by Google.
11350 * Add initial support for TLS extensions, specifically for the server_name
11351 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11352 have new members for a hostname. The SSL data structure has an
11353 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11354 stored in that context to allow for session resumption, even after the
11355 SSL has been switched to a new SSL_CTX in reaction to a client's
11356 server_name extension.
11358 New functions (subject to change):
11360 SSL_get_servername()
11361 SSL_get_servername_type()
11364 New CTRL codes and macros (subject to change):
11366 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11367 - SSL_CTX_set_tlsext_servername_callback()
11368 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11369 - SSL_CTX_set_tlsext_servername_arg()
11370 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11372 openssl s_client has a new '-servername ...' option.
11374 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11375 '-key2 ...', '-servername_fatal' (subject to change). This allows
11376 testing the HostName extension for a specific single hostname ('-cert'
11377 and '-key' remain fallbacks for handshakes without HostName
11378 negotiation). If the unrecognized_name alert has to be sent, this by
11379 default is a warning; it becomes fatal with the '-servername_fatal'
11382 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11384 * Add AES and SSE2 assembly language support to VC++ build.
11388 * Mitigate attack on final subtraction in Montgomery reduction.
11392 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11393 (which previously caused an internal error).
11397 * Squeeze another 10% out of IGE mode when in != out.
11401 * AES IGE mode speedup.
11403 *Dean Gaudet (Google)*
11405 * Add the Korean symmetric 128-bit cipher SEED (see
11406 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11407 add SEED ciphersuites from RFC 4162:
11409 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11410 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11411 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11412 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11414 To minimize changes between patchlevels in the OpenSSL 0.9.8
11415 series, SEED remains excluded from compilation unless OpenSSL
11416 is configured with 'enable-seed'.
11418 *KISA, Bodo Moeller*
11420 * Mitigate branch prediction attacks, which can be practical if a
11421 single processor is shared, allowing a spy process to extract
11422 information. For detailed background information, see
11423 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11424 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11425 and Necessary Software Countermeasures"). The core of the change
11426 are new versions BN_div_no_branch() and
11427 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11428 respectively, which are slower, but avoid the security-relevant
11429 conditional branches. These are automatically called by BN_div()
11430 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11431 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11432 remove a conditional branch.
11434 BN_FLG_CONSTTIME is the new name for the previous
11435 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11436 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11437 in the exponent causes BN_mod_exp_mont() to use the alternative
11438 implementation in BN_mod_exp_mont_consttime().) The old name
11439 remains as a deprecated alias.
11441 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11442 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11443 constant-time implementations for more than just exponentiation.
11444 Here too the old name is kept as a deprecated alias.
11446 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11447 the BN_BLINDING structure gets an independent copy of the
11448 modulus. This means that the previous `BIGNUM *m` argument to
11449 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11450 essentially becomes `const BIGNUM *m`, although we can't actually
11451 change this in the header file before 0.9.9. It allows
11452 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11453 enable BN_FLG_CONSTTIME.
11455 *Matthew D Wood (Intel Corp)*
11457 * In the SSL/TLS server implementation, be strict about session ID
11458 context matching (which matters if an application uses a single
11459 external cache for different purposes). Previously,
11460 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11461 set. This did ensure strict client verification, but meant that,
11462 with applications using a single external cache for quite
11463 different requirements, clients could circumvent ciphersuite
11464 restrictions for a given session ID context by starting a session
11465 in a different context.
11469 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11470 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11471 authentication-only ciphersuites.
11475 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11476 not complete and could lead to a possible single byte overflow
11477 ([CVE-2007-5135]) [Ben Laurie]
11479 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11481 * Since AES128 and AES256 (and similarly Camellia128 and
11482 Camellia256) share a single mask bit in the logic of
11483 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11484 kludge to work properly if AES128 is available and AES256 isn't
11485 (or if Camellia128 is available and Camellia256 isn't).
11489 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11490 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11491 When a point or a seed is encoded in a BIT STRING, we need to
11492 prevent the removal of trailing zero bits to get the proper DER
11493 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11494 of a NamedBitList, for which trailing 0 bits need to be removed.)
11498 * Have SSL/TLS server implementation tolerate "mismatched" record
11499 protocol version while receiving ClientHello even if the
11500 ClientHello is fragmented. (The server can't insist on the
11501 particular protocol version it has chosen before the ServerHello
11502 message has informed the client about his choice.)
11506 * Add RFC 3779 support.
11508 *Rob Austein for ARIN, Ben Laurie*
11510 * Load error codes if they are not already present instead of using a
11511 static variable. This allows them to be cleanly unloaded and reloaded.
11512 Improve header file function name parsing.
11516 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11517 or CAPABILITY handshake as required by RFCs.
11519 *Goetz Babin-Ebell*
11521 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11523 * Introduce limits to prevent malicious keys being able to
11524 cause a denial of service. ([CVE-2006-2940])
11526 *Steve Henson, Bodo Moeller*
11528 * Fix ASN.1 parsing of certain invalid structures that can result
11529 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11531 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11532 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11534 * Fix SSL client code which could crash if connecting to a
11535 malicious SSLv2 server. ([CVE-2006-4343])
11537 *Tavis Ormandy and Will Drewry, Google Security Team*
11539 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11540 match only those. Before that, "AES256-SHA" would be interpreted
11541 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11542 the same strength classification in 0.9.7h) as we currently only
11543 have a single AES bit in the ciphersuite description bitmap.
11544 That change, however, also applied to ciphersuite strings such as
11545 "RC4-MD5" that intentionally matched multiple ciphersuites --
11546 namely, SSL 2.0 ciphersuites in addition to the more common ones
11547 from SSL 3.0/TLS 1.0.
11549 So we change the selection algorithm again: Naming an explicit
11550 ciphersuite selects this one ciphersuite, and any other similar
11551 ciphersuite (same bitmap) from *other* protocol versions.
11552 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11553 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11555 Since SSL 2.0 does not have any ciphersuites for which the
11556 128/256 bit distinction would be relevant, this works for now.
11557 The proper fix will be to use different bits for AES128 and
11558 AES256, which would have avoided the problems from the beginning;
11559 however, bits are scarce, so we can only do this in a new release
11560 (not just a patchlevel) when we can change the SSL_CIPHER
11561 definition to split the single 'unsigned long mask' bitmap into
11562 multiple values to extend the available space.
11566 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11568 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11569 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11571 * Add AES IGE and biIGE modes.
11575 * Change the Unix randomness entropy gathering to use poll() when
11576 possible instead of select(), since the latter has some
11577 undesirable limitations.
11579 *Darryl Miles via Richard Levitte and Bodo Moeller*
11581 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11582 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11583 cannot be implicitly activated as part of, e.g., the "AES" alias.
11584 However, please upgrade to OpenSSL 0.9.9[-dev] for
11585 non-experimental use of the ECC ciphersuites to get TLS extension
11586 support, which is required for curve and point format negotiation
11587 to avoid potential handshake problems.
11591 * Disable rogue ciphersuites:
11593 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11594 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11595 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11597 The latter two were purportedly from
11598 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11601 Also deactivate the remaining ciphersuites from
11602 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11603 unofficial, and the ID has long expired.
11607 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11608 dual-core machines) and other potential thread-safety issues.
11612 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11613 versions), which is now available for royalty-free use
11614 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11615 Also, add Camellia TLS ciphersuites from RFC 4132.
11617 To minimize changes between patchlevels in the OpenSSL 0.9.8
11618 series, Camellia remains excluded from compilation unless OpenSSL
11619 is configured with 'enable-camellia'.
11623 * Disable the padding bug check when compression is in use. The padding
11624 bug check assumes the first packet is of even length, this is not
11625 necessarily true if compression is enabled and can result in false
11626 positives causing handshake failure. The actual bug test is ancient
11627 code so it is hoped that implementations will either have fixed it by
11628 now or any which still have the bug do not support compression.
11632 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11634 * When applying a cipher rule check to see if string match is an explicit
11635 cipher suite and only match that one cipher suite if it is.
11639 * Link in manifests for VC++ if needed.
11641 *Austin Ziegler <halostatue@gmail.com>*
11643 * Update support for ECC-based TLS ciphersuites according to
11644 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11645 TLS extensions, which are supported starting with the 0.9.9
11646 branch, not in the OpenSSL 0.9.8 branch).
11650 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11651 opaque EVP_CIPHER_CTX handling.
11655 * Fixes and enhancements to zlib compression code. We now only use
11656 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11657 to conform with the standards mentioned here:
11658 <http://www.zlib.net/DLL_FAQ.txt>
11659 Static zlib linking now works on Windows and the new --with-zlib-include
11660 --with-zlib-lib options to Configure can be used to supply the location
11661 of the headers and library. Gracefully handle case where zlib library
11666 * Several fixes and enhancements to the OID generation code. The old code
11667 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11668 handle numbers larger than ULONG_MAX, truncated printing and had a
11669 non standard OBJ_obj2txt() behaviour.
11673 * Add support for building of engines under engine/ as shared libraries
11674 under VC++ build system.
11678 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11679 Hopefully, we will not see any false combination of paths any more.
11683 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11685 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11686 (part of SSL_OP_ALL). This option used to disable the
11687 countermeasure against man-in-the-middle protocol-version
11688 rollback in the SSL 2.0 server implementation, which is a bad
11689 idea. ([CVE-2005-2969])
11691 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11692 for Information Security, National Institute of Advanced Industrial
11693 Science and Technology [AIST], Japan)*
11695 * Add two function to clear and return the verify parameter flags.
11699 * Keep cipherlists sorted in the source instead of sorting them at
11700 runtime, thus removing the need for a lock.
11704 * Avoid some small subgroup attacks in Diffie-Hellman.
11706 *Nick Mathewson and Ben Laurie*
11708 * Add functions for well-known primes.
11712 * Extended Windows CE support.
11714 *Satoshi Nakamura and Andy Polyakov*
11716 * Initialize SSL_METHOD structures at compile time instead of during
11717 runtime, thus removing the need for a lock.
11721 * Make PKCS7_decrypt() work even if no certificate is supplied by
11722 attempting to decrypt each encrypted key in turn. Add support to
11727 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11729 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11732 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11736 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11737 key into the same file any more.
11741 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11745 * Add -utf8 command line and config file option to 'ca'.
11747 *Stefan <stf@udoma.org*
11749 * Removed the macro des_crypt(), as it seems to conflict with some
11750 libraries. Use DES_crypt().
11754 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11755 involves renaming the source and generated shared-libs for
11756 both. The engines will accept the corrected or legacy ids
11757 ('ncipher' and '4758_cca' respectively) when binding. NB,
11758 this only applies when building 'shared'.
11760 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11762 * Add attribute functions to EVP_PKEY structure. Modify
11763 PKCS12_create() to recognize a CSP name attribute and
11764 use it. Make -CSP option work again in pkcs12 utility.
11768 * Add new functionality to the bn blinding code:
11769 - automatic re-creation of the BN_BLINDING parameters after
11770 a fixed number of uses (currently 32)
11771 - add new function for parameter creation
11772 - introduce flags to control the update behaviour of the
11773 BN_BLINDING parameters
11774 - hide BN_BLINDING structure
11775 Add a second BN_BLINDING slot to the RSA structure to improve
11776 performance when a single RSA object is shared among several
11781 * Add support for DTLS.
11783 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11785 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11786 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11790 * Remove buggy and incomplete DH cert support from
11791 ssl/ssl_rsa.c and ssl/s3_both.c
11795 * Use SHA-1 instead of MD5 as the default digest algorithm for
11796 the `apps/openssl` commands.
11800 * Compile clean with "-Wall -Wmissing-prototypes
11801 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11802 DEBUG_SAFESTACK must also be set.
11806 * Change ./Configure so that certain algorithms can be disabled by default.
11807 The new counterpiece to "no-xxx" is "enable-xxx".
11809 The patented RC5 and MDC2 algorithms will now be disabled unless
11810 "enable-rc5" and "enable-mdc2", respectively, are specified.
11812 (IDEA remains enabled despite being patented. This is because IDEA
11813 is frequently required for interoperability, and there is no license
11814 fee for non-commercial use. As before, "no-idea" can be used to
11815 avoid this algorithm.)
11819 * Add processing of proxy certificates (see RFC 3820). This work was
11820 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11821 EGEE (Enabling Grids for E-science in Europe).
11825 * RC4 performance overhaul on modern architectures/implementations, such
11826 as Intel P4, IA-64 and AMD64.
11830 * New utility extract-section.pl. This can be used specify an alternative
11831 section number in a pod file instead of having to treat each file as
11832 a separate case in Makefile. This can be done by adding two lines to the
11835 =for comment openssl_section:XXX
11837 The blank line is mandatory.
11841 * New arguments -certform, -keyform and -pass for s_client and s_server
11842 to allow alternative format key and certificate files and passphrase
11847 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11848 update associated structures and add various utility functions.
11850 Add new policy related verify parameters, include policy checking in
11851 standard verify code. Enhance 'smime' application with extra parameters
11852 to support policy checking and print out.
11856 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11857 Nehemiah processors. These extensions support AES encryption in hardware
11858 as well as RNG (though RNG support is currently disabled).
11860 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11862 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11866 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11868 *Andy Polyakov and a number of other people*
11870 * Improved PowerPC platform support. Most notably BIGNUM assembler
11871 implementation contributed by IBM.
11873 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11875 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11876 exponent rather than 'unsigned long'. There is a corresponding change to
11877 the new 'rsa_keygen' element of the RSA_METHOD structure.
11879 *Jelte Jansen, Geoff Thorpe*
11881 * Functionality for creating the initial serial number file is now
11882 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11884 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11885 number file to 1, which is bound to cause problems. To avoid
11886 the problems while respecting compatibility between different 0.9.7
11887 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11888 CA.pl for serial number initialization. With the new release 0.9.8,
11889 we can fix the problem directly in the 'ca' utility.)
11893 * Reduced header interdependencies by declaring more opaque objects in
11894 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11895 give fewer recursive includes, which could break lazy source code - so
11896 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11897 developers should define this symbol when building and using openssl to
11898 ensure they track the recommended behaviour, interfaces, [etc], but
11899 backwards-compatible behaviour prevails when this isn't defined.
11903 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11907 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11908 This will generate a random key of the appropriate length based on the
11909 cipher context. The EVP_CIPHER can provide its own random key generation
11910 routine to support keys of a specific form. This is used in the des and
11911 3des routines to generate a key of the correct parity. Update S/MIME
11912 code to use new functions and hence generate correct parity DES keys.
11913 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11914 valid (weak or incorrect parity).
11918 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11919 as looking them up. This is useful when the verified structure may contain
11920 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11921 present unless the new PKCS7_NO_CRL flag is asserted.
11925 * Extend ASN1 oid configuration module. It now additionally accepts the
11928 shortName = some long name, 1.2.3.4
11932 * Reimplemented the BN_CTX implementation. There is now no more static
11933 limitation on the number of variables it can handle nor the depth of the
11934 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11935 information can now expand as required, and rather than having a single
11936 static array of bignums, BN_CTX now uses a linked-list of such arrays
11937 allowing it to expand on demand whilst maintaining the usefulness of
11938 BN_CTX's "bundling".
11942 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11943 to allow all RSA operations to function using a single BN_CTX.
11947 * Preliminary support for certificate policy evaluation and checking. This
11948 is initially intended to pass the tests outlined in "Conformance Testing
11949 of Relying Party Client Certificate Path Processing Logic" v1.07.
11953 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11954 remained unused and not that useful. A variety of other little bignum
11955 tweaks and fixes have also been made continuing on from the audit (see
11960 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11961 associated ASN1, EVP and SSL functions and old ASN1 macros.
11965 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11966 and this should never fail. So the return value from the use of
11967 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11968 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11972 * BN_CTX_get() should return zero-valued bignums, providing the same
11973 initialised value as BN_new().
11975 *Geoff Thorpe, suggested by Ulf Möller*
11977 * Support for inhibitAnyPolicy certificate extension.
11981 * An audit of the BIGNUM code is underway, for which debugging code is
11982 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11983 is considered valid when processing BIGNUMs, and causes execution to
11984 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11985 further steps are taken to deliberately pollute unused data in BIGNUM
11986 structures to try and expose faulty code further on. For now, openssl will
11987 (in its default mode of operation) continue to tolerate the inconsistent
11988 forms that it has tolerated in the past, but authors and packagers should
11989 consider trying openssl and their own applications when compiled with
11990 these debugging symbols defined. It will help highlight potential bugs in
11991 their own code, and will improve the test coverage for OpenSSL itself. At
11992 some point, these tighter rules will become openssl's default to improve
11993 maintainability, though the assert()s and other overheads will remain only
11994 in debugging configurations. See bn.h for more details.
11996 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11998 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11999 that can only be obtained through BN_CTX_new() (which implicitly
12000 initialises it). The presence of this function only made it possible
12001 to overwrite an existing structure (and cause memory leaks).
12005 * Because of the callback-based approach for implementing LHASH as a
12006 template type, lh_insert() adds opaque objects to hash-tables and
12007 lh_doall() or lh_doall_arg() are typically used with a destructor callback
12008 to clean up those corresponding objects before destroying the hash table
12009 (and losing the object pointers). So some over-zealous constifications in
12010 LHASH have been relaxed so that lh_insert() does not take (nor store) the
12011 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
12012 prototyped to have "const" restrictions on the object pointers they are
12013 given (and so aren't required to cast them away any more).
12017 * The tmdiff.h API was so ugly and minimal that our own timing utility
12018 (speed) prefers to use its own implementation. The two implementations
12019 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
12020 its object type properly exposed (MS_TM) instead of casting to/from
12021 `char *`. This may still change yet if someone realises MS_TM and
12023 aren't necessarily the greatest nomenclatures - but this is what was used
12024 internally to the implementation so I've used that for now.
12028 * Ensure that deprecated functions do not get compiled when
12029 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
12030 the self-tests were still using deprecated key-generation functions so
12031 these have been updated also.
12035 * Reorganise PKCS#7 code to separate the digest location functionality
12036 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
12037 New function PKCS7_set_digest() to set the digest type for PKCS#7
12038 digestedData type. Add additional code to correctly generate the
12039 digestedData type and add support for this type in PKCS7 initialization
12044 * New function PKCS7_set0_type_other() this initializes a PKCS7
12045 structure of type "other".
12049 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
12050 sure the loop does correctly stop and breaking ("division by zero")
12051 modulus operations are not performed. The (pre-generated) prime
12052 table crypto/bn/bn_prime.h was already correct, but it could not be
12053 re-generated on some platforms because of the "division by zero"
12054 situation in the script.
12056 *Ralf S. Engelschall*
12058 * Update support for ECC-based TLS ciphersuites according to
12059 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
12060 SHA-1 now is only used for "small" curves (where the
12061 representation of a field element takes up to 24 bytes); for
12062 larger curves, the field element resulting from ECDH is directly
12063 used as premaster secret.
12065 *Douglas Stebila (Sun Microsystems Laboratories)*
12067 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
12068 curve secp160r1 to the tests.
12070 *Douglas Stebila (Sun Microsystems Laboratories)*
12072 * Add the possibility to load symbols globally with DSO.
12074 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
12076 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
12077 control of the error stack.
12081 * Add support for STORE in ENGINE.
12085 * Add the STORE type. The intention is to provide a common interface
12086 to certificate and key stores, be they simple file-based stores, or
12087 HSM-type store, or LDAP stores, or...
12088 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12092 * Add a generic structure called OPENSSL_ITEM. This can be used to
12093 pass a list of arguments to any function as well as provide a way
12094 for a function to pass data back to the caller.
12098 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12099 works like BUF_strdup() but can be used to duplicate a portion of
12100 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12105 * Add the function sk_find_ex() which works like sk_find(), but will
12106 return an index to an element even if an exact match couldn't be
12107 found. The index is guaranteed to point at the element where the
12108 searched-for key would be inserted to preserve sorting order.
12112 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12113 takes an extra flags argument for optional functionality. Currently,
12114 the following flags are defined:
12116 OBJ_BSEARCH_VALUE_ON_NOMATCH
12117 This one gets OBJ_bsearch_ex() to return a pointer to the first
12118 element where the comparing function returns a negative or zero
12121 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12122 This one gets OBJ_bsearch_ex() to return a pointer to the first
12123 element where the comparing function returns zero. This is useful
12124 if there are more than one element where the comparing function
12129 * Make it possible to create self-signed certificates with 'openssl ca'
12130 in such a way that the self-signed certificate becomes part of the
12131 CA database and uses the same mechanisms for serial number generation
12132 as all other certificate signing. The new flag '-selfsign' enables
12133 this functionality. Adapt CA.sh and CA.pl.in.
12137 * Add functionality to check the public key of a certificate request
12138 against a given private. This is useful to check that a certificate
12139 request can be signed by that key (self-signing).
12143 * Make it possible to have multiple active certificates with the same
12144 subject in the CA index file. This is done only if the keyword
12145 'unique_subject' is set to 'no' in the main CA section (default
12146 if 'CA_default') of the configuration file. The value is saved
12147 with the database itself in a separate index attribute file,
12148 named like the index file with '.attr' appended to the name.
12152 * Generate multi-valued AVAs using '+' notation in config files for
12157 * Support for nameConstraints certificate extension.
12161 * Support for policyConstraints certificate extension.
12165 * Support for policyMappings certificate extension.
12169 * Make sure the default DSA_METHOD implementation only uses its
12170 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12171 and change its own handlers to be NULL so as to remove unnecessary
12172 indirection. This lets alternative implementations fallback to the
12173 default implementation more easily.
12177 * Support for directoryName in GeneralName related extensions
12182 * Make it possible to link applications using Makefile.shared.
12183 Make that possible even when linking against static libraries!
12187 * Support for single pass processing for S/MIME signing. This now
12188 means that S/MIME signing can be done from a pipe, in addition
12189 cleartext signing (multipart/signed type) is effectively streaming
12190 and the signed data does not need to be all held in memory.
12192 This is done with a new flag PKCS7_STREAM. When this flag is set
12193 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12194 is done after the data is output (and digests calculated) in
12195 SMIME_write_PKCS7().
12199 * Add full support for -rpath/-R, both in shared libraries and
12200 applications, at least on the platforms where it's known how
12205 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12206 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12207 will now compute a table of multiples of the generator that
12208 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12209 faster (notably in the case of a single point multiplication,
12210 scalar * generator).
12212 *Nils Larsch, Bodo Moeller*
12214 * IPv6 support for certificate extensions. The various extensions
12215 which use the IP:a.b.c.d can now take IPv6 addresses using the
12216 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12221 * Added an ENGINE that implements RSA by performing private key
12222 exponentiations with the GMP library. The conversions to and from
12223 GMP's mpz_t format aren't optimised nor are any montgomery forms
12224 cached, and on x86 it appears OpenSSL's own performance has caught up.
12225 However there are likely to be other architectures where GMP could
12226 provide a boost. This ENGINE is not built in by default, but it can be
12227 specified at Configure time and should be accompanied by the necessary
12228 linker additions, eg;
12229 ./config -DOPENSSL_USE_GMP -lgmp
12233 * "openssl engine" will not display ENGINE/DSO load failure errors when
12234 testing availability of engines with "-t" - the old behaviour is
12235 produced by increasing the feature's verbosity with "-tt".
12239 * ECDSA routines: under certain error conditions uninitialized BN objects
12240 could be freed. Solution: make sure initialization is performed early
12241 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12246 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12247 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12248 software implementations. For DSA and DH, parameter generation can
12249 also be overridden by providing the appropriate method callbacks.
12253 * Change the "progress" mechanism used in key-generation and
12254 primality testing to functions that take a new BN_GENCB pointer in
12255 place of callback/argument pairs. The new API functions have `_ex`
12256 postfixes and the older functions are reimplemented as wrappers for
12257 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12258 declarations of the old functions to help (graceful) attempts to
12259 migrate to the new functions. Also, the new key-generation API
12260 functions operate on a caller-supplied key-structure and return
12261 success/failure rather than returning a key or NULL - this is to
12262 help make "keygen" another member function of RSA_METHOD etc.
12264 Example for using the new callback interface:
12266 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12267 void *my_arg = ...;
12270 BN_GENCB_set(&my_cb, my_callback, my_arg);
12272 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12273 /* For the meaning of a, b in calls to my_callback(), see the
12274 * documentation of the function that calls the callback.
12275 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12276 * my_callback should return 1 if it wants BN_is_prime_ex()
12277 * to continue, or 0 to stop.
12282 * Change the ZLIB compression method to be stateful, and make it
12283 available to TLS with the number defined in
12284 draft-ietf-tls-compression-04.txt.
12288 * Add the ASN.1 structures and functions for CertificatePair, which
12289 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12291 CertificatePair ::= SEQUENCE {
12292 forward [0] Certificate OPTIONAL,
12293 reverse [1] Certificate OPTIONAL,
12294 -- at least one of the pair shall be present -- }
12296 Also implement the PEM functions to read and write certificate
12297 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12299 This needed to be defined, mostly for the sake of the LDAP
12300 attribute crossCertificatePair, but may prove useful elsewhere as
12305 * Make it possible to inhibit symlinking of shared libraries in
12306 Makefile.shared, for Cygwin's sake.
12310 * Extend the BIGNUM API by creating a function
12311 void BN_set_negative(BIGNUM *a, int neg);
12312 and a macro that behave like
12313 int BN_is_negative(const BIGNUM *a);
12315 to avoid the need to access 'a->neg' directly in applications.
12319 * Implement fast modular reduction for pseudo-Mersenne primes
12320 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12321 EC_GROUP_new_curve_GFp() will now automatically use this
12324 *Nils Larsch <nla@trustcenter.de>*
12326 * Add new lock type (CRYPTO_LOCK_BN).
12330 * Change the ENGINE framework to automatically load engines
12331 dynamically from specific directories unless they could be
12332 found to already be built in or loaded. Move all the
12333 current engines except for the cryptodev one to a new
12334 directory engines/.
12335 The engines in engines/ are built as shared libraries if
12336 the "shared" options was given to ./Configure or ./config.
12337 Otherwise, they are inserted in libcrypto.a.
12338 /usr/local/ssl/engines is the default directory for dynamic
12339 engines, but that can be overridden at configure time through
12340 the usual use of --prefix and/or --openssldir, and at run
12341 time with the environment variable OPENSSL_ENGINES.
12343 *Geoff Thorpe and Richard Levitte*
12345 * Add Makefile.shared, a helper makefile to build shared
12346 libraries. Adapt Makefile.org.
12350 * Add version info to Win32 DLLs.
12352 *Peter 'Luna' Runestig" <peter@runestig.com>*
12354 * Add new 'medium level' PKCS#12 API. Certificates and keys
12355 can be added using this API to created arbitrary PKCS#12
12356 files while avoiding the low-level API.
12358 New options to PKCS12_create(), key or cert can be NULL and
12359 will then be omitted from the output file. The encryption
12360 algorithm NIDs can be set to -1 for no encryption, the mac
12361 iteration count can be set to 0 to omit the mac.
12363 Enhance pkcs12 utility by making the -nokeys and -nocerts
12364 options work when creating a PKCS#12 file. New option -nomac
12365 to omit the mac, NONE can be set for an encryption algorithm.
12366 New code is modified to use the enhanced PKCS12_create()
12367 instead of the low-level API.
12371 * Extend ASN1 encoder to support indefinite length constructed
12372 encoding. This can output sequences tags and octet strings in
12373 this form. Modify pk7_asn1.c to support indefinite length
12374 encoding. This is experimental and needs additional code to
12375 be useful, such as an ASN1 bio and some enhanced streaming
12378 Extend template encode functionality so that tagging is passed
12379 down to the template encoder.
12383 * Let 'openssl req' fail if an argument to '-newkey' is not
12384 recognized instead of using RSA as a default.
12388 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12389 As these are not official, they are not included in "ALL";
12390 the "ECCdraft" ciphersuite group alias can be used to select them.
12392 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12394 * Add ECDH engine support.
12396 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12398 * Add ECDH in new directory crypto/ecdh/.
12400 *Douglas Stebila (Sun Microsystems Laboratories)*
12402 * Let BN_rand_range() abort with an error after 100 iterations
12403 without success (which indicates a broken PRNG).
12407 * Change BN_mod_sqrt() so that it verifies that the input value
12408 is really the square of the return value. (Previously,
12409 BN_mod_sqrt would show GIGO behaviour.)
12413 * Add named elliptic curves over binary fields from X9.62, SECG,
12414 and WAP/WTLS; add OIDs that were still missing.
12416 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12418 * Extend the EC library for elliptic curves over binary fields
12419 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12422 EC_GF2m_simple_method
12426 EC_GROUP_new_curve_GF2m
12427 EC_GROUP_set_curve_GF2m
12428 EC_GROUP_get_curve_GF2m
12429 EC_POINT_set_affine_coordinates_GF2m
12430 EC_POINT_get_affine_coordinates_GF2m
12431 EC_POINT_set_compressed_coordinates_GF2m
12433 Point compression for binary fields is disabled by default for
12434 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12437 As binary polynomials are represented as BIGNUMs, various members
12438 of the EC_GROUP and EC_POINT data structures can be shared
12439 between the implementations for prime fields and binary fields;
12440 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12441 are essentially identical to their `..._GFp` counterparts.
12442 (For simplicity, the `..._GFp` prefix has been dropped from
12443 various internal method names.)
12445 An internal 'field_div' method (similar to 'field_mul' and
12446 'field_sqr') has been added; this is used only for binary fields.
12448 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12450 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12451 through methods ('mul', 'precompute_mult').
12453 The generic implementations (now internally called 'ec_wNAF_mul'
12454 and 'ec_wNAF_precomputed_mult') remain the default if these
12455 methods are undefined.
12457 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12459 * New function EC_GROUP_get_degree, which is defined through
12460 EC_METHOD. For curves over prime fields, this returns the bit
12461 length of the modulus.
12463 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12465 * New functions EC_GROUP_dup, EC_POINT_dup.
12466 (These simply call ..._new and ..._copy).
12468 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12470 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12471 Polynomials are represented as BIGNUMs (where the sign bit is not
12472 used) in the following functions [macros]:
12475 BN_GF2m_sub [= BN_GF2m_add]
12476 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12477 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12478 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12480 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12481 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12482 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12483 BN_GF2m_cmp [= BN_ucmp]
12485 (Note that only the 'mod' functions are actually for fields GF(2^m).
12486 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12488 For some functions, an the irreducible polynomial defining a
12489 field can be given as an 'unsigned int[]' with strictly
12490 decreasing elements giving the indices of those bits that are set;
12491 i.e., p[] represents the polynomial
12492 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12494 p[0] > p[1] > ... > p[k] = 0.
12495 This applies to the following functions:
12498 BN_GF2m_mod_mul_arr
12499 BN_GF2m_mod_sqr_arr
12500 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12501 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12502 BN_GF2m_mod_exp_arr
12503 BN_GF2m_mod_sqrt_arr
12504 BN_GF2m_mod_solve_quad_arr
12508 Conversion can be performed by the following functions:
12513 bntest.c has additional tests for binary polynomial arithmetic.
12515 Two implementations for BN_GF2m_mod_div() are available.
12516 The default algorithm simply uses BN_GF2m_mod_inv() and
12517 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12518 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12519 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12521 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12523 * Add new error code 'ERR_R_DISABLED' that can be used when some
12524 functionality is disabled at compile-time.
12526 *Douglas Stebila <douglas.stebila@sun.com>*
12528 * Change default behaviour of 'openssl asn1parse' so that more
12529 information is visible when viewing, e.g., a certificate:
12531 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12532 mode the content of non-printable OCTET STRINGs is output in a
12533 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12534 avoid the appearance of a printable string.
12536 *Nils Larsch <nla@trustcenter.de>*
12538 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12540 EC_GROUP_set_asn1_flag()
12541 EC_GROUP_get_asn1_flag()
12542 EC_GROUP_set_point_conversion_form()
12543 EC_GROUP_get_point_conversion_form()
12544 These control ASN1 encoding details:
12545 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12546 has been set to OPENSSL_EC_NAMED_CURVE.
12547 - Points are encoded in uncompressed form by default; options for
12548 asn1_for are as for point2oct, namely
12549 POINT_CONVERSION_COMPRESSED
12550 POINT_CONVERSION_UNCOMPRESSED
12551 POINT_CONVERSION_HYBRID
12553 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12555 EC_GROUP_set_seed()
12556 EC_GROUP_get0_seed()
12557 EC_GROUP_get_seed_len()
12558 This is used only for ASN1 purposes (so far).
12560 *Nils Larsch <nla@trustcenter.de>*
12562 * Add 'field_type' member to EC_METHOD, which holds the NID
12563 of the appropriate field type OID. The new function
12564 EC_METHOD_get_field_type() returns this value.
12566 *Nils Larsch <nla@trustcenter.de>*
12569 EC_POINT_point2bn()
12570 EC_POINT_bn2point()
12571 EC_POINT_point2hex()
12572 EC_POINT_hex2point()
12573 providing useful interfaces to EC_POINT_point2oct() and
12574 EC_POINT_oct2point().
12576 *Nils Larsch <nla@trustcenter.de>*
12578 * Change internals of the EC library so that the functions
12579 EC_GROUP_set_generator()
12580 EC_GROUP_get_generator()
12581 EC_GROUP_get_order()
12582 EC_GROUP_get_cofactor()
12583 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12584 to methods, which would lead to unnecessary code duplication when
12585 adding different types of curves.
12587 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12589 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12590 arithmetic, and such that modified wNAFs are generated
12591 (which avoid length expansion in many cases).
12595 * Add a function EC_GROUP_check_discriminant() (defined via
12596 EC_METHOD) that verifies that the curve discriminant is non-zero.
12598 Add a function EC_GROUP_check() that makes some sanity tests
12599 on a EC_GROUP, its generator and order. This includes
12600 EC_GROUP_check_discriminant().
12602 *Nils Larsch <nla@trustcenter.de>*
12604 * Add ECDSA in new directory crypto/ecdsa/.
12606 Add applications 'openssl ecparam' and 'openssl ecdsa'
12607 (these are based on 'openssl dsaparam' and 'openssl dsa').
12609 ECDSA support is also included in various other files across the
12610 library. Most notably,
12611 - 'openssl req' now has a '-newkey ecdsa:file' option;
12612 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12613 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12614 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12615 them suitable for ECDSA where domain parameters must be
12616 extracted before the specific public key;
12617 - ECDSA engine support has been added.
12619 *Nils Larsch <nla@trustcenter.de>*
12621 * Include some named elliptic curves, and add OIDs from X9.62,
12622 SECG, and WAP/WTLS. Each curve can be obtained from the new
12624 EC_GROUP_new_by_curve_name(),
12625 and the list of available named curves can be obtained with
12626 EC_get_builtin_curves().
12627 Also add a 'curve_name' member to EC_GROUP objects, which can be
12629 EC_GROUP_set_curve_name()
12630 EC_GROUP_get_curve_name()
12632 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12634 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12635 was actually never needed) and in BN_mul(). The removal in BN_mul()
12636 required a small change in bn_mul_part_recursive() and the addition
12637 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12638 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12639 bn_sub_words() and bn_add_words() except they take arrays with
12644 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12646 * Cleanse PEM buffers before freeing them since they may contain
12649 *Benjamin Bennett <ben@psc.edu>*
12651 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12652 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12653 authentication-only ciphersuites.
12657 * Since AES128 and AES256 share a single mask bit in the logic of
12658 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12659 kludge to work properly if AES128 is available and AES256 isn't.
12663 * Expand security boundary to match 1.1.1 module.
12667 * Remove redundant features: hash file source, editing of test vectors
12668 modify fipsld to use external fips_premain.c signature.
12672 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12673 run algorithm test programs.
12677 * Make algorithm test programs more tolerant of whitespace.
12681 * Have SSL/TLS server implementation tolerate "mismatched" record
12682 protocol version while receiving ClientHello even if the
12683 ClientHello is fragmented. (The server can't insist on the
12684 particular protocol version it has chosen before the ServerHello
12685 message has informed the client about his choice.)
12689 * Load error codes if they are not already present instead of using a
12690 static variable. This allows them to be cleanly unloaded and reloaded.
12694 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12696 * Introduce limits to prevent malicious keys being able to
12697 cause a denial of service. ([CVE-2006-2940])
12699 *Steve Henson, Bodo Moeller*
12701 * Fix ASN.1 parsing of certain invalid structures that can result
12702 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12704 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12705 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12707 * Fix SSL client code which could crash if connecting to a
12708 malicious SSLv2 server. ([CVE-2006-4343])
12710 *Tavis Ormandy and Will Drewry, Google Security Team*
12712 * Change ciphersuite string processing so that an explicit
12713 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12714 will no longer include "AES128-SHA"), and any other similar
12715 ciphersuite (same bitmap) from *other* protocol versions (so that
12716 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12717 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12718 changes from 0.9.8b and 0.9.8d.
12722 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12724 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12725 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12727 * Change the Unix randomness entropy gathering to use poll() when
12728 possible instead of select(), since the latter has some
12729 undesirable limitations.
12731 *Darryl Miles via Richard Levitte and Bodo Moeller*
12733 * Disable rogue ciphersuites:
12735 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12736 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12737 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12739 The latter two were purportedly from
12740 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12743 Also deactivate the remaining ciphersuites from
12744 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12745 unofficial, and the ID has long expired.
12749 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12750 dual-core machines) and other potential thread-safety issues.
12754 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12756 * Adapt fipsld and the build system to link against the validated FIPS
12757 module in FIPS mode.
12761 * Fixes for VC++ 2005 build under Windows.
12765 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12766 from a Windows bash shell such as MSYS. It is autodetected from the
12767 "config" script when run from a VC++ environment. Modify standard VC++
12768 build to use fipscanister.o from the GNU make build.
12772 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12774 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12775 The value now differs depending on if you build for FIPS or not.
12776 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12777 safely run with a non-FIPSed libcrypto, as it may crash because of
12778 the difference induced by this change.
12782 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12784 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12785 (part of SSL_OP_ALL). This option used to disable the
12786 countermeasure against man-in-the-middle protocol-version
12787 rollback in the SSL 2.0 server implementation, which is a bad
12788 idea. ([CVE-2005-2969])
12790 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12791 for Information Security, National Institute of Advanced Industrial
12792 Science and Technology [AIST, Japan)]*
12794 * Minimal support for X9.31 signatures and PSS padding modes. This is
12795 mainly for FIPS compliance and not fully integrated at this stage.
12799 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12800 the exponentiation using a fixed-length exponent. (Otherwise,
12801 the information leaked through timing could expose the secret key
12802 after many signatures; cf. Bleichenbacher's attack on DSA with
12807 * Make a new fixed-window mod_exp implementation the default for
12808 RSA, DSA, and DH private-key operations so that the sequence of
12809 squares and multiplies and the memory access pattern are
12810 independent of the particular secret key. This will mitigate
12811 cache-timing and potential related attacks.
12813 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12814 and this is automatically used by BN_mod_exp_mont() if the new flag
12815 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12816 will use this BN flag for private exponents unless the flag
12817 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12818 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12820 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12822 * Change the client implementation for SSLv23_method() and
12823 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12824 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12825 (Previously, the SSL 2.0 backwards compatible Client Hello
12826 message format would be used even with SSL_OP_NO_SSLv2.)
12830 * Add support for smime-type MIME parameter in S/MIME messages which some
12835 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12836 a threadsafe manner. Modify rsa code to use new function and add calls
12837 to dsa and dh code (which had race conditions before).
12841 * Include the fixed error library code in the C error file definitions
12842 instead of fixing them up at runtime. This keeps the error code
12843 structures constant.
12847 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12849 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12852 * Fixes for newer kerberos headers. NB: the casts are needed because
12853 the 'length' field is signed on one version and unsigned on another
12854 with no (?) obvious way to tell the difference, without these VC++
12855 complains. Also the "definition" of FAR (blank) is no longer included
12856 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12857 some needed definitions.
12861 * Undo Cygwin change.
12865 * Added support for proxy certificates according to RFC 3820.
12866 Because they may be a security thread to unaware applications,
12867 they must be explicitly allowed in run-time. See
12868 docs/HOWTO/proxy_certificates.txt for further information.
12872 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12874 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12875 server and client random values. Previously
12876 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12877 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12879 This change has negligible security impact because:
12881 1. Server and client random values still have 24 bytes of pseudo random
12884 2. Server and client random values are sent in the clear in the initial
12887 3. The master secret is derived using the premaster secret (48 bytes in
12888 size for static RSA ciphersuites) as well as client server and random
12891 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12894 *Stephen Henson, reported by UK NISCC*
12896 * Use Windows randomness collection on Cygwin.
12900 * Fix hang in EGD/PRNGD query when communication socket is closed
12901 prematurely by EGD/PRNGD.
12903 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12905 * Prompt for pass phrases when appropriate for PKCS12 input format.
12909 * Back-port of selected performance improvements from development
12910 branch, as well as improved support for PowerPC platforms.
12914 * Add lots of checks for memory allocation failure, error codes to indicate
12915 failure and freeing up memory if a failure occurs.
12917 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12919 * Add new -passin argument to dgst.
12923 * Perform some character comparisons of different types in X509_NAME_cmp:
12924 this is needed for some certificates that re-encode DNs into UTF8Strings
12925 (in violation of RFC3280) and can't or won't issue name rollover
12930 * Make an explicit check during certificate validation to see that
12931 the CA setting in each certificate on the chain is correct. As a
12932 side effect always do the following basic checks on extensions,
12933 not just when there's an associated purpose to the check:
12935 - if there is an unhandled critical extension (unless the user
12936 has chosen to ignore this fault)
12937 - if the path length has been exceeded (if one is set at all)
12938 - that certain extensions fit the associated purpose (if one has
12943 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12945 * Avoid a race condition when CRLs are checked in a multi threaded
12946 environment. This would happen due to the reordering of the revoked
12947 entries during signature checking and serial number lookup. Now the
12948 encoding is cached and the serial number sort performed under a lock.
12949 Add new STACK function sk_is_sorted().
12953 * Add Delta CRL to the extension code.
12957 * Various fixes to s3_pkt.c so alerts are sent properly.
12959 *David Holmes <d.holmes@f5.com>*
12961 * Reduce the chances of duplicate issuer name and serial numbers (in
12962 violation of RFC3280) using the OpenSSL certificate creation utilities.
12963 This is done by creating a random 64 bit value for the initial serial
12964 number when a serial number file is created or when a self signed
12965 certificate is created using 'openssl req -x509'. The initial serial
12966 number file is created using 'openssl x509 -next_serial' in CA.pl
12967 rather than being initialized to 1.
12971 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12973 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12974 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12976 *Joe Orton, Steve Henson*
12978 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12981 *Joe Orton, Steve Henson*
12983 * Make it possible to have multiple active certificates with the same
12984 subject in the CA index file. This is done only if the keyword
12985 'unique_subject' is set to 'no' in the main CA section (default
12986 if 'CA_default') of the configuration file. The value is saved
12987 with the database itself in a separate index attribute file,
12988 named like the index file with '.attr' appended to the name.
12992 * X509 verify fixes. Disable broken certificate workarounds when
12993 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12994 keyUsage extension present. Don't accept CRLs with unhandled critical
12995 extensions: since verify currently doesn't process CRL extensions this
12996 rejects a CRL with *any* critical extensions. Add new verify error codes
13001 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
13002 A clarification of RFC2560 will require the use of OCTET STRINGs and
13003 some implementations cannot handle the current raw format. Since OpenSSL
13004 copies and compares OCSP nonces as opaque blobs without any attempt at
13005 parsing them this should not create any compatibility issues.
13009 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
13010 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
13011 this HMAC (and other) operations are several times slower than OpenSSL
13016 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
13018 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13020 * Use the correct content when signing type "other".
13024 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
13026 * Fix various bugs revealed by running the NISCC test suite:
13028 Stop out of bounds reads in the ASN1 code when presented with
13029 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13031 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
13033 If verify callback ignores invalid public key errors don't try to check
13034 certificate signature with the NULL public key.
13038 * New -ignore_err option in ocsp application to stop the server
13039 exiting on the first error in a request.
13043 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13044 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13049 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13050 extra data after the compression methods not only for TLS 1.0
13051 but also for SSL 3.0 (as required by the specification).
13053 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13055 * Change X509_certificate_type() to mark the key as exported/exportable
13056 when it's 512 *bits* long, not 512 bytes.
13060 * Change AES_cbc_encrypt() so it outputs exact multiple of
13061 blocks during encryption.
13065 * Various fixes to base64 BIO and non blocking I/O. On write
13066 flushes were not handled properly if the BIO retried. On read
13067 data was not being buffered properly and had various logic bugs.
13068 This also affects blocking I/O when the data being decoded is a
13073 * Various S/MIME bugfixes and compatibility changes:
13074 output correct application/pkcs7 MIME type if
13075 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
13076 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
13077 of files as .eml work). Correctly handle very long lines in MIME
13082 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
13084 * Countermeasure against the Klima-Pokorny-Rosa extension of
13085 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13086 a protocol version number mismatch like a decryption error
13087 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13091 * Turn on RSA blinding by default in the default implementation
13092 to avoid a timing attack. Applications that don't want it can call
13093 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13094 They would be ill-advised to do so in most cases.
13096 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13098 * Change RSA blinding code so that it works when the PRNG is not
13099 seeded (in this case, the secret RSA exponent is abused as
13100 an unpredictable seed -- if it is not unpredictable, there
13101 is no point in blinding anyway). Make RSA blinding thread-safe
13102 by remembering the creator's thread ID in rsa->blinding and
13103 having all other threads use local one-time blinding factors
13104 (this requires more computation than sharing rsa->blinding, but
13105 avoids excessive locking; and if an RSA object is not shared
13106 between threads, blinding will still be very fast).
13110 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13111 ENGINE as defaults for all supported algorithms irrespective of
13112 the 'flags' parameter. 'flags' is now honoured, so applications
13113 should make sure they are passing it correctly.
13117 * Target "mingw" now allows native Windows code to be generated in
13118 the Cygwin environment as well as with the MinGW compiler.
13122 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13124 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13125 via timing by performing a MAC computation even if incorrect
13126 block cipher padding has been found. This is a countermeasure
13127 against active attacks where the attacker has to distinguish
13128 between bad padding and a MAC verification error. ([CVE-2003-0078])
13130 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13131 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13132 Martin Vuagnoux (EPFL, Ilion)*
13134 * Make the no-err option work as intended. The intention with no-err
13135 is not to have the whole error stack handling routines removed from
13136 libcrypto, it's only intended to remove all the function name and
13137 reason texts, thereby removing some of the footprint that may not
13138 be interesting if those errors aren't displayed anyway.
13140 NOTE: it's still possible for any application or module to have its
13141 own set of error texts inserted. The routines are there, just not
13142 used by default when no-err is given.
13146 * Add support for FreeBSD on IA64.
13148 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13150 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13151 Kerberos function mit_des_cbc_cksum(). Before this change,
13152 the value returned by DES_cbc_cksum() was like the one from
13153 mit_des_cbc_cksum(), except the bytes were swapped.
13155 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13157 * Allow an application to disable the automatic SSL chain building.
13158 Before this a rather primitive chain build was always performed in
13159 ssl3_output_cert_chain(): an application had no way to send the
13160 correct chain if the automatic operation produced an incorrect result.
13162 Now the chain builder is disabled if either:
13164 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13166 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13168 The reasoning behind this is that an application would not want the
13169 auto chain building to take place if extra chain certificates are
13170 present and it might also want a means of sending no additional
13171 certificates (for example the chain has two certificates and the
13176 * Add the possibility to build without the ENGINE framework.
13178 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13180 * Under Win32 gmtime() can return NULL: check return value in
13181 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13185 * DSA routines: under certain error conditions uninitialized BN objects
13186 could be freed. Solution: make sure initialization is performed early
13187 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13188 Nils Larsch <nla@trustcenter.de> via PR#459)
13192 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13193 checked on reconnect on the client side, therefore session resumption
13194 could still fail with a "ssl session id is different" error. This
13195 behaviour is masked when SSL_OP_ALL is used due to
13196 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13197 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13198 followup to PR #377.
13202 * IA-32 assembler support enhancements: unified ELF targets, support
13203 for SCO/Caldera platforms, fix for Cygwin shared build.
13207 * Add support for FreeBSD on sparc64. As a consequence, support for
13208 FreeBSD on non-x86 processors is separate from x86 processors on
13209 the config script, much like the NetBSD support.
13211 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13213 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13215 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13218 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13219 code (06) was taken as the first octet of the session ID and the last
13220 octet was ignored consequently. As a result SSLv2 client side session
13221 caching could not have worked due to the session ID mismatch between
13223 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13228 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13229 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13234 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13235 seems that in spite of existing for more than a year, many application
13236 author have done nothing to provide the necessary callbacks, which
13237 means that this particular engine will not work properly anywhere.
13238 This is a very unfortunate situation which forces us, in the name
13239 of usability, to give the hw_ncipher.c a static lock, which is part
13241 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13242 appear in 0.9.8 or later. We EXPECT application authors to have
13243 dealt properly with this when 0.9.8 is released (unless we actually
13244 make such changes in the libcrypto locking code that changes will
13245 have to be made anyway).
13249 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13250 octets have been read, EOF or an error occurs. Without this change
13251 some truncated ASN1 structures will not produce an error.
13255 * Disable Heimdal support, since it hasn't been fully implemented.
13256 Still give the possibility to force the use of Heimdal, but with
13257 warnings and a request that patches get sent to openssl-dev.
13261 * Add the VC-CE target, introduce the WINCE sysname, and add
13262 INSTALL.WCE and appropriate conditionals to make it build.
13264 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13266 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13267 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13268 edit numbers of the version.
13270 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13272 * Introduce safe string copy and catenation functions
13273 (BUF_strlcpy() and BUF_strlcat()).
13275 *Ben Laurie (CHATS) and Richard Levitte*
13277 * Avoid using fixed-size buffers for one-line DNs.
13279 *Ben Laurie (CHATS)*
13281 * Add BUF_MEM_grow_clean() to avoid information leakage when
13282 resizing buffers containing secrets, and use where appropriate.
13284 *Ben Laurie (CHATS)*
13286 * Avoid using fixed size buffers for configuration file location.
13288 *Ben Laurie (CHATS)*
13290 * Avoid filename truncation for various CA files.
13292 *Ben Laurie (CHATS)*
13294 * Use sizeof in preference to magic numbers.
13296 *Ben Laurie (CHATS)*
13298 * Avoid filename truncation in cert requests.
13300 *Ben Laurie (CHATS)*
13302 * Add assertions to check for (supposedly impossible) buffer
13305 *Ben Laurie (CHATS)*
13307 * Don't cache truncated DNS entries in the local cache (this could
13308 potentially lead to a spoofing attack).
13310 *Ben Laurie (CHATS)*
13312 * Fix various buffers to be large enough for hex/decimal
13313 representations in a platform independent manner.
13315 *Ben Laurie (CHATS)*
13317 * Add CRYPTO_realloc_clean() to avoid information leakage when
13318 resizing buffers containing secrets, and use where appropriate.
13320 *Ben Laurie (CHATS)*
13322 * Add BIO_indent() to avoid much slightly worrying code to do
13325 *Ben Laurie (CHATS)*
13327 * Convert sprintf()/BIO_puts() to BIO_printf().
13329 *Ben Laurie (CHATS)*
13331 * buffer_gets() could terminate with the buffer only half
13334 *Ben Laurie (CHATS)*
13336 * Add assertions to prevent user-supplied crypto functions from
13337 overflowing internal buffers by having large block sizes, etc.
13339 *Ben Laurie (CHATS)*
13341 * New OPENSSL_assert() macro (similar to assert(), but enabled
13344 *Ben Laurie (CHATS)*
13346 * Eliminate unused copy of key in RC4.
13348 *Ben Laurie (CHATS)*
13350 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13352 *Ben Laurie (CHATS)*
13354 * Fix off-by-one error in EGD path.
13356 *Ben Laurie (CHATS)*
13358 * If RANDFILE path is too long, ignore instead of truncating.
13360 *Ben Laurie (CHATS)*
13362 * Eliminate unused and incorrectly sized X.509 structure
13365 *Ben Laurie (CHATS)*
13367 * Eliminate unused and dangerous function knumber().
13369 *Ben Laurie (CHATS)*
13371 * Eliminate unused and dangerous structure, KSSL_ERR.
13373 *Ben Laurie (CHATS)*
13375 * Protect against overlong session ID context length in an encoded
13376 session object. Since these are local, this does not appear to be
13379 *Ben Laurie (CHATS)*
13381 * Change from security patch (see 0.9.6e below) that did not affect
13382 the 0.9.6 release series:
13384 Remote buffer overflow in SSL3 protocol - an attacker could
13385 supply an oversized master key in Kerberos-enabled versions.
13388 *Ben Laurie (CHATS)*
13390 * Change the SSL kerb5 codes to match RFC 2712.
13394 * Make -nameopt work fully for req and add -reqopt switch.
13396 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13398 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13400 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13402 * Make sure tests can be performed even if the corresponding algorithms
13403 have been removed entirely. This was also the last step to make
13404 OpenSSL compilable with DJGPP under all reasonable conditions.
13406 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13408 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13409 to allow version independent disabling of normally unselected ciphers,
13410 which may be activated as a side-effect of selecting a single cipher.
13412 (E.g., cipher list string "RSA" enables ciphersuites that are left
13413 out of "ALL" because they do not provide symmetric encryption.
13414 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13416 *Lutz Jaenicke, Bodo Moeller*
13418 * Add appropriate support for separate platform-dependent build
13419 directories. The recommended way to make a platform-dependent
13420 build directory is the following (tested on Linux), maybe with
13423 # Place yourself outside of the OpenSSL source tree. In
13424 # this example, the environment variable OPENSSL_SOURCE
13425 # is assumed to contain the absolute OpenSSL source directory.
13426 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13427 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13428 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13429 mkdir -p `dirname $F`
13430 ln -s $OPENSSL_SOURCE/$F $F
13433 To be absolutely sure not to disturb the source tree, a "make clean"
13434 is a good thing. If it isn't successful, don't worry about it,
13435 it probably means the source directory is very clean.
13439 * Make sure any ENGINE control commands make local copies of string
13440 pointers passed to them whenever necessary. Otherwise it is possible
13441 the caller may have overwritten (or deallocated) the original string
13442 data when a later ENGINE operation tries to use the stored values.
13444 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13446 * Improve diagnostics in file reading and command-line digests.
13448 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13450 * Add AES modes CFB and OFB to the object database. Correct an
13451 error in AES-CFB decryption.
13455 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13456 allows existing EVP_CIPHER_CTX structures to be reused after
13457 calling `EVP_*Final()`. This behaviour is used by encryption
13458 BIOs and some applications. This has the side effect that
13459 applications must explicitly clean up cipher contexts with
13460 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13464 * Check the values of dna and dnb in bn_mul_recursive before calling
13465 bn_mul_comba (a non zero value means the a or b arrays do not contain
13466 n2 elements) and fallback to bn_mul_normal if either is not zero.
13470 * Fix escaping of non-ASCII characters when using the -subj option
13471 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13475 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13476 form for "surname", serialNumber has no short form.
13477 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13478 therefore remove "mail" short name for "internet 7".
13479 The OID for unique identifiers in X509 certificates is
13480 x500UniqueIdentifier, not uniqueIdentifier.
13481 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13485 * Add an "init" command to the ENGINE config module and auto initialize
13486 ENGINEs. Without any "init" command the ENGINE will be initialized
13487 after all ctrl commands have been executed on it. If init=1 the
13488 ENGINE is initialized at that point (ctrls before that point are run
13489 on the uninitialized ENGINE and after on the initialized one). If
13490 init=0 then the ENGINE will not be initialized at all.
13494 * Fix the 'app_verify_callback' interface so that the user-defined
13495 argument is actually passed to the callback: In the
13496 SSL_CTX_set_cert_verify_callback() prototype, the callback
13497 declaration has been changed from
13500 int (*cb)(X509_STORE_CTX *,void *);
13501 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13502 i=s->ctx->app_verify_callback(&ctx)
13503 has been changed into
13504 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13506 To update applications using SSL_CTX_set_cert_verify_callback(),
13507 a dummy argument can be added to their callback functions.
13509 *D. K. Smetters <smetters@parc.xerox.com>*
13511 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13513 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13515 * Add and OPENSSL_LOAD_CONF define which will cause
13516 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13517 This allows older applications to transparently support certain
13518 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13519 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13520 load the config file and OPENSSL_add_all_algorithms_conf() which will
13521 always load it have also been added.
13525 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13526 Adjust NIDs and EVP layer.
13528 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13530 * Config modules support in openssl utility.
13532 Most commands now load modules from the config file,
13533 though in a few (such as version) this isn't done
13534 because it couldn't be used for anything.
13536 In the case of ca and req the config file used is
13537 the same as the utility itself: that is the -config
13538 command line option can be used to specify an
13543 * Move default behaviour from OPENSSL_config(). If appname is NULL
13544 use "openssl_conf" if filename is NULL use default openssl config file.
13548 * Add an argument to OPENSSL_config() to allow the use of an alternative
13549 config section name. Add a new flag to tolerate a missing config file
13550 and move code to CONF_modules_load_file().
13554 * Support for crypto accelerator cards from Accelerated Encryption
13555 Processing, www.aep.ie. (Use engine 'aep')
13556 The support was copied from 0.9.6c [engine] and adapted/corrected
13557 to work with the new engine framework.
13559 *AEP Inc. and Richard Levitte*
13561 * Support for SureWare crypto accelerator cards from Baltimore
13562 Technologies. (Use engine 'sureware')
13563 The support was copied from 0.9.6c [engine] and adapted
13564 to work with the new engine framework.
13568 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13569 make the newer ENGINE framework commands for the CHIL engine work.
13571 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13573 * Make it possible to produce shared libraries on ReliantUNIX.
13575 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13577 * Add the configuration target debug-linux-ppro.
13578 Make 'openssl rsa' use the general key loading routines
13579 implemented in `apps.c`, and make those routines able to
13580 handle the key format FORMAT_NETSCAPE and the variant
13583 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13585 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13587 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13589 * Add -keyform to rsautl, and document -engine.
13591 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13593 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13594 BIO_R_NO_SUCH_FILE error code rather than the generic
13595 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13599 * Add new functions
13600 ERR_peek_last_error
13601 ERR_peek_last_error_line
13602 ERR_peek_last_error_line_data.
13603 These are similar to
13605 ERR_peek_error_line
13606 ERR_peek_error_line_data,
13607 but report on the latest error recorded rather than the first one
13608 still in the error queue.
13610 *Ben Laurie, Bodo Moeller*
13612 * default_algorithms option in ENGINE config module. This allows things
13614 default_algorithms = ALL
13615 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13619 * Preliminary ENGINE config module.
13623 * New experimental application configuration code.
13627 * Change the AES code to follow the same name structure as all other
13628 symmetric ciphers, and behave the same way. Move everything to
13629 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13631 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13633 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13635 *Ben Laurie and Theo de Raadt*
13637 * Add option to output public keys in req command.
13639 *Massimiliano Pala madwolf@openca.org*
13641 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13642 (up to about 10% better than before for P-192 and P-224).
13646 * New functions/macros
13648 SSL_CTX_set_msg_callback(ctx, cb)
13649 SSL_CTX_set_msg_callback_arg(ctx, arg)
13650 SSL_set_msg_callback(ssl, cb)
13651 SSL_set_msg_callback_arg(ssl, arg)
13653 to request calling a callback function
13655 void cb(int write_p, int version, int content_type,
13656 const void *buf, size_t len, SSL *ssl, void *arg)
13658 whenever a protocol message has been completely received
13659 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13660 protocol version according to which the SSL library interprets
13661 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13662 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13663 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13664 specification (change_cipher_spec(20), alert(21), handshake(22)).
13665 'buf' and 'len' point to the actual message, 'ssl' to the
13666 SSL object, and 'arg' is the application-defined value set by
13667 SSL[_CTX]_set_msg_callback_arg().
13669 'openssl s_client' and 'openssl s_server' have new '-msg' options
13670 to enable a callback that displays all protocol messages.
13674 * Change the shared library support so shared libraries are built as
13675 soon as the corresponding static library is finished, and thereby get
13676 openssl and the test programs linked against the shared library.
13677 This still only happens when the keyword "shard" has been given to
13678 the configuration scripts.
13680 NOTE: shared library support is still an experimental thing, and
13681 backward binary compatibility is still not guaranteed.
13683 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13685 * Add support for Subject Information Access extension.
13687 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13689 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13690 additional bytes when new memory had to be allocated, not just
13691 when reusing an existing buffer.
13695 * New command line and configuration option 'utf8' for the req command.
13696 This allows field values to be specified as UTF8 strings.
13700 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13701 runs for the former and machine-readable output for the latter.
13705 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13706 of the e-mail address in the DN (i.e., it will go into a certificate
13707 extension only). The new configuration file option 'email_in_dn = no'
13708 has the same effect.
13710 *Massimiliano Pala madwolf@openca.org*
13712 * Change all functions with names starting with `des_` to be starting
13713 with `DES_` instead. Add wrappers that are compatible with libdes,
13714 but are named `_ossl_old_des_*`. Finally, add macros that map the
13715 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13716 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13717 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13720 Since we provide two compatibility mappings, the user needs to
13721 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13722 compatibility is desired. The default (i.e., when that macro
13723 isn't defined) is OpenSSL 0.9.6c compatibility.
13725 There are also macros that enable and disable the support of old
13726 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13727 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13728 are defined, the default will apply: to support the old des routines.
13730 In either case, one must include openssl/des.h to get the correct
13731 definitions. Do not try to just include openssl/des_old.h, that
13734 NOTE: This is a major break of an old API into a new one. Software
13735 authors are encouraged to switch to the `DES_` style functions. Some
13736 time in the future, des_old.h and the libdes compatibility functions
13737 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13738 default), and then completely removed.
13742 * Test for certificates which contain unsupported critical extensions.
13743 If such a certificate is found during a verify operation it is
13744 rejected by default: this behaviour can be overridden by either
13745 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13746 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13747 X509_supported_extension() has also been added which returns 1 if a
13748 particular extension is supported.
13752 * Modify the behaviour of EVP cipher functions in similar way to digests
13753 to retain compatibility with existing code.
13757 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13758 compatibility with existing code. In particular the 'ctx' parameter does
13759 not have to be to be initialized before the call to EVP_DigestInit() and
13760 it is tidied up after a call to EVP_DigestFinal(). New function
13761 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13762 EVP_MD_CTX_copy() changed to not require the destination to be
13763 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13764 requires the destination to be valid.
13766 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13767 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13771 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13772 so that complete 'Handshake' protocol structures are kept in memory
13773 instead of overwriting 'msg_type' and 'length' with 'body' data.
13777 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13779 *Massimo Santin via Richard Levitte*
13781 * Major restructuring to the underlying ENGINE code. This includes
13782 reduction of linker bloat, separation of pure "ENGINE" manipulation
13783 (initialisation, etc) from functionality dealing with implementations
13784 of specific crypto interfaces. This change also introduces integrated
13785 support for symmetric ciphers and digest implementations - so ENGINEs
13786 can now accelerate these by providing EVP_CIPHER and EVP_MD
13787 implementations of their own. This is detailed in
13788 [crypto/engine/README.md](crypto/engine/README.md)
13789 as it couldn't be adequately described here. However, there are a few
13790 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13791 were changed in the original introduction of ENGINE code have now
13792 reverted back - the hooking from this code to ENGINE is now a good
13793 deal more passive and at run-time, operations deal directly with
13794 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13795 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13796 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13797 they were not being used by the framework as there is no concept of a
13798 BIGNUM_METHOD and they could not be generalised to the new
13799 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13800 ENGINE_cpy() has been removed as it cannot be consistently defined in
13805 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13809 * Change mkdef.pl to sort symbols that get the same entry number,
13810 and make sure the automatically generated functions `ERR_load_*`
13811 become part of libeay.num as well.
13815 * New function SSL_renegotiate_pending(). This returns true once
13816 renegotiation has been requested (either SSL_renegotiate() call
13817 or HelloRequest/ClientHello received from the peer) and becomes
13818 false once a handshake has been completed.
13819 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13820 sends a HelloRequest, but does not ensure that a handshake takes
13821 place. SSL_renegotiate_pending() is useful for checking if the
13822 client has followed the request.)
13826 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13827 By default, clients may request session resumption even during
13828 renegotiation (if session ID contexts permit); with this option,
13829 session resumption is possible only in the first handshake.
13831 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13832 more bits available for options that should not be part of
13833 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13837 * Add some demos for certificate and certificate request creation.
13841 * Make maximum certificate chain size accepted from the peer application
13842 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13843 "Douglas E. Engert" <deengert@anl.gov>.
13847 * Add support for shared libraries for Unixware-7
13848 (Boyd Lynn Gerber <gerberb@zenez.com>).
13852 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13853 be done prior to destruction. Use this to unload error strings from
13854 ENGINEs that load their own error strings. NB: This adds two new API
13855 functions to "get" and "set" this destroy handler in an ENGINE.
13859 * Alter all existing ENGINE implementations (except "openssl" and
13860 "openbsd") to dynamically instantiate their own error strings. This
13861 makes them more flexible to be built both as statically-linked ENGINEs
13862 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13863 Also, add stub code to each that makes building them as self-contained
13864 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13868 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13869 implementations into applications that are completely implemented in
13870 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13871 commands that can be used to configure what shared-library to load and
13872 to control aspects of the way it is handled. Also, made an update to
13873 the [README-Engine.md](README-Engine.md) file
13874 that brings its information up-to-date and
13875 provides some information and instructions on the "dynamic" ENGINE
13876 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13880 * Make it possible to unload ranges of ERR strings with a new
13881 "ERR_unload_strings" function.
13885 * Add a copy() function to EVP_MD.
13889 * Make EVP_MD routines take a context pointer instead of just the
13890 md_data void pointer.
13894 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13895 that the digest can only process a single chunk of data
13896 (typically because it is provided by a piece of
13897 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13898 is only going to provide a single chunk of data, and hence the
13899 framework needn't accumulate the data for oneshot drivers.
13903 * As with "ERR", make it possible to replace the underlying "ex_data"
13904 functions. This change also alters the storage and management of global
13905 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13906 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13907 index counters. The API functions that use this state have been changed
13908 to take a "class_index" rather than pointers to the class's local STACK
13909 and counter, and there is now an API function to dynamically create new
13910 classes. This centralisation allows us to (a) plug a lot of the
13911 thread-safety problems that existed, and (b) makes it possible to clean
13912 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13913 such data would previously have always leaked in application code and
13914 workarounds were in place to make the memory debugging turn a blind eye
13915 to it. Application code that doesn't use this new function will still
13916 leak as before, but their memory debugging output will announce it now
13917 rather than letting it slide.
13919 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13920 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13921 has a return value to indicate success or failure.
13925 * Make it possible to replace the underlying "ERR" functions such that the
13926 global state (2 LHASH tables and 2 locks) is only used by the "default"
13927 implementation. This change also adds two functions to "get" and "set"
13928 the implementation prior to it being automatically set the first time
13929 any other ERR function takes place. Ie. an application can call "get",
13930 pass the return value to a module it has just loaded, and that module
13931 can call its own "set" function using that value. This means the
13932 module's "ERR" operations will use (and modify) the error state in the
13933 application and not in its own statically linked copy of OpenSSL code.
13937 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13938 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13939 the operation, and provides a more encapsulated way for external code
13940 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13941 to use these functions rather than manually incrementing the counts.
13943 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13947 * Add EVP test program.
13951 * Add symmetric cipher support to ENGINE. Expect the API to change!
13955 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13956 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13957 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13958 These allow a CRL to be built without having to access X509_CRL fields
13959 directly. Modify 'ca' application to use new functions.
13963 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13964 bug workarounds. Rollback attack detection is a security feature.
13965 The problem will only arise on OpenSSL servers when TLSv1 is not
13966 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13967 Software authors not wanting to support TLSv1 will have special reasons
13968 for their choice and can explicitly enable this option.
13970 *Bodo Moeller, Lutz Jaenicke*
13972 * Rationalise EVP so it can be extended: don't include a union of
13973 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13974 (similar to those existing for EVP_CIPHER_CTX).
13979 EVP_MD_CTX_init(&md); /* new function call */
13980 EVP_DigestInit(&md, EVP_sha1());
13981 EVP_DigestUpdate(&md, in, len);
13982 EVP_DigestFinal(&md, out, NULL);
13983 EVP_MD_CTX_cleanup(&md); /* new function call */
13987 * Make DES key schedule conform to the usual scheme, as well as
13988 correcting its structure. This means that calls to DES functions
13989 now have to pass a pointer to a des_key_schedule instead of a
13990 plain des_key_schedule (which was actually always a pointer
13993 des_key_schedule ks;
13995 des_set_key_checked(..., &ks);
13996 des_ncbc_encrypt(..., &ks, ...);
13998 (Note that a later change renames 'des_...' into 'DES_...'.)
14002 * Initial reduction of linker bloat: the use of some functions, such as
14003 PEM causes large amounts of unused functions to be linked in due to
14004 poor organisation. For example pem_all.c contains every PEM function
14005 which has a knock on effect of linking in large amounts of (unused)
14006 ASN1 code. Grouping together similar functions and splitting unrelated
14007 functions prevents this.
14011 * Cleanup of EVP macros.
14015 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
14016 correct `_ecb suffix`.
14020 * Add initial OCSP responder support to ocsp application. The
14021 revocation information is handled using the text based index
14022 use by the ca application. The responder can either handle
14023 requests generated internally, supplied in files (for example
14024 via a CGI script) or using an internal minimal server.
14028 * Add configuration choices to get zlib compression for TLS.
14032 * Changes to Kerberos SSL for RFC 2712 compliance:
14033 1. Implemented real KerberosWrapper, instead of just using
14034 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
14035 2. Implemented optional authenticator field of KerberosWrapper.
14037 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
14038 and authenticator structs; see crypto/krb5/.
14040 Generalized Kerberos calls to support multiple Kerberos libraries.
14041 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
14042 via Richard Levitte*
14044 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
14045 already does with RSA. testdsa.h now has 'priv_key/pub_key'
14046 values for each of the key sizes rather than having just
14047 parameters (and 'speed' generating keys each time).
14051 * Speed up EVP routines.
14054 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
14055 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
14056 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
14057 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
14059 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
14060 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
14061 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
14064 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
14066 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
14070 * Added the OS2-EMX target.
14072 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
14074 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
14075 New functions to support `NCONF` routines in extension code.
14076 New function `CONF_set_nconf()`
14077 to allow functions which take an `NCONF` to also handle the old `LHASH`
14078 structure: this means that the old `CONF` compatible routines can be
14079 retained (in particular w.rt. extensions) without having to duplicate the
14080 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
14084 * Enhance the general user interface with mechanisms for inner control
14085 and with possibilities to have yes/no kind of prompts.
14089 * Change all calls to low-level digest routines in the library and
14090 applications to use EVP. Add missing calls to HMAC_cleanup() and
14091 don't assume HMAC_CTX can be copied using memcpy().
14093 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14095 * Add the possibility to control engines through control names but with
14096 arbitrary arguments instead of just a string.
14097 Change the key loaders to take a UI_METHOD instead of a callback
14098 function pointer. NOTE: this breaks binary compatibility with earlier
14099 versions of OpenSSL [engine].
14100 Adapt the nCipher code for these new conditions and add a card insertion
14105 * Enhance the general user interface with mechanisms to better support
14106 dialog box interfaces, application-defined prompts, the possibility
14107 to use defaults (for example default passwords from somewhere else)
14108 and interrupts/cancellations.
14112 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14113 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14117 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14118 tidy up some unnecessarily weird code in 'sk_new()').
14120 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14122 * Change the key loading routines for ENGINEs to use the same kind
14123 callback (pem_password_cb) as all other routines that need this
14128 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14129 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14130 than this minimum value is recommended.
14134 * New random seeder for OpenVMS, using the system process statistics
14135 that are easily reachable.
14139 * Windows apparently can't transparently handle global
14140 variables defined in DLLs. Initialisations such as:
14142 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14144 won't compile. This is used by the any applications that need to
14145 declare their own ASN1 modules. This was fixed by adding the option
14146 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14147 needed for static libraries under Win32.
14151 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14152 setting of purpose and trust fields. New X509_STORE trust and
14153 purpose functions and tidy up setting in other SSL functions.
14157 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14158 structure. These are inherited by X509_STORE_CTX when it is
14159 initialised. This allows various defaults to be set in the
14160 X509_STORE structure (such as flags for CRL checking and custom
14161 purpose or trust settings) for functions which only use X509_STORE_CTX
14162 internally such as S/MIME.
14164 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14165 trust settings if they are not set in X509_STORE. This allows X509_STORE
14166 purposes and trust (in S/MIME for example) to override any set by default.
14168 Add command line options for CRL checking to smime, s_client and s_server
14173 * Initial CRL based revocation checking. If the CRL checking flag(s)
14174 are set then the CRL is looked up in the X509_STORE structure and
14175 its validity and signature checked, then if the certificate is found
14176 in the CRL the verify fails with a revoked error.
14178 Various new CRL related callbacks added to X509_STORE_CTX structure.
14180 Command line options added to 'verify' application to support this.
14182 This needs some additional work, such as being able to handle multiple
14183 CRLs with different times, extension based lookup (rather than just
14184 by subject name) and ultimately more complete V2 CRL extension
14189 * Add a general user interface API (crypto/ui/). This is designed
14190 to replace things like des_read_password and friends (backward
14191 compatibility functions using this new API are provided).
14192 The purpose is to remove prompting functions from the DES code
14193 section as well as provide for prompting through dialog boxes in
14194 a window system and the like.
14198 * Add "ex_data" support to ENGINE so implementations can add state at a
14199 per-structure level rather than having to store it globally.
14203 * Make it possible for ENGINE structures to be copied when retrieved by
14204 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14205 This causes the "original" ENGINE structure to act like a template,
14206 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14207 operational state can be localised to each ENGINE structure, despite the
14208 fact they all share the same "methods". New ENGINE structures returned in
14209 this case have no functional references and the return value is the single
14210 structural reference. This matches the single structural reference returned
14211 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14216 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14217 needs to match any other type at all we need to manually clear the
14222 * Changes to the "openssl engine" utility to include;
14223 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14224 about an ENGINE's available control commands.
14225 - executing control commands from command line arguments using the
14226 '-pre' and '-post' switches. '-post' is only used if '-t' is
14227 specified and the ENGINE is successfully initialised. The syntax for
14228 the individual commands are colon-separated, for example;
14229 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14233 * New dynamic control command support for ENGINEs. ENGINEs can now
14234 declare their own commands (numbers), names (strings), descriptions,
14235 and input types for run-time discovery by calling applications. A
14236 subset of these commands are implicitly classed as "executable"
14237 depending on their input type, and only these can be invoked through
14238 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14239 can be based on user input, config files, etc). The distinction is
14240 that "executable" commands cannot return anything other than a boolean
14241 result and can only support numeric or string input, whereas some
14242 discoverable commands may only be for direct use through
14243 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14244 pointers, or other custom uses. The "executable" commands are to
14245 support parameterisations of ENGINE behaviour that can be
14246 unambiguously defined by ENGINEs and used consistently across any
14247 OpenSSL-based application. Commands have been added to all the
14248 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14249 control over shared-library paths without source code alterations.
14253 * Changed all ENGINE implementations to dynamically allocate their
14254 ENGINEs rather than declaring them statically. Apart from this being
14255 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14256 this also allows the implementations to compile without using the
14257 internal engine_int.h header.
14261 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14262 'const' value. Any code that should be able to modify a RAND_METHOD
14263 should already have non-const pointers to it (ie. they should only
14264 modify their own ones).
14268 * Made a variety of little tweaks to the ENGINE code.
14269 - "atalla" and "ubsec" string definitions were moved from header files
14270 to C code. "nuron" string definitions were placed in variables
14271 rather than hard-coded - allowing parameterisation of these values
14272 later on via ctrl() commands.
14273 - Removed unused "#if 0"'d code.
14274 - Fixed engine list iteration code so it uses ENGINE_free() to release
14275 structural references.
14276 - Constified the RAND_METHOD element of ENGINE structures.
14277 - Constified various get/set functions as appropriate and added
14278 missing functions (including a catch-all ENGINE_cpy that duplicates
14279 all ENGINE values onto a new ENGINE except reference counts/state).
14280 - Removed NULL parameter checks in get/set functions. Setting a method
14281 or function to NULL is a way of cancelling out a previously set
14282 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14283 and doesn't justify the extra error symbols and code.
14284 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14285 flags from engine_int.h to engine.h.
14286 - Changed prototypes for ENGINE handler functions (init(), finish(),
14287 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14291 * Implement binary inversion algorithm for BN_mod_inverse in addition
14292 to the algorithm using long division. The binary algorithm can be
14293 used only if the modulus is odd. On 32-bit systems, it is faster
14294 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14295 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14296 up to 450 bits. In 64-bit environments, the binary algorithm
14297 appears to be advantageous for much longer moduli; here we use it
14298 for moduli up to 2048 bits.
14302 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14303 could not support the combine flag in choice fields.
14307 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14308 extensions from a certificate request to the certificate.
14312 * Allow multiple 'certopt' and 'nameopt' options to be separated
14313 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14314 file: this allows the display of the certificate about to be
14315 signed to be customised, to allow certain fields to be included
14316 or excluded and extension details. The old system didn't display
14317 multicharacter strings properly, omitted fields not in the policy
14318 and couldn't display additional details such as extensions.
14322 * Function EC_POINTs_mul for multiple scalar multiplication
14323 of an arbitrary number of elliptic curve points
14324 \sum scalars[i]*points[i],
14325 optionally including the generator defined for the EC_GROUP:
14326 scalar*generator + \sum scalars[i]*points[i].
14328 EC_POINT_mul is a simple wrapper function for the typical case
14329 that the point list has just one item (besides the optional
14334 * First EC_METHODs for curves over GF(p):
14336 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14337 operations and provides various method functions that can also
14338 operate with faster implementations of modular arithmetic.
14340 EC_GFp_mont_method() reuses most functions that are part of
14341 EC_GFp_simple_method, but uses Montgomery arithmetic.
14343 *Bodo Moeller; point addition and point doubling
14344 implementation directly derived from source code provided by
14345 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14347 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14348 crypto/ec/ec_lib.c):
14350 Curves are EC_GROUP objects (with an optional group generator)
14351 based on EC_METHODs that are built into the library.
14353 Points are EC_POINT objects based on EC_GROUP objects.
14355 Most of the framework would be able to handle curves over arbitrary
14356 finite fields, but as there are no obvious types for fields other
14357 than GF(p), some functions are limited to that for now.
14361 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14362 that the file contains a complete HTTP response.
14366 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14367 change the def and num file printf format specifier from "%-40sXXX"
14368 to "%-39s XXX". The latter will always guarantee a space after the
14369 field while the former will cause them to run together if the field
14370 is 40 of more characters long.
14374 * Constify the cipher and digest 'method' functions and structures
14375 and modify related functions to take constant EVP_MD and EVP_CIPHER
14380 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14381 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14385 * Modify `EVP_Digest*()` routines so they now return values. Although the
14386 internal software routines can never fail additional hardware versions
14391 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14393 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14394 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14397 ERR_R_NESTED_ASN1_ERROR
14399 ERR_R_MISSING_ASN1_EOS
14400 were 4 .. 9, conflicting with
14401 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14403 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14404 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14406 Add new error code 'ERR_R_INTERNAL_ERROR'.
14410 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14415 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14416 sets the subject name for a new request or supersedes the
14417 subject name in a given request. Formats that can be parsed are
14418 'CN=Some Name, OU=myOU, C=IT'
14420 'CN=Some Name/OU=myOU/C=IT'.
14422 Add options '-batch' and '-verbose' to 'openssl req'.
14424 *Massimiliano Pala <madwolf@hackmasters.net>*
14426 * Introduce the possibility to access global variables through
14427 functions on platform were that's the best way to handle exporting
14428 global variables in shared libraries. To enable this functionality,
14429 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14430 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14431 is normally done by Configure or something similar).
14433 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14434 in the source file (foo.c) like this:
14436 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14437 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14439 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14440 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14442 OPENSSL_DECLARE_GLOBAL(int,foo);
14443 #define foo OPENSSL_GLOBAL_REF(foo)
14444 OPENSSL_DECLARE_GLOBAL(double,bar);
14445 #define bar OPENSSL_GLOBAL_REF(bar)
14447 The #defines are very important, and therefore so is including the
14448 header file everywhere where the defined globals are used.
14450 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14451 of ASN.1 items, but that structure is a bit different.
14453 The largest change is in util/mkdef.pl which has been enhanced with
14454 better and easier to understand logic to choose which symbols should
14455 go into the Windows .def files as well as a number of fixes and code
14456 cleanup (among others, algorithm keywords are now sorted
14457 lexicographically to avoid constant rewrites).
14461 * In BN_div() keep a copy of the sign of 'num' before writing the
14462 result to 'rm' because if rm==num the value will be overwritten
14463 and produce the wrong result if 'num' is negative: this caused
14464 problems with BN_mod() and BN_nnmod().
14468 * Function OCSP_request_verify(). This checks the signature on an
14469 OCSP request and verifies the signer certificate. The signer
14470 certificate is just checked for a generic purpose and OCSP request
14475 * Add OCSP_check_validity() function to check the validity of OCSP
14476 responses. OCSP responses are prepared in real time and may only
14477 be a few seconds old. Simply checking that the current time lies
14478 between thisUpdate and nextUpdate max reject otherwise valid responses
14479 caused by either OCSP responder or client clock inaccuracy. Instead
14480 we allow thisUpdate and nextUpdate to fall within a certain period of
14481 the current time. The age of the response can also optionally be
14482 checked. Two new options -validity_period and -status_age added to
14487 * If signature or public key algorithm is unrecognized print out its
14488 OID rather that just UNKNOWN.
14492 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14493 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14494 ID to be generated from the issuer certificate alone which can then be
14495 passed to OCSP_id_issuer_cmp().
14499 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14500 ASN1 modules to export functions returning ASN1_ITEM pointers
14501 instead of the ASN1_ITEM structures themselves. This adds several
14502 new macros which allow the underlying ASN1 function/structure to
14503 be accessed transparently. As a result code should not use ASN1_ITEM
14504 references directly (such as &X509_it) but instead use the relevant
14505 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14506 use of the new ASN1 code on platforms where exporting structures
14507 is problematical (for example in shared libraries) but exporting
14508 functions returning pointers to structures is not.
14512 * Add support for overriding the generation of SSL/TLS session IDs.
14513 These callbacks can be registered either in an SSL_CTX or per SSL.
14514 The purpose of this is to allow applications to control, if they wish,
14515 the arbitrary values chosen for use as session IDs, particularly as it
14516 can be useful for session caching in multiple-server environments. A
14517 command-line switch for testing this (and any client code that wishes
14518 to use such a feature) has been added to "s_server".
14520 *Geoff Thorpe, Lutz Jaenicke*
14522 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14523 of the form `#if defined(...) || defined(...) || ...` and
14524 `#if !defined(...) && !defined(...) && ...`. This also avoids
14525 the growing number of special cases it was previously handling.
14529 * Make all configuration macros available for application by making
14530 sure they are available in opensslconf.h, by giving them names starting
14531 with `OPENSSL_` to avoid conflicts with other packages and by making
14532 sure e_os2.h will cover all platform-specific cases together with
14534 Additionally, it is now possible to define configuration/platform-
14535 specific names (called "system identities"). In the C code, these
14536 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14537 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14538 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14543 * New option -set_serial to 'req' and 'x509' this allows the serial
14544 number to use to be specified on the command line. Previously self
14545 signed certificates were hard coded with serial number 0 and the
14546 CA options of 'x509' had to use a serial number in a file which was
14551 * New options to 'ca' utility to support V2 CRL entry extensions.
14552 Currently CRL reason, invalidity date and hold instruction are
14553 supported. Add new CRL extensions to V3 code and some new objects.
14557 * New function EVP_CIPHER_CTX_set_padding() this is used to
14558 disable standard block padding (aka PKCS#5 padding) in the EVP
14559 API, which was previously mandatory. This means that the data is
14560 not padded in any way and so the total length much be a multiple
14561 of the block size, otherwise an error occurs.
14565 * Initial (incomplete) OCSP SSL support.
14569 * New function OCSP_parse_url(). This splits up a URL into its host,
14570 port and path components: primarily to parse OCSP URLs. New -url
14571 option to ocsp utility.
14575 * New nonce behavior. The return value of OCSP_check_nonce() now
14576 reflects the various checks performed. Applications can decide
14577 whether to tolerate certain situations such as an absent nonce
14578 in a response when one was present in a request: the ocsp application
14579 just prints out a warning. New function OCSP_add1_basic_nonce()
14580 this is to allow responders to include a nonce in a response even if
14581 the request is nonce-less.
14585 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14586 skipped when using openssl x509 multiple times on a single input file,
14587 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14591 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14592 set string type: to handle setting ASN1_TIME structures. Fix ca
14593 utility to correctly initialize revocation date of CRLs.
14597 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14598 the clients preferred ciphersuites and rather use its own preferences.
14599 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14600 Internet Explorer by ensuring unchanged hash method during stepup.
14601 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14605 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14606 to aes and add a new 'exist' option to print out symbols that don't
14611 * Additional options to ocsp utility to allow flags to be set and
14612 additional certificates supplied.
14616 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14617 OCSP client a number of certificate to only verify the response
14622 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14623 handle the new API. Currently only ECB, CBC modes supported. Add new
14626 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14627 Encryption Standard (AES) Ciphersuites for Transport Layer
14628 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14629 not enabled by default and were not part of the "ALL" ciphersuite
14630 alias because they were not yet official; they could be
14631 explicitly requested by specifying the "AESdraft" ciphersuite
14632 group alias. In the final release of OpenSSL 0.9.7, the group
14633 alias is called "AES" and is part of "ALL".)
14635 *Ben Laurie, Steve Henson, Bodo Moeller*
14637 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14638 request to response.
14642 * Functions for OCSP responders. OCSP_request_onereq_count(),
14643 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14644 extract information from a certificate request. OCSP_response_create()
14645 creates a response and optionally adds a basic response structure.
14646 OCSP_basic_add1_status() adds a complete single response to a basic
14647 response and returns the OCSP_SINGLERESP structure just added (to allow
14648 extensions to be included for example). OCSP_basic_add1_cert() adds a
14649 certificate to a basic response and OCSP_basic_sign() signs a basic
14650 response with various flags. New helper functions ASN1_TIME_check()
14651 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14652 (converts ASN1_TIME to GeneralizedTime).
14656 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14657 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14658 structure from a certificate. X509_pubkey_digest() digests the public_key
14659 contents: this is used in various key identifiers.
14663 * Make sk_sort() tolerate a NULL argument.
14665 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14667 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14668 passed by the function are trusted implicitly. If any of them signed the
14669 response then it is assumed to be valid and is not verified.
14673 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14674 to data. This was previously part of the PKCS7 ASN1 code. This
14675 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14676 *Steve Henson, reported by Kenneth R. Robinette
14677 <support@securenetterm.com>*
14679 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14680 routines: without these tracing memory leaks is very painful.
14681 Fix leaks in PKCS12 and PKCS7 routines.
14685 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14686 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14687 effectively meant GeneralizedTime would never be used. Now it
14688 is initialised to -1 but X509_time_adj() now has to check the value
14689 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14690 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14691 *Steve Henson, reported by Kenneth R. Robinette
14692 <support@securenetterm.com>*
14694 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14695 result in a zero length in the ASN1_INTEGER structure which was
14696 not consistent with the structure when d2i_ASN1_INTEGER() was used
14697 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14698 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14699 where it did not print out a minus for negative ASN1_INTEGER.
14703 * Add summary printout to ocsp utility. The various functions which
14704 convert status values to strings have been renamed to:
14705 OCSP_response_status_str(), OCSP_cert_status_str() and
14706 OCSP_crl_reason_str() and are no longer static. New options
14707 to verify nonce values and to disable verification. OCSP response
14708 printout format cleaned up.
14712 * Add additional OCSP certificate checks. These are those specified
14713 in RFC2560. This consists of two separate checks: the CA of the
14714 certificate being checked must either be the OCSP signer certificate
14715 or the issuer of the OCSP signer certificate. In the latter case the
14716 OCSP signer certificate must contain the OCSP signing extended key
14717 usage. This check is performed by attempting to match the OCSP
14718 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14719 in the OCSP_CERTID structures of the response.
14723 * Initial OCSP certificate verification added to OCSP_basic_verify()
14724 and related routines. This uses the standard OpenSSL certificate
14725 verify routines to perform initial checks (just CA validity) and
14726 to obtain the certificate chain. Then additional checks will be
14727 performed on the chain. Currently the root CA is checked to see
14728 if it is explicitly trusted for OCSP signing. This is used to set
14729 a root CA as a global signing root: that is any certificate that
14730 chains to that CA is an acceptable OCSP signing certificate.
14734 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14735 extensions from a separate configuration file.
14736 As when reading extensions from the main configuration file,
14737 the '-extensions ...' option may be used for specifying the
14740 *Massimiliano Pala <madwolf@comune.modena.it>*
14742 * New OCSP utility. Allows OCSP requests to be generated or
14743 read. The request can be sent to a responder and the output
14744 parsed, outputted or printed in text form. Not complete yet:
14745 still needs to check the OCSP response validity.
14749 * New subcommands for 'openssl ca':
14750 `openssl ca -status <serial>` prints the status of the cert with
14751 the given serial number (according to the index file).
14752 `openssl ca -updatedb` updates the expiry status of certificates
14755 *Massimiliano Pala <madwolf@comune.modena.it>*
14757 * New '-newreq-nodes' command option to CA.pl. This is like
14758 '-newreq', but calls 'openssl req' with the '-nodes' option
14759 so that the resulting key is not encrypted.
14761 *Damien Miller <djm@mindrot.org>*
14763 * New configuration for the GNU Hurd.
14765 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14767 * Initial code to implement OCSP basic response verify. This
14768 is currently incomplete. Currently just finds the signer's
14769 certificate and verifies the signature on the response.
14773 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14774 value of OPENSSLDIR. This is available via the new '-d' option
14775 to 'openssl version', and is also included in 'openssl version -a'.
14779 * Allowing defining memory allocation callbacks that will be given
14780 file name and line number information in additional arguments
14781 (a `const char*` and an int). The basic functionality remains, as
14782 well as the original possibility to just replace malloc(),
14783 realloc() and free() by functions that do not know about these
14784 additional arguments. To register and find out the current
14785 settings for extended allocation functions, the following
14786 functions are provided:
14788 CRYPTO_set_mem_ex_functions
14789 CRYPTO_set_locked_mem_ex_functions
14790 CRYPTO_get_mem_ex_functions
14791 CRYPTO_get_locked_mem_ex_functions
14793 These work the same way as CRYPTO_set_mem_functions and friends.
14794 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14795 extended allocation function is enabled.
14796 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14797 a conventional allocation function is enabled.
14799 *Richard Levitte, Bodo Moeller*
14801 * Finish off removing the remaining LHASH function pointer casts.
14802 There should no longer be any prototype-casting required when using
14803 the LHASH abstraction, and any casts that remain are "bugs". See
14804 the callback types and macros at the head of lhash.h for details
14805 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14809 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14810 If /dev/[u]random devices are not available or do not return enough
14811 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14813 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14814 /etc/entropy will be queried once each in this sequence, querying stops
14815 when enough entropy was collected without querying more sockets.
14819 * Change the Unix RAND_poll() variant to be able to poll several
14820 random devices, as specified by DEVRANDOM, until a sufficient amount
14821 of data has been collected. We spend at most 10 ms on each file
14822 (select timeout) and read in non-blocking mode. DEVRANDOM now
14823 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14824 (previously it was just the string "/dev/urandom"), so on typical
14825 platforms the 10 ms delay will never occur.
14826 Also separate out the Unix variant to its own file, rand_unix.c.
14827 For VMS, there's a currently-empty rand_vms.c.
14831 * Move OCSP client related routines to ocsp_cl.c. These
14832 provide utility functions which an application needing
14833 to issue a request to an OCSP responder and analyse the
14834 response will typically need: as opposed to those which an
14835 OCSP responder itself would need which will be added later.
14837 OCSP_request_sign() signs an OCSP request with an API similar
14838 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14839 response. OCSP_response_get1_basic() extracts basic response
14840 from response. OCSP_resp_find_status(): finds and extracts status
14841 information from an OCSP_CERTID structure (which will be created
14842 when the request structure is built). These are built from lower
14843 level functions which work on OCSP_SINGLERESP structures but
14844 won't normally be used unless the application wishes to examine
14845 extensions in the OCSP response for example.
14847 Replace nonce routines with a pair of functions.
14848 OCSP_request_add1_nonce() adds a nonce value and optionally
14849 generates a random value. OCSP_check_nonce() checks the
14850 validity of the nonce in an OCSP response.
14854 * Change function OCSP_request_add() to OCSP_request_add0_id().
14855 This doesn't copy the supplied OCSP_CERTID and avoids the
14856 need to free up the newly created id. Change return type
14857 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14858 This can then be used to add extensions to the request.
14859 Deleted OCSP_request_new(), since most of its functionality
14860 is now in OCSP_REQUEST_new() (and the case insensitive name
14861 clash) apart from the ability to set the request name which
14862 will be added elsewhere.
14866 * Update OCSP API. Remove obsolete extensions argument from
14867 various functions. Extensions are now handled using the new
14868 OCSP extension code. New simple OCSP HTTP function which
14869 can be used to send requests and parse the response.
14873 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14874 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14875 uses the special reorder version of SET OF to sort the attributes
14876 and reorder them to match the encoded order. This resolves a long
14877 standing problem: a verify on a PKCS7 structure just after signing
14878 it used to fail because the attribute order did not match the
14879 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14880 it uses the received order. This is necessary to tolerate some broken
14881 software that does not order SET OF. This is handled by encoding
14882 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14883 to produce the required SET OF.
14887 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14888 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14889 files to get correct declarations of the ASN.1 item variables.
14893 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14894 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14895 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14896 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14897 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14898 ASN1_ITEM and no wrapper functions.
14902 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14903 replace the old function pointer based I/O routines. Change most of
14904 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14908 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14909 lines, recognize more "algorithms" that can be deselected, and make
14910 it complain about algorithm deselection that isn't recognised.
14914 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14915 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14916 to use new functions. Add NO_ASN1_OLD which can be set to remove
14917 some old style ASN1 functions: this can be used to determine if old
14918 code will still work when these eventually go away.
14922 * New extension functions for OCSP structures, these follow the
14923 same conventions as certificates and CRLs.
14927 * New function X509V3_add1_i2d(). This automatically encodes and
14928 adds an extension. Its behaviour can be customised with various
14929 flags to append, replace or delete. Various wrappers added for
14930 certificates and CRLs.
14934 * Fix to avoid calling the underlying ASN1 print routine when
14935 an extension cannot be parsed. Correct a typo in the
14936 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14940 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14941 entries for variables.
14945 * Add functionality to `apps/openssl.c` for detecting locking
14946 problems: As the program is single-threaded, all we have
14947 to do is register a locking callback using an array for
14948 storing which locks are currently held by the program.
14952 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14953 SSL_get_ex_data_X509_STORE_idx(), which is used in
14954 ssl_verify_cert_chain() and thus can be called at any time
14955 during TLS/SSL handshakes so that thread-safety is essential.
14956 Unfortunately, the ex_data design is not at all suited
14957 for multi-threaded use, so it probably should be abolished.
14961 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14963 *Broadcom, tweaked and integrated by Geoff Thorpe*
14965 * Move common extension printing code to new function
14966 X509V3_print_extensions(). Reorganise OCSP print routines and
14967 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14971 * New function X509_signature_print() to remove duplication in some
14976 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14977 set (this was treated exactly the same as SET OF previously). This
14978 is used to reorder the STACK representing the structure to match the
14979 encoding. This will be used to get round a problem where a PKCS7
14980 structure which was signed could not be verified because the STACK
14981 order did not reflect the encoded order.
14985 * Reimplement the OCSP ASN1 module using the new code.
14989 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14990 for its ASN1 operations. The old style function pointers still exist
14991 for now but they will eventually go away.
14995 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14996 completely replaces the old ASN1 functionality with a table driven
14997 encoder and decoder which interprets an ASN1_ITEM structure describing
14998 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14999 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
15000 has also been converted to the new form.
15004 * Change BN_mod_exp_recp so that negative moduli are tolerated
15005 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
15006 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
15007 for negative moduli.
15011 * Fix BN_uadd and BN_usub: Always return non-negative results instead
15012 of not touching the result's sign bit.
15016 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
15021 * Changed the LHASH code to use prototypes for callbacks, and created
15022 macros to declare and implement thin (optionally static) functions
15023 that provide type-safety and avoid function pointer casting for the
15024 type-specific callbacks.
15028 * Added Kerberos Cipher Suites to be used with TLS, as written in
15030 *Veers Staats <staatsvr@asc.hpc.mil>,
15031 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
15033 * Reformat the FAQ so the different questions and answers can be divided
15034 in sections depending on the subject.
15038 * Have the zlib compression code load ZLIB.DLL dynamically under
15043 * New function BN_mod_sqrt for computing square roots modulo a prime
15044 (using the probabilistic Tonelli-Shanks algorithm unless
15045 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
15046 be handled deterministically).
15048 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15050 * Make BN_mod_inverse faster by explicitly handling small quotients
15051 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
15052 512 bits], about 30% for larger ones [1024 or 2048 bits].)
15056 * New function BN_kronecker.
15060 * Fix BN_gcd so that it works on negative inputs; the result is
15061 positive unless both parameters are zero.
15062 Previously something reasonably close to an infinite loop was
15063 possible because numbers could be growing instead of shrinking
15064 in the implementation of Euclid's algorithm.
15068 * Fix BN_is_word() and BN_is_one() macros to take into account the
15069 sign of the number in question.
15071 Fix BN_is_word(a,w) to work correctly for w == 0.
15073 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
15074 because its test if the absolute value of 'a' equals 'w'.
15075 Note that BN_abs_is_word does *not* handle w == 0 reliably;
15076 it exists mostly for use in the implementations of BN_is_zero(),
15077 BN_is_one(), and BN_is_word().
15081 * New function BN_swap.
15085 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
15086 the exponentiation functions are more likely to produce reasonable
15087 results on negative inputs.
15091 * Change BN_mod_mul so that the result is always non-negative.
15092 Previously, it could be negative if one of the factors was negative;
15093 I don't think anyone really wanted that behaviour.
15097 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15098 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15099 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15100 and add new functions:
15109 BN_mod_lshift1_quick
15111 BN_mod_lshift_quick
15113 These functions always generate non-negative results.
15115 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15116 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15118 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15119 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15120 be reduced modulo `m`.
15122 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15125 The following entry accidentally appeared in the CHANGES file
15126 distributed with OpenSSL 0.9.7. The modifications described in
15127 it do *not* apply to OpenSSL 0.9.7.
15129 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15130 was actually never needed) and in BN_mul(). The removal in BN_mul()
15131 required a small change in bn_mul_part_recursive() and the addition
15132 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15133 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15134 bn_sub_words() and bn_add_words() except they take arrays with
15140 * In 'openssl passwd', verify passwords read from the terminal
15141 unless the '-salt' option is used (which usually means that
15142 verification would just waste user's time since the resulting
15143 hash is going to be compared with some given password hash)
15144 or the new '-noverify' option is used.
15146 This is an incompatible change, but it does not affect
15147 non-interactive use of 'openssl passwd' (passwords on the command
15148 line, '-stdin' option, '-in ...' option) and thus should not
15149 cause any problems.
15153 * Remove all references to RSAref, since there's no more need for it.
15157 * Make DSO load along a path given through an environment variable
15158 (SHLIB_PATH) with shl_load().
15162 * Constify the ENGINE code as a result of BIGNUM constification.
15163 Also constify the RSA code and most things related to it. In a
15164 few places, most notable in the depth of the ASN.1 code, ugly
15165 casts back to non-const were required (to be solved at a later
15170 * Make it so the openssl application has all engines loaded by default.
15174 * Constify the BIGNUM routines a little more.
15178 * Add the following functions:
15180 ENGINE_load_cswift()
15182 ENGINE_load_atalla()
15183 ENGINE_load_nuron()
15184 ENGINE_load_builtin_engines()
15186 That way, an application can itself choose if external engines that
15187 are built-in in OpenSSL shall ever be used or not. The benefit is
15188 that applications won't have to be linked with libdl or other dso
15189 libraries unless it's really needed.
15191 Changed 'openssl engine' to load all engines on demand.
15192 Changed the engine header files to avoid the duplication of some
15193 declarations (they differed!).
15197 * 'openssl engine' can now list capabilities.
15201 * Better error reporting in 'openssl engine'.
15205 * Never call load_dh_param(NULL) in s_server.
15209 * Add engine application. It can currently list engines by name and
15210 identity, and test if they are actually available.
15214 * Improve RPM specification file by forcing symbolic linking and making
15215 sure the installed documentation is also owned by root.root.
15217 *Damien Miller <djm@mindrot.org>*
15219 * Give the OpenSSL applications more possibilities to make use of
15220 keys (public as well as private) handled by engines.
15224 * Add OCSP code that comes from CertCo.
15228 * Add VMS support for the Rijndael code.
15232 * Added untested support for Nuron crypto accelerator.
15236 * Add support for external cryptographic devices. This code was
15237 previously distributed separately as the "engine" branch.
15239 *Geoff Thorpe, Richard Levitte*
15241 * Rework the filename-translation in the DSO code. It is now possible to
15242 have far greater control over how a "name" is turned into a filename
15243 depending on the operating environment and any oddities about the
15244 different shared library filenames on each system.
15248 * Support threads on FreeBSD-elf in Configure.
15252 * Fix for SHA1 assembly problem with MASM: it produces
15253 warnings about corrupt line number information when assembling
15254 with debugging information. This is caused by the overlapping
15257 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15260 NCONF_get_number() has no error checking at all. As a replacement,
15261 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15262 promoted strongly. The old NCONF_get_number is kept around for
15263 binary backward compatibility.
15264 Make it possible for methods to load from something other than a BIO,
15265 by providing a function pointer that is given a name instead of a BIO.
15266 For example, this could be used to load configuration data from an
15271 * Fix for non blocking accept BIOs. Added new I/O special reason
15272 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15273 with non blocking I/O was not possible because no retry code was
15274 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15279 * Added the beginnings of Rijndael support.
15283 * Fix for bug in DirectoryString mask setting. Add support for
15284 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15285 to allow certificate printing to more controllable, additional
15286 'certopt' option to 'x509' to allow new printing options to be
15291 * Clean old EAY MD5 hack from e_os.h.
15295 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15297 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15298 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15300 *Joe Orton, Steve Henson*
15302 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15304 * Fix additional bug revealed by the NISCC test suite:
15306 Stop bug triggering large recursion when presented with
15307 certain ASN.1 tags ([CVE-2003-0851])
15311 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15313 * Fix various bugs revealed by running the NISCC test suite:
15315 Stop out of bounds reads in the ASN1 code when presented with
15316 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15318 If verify callback ignores invalid public key errors don't try to check
15319 certificate signature with the NULL public key.
15323 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15324 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15329 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15330 extra data after the compression methods not only for TLS 1.0
15331 but also for SSL 3.0 (as required by the specification).
15333 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15335 * Change X509_certificate_type() to mark the key as exported/exportable
15336 when it's 512 *bits* long, not 512 bytes.
15340 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15342 * Countermeasure against the Klima-Pokorny-Rosa extension of
15343 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15344 a protocol version number mismatch like a decryption error
15345 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15349 * Turn on RSA blinding by default in the default implementation
15350 to avoid a timing attack. Applications that don't want it can call
15351 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15352 They would be ill-advised to do so in most cases.
15354 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15356 * Change RSA blinding code so that it works when the PRNG is not
15357 seeded (in this case, the secret RSA exponent is abused as
15358 an unpredictable seed -- if it is not unpredictable, there
15359 is no point in blinding anyway). Make RSA blinding thread-safe
15360 by remembering the creator's thread ID in rsa->blinding and
15361 having all other threads use local one-time blinding factors
15362 (this requires more computation than sharing rsa->blinding, but
15363 avoids excessive locking; and if an RSA object is not shared
15364 between threads, blinding will still be very fast).
15368 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15370 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15371 via timing by performing a MAC computation even if incorrect
15372 block cipher padding has been found. This is a countermeasure
15373 against active attacks where the attacker has to distinguish
15374 between bad padding and a MAC verification error. ([CVE-2003-0078])
15376 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15377 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15378 Martin Vuagnoux (EPFL, Ilion)*
15380 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15382 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15383 memory from its contents. This is done with a counter that will
15384 place alternating values in each byte. This can be used to solve
15385 two issues: 1) the removal of calls to memset() by highly optimizing
15386 compilers, and 2) cleansing with other values than 0, since those can
15387 be read through on certain media, for example a swap space on disk.
15391 * Bugfix: client side session caching did not work with external caching,
15392 because the session->cipher setting was not restored when reloading
15393 from the external cache. This problem was masked, when
15394 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15395 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15399 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15400 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15402 *Zeev Lieber <zeev-l@yahoo.com>*
15404 * Undo an undocumented change introduced in 0.9.6e which caused
15405 repeated calls to OpenSSL_add_all_ciphers() and
15406 OpenSSL_add_all_digests() to be ignored, even after calling
15411 * Change the default configuration reader to deal with last line not
15412 being properly terminated.
15416 * Change X509_NAME_cmp() so it applies the special rules on handling
15417 DN values that are of type PrintableString, as well as RDNs of type
15418 emailAddress where the value has the type ia5String.
15420 *stefank@valicert.com via Richard Levitte*
15422 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15423 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15424 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15425 the bitwise-OR of the two for use by the majority of applications
15426 wanting this behaviour, and update the docs. The documented
15427 behaviour and actual behaviour were inconsistent and had been
15428 changing anyway, so this is more a bug-fix than a behavioural
15431 *Geoff Thorpe, diagnosed by Nadav Har'El*
15433 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15434 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15438 * Fix initialization code race conditions in
15439 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15440 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15441 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15442 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15443 ssl2_get_cipher_by_char(),
15444 ssl3_get_cipher_by_char().
15446 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15448 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15449 the cached sessions are flushed, as the remove_cb() might use ex_data
15450 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15451 (see [openssl.org #212]).
15453 *Geoff Thorpe, Lutz Jaenicke*
15455 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15456 length, instead of the encoding length to d2i_ASN1_OBJECT.
15460 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15462 * [In 0.9.6g-engine release:]
15463 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15465 *Lynn Gazis <lgazis@rainbow.com>*
15467 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15469 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15470 and get fix the header length calculation.
15471 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15472 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15474 * Use proper error handling instead of 'assertions' in buffer
15475 overflow checks added in 0.9.6e. This prevents DoS (the
15476 assertions could call abort()).
15478 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15480 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15482 * Add various sanity checks to asn1_get_length() to reject
15483 the ASN1 length bytes if they exceed sizeof(long), will appear
15484 negative or the content length exceeds the length of the
15487 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15489 * Fix cipher selection routines: ciphers without encryption had no flags
15490 for the cipher strength set and where therefore not handled correctly
15491 by the selection routines (PR #130).
15495 * Fix EVP_dsa_sha macro.
15500 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15501 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15502 that was added in OpenSSL 0.9.6d.
15504 As the countermeasure turned out to be incompatible with some
15505 broken SSL implementations, the new option is part of SSL_OP_ALL.
15506 SSL_OP_ALL is usually employed when compatibility with weird SSL
15507 implementations is desired (e.g. '-bugs' option to 's_client' and
15508 's_server'), so the new option is automatically set in many
15513 * Changes in security patch:
15515 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15516 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15517 Air Force Materiel Command, USAF, under agreement number
15520 * Add various sanity checks to asn1_get_length() to reject
15521 the ASN1 length bytes if they exceed sizeof(long), will appear
15522 negative or the content length exceeds the length of the
15523 supplied buffer. ([CVE-2002-0659])
15525 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15527 * Assertions for various potential buffer overflows, not known to
15528 happen in practice.
15530 *Ben Laurie (CHATS)*
15532 * Various temporary buffers to hold ASCII versions of integers were
15533 too small for 64 bit platforms. ([CVE-2002-0655])
15534 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15536 * Remote buffer overflow in SSL3 protocol - an attacker could
15537 supply an oversized session ID to a client. ([CVE-2002-0656])
15539 *Ben Laurie (CHATS)*
15541 * Remote buffer overflow in SSL2 protocol - an attacker could
15542 supply an oversized client master key. ([CVE-2002-0656])
15544 *Ben Laurie (CHATS)*
15546 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15548 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15549 encoded as NULL) with id-dsa-with-sha1.
15551 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15553 * Check various `X509_...()` return values in `apps/req.c`.
15555 *Nils Larsch <nla@trustcenter.de>*
15557 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15558 an end-of-file condition would erroneously be flagged, when the CRLF
15559 was just at the end of a processed block. The bug was discovered when
15560 processing data through a buffering memory BIO handing the data to a
15561 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15562 <ptsekov@syntrex.com> and Nedelcho Stanev.
15566 * Implement a countermeasure against a vulnerability recently found
15567 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15568 before application data chunks to avoid the use of known IVs
15569 with data potentially chosen by the attacker.
15573 * Fix length checks in ssl3_get_client_hello().
15577 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15578 to prevent ssl3_read_internal() from incorrectly assuming that
15579 ssl3_read_bytes() found application data while handshake
15580 processing was enabled when in fact s->s3->in_read_app_data was
15581 merely automatically cleared during the initial handshake.
15583 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15585 * Fix object definitions for Private and Enterprise: they were not
15586 recognized in their shortname (=lowercase) representation. Extend
15587 obj_dat.pl to issue an error when using undefined keywords instead
15588 of silently ignoring the problem (Svenning Sorensen
15589 <sss@sss.dnsalias.net>).
15593 * Fix DH_generate_parameters() so that it works for 'non-standard'
15594 generators, i.e. generators other than 2 and 5. (Previously, the
15595 code did not properly initialise the 'add' and 'rem' values to
15596 BN_generate_prime().)
15598 In the new general case, we do not insist that 'generator' is
15599 actually a primitive root: This requirement is rather pointless;
15600 a generator of the order-q subgroup is just as good, if not
15605 * Map new X509 verification errors to alerts. Discovered and submitted by
15606 Tom Wu <tom@arcot.com>.
15610 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15611 returning non-zero before the data has been completely received
15612 when using non-blocking I/O.
15614 *Bodo Moeller; problem pointed out by John Hughes*
15616 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15618 *Ben Laurie, Lutz Jaenicke*
15620 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15621 Yoram Zahavi <YoramZ@gilian.com>).
15625 * Add information about CygWin 1.3 and on, and preserve proper
15626 configuration for the versions before that.
15628 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15630 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15631 check whether we deal with a copy of a session and do not delete from
15632 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15633 <izhar@checkpoint.com>.
15637 * Do not store session data into the internal session cache, if it
15638 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15639 flag is set). Proposed by Aslam <aslam@funk.com>.
15643 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15648 * [In 0.9.6d-engine release:]
15649 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15651 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15653 * Add the configuration target linux-s390x.
15655 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15657 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15658 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15659 variable as an indication that a ClientHello message has been
15660 received. As the flag value will be lost between multiple
15661 invocations of ssl3_accept when using non-blocking I/O, the
15662 function may not be aware that a handshake has actually taken
15663 place, thus preventing a new session from being added to the
15666 To avoid this problem, we now set s->new_session to 2 instead of
15667 using a local variable.
15669 *Lutz Jaenicke, Bodo Moeller*
15671 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15672 if the SSL_R_LENGTH_MISMATCH error is detected.
15674 *Geoff Thorpe, Bodo Moeller*
15676 * New 'shared_ldflag' column in Configure platform table.
15680 * Fix EVP_CIPHER_mode macro.
15682 *"Dan S. Camper" <dan@bti.net>*
15684 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15685 type, we must throw them away by setting rr->length to 0.
15687 *D P Chang <dpc@qualys.com>*
15689 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15691 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15692 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15693 worked incorrectly for those cases where range = `10..._2` and
15694 `3*range` is two bits longer than range.)
15698 * Only add signing time to PKCS7 structures if it is not already
15703 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15704 OBJ_ld_ce should be OBJ_id_ce.
15705 Also some ip-pda OIDs in crypto/objects/objects.txt were
15706 incorrect (cf. RFC 3039).
15708 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15710 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15711 returns early because it has nothing to do.
15713 *Andy Schneider <andy.schneider@bjss.co.uk>*
15715 * [In 0.9.6c-engine release:]
15716 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15718 *Andy Schneider <andy.schneider@bjss.co.uk>*
15720 * [In 0.9.6c-engine release:]
15721 Add support for Cryptographic Appliance's keyserver technology.
15722 (Use engine 'keyclient')
15724 *Cryptographic Appliances and Geoff Thorpe*
15726 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15727 is called via tools/c89.sh because arguments have to be
15728 rearranged (all '-L' options must appear before the first object
15731 *Richard Shapiro <rshapiro@abinitio.com>*
15733 * [In 0.9.6c-engine release:]
15734 Add support for Broadcom crypto accelerator cards, backported
15737 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15739 * [In 0.9.6c-engine release:]
15740 Add support for SureWare crypto accelerator cards from
15741 Baltimore Technologies. (Use engine 'sureware')
15743 *Baltimore Technologies and Mark Cox*
15745 * [In 0.9.6c-engine release:]
15746 Add support for crypto accelerator cards from Accelerated
15747 Encryption Processing, www.aep.ie. (Use engine 'aep')
15749 *AEP Inc. and Mark Cox*
15751 * Add a configuration entry for gcc on UnixWare.
15753 *Gary Benson <gbenson@redhat.com>*
15755 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15756 messages are stored in a single piece (fixed-length part and
15757 variable-length part combined) and fix various bugs found on the way.
15761 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15762 instead. BIO_gethostbyname() does not know what timeouts are
15763 appropriate, so entries would stay in cache even when they have
15765 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15767 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15768 faced with a pathologically small ClientHello fragment that does
15769 not contain client_version: Instead of aborting with an error,
15770 simply choose the highest available protocol version (i.e.,
15771 TLS 1.0 unless it is disabled). In practice, ClientHello
15772 messages are never sent like this, but this change gives us
15773 strictly correct behaviour at least for TLS.
15777 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15778 never resets s->method to s->ctx->method when called from within
15779 one of the SSL handshake functions.
15781 *Bodo Moeller; problem pointed out by Niko Baric*
15783 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15784 (sent using the client's version number) if client_version is
15785 smaller than the protocol version in use. Also change
15786 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15787 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15788 the client will at least see that alert.
15792 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15797 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15798 client receives HelloRequest while in a handshake.
15800 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15802 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15803 should end in 'break', not 'goto end' which circumvents various
15804 cleanups done in state SSL_ST_OK. But session related stuff
15805 must be disabled for SSL_ST_OK in the case that we just sent a
15808 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15809 before just sending a HelloRequest.
15811 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15813 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15814 reveal whether illegal block cipher padding was found or a MAC
15815 verification error occurred. (Neither SSLerr() codes nor alerts
15816 are directly visible to potential attackers, but the information
15817 may leak via logfiles.)
15819 Similar changes are not required for the SSL 2.0 implementation
15820 because the number of padding bytes is sent in clear for SSL 2.0,
15821 and the extra bytes are just ignored. However ssl/s2_pkt.c
15822 failed to verify that the purported number of padding bytes is in
15827 * Add OpenUNIX-8 support including shared libraries
15828 (Boyd Lynn Gerber <gerberb@zenez.com>).
15832 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15833 'wristwatch attack' using huge encoding parameters (cf.
15834 James H. Manger's CRYPTO 2001 paper). Note that the
15835 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15836 encoding parameters and hence was not vulnerable.
15840 * BN_sqr() bug fix.
15842 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15844 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15845 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15846 followed by modular reduction.
15848 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15850 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15851 equivalent based on BN_pseudo_rand() instead of BN_rand().
15855 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15856 This function was broken, as the check for a new client hello message
15857 to handle SGC did not allow these large messages.
15858 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15862 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15866 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15867 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15871 * Rework the configuration and shared library support for Tru64 Unix.
15872 The configuration part makes use of modern compiler features and
15873 still retains old compiler behavior for those that run older versions
15874 of the OS. The shared library support part includes a variant that
15875 uses the RPATH feature, and is available through the special
15876 configuration target "alpha-cc-rpath", which will never be selected
15879 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15881 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15882 with the same message size as in ssl3_get_certificate_request().
15883 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15884 messages might inadvertently be reject as too long.
15886 *Petr Lampa <lampa@fee.vutbr.cz>*
15888 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15892 * Modified SSL library such that the verify_callback that has been set
15893 specifically for an SSL object with SSL_set_verify() is actually being
15894 used. Before the change, a verify_callback set with this function was
15895 ignored and the verify_callback() set in the SSL_CTX at the time of
15896 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15897 to allow the necessary settings.
15901 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15902 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15903 done automatically (in contradiction to the requirements of the C
15904 standard). This made problems when used from OpenSSH.
15908 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15909 dh->length and always used
15911 BN_rand_range(priv_key, dh->p).
15913 BN_rand_range() is not necessary for Diffie-Hellman, and this
15914 specific range makes Diffie-Hellman unnecessarily inefficient if
15915 dh->length (recommended exponent length) is much smaller than the
15916 length of dh->p. We could use BN_rand_range() if the order of
15917 the subgroup was stored in the DH structure, but we only have
15922 BN_rand(priv_key, l, ...)
15924 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15931 RSA_eay_public_encrypt
15932 RSA_eay_private_decrypt
15933 RSA_eay_private_encrypt (signing)
15934 RSA_eay_public_decrypt (signature verification)
15936 (default implementations for RSA_public_encrypt,
15937 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15938 always reject numbers >= n.
15942 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15943 to synchronize access to 'locking_thread'. This is necessary on
15944 systems where access to 'locking_thread' (an 'unsigned long'
15945 variable) is not atomic.
15949 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15950 *before* setting the 'crypto_lock_rand' flag. The previous code had
15951 a race condition if 0 is a valid thread ID.
15953 *Travis Vitek <vitek@roguewave.com>*
15955 * Add support for shared libraries under Irix.
15957 *Albert Chin-A-Young <china@thewrittenword.com>*
15959 * Add configuration option to build on Linux on both big-endian and
15960 little-endian MIPS.
15962 *Ralf Baechle <ralf@uni-koblenz.de>*
15964 * Add the possibility to create shared libraries on HP-UX.
15968 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15970 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15971 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15972 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15973 PRNG state recovery was possible based on the output of
15974 one PRNG request appropriately sized to gain knowledge on
15975 'md' followed by enough consecutive 1-byte PRNG requests
15976 to traverse all of 'state'.
15978 1. When updating 'md_local' (the current thread's copy of 'md')
15979 during PRNG output generation, hash all of the previous
15980 'md_local' value, not just the half used for PRNG output.
15982 2. Make the number of bytes from 'state' included into the hash
15983 independent from the number of PRNG bytes requested.
15985 The first measure alone would be sufficient to avoid
15986 Markku-Juhani's attack. (Actually it had never occurred
15987 to me that the half of 'md_local' used for chaining was the
15988 half from which PRNG output bytes were taken -- I had always
15989 assumed that the secret half would be used.) The second
15990 measure makes sure that additional data from 'state' is never
15991 mixed into 'md_local' in small portions; this heuristically
15992 further strengthens the PRNG.
15996 * Fix crypto/bn/asm/mips3.s.
16000 * When only the key is given to "enc", the IV is undefined. Print out
16001 an error message in this case.
16005 * Handle special case when X509_NAME is empty in X509 printing routines.
16009 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
16010 positive and less than q.
16014 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
16015 used: it isn't thread safe and the add_lock_callback should handle
16018 *Paul Rose <Paul.Rose@bridge.com>*
16020 * Verify that incoming data obeys the block size in
16021 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
16027 *Ulf Möller, Bodo Möller*
16029 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
16030 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
16031 when fixing the server behaviour for backwards-compatible 'client
16032 hello' messages. (Note that the attack is impractical against
16033 SSL 3.0 and TLS 1.0 anyway because length and version checking
16034 means that the probability of guessing a valid ciphertext is
16035 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
16038 Before 0.9.5, the countermeasure (hide the error by generating a
16039 random 'decryption result') did not work properly because
16040 ERR_clear_error() was missing, meaning that SSL_get_error() would
16041 detect the supposedly ignored error.
16043 Both problems are now fixed.
16047 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
16048 (previously it was 1024).
16052 * Fix for compatibility mode trust settings: ignore trust settings
16053 unless some valid trust or reject settings are present.
16057 * Fix for blowfish EVP: its a variable length cipher.
16061 * Fix various bugs related to DSA S/MIME verification. Handle missing
16062 parameters in DSA public key structures and return an error in the
16063 DSA routines if parameters are absent.
16067 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
16068 in the current directory if neither $RANDFILE nor $HOME was set.
16069 RAND_file_name() in 0.9.6a returned NULL in this case. This has
16070 caused some confusion to Windows users who haven't defined $HOME.
16071 Thus RAND_file_name() is changed again: e_os.h can define a
16072 DEFAULT_HOME, which will be used if $HOME is not set.
16073 For Windows, we use "C:"; on other platforms, we still require
16074 environment variables.
16076 * Move 'if (!initialized) RAND_poll()' into regions protected by
16077 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
16078 having multiple threads call RAND_poll() concurrently.
16082 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
16083 combination of a flag and a thread ID variable.
16084 Otherwise while one thread is in ssleay_rand_bytes (which sets the
16085 flag), *other* threads can enter ssleay_add_bytes without obeying
16086 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
16087 that they do not hold after the first thread unsets add_do_not_lock).
16091 * Change bctest again: '-x' expressions are not available in all
16092 versions of 'test'.
16096 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16098 * Fix a couple of memory leaks in PKCS7_dataDecode()
16100 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16102 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16103 the default extension for executables, if any. Also, make the perl
16104 scripts that use symlink() to test if it really exists and use "cp"
16105 if it doesn't. All this made OpenSSL compilable and installable in
16110 * Fix for asn1_GetSequence() for indefinite length constructed data.
16111 If SEQUENCE is length is indefinite just set c->slen to the total
16112 amount of data available.
16114 *Steve Henson, reported by shige@FreeBSD.org*
16116 *This change does not apply to 0.9.7.*
16118 * Change bctest to avoid here-documents inside command substitution
16119 (workaround for FreeBSD /bin/sh bug).
16120 For compatibility with Ultrix, avoid shell functions (introduced
16121 in the bctest version that searches along $PATH).
16125 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16126 with des_encrypt() defined on some operating systems, like Solaris
16131 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16132 On the Importance of Eliminating Errors in Cryptographic
16133 Computations, J. Cryptology 14 (2001) 2, 101-119,
16134 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16138 * MIPS assembler BIGNUM division bug fix.
16142 * Disabled incorrect Alpha assembler code.
16146 * Fix PKCS#7 decode routines so they correctly update the length
16147 after reading an EOC for the EXPLICIT tag.
16151 *This change does not apply to 0.9.7.*
16153 * Fix bug in PKCS#12 key generation routines. This was triggered
16154 if a 3DES key was generated with a 0 initial byte. Include
16155 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16156 (but broken) behaviour.
16160 * Enhance bctest to search for a working bc along $PATH and print
16163 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16165 * Fix memory leaks in err.c: free err_data string if necessary;
16166 don't write to the wrong index in ERR_set_error_data.
16170 * Implement ssl23_peek (analogous to ssl23_read), which previously
16175 * Replace rdtsc with `_emit` statements for VC++ version 5.
16177 *Jeremy Cooper <jeremy@baymoo.org>*
16179 * Make it possible to reuse SSLv2 sessions.
16183 * In copy_email() check for >= 0 as a return value for
16184 X509_NAME_get_index_by_NID() since 0 is a valid index.
16186 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16188 * Avoid coredump with unsupported or invalid public keys by checking if
16189 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16190 PKCS7_verify() fails with non detached data.
16194 * Don't use getenv in library functions when run as setuid/setgid.
16195 New function OPENSSL_issetugid().
16199 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16200 due to incorrect handling of multi-threading:
16202 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16204 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16206 3. Count how many times MemCheck_off() has been called so that
16207 nested use can be treated correctly. This also avoids
16208 inband-signalling in the previous code (which relied on the
16209 assumption that thread ID 0 is impossible).
16213 * Add "-rand" option also to s_client and s_server.
16217 * Fix CPU detection on Irix 6.x.
16218 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16219 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16221 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16226 *This change does not apply to 0.9.7.*
16228 * Use the cached encoding of an X509_NAME structure rather than
16229 copying it. This is apparently the reason for the libsafe "errors"
16230 but the code is actually correct.
16234 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16235 Bleichenbacher's DSA attack.
16236 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16237 to be set and top=0 forces the highest bit to be set; top=-1 is new
16238 and leaves the highest bit random.
16240 *Ulf Moeller, Bodo Moeller*
16242 * In the `NCONF_...`-based implementations for `CONF_...` queries
16243 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16244 a temporary CONF structure with the data component set to NULL
16245 (which gives segmentation faults in lh_retrieve).
16246 Instead, use NULL for the CONF pointer in CONF_get_string and
16247 CONF_get_number (which may use environment variables) and directly
16248 return NULL from CONF_get_section.
16252 * Fix potential buffer overrun for EBCDIC.
16256 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16257 keyUsage if basicConstraints absent for a CA.
16261 * Make SMIME_write_PKCS7() write mail header values with a format that
16262 is more generally accepted (no spaces before the semicolon), since
16263 some programs can't parse those values properly otherwise. Also make
16264 sure BIO's that break lines after each write do not create invalid
16269 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16270 macros previously used would not encode an empty SEQUENCE OF
16271 and break the signature.
16275 *This change does not apply to 0.9.7.*
16277 * Zero the premaster secret after deriving the master secret in
16282 * Add some EVP_add_digest_alias registrations (as found in
16283 OpenSSL_add_all_digests()) to SSL_library_init()
16284 aka OpenSSL_add_ssl_algorithms(). This provides improved
16285 compatibility with peers using X.509 certificates
16286 with unconventional AlgorithmIdentifier OIDs.
16290 * Fix for Irix with NO_ASM.
16292 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16294 * ./config script fixes.
16296 *Ulf Moeller, Richard Levitte*
16298 * Fix 'openssl passwd -1'.
16302 * Change PKCS12_key_gen_asc() so it can cope with non null
16303 terminated strings whose length is passed in the passlen
16304 parameter, for example from PEM callbacks. This was done
16305 by adding an extra length parameter to asc2uni().
16307 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16309 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16310 call failed, free the DSA structure.
16314 * Fix to uni2asc() to cope with zero length Unicode strings.
16315 These are present in some PKCS#12 files.
16319 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16320 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16321 when writing a 32767 byte record.
16323 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16325 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16326 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16328 (RSA objects have a reference count access to which is protected
16329 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16330 so they are meant to be shared between threads.)
16331 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16332 "Reddie, Steven" <Steven.Reddie@ca.com>*
16334 * Fix a deadlock in CRYPTO_mem_leaks().
16338 * Use better test patterns in bntest.
16342 * rand_win.c fix for Borland C.
16346 * BN_rshift bugfix for n == 0.
16350 * Add a 'bctest' script that checks for some known 'bc' bugs
16351 so that 'make test' does not abort just because 'bc' is broken.
16355 * Store verify_result within SSL_SESSION also for client side to
16356 avoid potential security hole. (Re-used sessions on the client side
16357 always resulted in verify_result==X509_V_OK, not using the original
16358 result of the server certificate verification.)
16362 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16363 SSL3_RT_APPLICATION_DATA, return 0.
16364 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16369 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16370 releases, have been re-implemented by renaming the previous
16371 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16372 and ssl3_read_internal, respectively, and adding 'peek' parameters
16373 to them. The new ssl[23]_{read,peek} functions are calls to
16374 ssl[23]_read_internal with the 'peek' flag set appropriately.
16375 A 'peek' parameter has also been added to ssl3_read_bytes, which
16376 does the actual work for ssl3_read_internal.
16380 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16381 the method-specific "init()" handler. Also clean up ex_data after
16382 calling the method-specific "finish()" handler. Previously, this was
16383 happening the other way round.
16387 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16388 The previous value, 12, was not always sufficient for BN_mod_exp().
16392 * Make sure that shared libraries get the internal name engine with
16393 the full version number and not just 0. This should mark the
16394 shared libraries as not backward compatible. Of course, this should
16395 be changed again when we can guarantee backward binary compatibility.
16399 * Fix typo in get_cert_by_subject() in by_dir.c
16401 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16403 * Rework the system to generate shared libraries:
16405 - Make note of the expected extension for the shared libraries and
16406 if there is a need for symbolic links from for example libcrypto.so.0
16407 to libcrypto.so.0.9.7. There is extended info in Configure for
16410 - Make as few rebuilds of the shared libraries as possible.
16412 - Still avoid linking the OpenSSL programs with the shared libraries.
16414 - When installing, install the shared libraries separately from the
16419 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16421 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16422 and not in SSL_clear because the latter is also used by the
16423 accept/connect functions; previously, the settings made by
16424 SSL_set_read_ahead would be lost during the handshake.
16426 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16428 * Correct util/mkdef.pl to be selective about disabled algorithms.
16429 Previously, it would create entries for disabled algorithms no
16434 * Added several new manual pages for SSL_* function.
16438 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16440 * In ssl23_get_client_hello, generate an error message when faced
16441 with an initial SSL 3.0/TLS record that is too small to contain the
16442 first two bytes of the ClientHello message, i.e. client_version.
16443 (Note that this is a pathologic case that probably has never happened
16444 in real life.) The previous approach was to use the version number
16445 from the record header as a substitute; but our protocol choice
16446 should not depend on that one because it is not authenticated
16447 by the Finished messages.
16451 * More robust randomness gathering functions for Windows.
16453 *Jeffrey Altman <jaltman@columbia.edu>*
16455 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16456 not set then we don't setup the error code for issuer check errors
16457 to avoid possibly overwriting other errors which the callback does
16458 handle. If an application does set the flag then we assume it knows
16459 what it is doing and can handle the new informational codes
16464 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16465 a general "ANY" type, as such it should be able to decode anything
16466 including tagged types. However it didn't check the class so it would
16467 wrongly interpret tagged types in the same way as their universal
16468 counterpart and unknown types were just rejected. Changed so that the
16469 tagged and unknown types are handled in the same way as a SEQUENCE:
16470 that is the encoding is stored intact. There is also a new type
16471 "V_ASN1_OTHER" which is used when the class is not universal, in this
16472 case we have no idea what the actual type is so we just lump them all
16477 * On VMS, stdout may very well lead to a file that is written to
16478 in a record-oriented fashion. That means that every write() will
16479 write a separate record, which will be read separately by the
16480 programs trying to read from it. This can be very confusing.
16482 The solution is to put a BIO filter in the way that will buffer
16483 text until a linefeed is reached, and then write everything a
16484 line at a time, so every record written will be an actual line,
16485 not chunks of lines and not (usually doesn't happen, but I've
16486 seen it once) several lines in one record. BIO_f_linebuffer() is
16489 Currently, it's a VMS-only method, because that's where it has
16490 been tested well enough.
16494 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16495 it can return incorrect results.
16496 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16497 but it was in 0.9.6-beta[12].)
16501 * Disable the check for content being present when verifying detached
16502 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16503 include zero length content when signing messages.
16507 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16508 BIO_ctrl (for BIO pairs).
16512 * Add DSO method for VMS.
16516 * Bug fix: Montgomery multiplication could produce results with the
16521 * Add RPM specification openssl.spec and modify it to build three
16522 packages. The default package contains applications, application
16523 documentation and run-time libraries. The devel package contains
16524 include files, static libraries and function documentation. The
16525 doc package contains the contents of the doc directory. The original
16526 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16530 * Add a large number of documentation files for many SSL routines.
16532 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16534 * Add a configuration entry for Sony News 4.
16536 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16538 * Don't set the two most significant bits to one when generating a
16539 random number < q in the DSA library.
16543 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16544 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16545 the underlying transport is blocking) if a handshake took place.
16546 (The default behaviour is needed by applications such as s_client
16547 and s_server that use select() to determine when to use SSL_read;
16548 but for applications that know in advance when to expect data, it
16549 just makes things more complicated.)
16553 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16558 * Add a few more EBCDIC conditionals that make `req` and `x509`
16559 work better on such systems.
16561 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16563 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16564 Update PKCS12_parse() so it copies the friendlyName and the
16565 keyid to the certificates aux info.
16569 * Fix bug in PKCS7_verify() which caused an infinite loop
16570 if there was more than one signature.
16572 *Sven Uszpelkat <su@celocom.de>*
16574 * Major change in util/mkdef.pl to include extra information
16575 about each symbol, as well as presenting variables as well
16576 as functions. This change means that there's n more need
16577 to rebuild the .num files when some algorithms are excluded.
16581 * Allow the verify time to be set by an application,
16582 rather than always using the current time.
16586 * Phase 2 verify code reorganisation. The certificate
16587 verify code now looks up an issuer certificate by a
16588 number of criteria: subject name, authority key id
16589 and key usage. It also verifies self signed certificates
16590 by the same criteria. The main comparison function is
16591 X509_check_issued() which performs these checks.
16593 Lot of changes were necessary in order to support this
16594 without completely rewriting the lookup code.
16596 Authority and subject key identifier are now cached.
16598 The LHASH 'certs' is X509_STORE has now been replaced
16599 by a STACK_OF(X509_OBJECT). This is mainly because an
16600 LHASH can't store or retrieve multiple objects with
16601 the same hash value.
16603 As a result various functions (which were all internal
16604 use only) have changed to handle the new X509_STORE
16605 structure. This will break anything that messed round
16606 with X509_STORE internally.
16608 The functions X509_STORE_add_cert() now checks for an
16609 exact match, rather than just subject name.
16611 The X509_STORE API doesn't directly support the retrieval
16612 of multiple certificates matching a given criteria, however
16613 this can be worked round by performing a lookup first
16614 (which will fill the cache with candidate certificates)
16615 and then examining the cache for matches. This is probably
16616 the best we can do without throwing out X509_LOOKUP
16617 entirely (maybe later...).
16619 The X509_VERIFY_CTX structure has been enhanced considerably.
16621 All certificate lookup operations now go via a get_issuer()
16622 callback. Although this currently uses an X509_STORE it
16623 can be replaced by custom lookups. This is a simple way
16624 to bypass the X509_STORE hackery necessary to make this
16625 work and makes it possible to use more efficient techniques
16626 in future. A very simple version which uses a simple
16627 STACK for its trusted certificate store is also provided
16628 using X509_STORE_CTX_trusted_stack().
16630 The verify_cb() and verify() callbacks now have equivalents
16631 in the X509_STORE_CTX structure.
16633 X509_STORE_CTX also has a 'flags' field which can be used
16634 to customise the verify behaviour.
16638 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16639 excludes S/MIME capabilities.
16643 * When a certificate request is read in keep a copy of the
16644 original encoding of the signed data and use it when outputting
16645 again. Signatures then use the original encoding rather than
16646 a decoded, encoded version which may cause problems if the
16647 request is improperly encoded.
16651 * For consistency with other BIO_puts implementations, call
16652 buffer_write(b, ...) directly in buffer_puts instead of calling
16655 In BIO_puts, increment b->num_write as in BIO_write.
16657 *Peter.Sylvester@EdelWeb.fr*
16659 * Fix BN_mul_word for the case where the word is 0. (We have to use
16660 BN_zero, we may not return a BIGNUM with an array consisting of
16661 words set to zero.)
16665 * Avoid calling abort() from within the library when problems are
16666 detected, except if preprocessor symbols have been defined
16667 (such as REF_CHECK, BN_DEBUG etc.).
16671 * New openssl application 'rsautl'. This utility can be
16672 used for low-level RSA operations. DER public key
16673 BIO/fp routines also added.
16677 * New Configure entry and patches for compiling on QNX 4.
16679 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16681 * A demo state-machine implementation was sponsored by
16682 Nuron (<http://www.nuron.com/>) and is now available in
16683 demos/state_machine.
16687 * New options added to the 'dgst' utility for signature
16688 generation and verification.
16692 * Unrecognized PKCS#7 content types are now handled via a
16693 catch all ASN1_TYPE structure. This allows unsupported
16694 types to be stored as a "blob" and an application can
16695 encode and decode it manually.
16699 * Fix various signed/unsigned issues to make a_strex.c
16700 compile under VC++.
16702 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16704 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16705 length if passed a buffer. ASN1_INTEGER_to_BN failed
16706 if passed a NULL BN and its argument was negative.
16708 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16710 * Modification to PKCS#7 encoding routines to output definite
16711 length encoding. Since currently the whole structures are in
16712 memory there's not real point in using indefinite length
16713 constructed encoding. However if OpenSSL is compiled with
16714 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16718 * Added BIO_vprintf() and BIO_vsnprintf().
16722 * Added more prefixes to parse for in the strings written
16723 through a logging bio, to cover all the levels that are available
16724 through syslog. The prefixes are now:
16726 PANIC, EMERG, EMR => LOG_EMERG
16727 ALERT, ALR => LOG_ALERT
16728 CRIT, CRI => LOG_CRIT
16729 ERROR, ERR => LOG_ERR
16730 WARNING, WARN, WAR => LOG_WARNING
16731 NOTICE, NOTE, NOT => LOG_NOTICE
16732 INFO, INF => LOG_INFO
16733 DEBUG, DBG => LOG_DEBUG
16735 and as before, if none of those prefixes are present at the
16736 beginning of the string, LOG_ERR is chosen.
16738 On Win32, the `LOG_*` levels are mapped according to this:
16740 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16741 LOG_WARNING => EVENTLOG_WARNING_TYPE
16742 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16746 * Made it possible to reconfigure with just the configuration
16747 argument "reconf" or "reconfigure". The command line arguments
16748 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16749 and are retrieved from there when reconfiguring.
16755 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16757 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16761 * The obj_dat.pl script was messing up the sorting of object
16762 names. The reason was that it compared the quoted version
16763 of strings as a result "OCSP" > "OCSP Signing" because
16764 " > SPACE. Changed script to store unquoted versions of
16765 names and add quotes on output. It was also omitting some
16766 names from the lookup table if they were given a default
16767 value (that is if SN is missing it is given the same
16768 value as LN and vice versa), these are now added on the
16769 grounds that if an object has a name we should be able to
16770 look it up. Finally added warning output when duplicate
16771 short or long names are found.
16775 * Changes needed for Tandem NSK.
16777 *Scott Uroff <scott@xypro.com>*
16779 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16780 RSA_padding_check_SSLv23(), special padding was never detected
16781 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16782 version rollback attacks was not effective.
16784 In s23_clnt.c, don't use special rollback-attack detection padding
16785 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16786 client; similarly, in s23_srvr.c, don't do the rollback check if
16787 SSL 2.0 is the only protocol enabled in the server.
16791 * Make it possible to get hexdumps of unprintable data with 'openssl
16792 asn1parse'. By implication, the functions ASN1_parse_dump() and
16793 BIO_dump_indent() are added.
16797 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16798 these print out strings and name structures based on various
16799 flags including RFC2253 support and proper handling of
16800 multibyte characters. Added options to the 'x509' utility
16801 to allow the various flags to be set.
16805 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16806 Also change the functions X509_cmp_current_time() and
16807 X509_gmtime_adj() work with an ASN1_TIME structure,
16808 this will enable certificates using GeneralizedTime in validity
16809 dates to be checked.
16813 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16814 negative public key encodings) on by default,
16815 NO_NEG_PUBKEY_BUG can be set to disable it.
16819 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16820 content octets. An i2c_ASN1_OBJECT is unnecessary because
16821 the encoding can be trivially obtained from the structure.
16825 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16826 not read locks (`CRYPTO_r_[un]lock`).
16830 * A first attempt at creating official support for shared
16831 libraries through configuration. I've kept it so the
16832 default is static libraries only, and the OpenSSL programs
16833 are always statically linked for now, but there are
16834 preparations for dynamic linking in place.
16835 This has been tested on Linux and Tru64.
16839 * Randomness polling function for Win9x, as described in:
16840 Peter Gutmann, Software Generation of Practically Strong
16845 * Fix so PRNG is seeded in req if using an already existing
16850 * New options to smime application. -inform and -outform
16851 allow alternative formats for the S/MIME message including
16852 PEM and DER. The -content option allows the content to be
16853 specified separately. This should allow things like Netscape
16854 form signing output easier to verify.
16858 * Fix the ASN1 encoding of tags using the 'long form'.
16862 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16863 STRING types. These convert content octets to and from the
16864 underlying type. The actual tag and length octets are
16865 already assumed to have been read in and checked. These
16866 are needed because all other string types have virtually
16867 identical handling apart from the tag. By having versions
16868 of the ASN1 functions that just operate on content octets
16869 IMPLICIT tagging can be handled properly. It also allows
16870 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16871 and ASN1_INTEGER are identical apart from the tag.
16875 * Change the handling of OID objects as follows:
16877 - New object identifiers are inserted in objects.txt, following
16878 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16879 - objects.pl is used to process obj_mac.num and create a new
16881 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16884 This is currently kind of a hack, and the perl code in objects.pl
16885 isn't very elegant, but it works as I intended. The simplest way
16886 to check that it worked correctly is to look in obj_dat.h and
16887 check the array nid_objs and make sure the objects haven't moved
16888 around (this is important!). Additions are OK, as well as
16889 consistent name changes.
16893 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16897 * Addition of the command line parameter '-rand file' to 'openssl req'.
16898 The given file adds to whatever has already been seeded into the
16899 random pool through the RANDFILE configuration file option or
16900 environment variable, or the default random state file.
16904 * mkstack.pl now sorts each macro group into lexical order.
16905 Previously the output order depended on the order the files
16906 appeared in the directory, resulting in needless rewriting
16911 * Patches to make OpenSSL compile under Win32 again. Mostly
16912 work arounds for the VC++ problem that it treats func() as
16913 func(void). Also stripped out the parts of mkdef.pl that
16914 added extra typesafe functions: these no longer exist.
16918 * Reorganisation of the stack code. The macros are now all
16919 collected in safestack.h . Each macro is defined in terms of
16920 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16921 DEBUG_SAFESTACK is now handled in terms of function casts,
16922 this has the advantage of retaining type safety without the
16923 use of additional functions. If DEBUG_SAFESTACK is not defined
16924 then the non typesafe macros are used instead. Also modified the
16925 mkstack.pl script to handle the new form. Needs testing to see
16926 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16927 the default if no major problems. Similar behaviour for ASN1_SET_OF
16928 and PKCS12_STACK_OF.
16932 * When some versions of IIS use the 'NET' form of private key the
16933 key derivation algorithm is different. Normally MD5(password) is
16934 used as a 128 bit RC4 key. In the modified case
16935 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16936 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16937 as the old Netscape_RSA functions except they have an additional
16938 'sgckey' parameter which uses the modified algorithm. Also added
16939 an -sgckey command line option to the rsa utility. Thanks to
16940 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16941 algorithm to openssl-dev.
16945 * The evp_local.h macros were using 'c.##kname' which resulted in
16946 invalid expansion on some systems (SCO 5.0.5 for example).
16947 Corrected to 'c.kname'.
16949 *Phillip Porch <root@theporch.com>*
16951 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16952 a STACK of email addresses from a certificate or request, these look
16953 in the subject name and the subject alternative name extensions and
16954 omit any duplicate addresses.
16958 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16959 This makes DSA verification about 2 % faster.
16963 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16964 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16965 plus overhead for 1024 bit moduli).
16966 This makes exponentiations about 0.5 % faster for 1024 bit
16967 exponents (as measured by "openssl speed rsa2048").
16971 * Rename memory handling macros to avoid conflicts with other
16973 Malloc => OPENSSL_malloc
16974 Malloc_locked => OPENSSL_malloc_locked
16975 Realloc => OPENSSL_realloc
16976 Free => OPENSSL_free
16980 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16981 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16985 * CygWin32 support.
16987 *John Jarvie <jjarvie@newsguy.com>*
16989 * The type-safe stack code has been rejigged. It is now only compiled
16990 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16991 by default all type-specific stack functions are "#define"d back to
16992 standard stack functions. This results in more streamlined output
16993 but retains the type-safety checking possibilities of the original
16998 * The STACK code has been cleaned up, and certain type declarations
16999 that didn't make a lot of sense have been brought in line. This has
17000 also involved a cleanup of sorts in safestack.h to more correctly
17001 map type-safe stack functions onto their plain stack counterparts.
17002 This work has also resulted in a variety of "const"ifications of
17003 lots of the code, especially `_cmp` operations which should normally
17004 be prototyped with "const" parameters anyway.
17008 * When generating bytes for the first time in md_rand.c, 'stir the pool'
17009 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
17010 (The PRNG state consists of two parts, the large pool 'state' and 'md',
17011 where all of 'md' is used each time the PRNG is used, but 'state'
17012 is used only indexed by a cyclic counter. As entropy may not be
17013 well distributed from the beginning, 'md' is important as a
17014 chaining variable. However, the output function chains only half
17015 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
17016 all of 'md', and seeding with STATE_SIZE dummy bytes will result
17017 in all of 'state' being rewritten, with the new values depending
17018 on virtually all of 'md'. This overcomes the 80 bit limitation.)
17022 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
17023 the handshake is continued after ssl_verify_cert_chain();
17024 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
17025 can lead to 'unexplainable' connection aborts later.
17027 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
17029 * Major EVP API cipher revision.
17030 Add hooks for extra EVP features. This allows various cipher
17031 parameters to be set in the EVP interface. Support added for variable
17032 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
17033 setting of RC2 and RC5 parameters.
17035 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
17038 Remove lots of duplicated code from the EVP library. For example *every*
17039 cipher init() function handles the 'iv' in the same way according to the
17040 cipher mode. They also all do nothing if the 'key' parameter is NULL and
17041 for CFB and OFB modes they zero ctx->num.
17043 New functionality allows removal of S/MIME code RC2 hack.
17045 Most of the routines have the same form and so can be declared in terms
17048 By shifting this to the top level EVP_CipherInit() it can be removed from
17049 all individual ciphers. If the cipher wants to handle IVs or keys
17050 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
17053 Change lots of functions like EVP_EncryptUpdate() to now return a
17054 value: although software versions of the algorithms cannot fail
17055 any installed hardware versions can.
17059 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
17060 this option is set, tolerate broken clients that send the negotiated
17061 protocol version number instead of the requested protocol version
17066 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
17067 i.e. non-zero for export ciphersuites, zero otherwise.
17068 Previous versions had this flag inverted, inconsistent with
17069 rsa_tmp_cb (..._TMP_RSA_CB).
17071 *Bodo Moeller; problem reported by Amit Chopra*
17073 * Add missing DSA library text string. Work around for some IIS
17074 key files with invalid SEQUENCE encoding.
17078 * Add a document (doc/standards.txt) that list all kinds of standards
17079 and so on that are implemented in OpenSSL.
17083 * Enhance c_rehash script. Old version would mishandle certificates
17084 with the same subject name hash and wouldn't handle CRLs at all.
17085 Added -fingerprint option to crl utility, to support new c_rehash
17090 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17094 * Fix for SSL server purpose checking. Server checking was
17095 rejecting certificates which had extended key usage present
17096 but no ssl client purpose.
17098 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17100 * Make PKCS#12 code work with no password. The PKCS#12 spec
17101 is a little unclear about how a blank password is handled.
17102 Since the password in encoded as a BMPString with terminating
17103 double NULL a zero length password would end up as just the
17104 double NULL. However no password at all is different and is
17105 handled differently in the PKCS#12 key generation code. NS
17106 treats a blank password as zero length. MSIE treats it as no
17107 password on export: but it will try both on import. We now do
17108 the same: PKCS12_parse() tries zero length and no password if
17109 the password is set to "" or NULL (NULL is now a valid password:
17110 it wasn't before) as does the pkcs12 application.
17114 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17115 perror when PEM_read_bio_X509_REQ fails, the error message must
17116 be obtained from the error queue.
17120 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17121 it in ERR_remove_state if appropriate, and change ERR_get_state
17122 accordingly to avoid race conditions (this is necessary because
17123 thread_hash is no longer constant once set).
17127 * Bugfix for linux-elf makefile.one.
17131 * RSA_get_default_method() will now cause a default
17132 RSA_METHOD to be chosen if one doesn't exist already.
17133 Previously this was only set during a call to RSA_new()
17134 or RSA_new_method(NULL) meaning it was possible for
17135 RSA_get_default_method() to return NULL.
17139 * Added native name translation to the existing DSO code
17140 that will convert (if the flag to do so is set) filenames
17141 that are sufficiently small and have no path information
17142 into a canonical native form. Eg. "blah" converted to
17143 "libblah.so" or "blah.dll" etc.
17147 * New function ERR_error_string_n(e, buf, len) which is like
17148 ERR_error_string(e, buf), but writes at most 'len' bytes
17149 including the 0 terminator. For ERR_error_string_n, 'buf'
17152 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17154 * CONF library reworked to become more general. A new CONF
17155 configuration file reader "class" is implemented as well as a
17156 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17157 old `CONF_*` functions are still there, but are reimplemented to
17158 work in terms of the new functions. Also, a set of functions
17159 to handle the internal storage of the configuration data is
17160 provided to make it easier to write new configuration file
17161 reader "classes" (I can definitely see something reading a
17162 configuration file in XML format, for example), called `_CONF_*`,
17163 or "the configuration storage API"...
17165 The new configuration file reading functions are:
17167 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17168 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17170 NCONF_default, NCONF_WIN32
17172 NCONF_dump_fp, NCONF_dump_bio
17174 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17175 NCONF_new creates a new CONF object. This works in the same way
17176 as other interfaces in OpenSSL, like the BIO interface.
17177 `NCONF_dump_*` dump the internal storage of the configuration file,
17178 which is useful for debugging. All other functions take the same
17179 arguments as the old `CONF_*` functions with the exception of the
17180 first that must be a `CONF *` instead of a `LHASH *`.
17182 To make it easier to use the new classes with the old `CONF_*` functions,
17183 the function CONF_set_default_method is provided.
17187 * Add '-tls1' option to 'openssl ciphers', which was already
17188 mentioned in the documentation but had not been implemented.
17189 (This option is not yet really useful because even the additional
17190 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17194 * Initial DSO code added into libcrypto for letting OpenSSL (and
17195 OpenSSL-based applications) load shared libraries and bind to
17196 them in a portable way.
17198 *Geoff Thorpe, with contributions from Richard Levitte*
17200 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17202 * Make sure _lrotl and _lrotr are only used with MSVC.
17204 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17205 (the default implementation of RAND_status).
17207 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17208 to '-clrext' (= clear extensions), as intended and documented.
17209 *Bodo Moeller; inconsistency pointed out by Michael Attili
17210 <attili@amaxo.com>*
17212 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17213 was larger than the MD block size.
17215 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17217 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17218 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17219 using the passed key: if the passed key was a private key the result
17220 of X509_print(), for example, would be to print out all the private key
17225 * des_quad_cksum() byte order bug fix.
17226 *Ulf Möller, using the problem description in krb4-0.9.7, where
17227 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17229 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17232 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17234 * For easily testing in shell scripts whether some command
17235 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17236 returns with exit code 0 iff no command of the given name is available.
17237 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17238 the output goes to stdout and nothing is printed to stderr.
17239 Additional arguments are always ignored.
17241 Since for each cipher there is a command of the same name,
17242 the 'no-cipher' compilation switches can be tested this way.
17244 ('openssl no-XXX' is not able to detect pseudo-commands such
17245 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17249 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17253 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17254 is set; it will be thrown away anyway because each handshake creates
17256 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17257 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17258 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17259 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17263 * New s_client option -ign_eof: EOF at stdin is ignored, and
17264 'Q' and 'R' lose their special meanings (quit/renegotiate).
17265 This is part of what -quiet does; unlike -quiet, -ign_eof
17266 does not suppress any output.
17270 * Add compatibility options to the purpose and trust code. The
17271 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17272 accepts a certificate or CA, this was the previous behaviour,
17273 with all the associated security issues.
17275 X509_TRUST_COMPAT is the old trust behaviour: only and
17276 automatically trust self signed roots in certificate store. A
17277 new trust setting X509_TRUST_DEFAULT is used to specify that
17278 a purpose has no associated trust setting and it should instead
17279 use the value in the default purpose.
17283 * Fix the PKCS#8 DSA private key code so it decodes keys again
17284 and fix a memory leak.
17288 * In util/mkerr.pl (which implements 'make errors'), preserve
17289 reason strings from the previous version of the .c file, as
17290 the default to have only downcase letters (and digits) in
17291 automatically generated reasons codes is not always appropriate.
17295 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17296 using strerror. Previously, ERR_reason_error_string() returned
17297 library names as reason strings for SYSerr; but SYSerr is a special
17298 case where small numbers are errno values, not library numbers.
17302 * Add '-dsaparam' option to 'openssl dhparam' application. This
17303 converts DSA parameters into DH parameters. (When creating parameters,
17304 DSA_generate_parameters is used.)
17308 * Include 'length' (recommended exponent length) in C code generated
17309 by 'openssl dhparam -C'.
17313 * The second argument to set_label in perlasm was already being used
17314 so couldn't be used as a "file scope" flag. Moved to third argument
17319 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17320 instead of RAND_bytes for encryption IVs and salts.
17324 * Include RAND_status() into RAND_METHOD instead of implementing
17325 it only for md_rand.c Otherwise replacing the PRNG by calling
17326 RAND_set_rand_method would be impossible.
17330 * Don't let DSA_generate_key() enter an infinite loop if the random
17331 number generation fails.
17335 * New 'rand' application for creating pseudo-random output.
17339 * Added configuration support for Linux/IA64
17341 *Rolf Haberrecker <rolf@suse.de>*
17343 * Assembler module support for Mingw32.
17347 * Shared library support for HPUX (in shlib/).
17349 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17351 * Shared library support for Solaris gcc.
17353 *Lutz Behnke <behnke@trustcenter.de>*
17355 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17357 * PKCS7_encrypt() was adding text MIME headers twice because they
17358 were added manually and by SMIME_crlf_copy().
17362 * In bntest.c don't call BN_rand with zero bits argument.
17364 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17366 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17367 case was implemented. This caused BN_div_recp() to fail occasionally.
17371 * Add an optional second argument to the set_label() in the perl
17372 assembly language builder. If this argument exists and is set
17373 to 1 it signals that the assembler should use a symbol whose
17374 scope is the entire file, not just the current function. This
17375 is needed with MASM which uses the format label:: for this scope.
17377 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17379 * Change the ASN1 types so they are typedefs by default. Before
17380 almost all types were #define'd to ASN1_STRING which was causing
17381 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17386 * Change names of new functions to the new get1/get0 naming
17387 convention: After 'get1', the caller owns a reference count
17388 and has to call `..._free`; 'get0' returns a pointer to some
17389 data structure without incrementing reference counters.
17390 (Some of the existing 'get' functions increment a reference
17391 counter, some don't.)
17392 Similarly, 'set1' and 'add1' functions increase reference
17393 counters or duplicate objects.
17397 * Allow for the possibility of temp RSA key generation failure:
17398 the code used to assume it always worked and crashed on failure.
17402 * Fix potential buffer overrun problem in BIO_printf().
17403 *Ulf Möller, using public domain code by Patrick Powell; problem
17404 pointed out by David Sacerdote <das33@cornell.edu>*
17406 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17407 RAND_egd() and RAND_status(). In the command line application,
17408 the EGD socket can be specified like a seed file using RANDFILE
17413 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17414 Some CAs (e.g. Verisign) distribute certificates in this form.
17418 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17419 list to exclude them. This means that no special compilation option
17420 is needed to use anonymous DH: it just needs to be included in the
17425 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17426 EVP_MD_type. The old functionality is available in a new macro called
17427 EVP_MD_md(). Change code that uses it and update docs.
17431 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17432 where the `void *` argument is replaced by a function pointer argument.
17433 Previously `void *` was abused to point to functions, which works on
17434 many platforms, but is not correct. As these functions are usually
17435 called by macros defined in OpenSSL header files, most source code
17436 should work without changes.
17440 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17441 sections with information on -D... compiler switches used for
17442 compiling the library so that applications can see them. To enable
17443 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17444 must be defined. E.g.,
17445 #define OPENSSL_ALGORITHM_DEFINES
17446 #include <openssl/opensslconf.h>
17447 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17449 *Richard Levitte, Ulf and Bodo Möller*
17451 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17456 * Change the 'other' type in certificate aux info to a STACK_OF
17457 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17458 the required ASN1 format: arbitrary types determined by an OID.
17462 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17463 argument to 'req'. This is not because the function is newer or
17464 better than others it just uses the work 'NEW' in the certificate
17465 request header lines. Some software needs this.
17469 * Reorganise password command line arguments: now passwords can be
17470 obtained from various sources. Delete the PEM_cb function and make
17471 it the default behaviour: i.e. if the callback is NULL and the
17472 usrdata argument is not NULL interpret it as a null terminated pass
17473 phrase. If usrdata and the callback are NULL then the pass phrase
17474 is prompted for as usual.
17478 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17479 the support is automatically enabled. The resulting binaries will
17480 autodetect the card and use it if present.
17482 *Ben Laurie and Compaq Inc.*
17484 * Work around for Netscape hang bug. This sends certificate request
17485 and server done in one record. Since this is perfectly legal in the
17486 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17487 the bugs/SSLv3 entry for more info.
17491 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17495 * Add -rand argument to smime and pkcs12 applications and read/write
17500 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17504 * Add command line password options to the remaining applications.
17508 * Bug fix for BN_div_recp() for numerators with an even number of
17513 * More tests in bntest.c, and changed test_bn output.
17517 * ./config recognizes MacOS X now.
17521 * Bug fix for BN_div() when the first words of num and divisor are
17522 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17526 * Add support for various broken PKCS#8 formats, and command line
17527 options to produce them.
17531 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17532 get temporary BIGNUMs from a BN_CTX.
17536 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17541 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17542 include a #define from the old name to the new. The original intent
17543 was that statically linked binaries could for example just call
17544 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17545 link with digests. This never worked because SSLeay_add_all_digests()
17546 and SSLeay_add_all_ciphers() were in the same source file so calling
17547 one would link with the other. They are now in separate source files.
17551 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17555 * Use a less unusual form of the Miller-Rabin primality test (it used
17556 a binary algorithm for exponentiation integrated into the Miller-Rabin
17557 loop, our standard modexp algorithms are faster).
17561 * Support for the EBCDIC character set completed.
17563 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17565 * Source code cleanups: use const where appropriate, eliminate casts,
17566 use `void *` instead of `char *` in lhash.
17570 * Bugfix: ssl3_send_server_key_exchange was not restartable
17571 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17572 this the server could overwrite ephemeral keys that the client
17577 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17578 using 50 iterations of the Rabin-Miller test.
17580 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17581 iterations of the Rabin-Miller test as required by the appendix
17582 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17583 As BN_is_prime_fasttest includes trial division, DSA parameter
17584 generation becomes much faster.
17586 This implies a change for the callback functions in DSA_is_prime
17587 and DSA_generate_parameters: The callback function is called once
17588 for each positive witness in the Rabin-Miller test, not just
17589 occasionally in the inner loop; and the parameters to the
17590 callback function now provide an iteration count for the outer
17591 loop rather than for the current invocation of the inner loop.
17592 DSA_generate_parameters additionally can call the callback
17593 function with an 'iteration count' of -1, meaning that a
17594 candidate has passed the trial division test (when q is generated
17595 from an application-provided seed, trial division is skipped).
17599 * New function BN_is_prime_fasttest that optionally does trial
17600 division before starting the Rabin-Miller test and has
17601 an additional BN_CTX * argument (whereas BN_is_prime always
17602 has to allocate at least one BN_CTX).
17603 'callback(1, -1, cb_arg)' is called when a number has passed the
17604 trial division stage.
17608 * Fix for bug in CRL encoding. The validity dates weren't being handled
17613 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17617 * New function BN_pseudo_rand().
17621 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17622 bignum version of BN_from_montgomery() with the working code from
17623 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17628 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17629 made it impossible to use the same SSL_SESSION data structure in
17630 SSL2 clients in multiple threads.
17634 * The return value of RAND_load_file() no longer counts bytes obtained
17635 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17636 to seed the PRNG (previously an explicit byte count was required).
17638 *Ulf Möller, Bodo Möller*
17640 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17641 used `char *` instead of `void *` and had casts all over the place.
17645 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17649 * Retain source code compatibility for BN_prime_checks macro:
17650 BN_is_prime(..., BN_prime_checks, ...) now uses
17651 BN_prime_checks_for_size to determine the appropriate number of
17652 Rabin-Miller iterations.
17656 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17657 DH_CHECK_P_NOT_SAFE_PRIME.
17658 (Check if this is true? OpenPGP calls them "strong".)
17662 * Merge the functionality of "dh" and "gendh" programs into a new program
17663 "dhparam". The old programs are retained for now but will handle DH keys
17664 (instead of parameters) in future.
17668 * Make the ciphers, s_server and s_client programs check the return values
17669 when a new cipher list is set.
17673 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17674 ciphers. Before when the 56bit ciphers were enabled the sorting was
17677 The syntax for the cipher sorting has been extended to support sorting by
17678 cipher-strength (using the strength_bits hard coded in the tables).
17679 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17681 Fix a bug in the cipher-command parser: when supplying a cipher command
17682 string with an "undefined" symbol (neither command nor alphanumeric
17683 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17684 an error is flagged.
17686 Due to the strength-sorting extension, the code of the
17687 ssl_create_cipher_list() function was completely rearranged. I hope that
17688 the readability was also increased :-)
17690 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17692 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17693 for the first serial number and places 2 in the serial number file. This
17694 avoids problems when the root CA is created with serial number zero and
17695 the first user certificate has the same issuer name and serial number
17700 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17701 the new code. Add documentation for this stuff.
17705 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17706 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17707 structures and behave in an analogous way to the X509v3 functions:
17708 they shouldn't be called directly but wrapper functions should be used
17711 So we also now have some wrapper functions that call the X509at functions
17712 when passed certificate requests. (TO DO: similar things can be done with
17713 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17714 things. Some of these need some d2i or i2d and print functionality
17715 because they handle more complex structures.)
17719 * Add missing #ifndefs that caused missing symbols when building libssl
17720 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17721 NO_RSA in `ssl/s2*.c`.
17723 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17725 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17726 has a return value which indicates the quality of the random data
17727 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17728 error queue. New function RAND_pseudo_bytes() generates output that is
17729 guaranteed to be unique but not unpredictable. RAND_add is like
17730 RAND_seed, but takes an extra argument for an entropy estimate
17731 (RAND_seed always assumes full entropy).
17735 * Do more iterations of Rabin-Miller probable prime test (specifically,
17736 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17737 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17738 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17739 false-positive rate of at most 2^-80 for random input.
17743 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17747 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17748 in the 0.9.5 release), this returns the chain
17749 from an X509_CTX structure with a dup of the stack and all
17750 the X509 reference counts upped: so the stack will exist
17751 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17754 Also make SSL_SESSION_print() print out the verify return
17759 * Add manpage for the pkcs12 command. Also change the default
17760 behaviour so MAC iteration counts are used unless the new
17761 -nomaciter option is used. This improves file security and
17762 only older versions of MSIE (4.0 for example) need it.
17766 * Honor the no-xxx Configure options when creating .DEF files.
17770 * Add PKCS#10 attributes to field table: challengePassword,
17771 unstructuredName and unstructuredAddress. These are taken from
17772 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17773 international characters are used.
17775 More changes to X509_ATTRIBUTE code: allow the setting of types
17776 based on strings. Remove the 'loc' parameter when adding
17777 attributes because these will be a SET OF encoding which is sorted
17782 * Initial changes to the 'req' utility to allow request generation
17783 automation. This will allow an application to just generate a template
17784 file containing all the field values and have req construct the
17787 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17788 used all over the place including certificate requests and PKCS#7
17789 structures. They are currently handled manually where necessary with
17790 some primitive wrappers for PKCS#7. The new functions behave in a
17791 manner analogous to the X509 extension functions: they allow
17792 attributes to be looked up by NID and added.
17794 Later something similar to the X509V3 code would be desirable to
17795 automatically handle the encoding, decoding and printing of the
17796 more complex types. The string types like challengePassword can
17797 be handled by the string table functions.
17799 Also modified the multi byte string table handling. Now there is
17800 a 'global mask' which masks out certain types. The table itself
17801 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17802 is useful when for example there is only one permissible type
17803 (as in countryName) and using the mask might result in no valid
17808 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17809 SSL_get_peer_finished to allow applications to obtain the latest
17810 Finished messages sent to the peer or expected from the peer,
17811 respectively. (SSL_get_peer_finished is usually the Finished message
17812 actually received from the peer, otherwise the protocol will be aborted.)
17814 As the Finished message are message digests of the complete handshake
17815 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17816 be used for external authentication procedures when the authentication
17817 provided by SSL/TLS is not desired or is not enough.
17821 * Enhanced support for Alpha Linux is added. Now ./config checks if
17822 the host supports BWX extension and if Compaq C is present on the
17823 $PATH. Just exploiting of the BWX extension results in 20-30%
17824 performance kick for some algorithms, e.g. DES and RC4 to mention
17825 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17830 * Add support for MS "fast SGC". This is arguably a violation of the
17831 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17832 weak crypto and after checking the certificate is SGC a second one
17833 with strong crypto. MS SGC stops the first handshake after receiving
17834 the server certificate message and sends a second client hello. Since
17835 a server will typically do all the time consuming operations before
17836 expecting any further messages from the client (server key exchange
17837 is the most expensive) there is little difference between the two.
17839 To get OpenSSL to support MS SGC we have to permit a second client
17840 hello message after we have sent server done. In addition we have to
17841 reset the MAC if we do get this second client hello.
17845 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17846 if a DER encoded private key is RSA or DSA traditional format. Changed
17847 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17848 format DER encoded private key. Newer code should use PKCS#8 format which
17849 has the key type encoded in the ASN1 structure. Added DER private key
17850 support to pkcs8 application.
17854 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17855 ciphersuites has been selected (as required by the SSL 3/TLS 1
17856 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17857 is set, we interpret this as a request to violate the specification
17858 (the worst that can happen is a handshake failure, and 'correct'
17859 behaviour would result in a handshake failure anyway).
17863 * In SSL_CTX_add_session, take into account that there might be multiple
17864 SSL_SESSION structures with the same session ID (e.g. when two threads
17865 concurrently obtain them from an external cache).
17866 The internal cache can handle only one SSL_SESSION with a given ID,
17867 so if there's a conflict, we now throw out the old one to achieve
17872 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17873 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17874 some routines that use cipher OIDs: some ciphers do not have OIDs
17875 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17880 * Simplify the trust setting structure and code. Now we just have
17881 two sequences of OIDs for trusted and rejected settings. These will
17882 typically have values the same as the extended key usage extension
17883 and any application specific purposes.
17885 The trust checking code now has a default behaviour: it will just
17886 check for an object with the same NID as the passed id. Functions can
17887 be provided to override either the default behaviour or the behaviour
17888 for a given id. SSL client, server and email already have functions
17889 in place for compatibility: they check the NID and also return "trusted"
17890 if the certificate is self signed.
17894 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17895 traditional format into an EVP_PKEY structure.
17899 * Add a password callback function PEM_cb() which either prompts for
17900 a password if usr_data is NULL or otherwise assumes it is a null
17901 terminated password. Allow passwords to be passed on command line
17902 environment or config files in a few more utilities.
17906 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17907 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17908 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17909 Update documentation.
17913 * Support for ASN1 "NULL" type. This could be handled before by using
17914 ASN1_TYPE but there wasn't any function that would try to read a NULL
17915 and produce an error if it couldn't. For compatibility we also have
17916 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17917 don't allocate anything because they don't need to.
17921 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17924 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17926 * Rebuild of the memory allocation routines used by OpenSSL code and
17927 possibly others as well. The purpose is to make an interface that
17928 provide hooks so anyone can build a separate set of allocation and
17929 deallocation routines to be used by OpenSSL, for example memory
17930 pool implementations, or something else, which was previously hard
17931 since Malloc(), Realloc() and Free() were defined as macros having
17932 the values malloc, realloc and free, respectively (except for Win32
17933 compilations). The same is provided for memory debugging code.
17934 OpenSSL already comes with functionality to find memory leaks, but
17935 this gives people a chance to debug other memory problems.
17937 With these changes, a new set of functions and macros have appeared:
17939 CRYPTO_set_mem_debug_functions() [F]
17940 CRYPTO_get_mem_debug_functions() [F]
17941 CRYPTO_dbg_set_options() [F]
17942 CRYPTO_dbg_get_options() [F]
17943 CRYPTO_malloc_debug_init() [M]
17945 The memory debug functions are NULL by default, unless the library
17946 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17947 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17948 gives the standard debugging functions that come with OpenSSL) or
17949 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17950 provided by the library user) must be used. When the standard
17951 debugging functions are used, CRYPTO_dbg_set_options can be used to
17952 request additional information:
17953 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17954 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17956 Also, things like CRYPTO_set_mem_functions will always give the
17957 expected result (the new set of functions is used for allocation
17958 and deallocation) at all times, regardless of platform and compiler
17961 To finish it up, some functions that were never use in any other
17962 way than through macros have a new API and new semantic:
17964 CRYPTO_dbg_malloc()
17965 CRYPTO_dbg_realloc()
17968 All macros of value have retained their old syntax.
17970 *Richard Levitte and Bodo Moeller*
17972 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17973 ordering of SMIMECapabilities wasn't in "strength order" and there
17974 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17979 * Some ASN1 types with illegal zero length encoding (INTEGER,
17980 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17982 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17984 * Merge in my S/MIME library for OpenSSL. This provides a simple
17985 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17986 functionality to handle multipart/signed properly) and a utility
17987 called 'smime' to call all this stuff. This is based on code I
17988 originally wrote for Celo who have kindly allowed it to be
17989 included in OpenSSL.
17993 * Add variants des_set_key_checked and des_set_key_unchecked of
17994 des_set_key (aka des_key_sched). Global variable des_check_key
17995 decides which of these is called by des_set_key; this way
17996 des_check_key behaves as it always did, but applications and
17997 the library itself, which was buggy for des_check_key == 1,
17998 have a cleaner way to pick the version they need.
18002 * New function PKCS12_newpass() which changes the password of a
18007 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
18008 dynamic mix. In both cases the ids can be used as an index into the
18009 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
18010 functions so they accept a list of the field values and the
18011 application doesn't need to directly manipulate the X509_TRUST
18016 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
18021 * Modify the way the V3 extension code looks up extensions. This now
18022 works in a similar way to the object code: we have some "standard"
18023 extensions in a static table which is searched with OBJ_bsearch()
18024 and the application can add dynamic ones if needed. The file
18025 crypto/x509v3/ext_dat.h now has the info: this file needs to be
18026 updated whenever a new extension is added to the core code and kept
18027 in ext_nid order. There is a simple program 'tabtest.c' which checks
18028 this. New extensions are not added too often so this file can readily
18029 be maintained manually.
18031 There are two big advantages in doing things this way. The extensions
18032 can be looked up immediately and no longer need to be "added" using
18033 X509V3_add_standard_extensions(): this function now does nothing.
18034 Side note: I get *lots* of email saying the extension code doesn't
18035 work because people forget to call this function.
18036 Also no dynamic allocation is done unless new extensions are added:
18037 so if we don't add custom extensions there is no need to call
18038 X509V3_EXT_cleanup().
18042 * Modify enc utility's salting as follows: make salting the default. Add a
18043 magic header, so unsalted files fail gracefully instead of just decrypting
18044 to garbage. This is because not salting is a big security hole, so people
18045 should be discouraged from doing it.
18049 * Fixes and enhancements to the 'x509' utility. It allowed a message
18050 digest to be passed on the command line but it only used this
18051 parameter when signing a certificate. Modified so all relevant
18052 operations are affected by the digest parameter including the
18053 -fingerprint and -x509toreq options. Also -x509toreq choked if a
18054 DSA key was used because it didn't fix the digest.
18058 * Initial certificate chain verify code. Currently tests the untrusted
18059 certificates for consistency with the verify purpose (which is set
18060 when the X509_STORE_CTX structure is set up) and checks the pathlength.
18062 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
18063 this is because it will reject chains with invalid extensions whereas
18064 every previous version of OpenSSL and SSLeay made no checks at all.
18066 Trust code: checks the root CA for the relevant trust settings. Trust
18067 settings have an initial value consistent with the verify purpose: e.g.
18068 if the verify purpose is for SSL client use it expects the CA to be
18069 trusted for SSL client use. However the default value can be changed to
18070 permit custom trust settings: one example of this would be to only trust
18071 certificates from a specific "secure" set of CAs.
18073 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
18074 which should be used for version portability: especially since the
18075 verify structure is likely to change more often now.
18077 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
18078 to set them. If not set then assume SSL clients will verify SSL servers
18081 Two new options to the verify program: -untrusted allows a set of
18082 untrusted certificates to be passed in and -purpose which sets the
18083 intended purpose of the certificate. If a purpose is set then the
18084 new chain verify code is used to check extension consistency.
18088 * Support for the authority information access extension.
18092 * Modify RSA and DSA PEM read routines to transparently handle
18093 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18094 public keys in a format compatible with certificate
18095 SubjectPublicKeyInfo structures. Unfortunately there were already
18096 functions called *_PublicKey_* which used various odd formats so
18097 these are retained for compatibility: however the DSA variants were
18098 never in a public release so they have been deleted. Changed dsa/rsa
18099 utilities to handle the new format: note no releases ever handled public
18100 keys so we should be OK.
18102 The primary motivation for this change is to avoid the same fiasco
18103 that dogs private keys: there are several incompatible private key
18104 formats some of which are standard and some OpenSSL specific and
18105 require various evil hacks to allow partial transparent handling and
18106 even then it doesn't work with DER formats. Given the option anything
18107 other than PKCS#8 should be dumped: but the other formats have to
18108 stay in the name of compatibility.
18110 With public keys and the benefit of hindsight one standard format
18111 is used which works with EVP_PKEY, RSA or DSA structures: though
18112 it clearly returns an error if you try to read the wrong kind of key.
18114 Added a -pubkey option to the 'x509' utility to output the public key.
18115 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18116 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18117 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18118 that do the same as the `EVP_PKEY_assign_*()` except they up the
18119 reference count of the added key (they don't "swallow" the
18124 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18125 CRLs would fail if the file contained no certificates or no CRLs:
18126 added a new function to read in both types and return the number
18127 read: this means that if none are read it will be an error. The
18128 DER versions of the certificate and CRL reader would always fail
18129 because it isn't possible to mix certificates and CRLs in DER format
18130 without choking one or the other routine. Changed this to just read
18131 a certificate: this is the best we can do. Also modified the code
18132 in `apps/verify.c` to take notice of return codes: it was previously
18133 attempting to read in certificates from NULL pointers and ignoring
18134 any errors: this is one reason why the cert and CRL reader seemed
18135 to work. It doesn't check return codes from the default certificate
18136 routines: these may well fail if the certificates aren't installed.
18140 * Code to support otherName option in GeneralName.
18144 * First update to verify code. Change the verify utility
18145 so it warns if it is passed a self signed certificate:
18146 for consistency with the normal behaviour. X509_verify
18147 has been modified to it will now verify a self signed
18148 certificate if *exactly* the same certificate appears
18149 in the store: it was previously impossible to trust a
18150 single self signed certificate. This means that:
18151 openssl verify ss.pem
18152 now gives a warning about a self signed certificate but
18153 openssl verify -CAfile ss.pem ss.pem
18158 * For servers, store verify_result in SSL_SESSION data structure
18159 (and add it to external session representation).
18160 This is needed when client certificate verifications fails,
18161 but an application-provided verification callback (set by
18162 SSL_CTX_set_cert_verify_callback) allows accepting the session
18163 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18164 but returns 1): When the session is reused, we have to set
18165 ssl->verify_result to the appropriate error code to avoid
18168 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18170 * Fix a bug in the new PKCS#7 code: it didn't consider the
18171 case in PKCS7_dataInit() where the signed PKCS7 structure
18172 didn't contain any existing data because it was being created.
18174 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18176 * Add a salt to the key derivation routines in enc.c. This
18177 forms the first 8 bytes of the encrypted file. Also add a
18178 -S option to allow a salt to be input on the command line.
18182 * New function X509_cmp(). Oddly enough there wasn't a function
18183 to compare two certificates. We do this by working out the SHA1
18184 hash and comparing that. X509_cmp() will be needed by the trust
18189 * SSL_get1_session() is like SSL_get_session(), but increments
18190 the reference count in the SSL_SESSION returned.
18192 *Geoff Thorpe <geoff@eu.c2.net>*
18194 * Fix for 'req': it was adding a null to request attributes.
18195 Also change the X509_LOOKUP and X509_INFO code to handle
18196 certificate auxiliary information.
18200 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18205 * Add the possibility to add extra information to the memory leak
18206 detecting output, to form tracebacks, showing from where each
18207 allocation was originated: CRYPTO_push_info("constant string") adds
18208 the string plus current file name and line number to a per-thread
18209 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18210 is like calling CYRPTO_pop_info() until the stack is empty.
18211 Also updated memory leak detection code to be multi-thread-safe.
18215 * Add options -text and -noout to pkcs7 utility and delete the
18216 encryption options which never did anything. Update docs.
18220 * Add options to some of the utilities to allow the pass phrase
18221 to be included on either the command line (not recommended on
18222 OSes like Unix) or read from the environment. Update the
18223 manpages and fix a few bugs.
18227 * Add a few manpages for some of the openssl commands.
18231 * Fix the -revoke option in ca. It was freeing up memory twice,
18232 leaking and not finding already revoked certificates.
18236 * Extensive changes to support certificate auxiliary information.
18237 This involves the use of X509_CERT_AUX structure and X509_AUX
18238 functions. An X509_AUX function such as PEM_read_X509_AUX()
18239 can still read in a certificate file in the usual way but it
18240 will also read in any additional "auxiliary information". By
18241 doing things this way a fair degree of compatibility can be
18242 retained: existing certificates can have this information added
18243 using the new 'x509' options.
18245 Current auxiliary information includes an "alias" and some trust
18246 settings. The trust settings will ultimately be used in enhanced
18247 certificate chain verification routines: currently a certificate
18248 can only be trusted if it is self signed and then it is trusted
18253 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18254 The problem was that one of the replacement routines had not been working
18255 since SSLeay releases. For now the offending routine has been replaced
18256 with non-optimised assembler. Even so, this now gives around 95%
18257 performance improvement for 1024 bit RSA signs.
18261 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18262 handling. Most clients have the effective key size in bits equal to
18263 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18264 A few however don't do this and instead use the size of the decrypted key
18265 to determine the RC2 key length and the AlgorithmIdentifier to determine
18266 the effective key length. In this case the effective key length can still
18267 be 40 bits but the key length can be 168 bits for example. This is fixed
18268 by manually forcing an RC2 key into the EVP_PKEY structure because the
18269 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18270 the key length and effective key length are equal.
18274 * Add a bunch of functions that should simplify the creation of
18275 X509_NAME structures. Now you should be able to do:
18276 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18277 and have it automatically work out the correct field type and fill in
18278 the structures. The more adventurous can try:
18279 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18280 and it will (hopefully) work out the correct multibyte encoding.
18284 * Change the 'req' utility to use the new field handling and multibyte
18285 copy routines. Before the DN field creation was handled in an ad hoc
18286 way in req, ca, and x509 which was rather broken and didn't support
18287 BMPStrings or UTF8Strings. Since some software doesn't implement
18288 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18289 using the dirstring_type option. See the new comment in the default
18290 openssl.cnf for more info.
18294 * Make crypto/rand/md_rand.c more robust:
18295 - Assure unique random numbers after fork().
18296 - Make sure that concurrent threads access the global counter and
18297 md serializably so that we never lose entropy in them
18298 or use exactly the same state in multiple threads.
18299 Access to the large state is not always serializable because
18300 the additional locking could be a performance killer, and
18301 md should be large enough anyway.
18305 * New file `apps/app_rand.c` with commonly needed functionality
18306 for handling the random seed file.
18308 Use the random seed file in some applications that previously did not:
18310 dsaparam -genkey (which also ignored its '-rand' option),
18313 x509 (when signing).
18314 Except on systems with /dev/urandom, it is crucial to have a random
18315 seed file at least for key creation, DSA signing, and for DH exchanges;
18316 for RSA signatures we could do without one.
18318 gendh and gendsa (unlike genrsa) used to read only the first byte
18319 of each file listed in the '-rand' option. The function as previously
18320 found in genrsa is now in app_rand.c and is used by all programs
18321 that support '-rand'.
18325 * In RAND_write_file, use mode 0600 for creating files;
18326 don't just chmod when it may be too late.
18330 * Report an error from X509_STORE_load_locations
18331 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18335 * New function ASN1_mbstring_copy() this copies a string in either
18336 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18337 into an ASN1_STRING type. A mask of permissible types is passed
18338 and it chooses the "minimal" type to use or an error if not type
18343 * Add function equivalents to the various macros in asn1.h. The old
18344 macros are retained with an `M_` prefix. Code inside the library can
18345 use the `M_` macros. External code (including the openssl utility)
18346 should *NOT* in order to be "shared library friendly".
18350 * Add various functions that can check a certificate's extensions
18351 to see if it usable for various purposes such as SSL client,
18352 server or S/MIME and CAs of these types. This is currently
18353 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18354 verification. Also added a -purpose flag to x509 utility to
18355 print out all the purposes.
18359 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18364 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18365 for, obtain and decode and extension and obtain its critical flag.
18366 This allows all the necessary extension code to be handled in a
18367 single function call.
18371 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18372 platforms. See crypto/rc4/rc4_enc.c for further details.
18376 * New -noout option to asn1parse. This causes no output to be produced
18377 its main use is when combined with -strparse and -out to extract data
18378 from a file (which may not be in ASN.1 format).
18382 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18383 when producing the local key id.
18385 *Richard Levitte <levitte@stacken.kth.se>*
18387 * New option -dhparam in s_server. This allows a DH parameter file to be
18388 stated explicitly. If it is not stated then it tries the first server
18389 certificate file. The previous behaviour hard coded the filename
18394 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18395 a public key to be input or output. For example:
18396 openssl rsa -in key.pem -pubout -out pubkey.pem
18397 Also added necessary DSA public key functions to handle this.
18401 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18402 in the message. This was handled by allowing
18403 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18405 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18407 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18408 to the end of the strings whereas this didn't. This would cause problems
18409 if strings read with d2i_ASN1_bytes() were later modified.
18411 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18413 * Fix for base64 decode bug. When a base64 bio reads only one line of
18414 data and it contains EOF it will end up returning an error. This is
18415 caused by input 46 bytes long. The cause is due to the way base64
18416 BIOs find the start of base64 encoded data. They do this by trying a
18417 trial decode on each line until they find one that works. When they
18418 do a flag is set and it starts again knowing it can pass all the
18419 data directly through the decoder. Unfortunately it doesn't reset
18420 the context it uses. This means that if EOF is reached an attempt
18421 is made to pass two EOFs through the context and this causes the
18422 resulting error. This can also cause other problems as well. As is
18423 usual with these problems it takes *ages* to find and the fix is
18424 trivial: move one line.
18426 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18428 * Ugly workaround to get s_client and s_server working under Windows. The
18429 old code wouldn't work because it needed to select() on sockets and the
18430 tty (for keypresses and to see if data could be written). Win32 only
18431 supports select() on sockets so we select() with a 1s timeout on the
18432 sockets and then see if any characters are waiting to be read, if none
18433 are present then we retry, we also assume we can always write data to
18434 the tty. This isn't nice because the code then blocks until we've
18435 received a complete line of data and it is effectively polling the
18436 keyboard at 1s intervals: however it's quite a bit better than not
18437 working at all :-) A dedicated Windows application might handle this
18438 with an event loop for example.
18442 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18443 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18444 will be called when RSA_sign() and RSA_verify() are used. This is useful
18445 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18446 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18447 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18448 This necessitated the support of an extra signature type NID_md5_sha1
18449 for SSL signatures and modifications to the SSL library to use it instead
18450 of calling RSA_public_decrypt() and RSA_private_encrypt().
18454 * Add new -verify -CAfile and -CApath options to the crl program, these
18455 will lookup a CRL issuers certificate and verify the signature in a
18456 similar way to the verify program. Tidy up the crl program so it
18457 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18458 less strict. It will now permit CRL extensions even if it is not
18459 a V2 CRL: this will allow it to tolerate some broken CRLs.
18463 * Initialize all non-automatic variables each time one of the openssl
18464 sub-programs is started (this is necessary as they may be started
18465 multiple times from the "OpenSSL>" prompt).
18467 *Lennart Bang, Bodo Moeller*
18469 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18470 removing all other RSA functionality (this is what NO_RSA does). This
18471 is so (for example) those in the US can disable those operations covered
18472 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18477 * Non-copying interface to BIO pairs.
18478 (still largely untested)
18482 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18483 ASCII string. This was handled independently in various places before.
18487 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18488 UTF8 strings a character at a time.
18492 * Use client_version from client hello to select the protocol
18493 (s23_srvr.c) and for RSA client key exchange verification
18494 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18498 * Add various utility functions to handle SPKACs, these were previously
18499 handled by poking round in the structure internals. Added new function
18500 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18501 print, verify and generate SPKACs. Based on an original idea from
18502 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18506 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18510 * Allow the config file extension section to be overwritten on the
18511 command line. Based on an original idea from Massimiliano Pala
18512 <madwolf@comune.modena.it>. The new option is called -extensions
18513 and can be applied to ca, req and x509. Also -reqexts to override
18514 the request extensions in req and -crlexts to override the crl extensions
18519 * Add new feature to the SPKAC handling in ca. Now you can include
18520 the same field multiple times by preceding it by "XXXX." for example:
18523 this is the same syntax as used in the req config file.
18527 * Allow certificate extensions to be added to certificate requests. These
18528 are specified in a 'req_extensions' option of the req section of the
18529 config file. They can be printed out with the -text option to req but
18530 are otherwise ignored at present.
18534 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18535 data read consists of only the final block it would not decrypted because
18536 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18537 A misplaced 'break' also meant the decrypted final block might not be
18538 copied until the next read.
18542 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18543 a few extra parameters to the DH structure: these will be useful if
18544 for example we want the value of 'q' or implement X9.42 DH.
18548 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18549 provides hooks that allow the default DSA functions or functions on a
18550 "per key" basis to be replaced. This allows hardware acceleration and
18551 hardware key storage to be handled without major modification to the
18552 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18553 associated functions.
18557 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18558 as "read only": it can't be written to and the buffer it points to will
18559 not be freed. Reading from a read only BIO is much more efficient than
18560 a normal memory BIO. This was added because there are several times when
18561 an area of memory needs to be read from a BIO. The previous method was
18562 to create a memory BIO and write the data to it, this results in two
18563 copies of the data and an O(n^2) reading algorithm. There is a new
18564 function BIO_new_mem_buf() which creates a read only memory BIO from
18565 an area of memory. Also modified the PKCS#7 routines to use read only
18570 * Bugfix: ssl23_get_client_hello did not work properly when called in
18571 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18572 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18573 but a retry condition occurred while trying to read the rest.
18577 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18578 NID_pkcs7_encrypted by default: this was wrong since this should almost
18579 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18580 the encrypted data type: this is a more sensible place to put it and it
18581 allows the PKCS#12 code to be tidied up that duplicated this
18586 * Changed obj_dat.pl script so it takes its input and output files on
18587 the command line. This should avoid shell escape redirection problems
18592 * Initial support for certificate extension requests, these are included
18593 in things like Xenroll certificate requests. Included functions to allow
18594 extensions to be obtained and added.
18598 * -crlf option to s_client and s_server for sending newlines as
18599 CRLF (as required by many protocols).
18603 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18605 * Install libRSAglue.a when OpenSSL is built with RSAref.
18607 *Ralf S. Engelschall*
18609 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18611 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18613 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18618 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18619 DH parameters/keys (q is lost during that conversion, but the resulting
18620 DH parameters contain its length).
18622 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18623 much faster than DH_generate_parameters (which creates parameters
18624 where `p = 2*q + 1`), and also the smaller q makes DH computations
18625 much more efficient (160-bit exponentiation instead of 1024-bit
18626 exponentiation); so this provides a convenient way to support DHE
18627 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18628 utter importance to use
18629 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18631 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18632 when such DH parameters are used, because otherwise small subgroup
18633 attacks may become possible!
18637 * Avoid memory leak in i2d_DHparams.
18641 * Allow the -k option to be used more than once in the enc program:
18642 this allows the same encrypted message to be read by multiple recipients.
18646 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18647 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18648 it will always use the numerical form of the OID, even if it has a short
18653 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18654 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18655 otherwise bn_mod_exp was called. In the case of hardware keys for example
18656 no private key components need be present and it might store extra data
18657 in the RSA structure, which cannot be accessed from bn_mod_exp.
18658 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18659 private key operations.
18663 * Added support for SPARC Linux.
18667 * pem_password_cb function type incompatibly changed from
18668 typedef int pem_password_cb(char *buf, int size, int rwflag);
18670 ....(char *buf, int size, int rwflag, void *userdata);
18671 so that applications can pass data to their callbacks:
18672 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18673 additional void * argument, which is just handed through whenever
18674 the password callback is called.
18676 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18678 New function SSL_CTX_set_default_passwd_cb_userdata.
18680 Compatibility note: As many C implementations push function arguments
18681 onto the stack in reverse order, the new library version is likely to
18682 interoperate with programs that have been compiled with the old
18683 pem_password_cb definition (PEM_whatever takes some data that
18684 happens to be on the stack as its last argument, and the callback
18685 just ignores this garbage); but there is no guarantee whatsoever that
18688 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18689 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18690 problems not only on Windows, but also on some Unix platforms.
18691 To avoid problematic command lines, these definitions are now in an
18692 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18693 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18697 * MIPS III/IV assembler module is reimplemented.
18701 * More DES library cleanups: remove references to srand/rand and
18702 delete an unused file.
18706 * Add support for the free Netwide assembler (NASM) under Win32,
18707 since not many people have MASM (ml) and it can be hard to obtain.
18708 This is currently experimental but it seems to work OK and pass all
18709 the tests. Check out INSTALL.W32 for info.
18713 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18714 without temporary keys kept an extra copy of the server key,
18715 and connections with temporary keys did not free everything in case
18720 * New function RSA_check_key and new openssl rsa option -check
18721 for verifying the consistency of RSA keys.
18723 *Ulf Moeller, Bodo Moeller*
18725 * Various changes to make Win32 compile work:
18726 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18727 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18728 comparison" warnings.
18729 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18733 * Add a debugging option to PKCS#5 v2 key generation function: when
18734 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18735 derived keys are printed to stderr.
18739 * Copy the flags in ASN1_STRING_dup().
18741 *Roman E. Pavlov <pre@mo.msk.ru>*
18743 * The x509 application mishandled signing requests containing DSA
18744 keys when the signing key was also DSA and the parameters didn't match.
18746 It was supposed to omit the parameters when they matched the signing key:
18747 the verifying software was then supposed to automatically use the CA's
18748 parameters if they were absent from the end user certificate.
18750 Omitting parameters is no longer recommended. The test was also
18751 the wrong way round! This was probably due to unusual behaviour in
18752 EVP_cmp_parameters() which returns 1 if the parameters match.
18753 This meant that parameters were omitted when they *didn't* match and
18754 the certificate was useless. Certificates signed with 'ca' didn't have
18757 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18759 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18760 The interface is as follows:
18761 Applications can use
18762 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18763 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18764 "off" is now the default.
18765 The library internally uses
18766 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18767 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18768 to disable memory-checking temporarily.
18770 Some inconsistent states that previously were possible (and were
18771 even the default) are now avoided.
18773 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18774 with each memory chunk allocated; this is occasionally more helpful
18775 than just having a counter.
18777 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18779 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18784 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18785 which largely parallels "options", but is for changing API behaviour,
18786 whereas "options" are about protocol behaviour.
18787 Initial "mode" flags are:
18789 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18790 a single record has been written.
18791 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18792 retries use the same buffer location.
18793 (But all of the contents must be
18798 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18801 * Fix problems with no-hmac etc.
18803 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18805 * New functions RSA_get_default_method(), RSA_set_method() and
18806 RSA_get_method(). These allows replacement of RSA_METHODs without having
18807 to mess around with the internals of an RSA structure.
18811 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18812 Also really enable memory leak checks in openssl.c and in some
18815 *Chad C. Mulligan, Bodo Moeller*
18817 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18818 up the length of negative integers. This has now been simplified to just
18819 store the length when it is first determined and use it later, rather
18820 than trying to keep track of where data is copied and updating it to
18822 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18824 * Add a new function PKCS7_signatureVerify. This allows the verification
18825 of a PKCS#7 signature but with the signing certificate passed to the
18826 function itself. This contrasts with PKCS7_dataVerify which assumes the
18827 certificate is present in the PKCS#7 structure. This isn't always the
18828 case: certificates can be omitted from a PKCS#7 structure and be
18829 distributed by "out of band" means (such as a certificate database).
18833 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18834 function prototypes in pem.h, also change util/mkdef.pl to add the
18835 necessary function names.
18839 * mk1mf.pl (used by Windows builds) did not properly read the
18840 options set by Configure in the top level Makefile, and Configure
18841 was not even able to write more than one option correctly.
18842 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18846 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18847 file to be loaded from a BIO or FILE pointer. The BIO version will
18848 for example allow memory BIOs to contain config info.
18852 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18853 Whoever hopes to achieve shared-library compatibility across versions
18854 must use this, not the compile-time macro.
18855 (Exercise 0.9.4: Which is the minimum library version required by
18857 Note: All this applies only to multi-threaded programs, others don't
18862 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18863 through a BIO pair triggered the default case, i.e.
18864 SSLerr(...,SSL_R_UNKNOWN_STATE).
18868 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18869 can use the SSL library even if none of the specific BIOs is
18874 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18875 for the encoded length.
18877 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18879 * Add initial documentation of the X509V3 functions.
18883 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18884 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18885 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18886 secure PKCS#8 private key format with a high iteration count.
18890 * Fix determination of Perl interpreter: A perl or perl5
18891 *directory* in $PATH was also accepted as the interpreter.
18893 *Ralf S. Engelschall*
18895 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18896 wrong with it but it was very old and did things like calling
18897 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18898 unusual formatting.
18902 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18903 to use the new extension code.
18907 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18908 with macros. This should make it easier to change their form, add extra
18909 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18914 * Add to configuration table a new entry that can specify an alternative
18915 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18916 according to Mark Crispin <MRC@Panda.COM>.
18920 * DES CBC did not update the IV. Weird.
18924 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18925 Changing the behaviour of the former might break existing programs --
18926 where IV updating is needed, des_ncbc_encrypt can be used.
18929 * When bntest is run from "make test" it drives bc to check its
18930 calculations, as well as internally checking them. If an internal check
18931 fails, it needs to cause bc to give a non-zero result or make test carries
18932 on without noticing the failure. Fixed.
18936 * DES library cleanups.
18940 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18941 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18942 ciphers. NOTE: although the key derivation function has been verified
18943 against some published test vectors it has not been extensively tested
18944 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18949 * Instead of "mkdir -p", which is not fully portable, use new
18950 Perl script "util/mkdir-p.pl".
18954 * Rewrite the way password based encryption (PBE) is handled. It used to
18955 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18956 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18957 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18958 the 'parameter' field of the AlgorithmIdentifier is passed to the
18959 underlying key generation function so it must do its own ASN1 parsing.
18960 This has also changed the EVP_PBE_CipherInit() function which now has a
18961 'parameter' argument instead of literal salt and iteration count values
18962 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18966 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18967 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18968 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18969 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18970 value was just used as a "magic string" and not used directly its
18971 value doesn't matter.
18975 * Introduce some semblance of const correctness to BN. Shame C doesn't
18980 * "linux-sparc64" configuration (ultrapenguin).
18982 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18983 "linux-sparc" configuration.
18985 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18987 * config now generates no-xxx options for missing ciphers.
18991 * Support the EBCDIC character set (work in progress).
18992 File ebcdic.c not yet included because it has a different license.
18994 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18996 * Support BS2000/OSD-POSIX.
18998 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
19000 * Make callbacks for key generation use `void *` instead of `char *`.
19004 * Make S/MIME samples compile (not yet tested).
19008 * Additional typesafe stacks.
19012 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
19016 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
19018 * New configuration variant "sco5-gcc".
19020 * Updated some demos.
19022 *Sean O Riordain, Wade Scholine*
19024 * Add missing BIO_free at exit of pkcs12 application.
19028 * Fix memory leak in conf.c.
19032 * Updates for Win32 to assembler version of MD5.
19036 * Set #! path to perl in `apps/der_chop` to where we found it
19037 instead of using a fixed path.
19041 * SHA library changes for irix64-mips4-cc.
19045 * Improvements for VMS support.
19049 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
19051 * Bignum library bug fix. IRIX 6 passes "make test" now!
19052 This also avoids the problems with SC4.2 and unpatched SC5.
19054 *Andy Polyakov <appro@fy.chalmers.se>*
19056 * New functions sk_num, sk_value and sk_set to replace the previous macros.
19057 These are required because of the typesafe stack would otherwise break
19058 existing code. If old code used a structure member which used to be STACK
19059 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
19060 sk_num or sk_value it would produce an error because the num, data members
19061 are not present in STACK_OF. Now it just produces a warning. sk_set
19062 replaces the old method of assigning a value to sk_value
19063 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
19064 that does this will no longer work (and should use sk_set instead) but
19065 this could be regarded as a "questionable" behaviour anyway.
19069 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
19070 correctly handle encrypted S/MIME data.
19074 * Change type of various DES function arguments from des_cblock
19075 (which means, in function argument declarations, pointer to char)
19076 to des_cblock * (meaning pointer to array with 8 char elements),
19077 which allows the compiler to do more typechecking; it was like
19078 that back in SSLeay, but with lots of ugly casts.
19080 Introduce new type const_des_cblock.
19084 * Reorganise the PKCS#7 library and get rid of some of the more obvious
19085 problems: find RecipientInfo structure that matches recipient certificate
19086 and initialise the ASN1 structures properly based on passed cipher.
19090 * Belatedly make the BN tests actually check the results.
19094 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19095 to and from BNs: it was completely broken. New compilation option
19096 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19097 key elements as negative integers.
19101 * Reorganize and speed up MD5.
19103 *Andy Polyakov <appro@fy.chalmers.se>*
19107 *Richard Levitte <richard@levitte.org>*
19109 * New option -out to asn1parse to allow the parsed structure to be
19110 output to a file. This is most useful when combined with the -strparse
19111 option to examine the output of things like OCTET STRINGS.
19115 * Make SSL library a little more fool-proof by not requiring any longer
19116 that `SSL_set_{accept,connect}_state` be called before
19117 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19118 in many applications because usually everything *appeared* to work as
19119 intended anyway -- now it really works as intended).
19123 * Move openssl.cnf out of lib/.
19127 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19128 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19129 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19131 *Ralf S. Engelschall*
19133 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19134 handle PKCS#7 enveloped data properly.
19136 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19138 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19139 copying pointers. The cert_st handling is changed by this in
19140 various ways (and thus what used to be known as ctx->default_cert
19141 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19142 any longer when s->cert does not give us what we need).
19143 ssl_cert_instantiate becomes obsolete by this change.
19144 As soon as we've got the new code right (possibly it already is?),
19145 we have solved a couple of bugs of the earlier code where s->cert
19146 was used as if it could not have been shared with other SSL structures.
19148 Note that using the SSL API in certain dirty ways now will result
19149 in different behaviour than observed with earlier library versions:
19150 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19151 does not influence s as it used to.
19153 In order to clean up things more thoroughly, inside SSL_SESSION
19154 we don't use CERT any longer, but a new structure SESS_CERT
19155 that holds per-session data (if available); currently, this is
19156 the peer's certificate chain and, for clients, the server's certificate
19157 and temporary key. CERT holds only those values that can have
19158 meaningful defaults in an SSL_CTX.
19162 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19163 from the internal representation. Various PKCS#7 fixes: remove some
19164 evil casts and set the enc_dig_alg field properly based on the signing
19169 * Allow PKCS#12 password to be set from the command line or the
19170 environment. Let 'ca' get its config file name from the environment
19171 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19176 * Allow certificate policies extension to use an IA5STRING for the
19177 organization field. This is contrary to the PKIX definition but
19178 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19183 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19184 without disallowing inline assembler and the like for non-pedantic builds.
19188 * Support Borland C++ builder.
19190 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19196 * SHA-1 cleanups and performance enhancements.
19198 *Andy Polyakov <appro@fy.chalmers.se>*
19200 * Sparc v8plus assembler for the bignum library.
19202 *Andy Polyakov <appro@fy.chalmers.se>*
19204 * Accept any -xxx and +xxx compiler options in Configure.
19208 * Update HPUX configuration.
19212 * Add missing `sk_<type>_unshift()` function to safestack.h
19214 *Ralf S. Engelschall*
19216 * New function SSL_CTX_use_certificate_chain_file that sets the
19217 "extra_cert"s in addition to the certificate. (This makes sense
19218 only for "PEM" format files, as chains as a whole are not
19223 * Support verify_depth from the SSL API.
19224 x509_vfy.c had what can be considered an off-by-one-error:
19225 Its depth (which was not part of the external interface)
19226 was actually counting the number of certificates in a chain;
19227 now it really counts the depth.
19231 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19232 instead of X509err, which often resulted in confusing error
19233 messages since the error codes are not globally unique
19234 (e.g. an alleged error in ssl3_accept when a certificate
19235 didn't match the private key).
19237 * New function SSL_CTX_set_session_id_context that allows to set a default
19238 value (so that you don't need SSL_set_session_id_context for each
19239 connection using the SSL_CTX).
19243 * OAEP decoding bug fix.
19247 * Support INSTALL_PREFIX for package builders, as proposed by
19252 * New Configure options "threads" and "no-threads". For systems
19253 where the proper compiler options are known (currently Solaris
19254 and Linux), "threads" is the default.
19258 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19262 * Install various scripts to $(OPENSSLDIR)/misc, not to
19263 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19264 such as /usr/local/bin.
19268 * "make linux-shared" to build shared libraries.
19270 *Niels Poppe <niels@netbox.org>*
19272 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19276 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19277 extension adding in x509 utility.
19281 * Remove NOPROTO sections and error code comments.
19285 * Partial rewrite of the DEF file generator to now parse the ANSI
19290 * New Configure options --prefix=DIR and --openssldir=DIR.
19294 * Complete rewrite of the error code script(s). It is all now handled
19295 by one script at the top level which handles error code gathering,
19296 header rewriting and C source file generation. It should be much better
19297 than the old method: it now uses a modified version of Ulf's parser to
19298 read the ANSI prototypes in all header files (thus the old K&R definitions
19299 aren't needed for error creation any more) and do a better job of
19300 translating function codes into names. The old 'ASN1 error code embedded
19301 in a comment' is no longer necessary and it doesn't use .err files which
19302 have now been deleted. Also the error code call doesn't have to appear all
19303 on one line (which resulted in some large lines...).
19307 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19311 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19312 0 (which usually indicates a closed connection), but continue reading.
19316 * Fix some race conditions.
19320 * Add support for CRL distribution points extension. Add Certificate
19321 Policies and CRL distribution points documentation.
19325 * Move the autogenerated header file parts to crypto/opensslconf.h.
19329 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19330 8 of keying material. Merlin has also confirmed interop with this fix
19331 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19333 *Merlin Hughes <merlin@baltimore.ie>*
19335 * Fix lots of warnings.
19337 *Richard Levitte <levitte@stacken.kth.se>*
19339 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19340 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19342 *Richard Levitte <levitte@stacken.kth.se>*
19344 * Fix problems with sizeof(long) == 8.
19346 *Andy Polyakov <appro@fy.chalmers.se>*
19348 * Change functions to ANSI C.
19352 * Fix typos in error codes.
19354 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19356 * Remove defunct assembler files from Configure.
19360 * SPARC v8 assembler BIGNUM implementation.
19362 *Andy Polyakov <appro@fy.chalmers.se>*
19364 * Support for Certificate Policies extension: both print and set.
19365 Various additions to support the r2i method this uses.
19369 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19370 return a const string when you are expecting an allocated buffer.
19374 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19375 types DirectoryString and DisplayText.
19379 * Add code to allow r2i extensions to access the configuration database,
19380 add an LHASH database driver and add several ctx helper functions.
19384 * Fix an evil bug in bn_expand2() which caused various BN functions to
19385 fail when they extended the size of a BIGNUM.
19389 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19390 support typesafe stack.
19394 * Fix typo in SSL_[gs]et_options().
19396 *Nils Frostberg <nils@medcom.se>*
19398 * Delete various functions and files that belonged to the (now obsolete)
19399 old X509V3 handling code.
19403 * New Configure option "rsaref".
19407 * Don't auto-generate pem.h.
19411 * Introduce type-safe ASN.1 SETs.
19415 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19417 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19419 * Introduce type-safe STACKs. This will almost certainly break lots of code
19420 that links with OpenSSL (well at least cause lots of warnings), but fear
19421 not: the conversion is trivial, and it eliminates loads of evil casts. A
19422 few STACKed things have been converted already. Feel free to convert more.
19423 In the fullness of time, I'll do away with the STACK type altogether.
19427 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19428 specified in `<certfile>` by updating the entry in the index.txt file.
19429 This way one no longer has to edit the index.txt file manually for
19430 revoking a certificate. The -revoke option does the gory details now.
19432 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19434 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19435 `-text` option at all and this way the `-noout -text` combination was
19436 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19438 *Ralf S. Engelschall*
19440 * Make sure a corresponding plain text error message exists for the
19441 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19442 verify callback function determined that a certificate was revoked.
19444 *Ralf S. Engelschall*
19446 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19447 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19448 all available ciphers including rc5, which was forgotten until now.
19449 In order to let the testing shell script know which algorithms
19450 are available, a new (up to now undocumented) command
19451 `openssl list-cipher-commands` is used.
19455 * Bugfix: s_client occasionally would sleep in select() when
19456 it should have checked SSL_pending() first.
19460 * New functions DSA_do_sign and DSA_do_verify to provide access to
19461 the raw DSA values prior to ASN.1 encoding.
19465 * Tweaks to Configure
19467 *Niels Poppe <niels@netbox.org>*
19469 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19474 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19478 * New config option to avoid instructions that are illegal on the 80386.
19479 The default code is faster, but requires at least a 486.
19483 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19484 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19485 same as SSL2_VERSION anyway.
19489 * New "-showcerts" option for s_client.
19493 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19494 application. Various cleanups and fixes.
19498 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19499 modify error routines to work internally. Add error codes and PBE init
19500 to library startup routines.
19504 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19505 packing functions to asn1 and evp. Changed function names and error
19506 codes along the way.
19510 * PKCS12 integration: and so it begins... First of several patches to
19511 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19512 objects to objects.h
19516 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19517 and display support for Thawte strong extranet extension.
19521 * Add LinuxPPC support.
19523 *Jeff Dubrule <igor@pobox.org>*
19525 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19526 bn_div_words in alpha.s.
19528 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19530 * Make sure the RSA OAEP test is skipped under -DRSAref because
19531 OAEP isn't supported when OpenSSL is built with RSAref.
19533 *Ulf Moeller <ulf@fitug.de>*
19535 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19536 so they no longer are missing under -DNOPROTO.
19538 *Soren S. Jorvang <soren@t.dk>*
19540 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19542 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19543 doesn't work when the session is reused. Coming soon!
19547 * Fix a security hole, that allows sessions to be reused in the wrong
19548 context thus bypassing client cert protection! All software that uses
19549 client certs and session caches in multiple contexts NEEDS PATCHING to
19550 allow session reuse! A fuller solution is in the works.
19552 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19554 * Some more source tree cleanups (removed obsolete files
19555 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19556 permission on "config" script to be executable) and a fix for the INSTALL
19559 *Ulf Moeller <ulf@fitug.de>*
19561 * Remove some legacy and erroneous uses of malloc, free instead of
19564 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19566 * Make rsa_oaep_test return non-zero on error.
19568 *Ulf Moeller <ulf@fitug.de>*
19570 * Add support for native Solaris shared libraries. Configure
19571 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19572 if someone would make that last step automatic.
19574 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19576 * ctx_size was not built with the right compiler during "make links". Fixed.
19580 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19581 except NULL ciphers". This means the default cipher list will no longer
19582 enable NULL ciphers. They need to be specifically enabled e.g. with
19583 the string "DEFAULT:eNULL".
19587 * Fix to RSA private encryption routines: if p < q then it would
19588 occasionally produce an invalid result. This will only happen with
19589 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19593 * Be less restrictive and allow also `perl util/perlpath.pl
19594 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19595 because this way one can also use an interpreter named `perl5` (which is
19596 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19597 installed as `perl`).
19599 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19601 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19603 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19605 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19606 advapi32.lib to Win32 build and change the pem test comparison
19607 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19608 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19609 and crypto/des/ede_cbcm_enc.c.
19613 * DES quad checksum was broken on big-endian architectures. Fixed.
19617 * Comment out two functions in bio.h that aren't implemented. Fix up the
19618 Win32 test batch file so it (might) work again. The Win32 test batch file
19619 is horrible: I feel ill....
19623 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19624 in e_os.h. Audit of header files to check ANSI and non ANSI
19625 sections: 10 functions were absent from non ANSI section and not exported
19626 from Windows DLLs. Fixed up libeay.num for new functions.
19630 * Make `openssl version` output lines consistent.
19632 *Ralf S. Engelschall*
19634 * Fix Win32 symbol export lists for BIO functions: Added
19635 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19636 to ms/libeay{16,32}.def.
19638 *Ralf S. Engelschall*
19640 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19641 fine under Unix and passes some trivial tests I've now added. But the
19642 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19643 added to make sure no one expects that this stuff really works in the
19644 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19645 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19648 *Ralf S. Engelschall*
19650 * Fix the generation of two part addresses in perl.
19652 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19654 * Add config entry for Linux on MIPS.
19656 *John Tobey <jtobey@channel1.com>*
19658 * Make links whenever Configure is run, unless we are on Windoze.
19662 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19663 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19668 * Add a useful kludge to allow package maintainers to specify compiler and
19669 other platforms details on the command line without having to patch the
19670 Configure script every time: One now can use
19671 `perl Configure <id>:<details>`,
19672 i.e. platform ids are allowed to have details appended
19673 to them (separated by colons). This is treated as there would be a static
19674 pre-configured entry in Configure's %table under key `<id>` with value
19675 `<details>` and `perl Configure <id>` is called. So, when you want to
19676 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19677 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19678 now, which overrides the FreeBSD-elf entry on-the-fly.
19680 *Ralf S. Engelschall*
19682 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19686 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19687 on the `perl Configure ...` command line. This way one can compile
19688 OpenSSL libraries with Position Independent Code (PIC) which is needed
19689 for linking it into DSOs.
19691 *Ralf S. Engelschall*
19693 * Remarkably, export ciphers were totally broken and no-one had noticed!
19698 * Cleaned up the LICENSE document: The official contact for any license
19699 questions now is the OpenSSL core team under openssl-core@openssl.org.
19700 And add a paragraph about the dual-license situation to make sure people
19701 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19702 to the OpenSSL toolkit.
19704 *Ralf S. Engelschall*
19706 * General source tree makefile cleanups: Made `making xxx in yyy...`
19707 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19708 Additionally cleaned up the `make links` target: Remove unnecessary
19709 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19710 to speed processing and no longer clutter the display with confusing
19711 stuff. Instead only the actually done links are displayed.
19713 *Ralf S. Engelschall*
19715 * Permit null encryption ciphersuites, used for authentication only. It used
19716 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19717 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19722 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19723 signed attributes when verifying signatures (this would break them),
19724 the detached data encoding was wrong and public keys obtained using
19725 X509_get_pubkey() weren't freed.
19729 * Add text documentation for the BUFFER functions. Also added a work around
19730 to a Win95 console bug. This was triggered by the password read stuff: the
19731 last character typed gets carried over to the next fread(). If you were
19732 generating a new cert request using 'req' for example then the last
19733 character of the passphrase would be CR which would then enter the first
19738 * Added the new 'Includes OpenSSL Cryptography Software' button as
19739 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19740 button and can be used by applications based on OpenSSL to show the
19741 relationship to the OpenSSL project.
19743 *Ralf S. Engelschall*
19745 * Remove confusing variables in function signatures in files
19746 ssl/ssl_lib.c and ssl/ssl.h.
19748 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19750 * Don't install bss_file.c under PREFIX/include/
19752 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19754 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19755 functions that return function pointers and has support for NT specific
19756 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19757 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19758 unsigned to signed types: this was killing the Win32 compile.
19762 * Add new certificate file to stack functions,
19763 SSL_add_dir_cert_subjects_to_stack() and
19764 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19765 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19766 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19767 This means that Apache-SSL and similar packages don't have to mess around
19768 to add as many CAs as they want to the preferred list.
19772 * Experiment with doxygen documentation. Currently only partially applied to
19774 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19775 openssl.doxy as the configuration file.
19779 * Get rid of remaining C++-style comments which strict C compilers hate.
19781 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19783 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19784 compiled in by default: it has problems with large keys.
19788 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19789 DH private keys and/or callback functions which directly correspond to
19790 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19791 is needed for applications which have to configure certificates on a
19792 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19794 For the RSA certificate situation is makes no difference, but
19795 for the DSA certificate situation this fixes the "no shared cipher"
19796 problem where the OpenSSL cipher selection procedure failed because the
19797 temporary keys were not overtaken from the context and the API provided
19798 no way to reconfigure them.
19799 The new functions now let applications reconfigure the stuff and they
19800 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19801 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19802 non-public-API function ssl_cert_instantiate() is used as a helper
19803 function and also to reduce code redundancy inside ssl_rsa.c.
19805 *Ralf S. Engelschall*
19807 * Move s_server -dcert and -dkey options out of the undocumented feature
19808 area because they are useful for the DSA situation and should be
19809 recognized by the users.
19811 *Ralf S. Engelschall*
19813 * Fix the cipher decision scheme for export ciphers: the export bits are
19814 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19815 SSL_EXP_MASK. So, the original variable has to be used instead of the
19816 already masked variable.
19818 *Richard Levitte <levitte@stacken.kth.se>*
19820 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19822 *Richard Levitte <levitte@stacken.kth.se>*
19824 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19825 from `int` to `unsigned int` because it is a length and initialized by
19826 EVP_DigestFinal() which expects an `unsigned int *`.
19828 *Richard Levitte <levitte@stacken.kth.se>*
19830 * Don't hard-code path to Perl interpreter on shebang line of Configure
19831 script. Instead use the usual Shell->Perl transition trick.
19833 *Ralf S. Engelschall*
19835 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19836 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19837 -noout -modulus` as it's already the case for `openssl rsa -noout
19838 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19839 currently the public key is printed (a decision which was already done by
19840 `openssl dsa -modulus` in the past) which serves a similar purpose.
19841 Additionally the NO_RSA no longer completely removes the whole -modulus
19842 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19845 *Ralf S. Engelschall*
19847 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19848 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19850 *Arne Ansper <arne@ats.cyber.ee>*
19852 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19853 to be added. Now both 'req' and 'ca' can use new objects defined in the
19858 * Add cool BIO that does syslog (or event log on NT).
19860 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19862 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19863 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19864 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19865 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19869 * Add preliminary config info for new extension code.
19873 * Make RSA_NO_PADDING really use no padding.
19875 *Ulf Moeller <ulf@fitug.de>*
19877 * Generate errors when private/public key check is done.
19881 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19882 for some CRL extensions and new objects added.
19886 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19887 key usage extension and fuller support for authority key id.
19891 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19892 padding method for RSA, which is recommended for new applications in PKCS
19893 #1 v2.0 (RFC 2437, October 1998).
19894 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19895 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19896 against Bleichbacher's attack on RSA.
19897 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19900 * Updates to the new SSL compression code
19902 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19904 * Fix so that the version number in the master secret, when passed
19905 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19906 (because the server will not accept higher), that the version number
19907 is 0x03,0x01, not 0x03,0x00
19909 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19911 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19912 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19913 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19917 * Support for RAW extensions where an arbitrary extension can be
19918 created by including its DER encoding. See `apps/openssl.cnf` for
19923 * Make sure latest Perl versions don't interpret some generated C array
19924 code as Perl array code in the crypto/err/err_genc.pl script.
19926 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19928 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19929 not many people have the assembler. Various Win32 compilation fixes and
19930 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19931 build instructions.
19935 * Modify configure script 'Configure' to automatically create crypto/date.h
19936 file under Win32 and also build pem.h from pem.org. New script
19937 util/mkfiles.pl to create the MINFO file on environments that can't do a
19938 'make files': perl util/mkfiles.pl >MINFO should work.
19942 * Major rework of DES function declarations, in the pursuit of correctness
19943 and purity. As a result, many evil casts evaporated, and some weirdness,
19944 too. You may find this causes warnings in your code. Zapping your evil
19945 casts will probably fix them. Mostly.
19949 * Fix for a typo in asn1.h. Bug fix to object creation script
19950 obj_dat.pl. It considered a zero in an object definition to mean
19951 "end of object": none of the objects in objects.h have any zeros
19952 so it wasn't spotted.
19954 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19956 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19957 Masking (CBCM). In the absence of test vectors, the best I have been able
19958 to do is check that the decrypt undoes the encrypt, so far. Send me test
19959 vectors if you have them.
19963 * Correct calculation of key length for export ciphers (too much space was
19964 allocated for null ciphers). This has not been tested!
19968 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19969 message is now correct (it understands "crypto" and "ssl" on its
19970 command line). There is also now an "update" option. This will update
19971 the util/ssleay.num and util/libeay.num files with any new functions.
19973 perl util/mkdef.pl crypto ssl update
19974 it will update them.
19978 * Overhauled the Perl interface:
19979 - ported BN stuff to OpenSSL's different BN library
19980 - made the perl/ source tree CVS-aware
19981 - renamed the package from SSLeay to OpenSSL (the files still contain
19982 their history because I've copied them in the repository)
19983 - removed obsolete files (the test scripts will be replaced
19984 by better Test::Harness variants in the future)
19986 *Ralf S. Engelschall*
19988 * First cut for a very conservative source tree cleanup:
19989 1. merge various obsolete readme texts into doc/ssleay.txt
19990 where we collect the old documents and readme texts.
19991 2. remove the first part of files where I'm already sure that we no
19992 longer need them because of three reasons: either they are just temporary
19993 files which were left by Eric or they are preserved original files where
19994 I've verified that the diff is also available in the CVS via "cvs diff
19995 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19996 the crypto/md/ stuff).
19998 *Ralf S. Engelschall*
20000 * More extension code. Incomplete support for subject and issuer alt
20001 name, issuer and authority key id. Change the i2v function parameters
20002 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
20003 what that's for :-) Fix to ASN1 macro which messed up
20004 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
20008 * Preliminary support for ENUMERATED type. This is largely copied from the
20013 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
20015 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20017 * Make sure `make rehash` target really finds the `openssl` program.
20019 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
20021 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
20022 like to hear about it if this slows down other processors.
20026 * Add CygWin32 platform information to Configure script.
20028 *Alan Batie <batie@aahz.jf.intel.com>*
20030 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
20032 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
20034 * New program nseq to manipulate netscape certificate sequences
20038 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
20043 * Fixes to BN code. Previously the default was to define BN_RECURSION
20044 but the BN code had some problems that would cause failures when
20045 doing certificate verification and some other functions.
20047 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
20049 * Add ASN1 and PEM code to support netscape certificate sequences.
20053 * Add ASN1 and PEM code to support netscape certificate sequences.
20057 * Add several PKIX and private extended key usage OIDs.
20061 * Modify the 'ca' program to handle the new extension code. Modify
20062 openssl.cnf for new extension format, add comments.
20066 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
20067 and add a sample to openssl.cnf so req -x509 now adds appropriate
20072 * Continued X509 V3 changes. Add to other makefiles, integrate with the
20073 error code, add initial support to X509_print() and x509 application.
20077 * Takes a deep breath and start adding X509 V3 extension support code. Add
20078 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
20079 stuff is currently isolated and isn't even compiled yet.
20083 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
20084 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
20085 Removed the versions check from X509 routines when loading extensions:
20086 this allows certain broken certificates that don't set the version
20087 properly to be processed.
20091 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20092 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20093 can still be regenerated with "make depend".
20097 * Spelling mistake in C version of CAST-128.
20099 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20101 * Changes to the error generation code. The perl script err-code.pl
20102 now reads in the old error codes and retains the old numbers, only
20103 adding new ones if necessary. It also only changes the .err files if new
20104 codes are added. The makefiles have been modified to only insert errors
20105 when needed (to avoid needlessly modifying header files). This is done
20106 by only inserting errors if the .err file is newer than the auto generated
20107 C file. To rebuild all the error codes from scratch (the old behaviour)
20108 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20109 or delete all the .err files.
20113 * CAST-128 was incorrectly implemented for short keys. The C version has
20114 been fixed, but is untested. The assembler versions are also fixed, but
20115 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20116 to regenerate it if needed.
20117 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20118 Hagino <itojun@kame.net>*
20120 * File was opened incorrectly in randfile.c.
20122 *Ulf Möller <ulf@fitug.de>*
20124 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20125 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20126 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20127 al: it's just almost always a UTCTime. Note this patch adds new error
20128 codes so do a "make errors" if there are problems.
20132 * Correct Linux 1 recognition in config.
20134 *Ulf Möller <ulf@fitug.de>*
20136 * Remove pointless MD5 hash when using DSA keys in ca.
20138 *Anonymous <nobody@replay.com>*
20140 * Generate an error if given an empty string as a cert directory. Also
20141 generate an error if handed NULL (previously returned 0 to indicate an
20142 error, but didn't set one).
20144 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20146 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20150 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20151 parameters. This was causing a warning which killed off the Win32 compile.
20155 * Remove C++ style comments from crypto/bn/bn_local.h.
20157 *Neil Costigan <neil.costigan@celocom.com>*
20159 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20160 based on a text string, looking up short and long names and finally
20161 "dot" format. The "dot" format stuff didn't work. Added new function
20162 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20163 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20164 OID is not part of the table.
20168 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20169 X509_LOOKUP_by_alias().
20173 * Sort openssl functions by name.
20177 * Get the `gendsa` command working and add it to the `list` command. Remove
20178 encryption from sample DSA keys (in case anyone is interested the password
20183 * Make *all* `*_free` functions accept a NULL pointer.
20185 *Frans Heymans <fheymans@isaserver.be>*
20187 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20190 *Anonymous <nobody@replay.com>*
20192 * s_server should send the CAfile as acceptable CAs, not its own cert.
20194 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20196 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20198 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20200 * Temp key "for export" tests were wrong in s3_srvr.c.
20202 *Anonymous <nobody@replay.com>*
20204 * Add prototype for temp key callback functions
20205 SSL_CTX_set_tmp_{rsa,dh}_callback().
20209 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20210 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20214 * X509_name_add_entry() freed the wrong thing after an error.
20216 *Arne Ansper <arne@ats.cyber.ee>*
20218 * rsa_eay.c would attempt to free a NULL context.
20220 *Arne Ansper <arne@ats.cyber.ee>*
20222 * BIO_s_socket() had a broken should_retry() on Windoze.
20224 *Arne Ansper <arne@ats.cyber.ee>*
20226 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20228 *Arne Ansper <arne@ats.cyber.ee>*
20230 * Make sure the already existing X509_STORE->depth variable is initialized
20231 in X509_STORE_new(), but document the fact that this variable is still
20232 unused in the certificate verification process.
20234 *Ralf S. Engelschall*
20236 * Fix the various library and `apps/` files to free up pkeys obtained from
20237 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20241 * Fix reference counting in X509_PUBKEY_get(). This makes
20242 demos/maurice/example2.c work, amongst others, probably.
20244 *Steve Henson and Ben Laurie*
20246 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20247 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20248 are no longer created. This way we have a single and consistent command
20249 line interface `openssl <command>`, similar to `cvs <command>`.
20251 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20253 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20254 BIT STRING wrapper always have zero unused bits.
20258 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20262 * Make the top-level INSTALL documentation easier to understand.
20266 * Makefiles updated to exit if an error occurs in a sub-directory
20267 make (including if user presses ^C) [Paul Sutton]
20269 * Make Montgomery context stuff explicit in RSA data structure.
20273 * Fix build order of pem and err to allow for generated pem.h.
20277 * Fix renumbering bug in X509_NAME_delete_entry().
20281 * Enhanced the err-ins.pl script so it makes the error library number
20282 global and can add a library name. This is needed for external ASN1 and
20283 other error libraries.
20287 * Fixed sk_insert which never worked properly.
20291 * Fix ASN1 macros so they can handle indefinite length constructed
20292 EXPLICIT tags. Some non standard certificates use these: they can now
20297 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20298 into a single doc/ssleay.txt bundle. This way the information is still
20299 preserved but no longer messes up this directory. Now it's new room for
20300 the new set of documentation files.
20302 *Ralf S. Engelschall*
20304 * SETs were incorrectly DER encoded. This was a major pain, because they
20305 shared code with SEQUENCEs, which aren't coded the same. This means that
20306 almost everything to do with SETs or SEQUENCEs has either changed name or
20307 number of arguments.
20309 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20311 * Fix test data to work with the above.
20315 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20316 was already fixed by Eric for 0.9.1 it seems.
20318 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20320 * Autodetect FreeBSD3.
20324 * Fix various bugs in Configure. This affects the following platforms:
20328 unixware-2.0-pentium
20333 * Eliminate generated files from CVS. Reorder tests to regenerate files
20334 before they are needed.
20338 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20342 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20344 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20345 changed SSLeay to OpenSSL in version strings.
20347 *Ralf S. Engelschall*
20349 * Some fixups to the top-level documents.
20353 * Fixed the nasty bug where rsaref.h was not found under compile-time
20354 because the symlink to include/ was missing.
20356 *Ralf S. Engelschall*
20358 * Incorporated the popular no-RSA/DSA-only patches
20359 which allow to compile an RSA-free SSLeay.
20361 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20363 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20364 when "ssleay" is still not found.
20366 *Ralf S. Engelschall*
20368 * Added more platforms to Configure: Cray T3E, HPUX 11,
20370 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20372 * Updated the README file.
20374 *Ralf S. Engelschall*
20376 * Added various .cvsignore files in the CVS repository subdirs
20377 to make a "cvs update" really silent.
20379 *Ralf S. Engelschall*
20381 * Recompiled the error-definition header files and added
20382 missing symbols to the Win32 linker tables.
20384 *Ralf S. Engelschall*
20386 * Cleaned up the top-level documents;
20387 o new files: CHANGES and LICENSE
20388 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20389 o merged COPYRIGHT into LICENSE
20390 o removed obsolete TODO file
20391 o renamed MICROSOFT to INSTALL.W32
20393 *Ralf S. Engelschall*
20395 * Removed dummy files from the 0.9.1b source tree:
20396 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20397 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20398 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20399 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20400 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20402 *Ralf S. Engelschall*
20404 * Added various platform portability fixes.
20408 * The Genesis of the OpenSSL rpject:
20409 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20410 Young and Tim J. Hudson created while they were working for C2Net until
20413 *The OpenSSL Project*
20415 ### Changes between 0.9.0b and 0.9.1b [not released]
20417 * Updated a few CA certificates under certs/
20421 * Changed some BIGNUM api stuff.
20425 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20426 DGUX x86, Linux Alpha, etc.
20430 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20431 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20436 * Add -strparse option to asn1pars program which parses nested
20439 *Dr Stephen Henson <shenson@bigfoot.com>*
20441 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20445 * DSA fix for "ca" program.
20449 * Added "-genkey" option to "dsaparam" program.
20453 * Added RIPE MD160 (rmd160) message digest.
20457 * Added -a (all) option to "ssleay version" command.
20461 * Added PLATFORM define which is the id given to Configure.
20465 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20469 * Extended the ASN.1 parser routines.
20473 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20477 * Added a BN_CTX to the BN library.
20481 * Fixed the weak key values in DES library
20485 * Changed API in EVP library for cipher aliases.
20489 * Added support for RC2/64bit cipher.
20493 * Converted the lhash library to the crypto/mem.c functions.
20497 * Added more recognized ASN.1 object ids.
20501 * Added more RSA padding checks for SSL/TLS.
20505 * Added BIO proxy/filter functionality.
20509 * Added extra_certs to SSL_CTX which can be used
20510 send extra CA certificates to the client in the CA cert chain sending
20511 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20515 * Now Fortezza is denied in the authentication phase because
20516 this is key exchange mechanism is not supported by SSLeay at all.
20520 * Additional PKCS1 checks.
20524 * Support the string "TLSv1" for all TLS v1 ciphers.
20528 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20529 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20533 * Fixed a few memory leaks.
20537 * Fixed various code and comment typos.
20541 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20542 bytes sent in the client random.
20544 *Edward Bishop <ebishop@spyglass.com>*
20548 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20549 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20550 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20551 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20552 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20553 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20554 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20555 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20556 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20557 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20558 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20559 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20560 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20561 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20562 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20563 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20564 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20565 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20566 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20567 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20568 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20569 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20570 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20571 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20572 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20573 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20574 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20575 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20576 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20577 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20578 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20579 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20580 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20581 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20582 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20583 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20584 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20585 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20586 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20587 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20588 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20589 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20590 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20591 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20592 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20593 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20594 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20595 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20596 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20597 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20598 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20599 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20600 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20601 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20602 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20603 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20604 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20605 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20606 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20607 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20608 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20609 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20610 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20611 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20612 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20613 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20614 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20615 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20616 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20617 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20618 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20619 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20620 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20621 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20622 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20623 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20624 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20625 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20626 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20627 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20628 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20629 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20630 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20631 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20632 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20633 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20634 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20635 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20636 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20637 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20638 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20639 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20640 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20641 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20642 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20643 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20644 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20645 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20646 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20647 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20648 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20649 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20650 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20651 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20652 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20653 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20654 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20655 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20656 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20657 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20658 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20659 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20660 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20661 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20662 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20663 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20664 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20665 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20666 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20667 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20668 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20669 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20670 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20671 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20672 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20673 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20674 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20675 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20676 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20677 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20678 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20679 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20680 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20681 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20682 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20683 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20684 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20685 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20686 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20687 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20688 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20689 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20690 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20691 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20692 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20693 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20694 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20695 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20696 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20697 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20698 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20699 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20700 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20701 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20702 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20703 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20704 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20705 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20706 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20707 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20708 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20709 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20710 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20711 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20712 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20713 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20714 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20715 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20716 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20717 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20718 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20719 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20720 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20721 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20722 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20723 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20724 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20725 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20726 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20727 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20728 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20729 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20730 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20731 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20732 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20733 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20734 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20735 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655